Global Financial Systems Stability and Risk by Danielsson

Global Financial Systems
Stability and Risk

Jon Danielsson
Under what circumstances have we achieved financial stability?

Which previous crises inform the current ones and in what way?
What are the common themes and lessons for policy, regulation and financial theory?
Global Financial Systems: Stability and Risk is an innovative textbook that explores the ‘why’ behind
global financial stability, providing insightful discussions on the international financial system and the
contemporary issues of today. Drawing on economic theory, finance, mathematical modelling and risk
theory, this book presents a comprehensive, coherent and current economic analysis of the inherent
instabilities of the financial system, and the design of optimal policy response.
Key features
• Up-to-date and thorough analysis of the 2007/08 financial crisis.
• Case studies and practical examples illustrate key arguments and apply the theory to the
real world.
• End-of-chapter questions provoke discussion and critical thinking, and provide
opportunities to test your understanding.
• Accompanied by instructor resources including PowerPoint slides, plus an author-hosted
website featuring regular updates on current events in the global financial system and links
to useful websites.
Jon Danielsson is Reader in Finance and member of the Financial

Markets Group at the London School of Economics and Political
Science, and co-director of the LSE’s Financial Markets Group’s
Systemic Risk Centre.
Front cover image:
© Getty Images
www.pearson-books.com
Global Financial
SYSTEMS
i
Global Financial
Systems
Stability and Risk
Jon Danielsson
Director, Systemic Risk Centre
London School of Economics
PEARSON EDUCATION LIMITED
Edinburgh Gate
Harlow CM20 2JE
United Kingdom
Tel: +44 (0)1279 623623
Web: www.pearson.com/uk
First published 2013 (print and electronic)
© Jon Danielsson 2013 (print and electronic)
The right of Jon Danielsson to be identified as author of this work has been asserted by him in accordance
with the Copyright, Designs and Patents Act 1988.
The print publication is protected by copyright. Prior to any prohibited reproduction, storage in a retrieval
system, distribution or transmission in any form or by any means, electronic, mechanical, recording or
otherwise, permission should be obtained from the publisher or, where applicable, a licence permitting
restricted copying in the United Kingdom should be obtained from the Copyright Licensing Agency Ltd,
Saffron House, 6–10 Kirby Street, London EC1N 8TS.
The ePublication is protected by copyright and must not be copied, reproduced, transferred, distributed,
leased, licensed or publicly performed or used in any way except as specifically permitted in writing by the
publishers, as allowed under the terms and conditions under which it was purchased, or as strictly permitted
by applicable copyright law. Any unauthorised distribution or use of this text may be a direct infringement of
the author's and the publishers' rights and those responsible may be liable in law accordingly.
All trademarks used herein are the property of their respective owners. The use of any trademark in this text
does not vest in the author or publisher any trademark ownership rights in such trademarks, nor does the use
of such trademarks imply any affiliation with or endorsement of this book by such owners.
Pearson Education is not responsible for the content of third-party internet sites.
ISBN: 978-0-273-77466-2 (print)

978-0-273-77471-6 (PDF)
978-0-273-77467-9 (eText)
British Library Cataloguing-in-Publication Data

A catalogue record for the print edition is available from the British Library
Library of Congress Cataloging-in-Publication Data

Daníelsson, Jón.
Global financial systems / Jon Danielsson.
pages cm
ISBN 978-0-273-77466-2
1. International finance. I. Title.
HG3881.D3266 2013
332'.042—dc23
2013012036
10 9 8 7 6 5 4 3 2 1
16 15 14 13
Print edition typeset in 9.25/13.5 Stone Humanist ITC Std by 75

Print edition printed and bound in Slovenia by Svet Print – Ljubljana d.o.o.
NOTE THAT ANY PAGE CROSS-REFERENCES REFER TO THE PRINT EDITION

Contents
Author’s acknowledgements xi
Publisher’s acknowledgements xii
Introduction xiii
1 Systemic risk 1
1.1 Case study: the 1914 crisis 2
1.2 The concept of systemic risk 4
1.3 Who creates systemic risk? 9
1.4 Fundamental origins of systemic risk 11
1.5 Summary 17
References 18
2 The Great Depression, 1929–1933 19

2.1 Build-up to a depression 20
2.2 The Great Depression 25
2.3 Causes of the Great Depression 29
2.4 Implications for future policy 35
2.5 Summary 37
References 38
3 Endogenous risk 39
3.1 Millennium Bridge 41
3.2 Dual role of prices 43
3.3 Risk 46
3.4 Dynamic trading strategies 48
3.5 Actual and perceived risk and bubbles 53
3.6 The LTCM crisis of 1998 56
3.7 Conclusion 59
References 60
4 Liquidity 61
4.1 The liquidity crisis of 1998 62
4.2 What is liquidity? 65
4.3 Liquidity models 70
4.4 Policy implications 73
4.5 Summary 75
References 76
v
Contents
5 The central bank 77

5.1 The origins of central banks 78
5.2 Banking supervision 80
5.3 Monetary policy 82
5.4 Financial stability 86
5.5 Bailing out governments 87
5.6 Challenges for central banking 90
5.7 Summary 93
Appendix: central bank interest rate 94
References 95
6 The Asian crisis of 1997 and the IMF 97

6.1 Building up to a crisis 99
6.2 The crisis in individual countries 102
6.3 Reasons for the crisis 106
6.4 Policy options for the crisis countries 109
6.5 Role of the IMF 112
6.6 Wider lessons 115
6.7 Summary 117
References 118
7 Banking crises 119

7.1 Money and early banking 121
7.2 Moral hazard 123
7.3 Costs of banking crises 124
7.4 Causes of banking crises 126
7.5 Bank and banking system failures 128
7.6 Summary 134
References 135
8 Bank runs and deposit insurance 137

8.1 Bank runs and crises 139
8.2 Modelling deposit insurance 142
8.3 Pros and cons of deposit insurance 147
8.4 Summary 149
References 150
9 Trading and speculation 151

9.1 Trading scandals and abuse 152
9.2 Trading and risk 154
9.3 Trading activities 156
vi
Contents
9.4 Policy issues 163

9.5 Summary 171
Appendix: basic terminology of trading 171
References 173
10 Credit markets 175

10.1 Market for credit 176
10.2 Credit rating agencies 179
10.3 Credit models 183
10.4 Margins, haircuts and mark-to-market 185
10.5 Securitisation 188
10.6 Summary 191
References 191
11 Currency markets 193

11.1 Fixed or floating 195
11.2 Foreign exchange interventions 197
11.3 Capital controls 200
11.4 Exchange rate regimes 203
11.5 Perils of overvaluation 208
11.6 Undervaluation and ‘currency wars’ 209
11.7 Reserve currency 210
11.8 Summary 212
Appendix: exchange rate regimes 213
References 215
12 Currency crisis models 217

12.1 First-generation models 219
12.2 The Argentinian crisis 222
12.3 Second-generation models 224
12.4 The European crisis, 1992–1993 228
12.5 Global games currency crisis model 231
12.6 Summary 234
References 236
13 Financial regulations 237

13.1 Banking regulations 239
13.2 Bank capital 245
13.3 International financial regulations: Basel 250
13.4 Summary 257
Appendix: Value-at-Risk 257
References 259
vii
Contents
14 Bailouts 261
14.1 Successful and unsuccessful bailouts 263
14.2 The historical origins of lending of last resort (LOLR) 265
14.3 What are bailouts? 267
14.4 Alternatives to bailouts 272
14.5 Bailouts in the crisis starting in 2007 273
14.6 Bailouts, moral hazard and politics 278
14.7 Model of asset bubbles 279
14.8 Summary 283
References 284
15 Dangerous financial instruments 285

15.1 Complexity kills 287
15.2 Derivatives 288
15.3 Credit default swaps 289
15.4 Collateralised debt obligations 293
15.5 Summary 301
Appendix A: mechanics of CDSs 301
Appendix B: CDO calculations 303
References 308
16 Failures in risk management and regulations

before the crisis 309
16.1 Regulatory failures 310
16.2 Capital and the crisis 317
16.3 Summary 323
References 323
17 The ongoing crisis: 2007–2009 phase 325

17.1 Build-up to a crisis 326
17.2 Hidden and ignored risk 330
17.3 The changing nature of banking 331
17.4 Crisis, 2007–2008 333
17.5 Was it a subprime crisis? 337
17.6 Policy response 338
17.7 Summary 339
References 339
18 Ongoing developments in financial regulation 341

18.1 New and changed institutions 344
18.2 Basel III 348
18.3 Liquidity 352
18.4 How much capital? 356
viii
Contents
18.5 Recovery and resolution 359

18.6 What about too big to fail? 361
18.7 Summary 364
References 365
19 Sovereign debt crises 367

19.1 Newfoundland 368
19.2 Sovereign debt 369
19.3 Enforcement 375
19.4 Background to the European sovereign debt crisis 377
19.5 Summary 389
References 390
Glossary 391
Bibliography 393
Index 399
Errata and online chapter

For errata and an online chapter on the latest regulation and crisis
development, please visit www.GlobalFinancialSystems.org.
ix
Author’s Acknowledgements
This book is based on my lecture notes for my LSE course ‘Global Financial Systems’. The
course was originally developed by Hyun Song Shin, who taught it from 2001 until 2004
and generously gave me access to his lecture notes which formed the skeleton for the first
time I taught the course in 2008.
I was fortunate to be able to employ two exemplary students, Jacqueline Li and Jing
Zeng, who took the course in 2008, to work with me in the summer of 2009 to help me
to develop the lecture material into what became this book. Their assistance was invalu-
able, and they made significant contributions both to the book version of the lecture
notes and to the slides. Their ability to master the computational aspects of the course
material (latex, mercurial, R) and the economic topics was impressive, as was the appar-
ent ease with which they mastered all the technicalities. Their work was generously
funded by the LSE Teaching Development Fund and the Department of Finance.
I was equally fortunate during the summer of 2010 to employ three former students of
the course, Georgia Lv, Eric Pashman and Nick Zeifang. Since then, Kyounghwan Lee has
been very helpful in developing some of the material. Several students have made valu-
able contributions, including Fan Gao, Radhika Saini, Yong Bin Ng, Murathan Kurt, Basak
Yeltekin, Janis Sussick and Stefan Doykin.
Several friends made invaluable contributions to the book. Robert Macrae of Arcus
Investment read the manuscript twice, and made innumerable comments. His insight into
the subject matter and his keen understanding of financial markets from an investor’s
point of view made the book much better than it otherwise would have been.
Con Keating of EFFAS read the whole manuscript, and by identifying the places where
my analysis and logic fell short, significantly improved the book.
Giovanni Bassini read the chapters on financial regulations and pointed out the most
up-to-date developments. David Schraa from IIF read the chapter on future developments
in regulations and gave me very useful insights into the fine points of the ongoing regula-
tory debate.
I would like to thank my wife Sigrun for her shrewd comments, her honest observations
and, above all, her patience.
Without all of these people, this book would not have seen the light of day. Any inac-
curacies and errors that remain in the book are the result of my not paying enough atten-
tion to their comments.
The author gratefully acknowledges the support of the Economic and

Social Research Council (UK) [grant number ES/K 002309/1].
xi
Publisher’s Acknowledgements
We are grateful to the following for permission to reproduce copyright material:
Logos
Logo on page xi from Economic and Social Research Council.
Tables
Table 5.1 from On the measurement of Zimbabwe’s hyperinflation, Cato Journal, Vol. 29,
No. 2, 356 (Hanke, S. H. and Kwok, A. K. F. Spring/Summer 2009), Copyright © Cato
Institute. All rights reserved; Tables 6.1, 6.2 after http://data.worldbank.org/, World Bank;
Table 11.2 from Forty Years’ Experience with the OECD Code of Liberalisation of Capital
Movements, OECD Publishing. http://dx.doi.org/10.1787/9789264176188-en.
Photographs
The publisher would like to thank the following for their kind permission to reproduce
their photographs:
Getty Images: Cate Gillon / Staff 173, Agus Lolong 147.
Cover images: Front: Getty Images
In some instances we have been unable to trace the owners of copyright material, and we
would appreciate any information that would enable us to do so.
xii
Introduction
The focus of this book is on how the world’s financial system functions, the various policy
choices governments have, and how the system has built-in vulnerabilities which lead to
crises. Financial crises have been our constant companion from the very first time human
beings created a financial system. This means that over time, we have accumulated deep
knowledge and understanding of the economic forces enabling such crises. This experi-
ence shows that financial crises are all fundamentally the same, only the details differ. This
is why financial crises are so hard to prevent and so costly to fight. Every time, we are faced
with new details that enforce age-old vulnerabilities.
The various types of financial-system fragilities are systematically analysed in this book,
lessons from past crisis events are used to study recent crises, and up-to-date research is
employed for the analysis of crises long past. In doing so, we make use of the rich body
of research that has emerged with the continuous crisis from 2007.
The focus of the book is on policy issues. It uses a number of case studies, aiming to
create a unified theme linking all the cases. It is written from a point of view of economics,
but does have relevance in other policy-oriented fields such as government, political sci-
ence and law.
The target level is intermediate to advanced undergraduate students and masters stu-
dents. Some knowledge of economics and financial markets is helpful but not essential.
Most of the chapters are non-mathematical, a few make some use of mathematics and
one is essentially dedicated to formal models. It is an open question what is the best way
to incorporate the technical material. It is quite helpful for students with the right back-
ground, but can be skipped by others.
The first part of the book presents basic concepts in financial stability, such as systemic
risk, endogenous risk, the fundamental role of the central bank and the multi-faceted
concept of liquidity. These chapters provide a foundation for more specialised analysis
later in the book. One important case study is contained in this first part, the Great
Depression of the 1930s, which is the worst financial and economic crisis we have ever
seen and, hence, has had a huge impact on financial and economic policy. The response
to the ongoing crisis is significantly shaped by the Great Depression.
After this, we provide more specific analysis of the various parts of the financial system,
and how they relate to financial stability. We start with a detailed case study of the Asian
crisis of 1997 which clearly demonstrates the various dimensions of modern financial
crises. This is followed by discussion of banking crises and bank runs. Thereafter the book
focuses on financial markets, speculation, trading and the market for credit. The final part
of the financial system that is discussed is currency markets, which are the focus of two
chapters, the first focused on policy issues and the second currency crisis models.
Governments have various ways of intervening in financial markets, and the next chap-
ters in the book discuss financial regulations and bailouts. Financial regulation is a rich
topic that is easily the subject of its own book, but here we limit the discussion to the
xiii
Introduction
main arguments for regulating, some of the challenges faced by the authorities and finally
international regulations on bank capital. The last part of the book focuses on the
ongoing crisis from 2007, starting with the dangerous financial instruments that were one
of the factors that enabled excessive risk-taking to take place out of sight. Failures in
financial regulation and risk management also played a role and are the subject of the
following chapter. After this, we bring together the main lessons from the book into
a three-chapter discussion on the ongoing crisis, first the credit phase from 2007 until
2009, then how the crisis is shaping financial regulations, and finally sovereign debt cri-
ses, both generally and with specific reference to the ongoing European sovereign debt
crisis. Because this crisis is ongoing, we maintain a final chapter online at www.
GlobalFinancialSystems.org addressing day-to-day developments in the crisis. This chap-
ter is updated regularly to reflect new information.
xiv
1 SysTemic risk
The world’s economy was on the brink of collapse in the autumn of 2008. Confidence,
the lifeblood of the financial system, was evaporating at an alarming rate, financial
institutions refused to do business with each other, people took their money out
of banks and it looked like the real economy might be heading for a second Great
Depression. Then, just as suddenly as the crisis materialised, it seemed like it was over.
What we experienced was a near-miss systemic crisis, generally defined as the col-
lapse of the entire financial system, followed by an economic depression. The full
crisis was only averted thanks to the swift actions of the authorities.
Over the past half-century, until a few years ago, systemic risk had been the pur-
view of a few academics and policy makers, very much a backwater discipline. The
prevailing approach was to study the risk of the individual parts of the financial sys-
tem separately, not in aggregate, since the objective of interest was the institution,
not the system. An example is how the Basel Accords – the main body of international
financial regulations – focus on individual prudential behaviour instead of the finan-
cial system in its entirety. The series of crises that started in 2007 demonstrated the
folly of such thinking.
2007 was not the first time we faced systemic risk; it has been present ever since
the first financial system was created, and is an inevitable part of any market-based
economic system. It was a real and recognised danger during the era of the first
globalism – 1873 to 1914. In the highly regulated financial world after the Second
World War (WWII), systemic crises were a relatively remote eventuality, only to
re-emerge with the collapse of the Bretton Woods system in the early 1970s.
Chapter 1 SysTemic risk
Systemic risk is most damaging for advanced financial markets; after all, a country
with a small and underdeveloped financial system is much more resilient to problems
in the financial sector. The key question for policy makers in countries with advanced
financial markets is how to limit the build-up of systemic risk and contain crisis events
when they happen. To answer this we need to identify and understand the different
aspects of systemic risk and the tools available to policy makers.
Links to other chapters

This chapter introduces the main concepts of financial stability and systemic risk, and
many of these topics will be discussed in considerable detail in later chapters. For
example, liquidity is discussed in Chapter 4, and the acceleration mechanisms for
pro-cyclicality in Chapter 3. Ultimately, these concepts are applied to the various case
studies of financial crises, most importantly the ongoing crisis, discussed in several
chapters towards the end of this book.
Key concepts
■ Systemic risk
■ Moral hazard
■ Liquidity
■ Bank runs
■ Pro-cyclicality
■ Firesale externality
Readings for this chapter

Few academic studies of systemic risk are available in the existing literature, but this
is now rapidly changing. Several authors have written about financial crisis from a his-
torical point of view, for example, Ferguson (2008), Kindleberger (1996) and Reinhart
and Rogoff (2009). On a more technical level, early analysis was provided by Minsky
(1992), with Bandt and Hartmann (2000) discussing the latest work on the topic prior
to the crises from 2007.
Notation specific to this chapter

G The money multiplier
D Reserve requirement
1.1 Case study: The 1914 crisis

The biggest systemic event in recent history may have occurred in 1914. Brown (1940) and
Ferguson (2008, p. 298) document the chain of events. The main crisis event was charac-
terised by a rapid loss of confidence, with leverage, liquidity and interconnectedness all
playing a major role.
In 1914, globalism amongst the world’s industrial nations was perhaps as advanced
as now, and maybe even more so. At the time, the major supplier of credit to the world
was the City of London which had developed a highly sophisticated financial industry.
2
1.1 Case study: The 1914 crisis
Unfortunately, that sophistication created particular vulnerabilities, such as interconnect-

edness, common to modern financial centres.
The crisis was triggered by an exogenous shock, the assassination of Archduke Franz
Ferdinand of Austria on 28 June 1914. This led to posturing amongst the great powers of
Europe and the build-up of the expectation of war. That expectation, in turn, created wor-
ries that financial institutions might experience difficulties in having cross-border loans
repaid — after all, if two countries are at war, it is difficult to enforce contracts across bor-
ders. Surprisingly, even after the war started, remittance between the Central Powers and
the Allies continued via neutral countries, especially Switzerland, helping to propel it into
the ranks of major financial centres.
Cross-border creditors started to repatriate funds. The immediate financial conse-
quence was that demand for sterling spiked, not least since London banks and finance
houses called in their overseas balances. The pound appreciated from $4.89 to $6.35 in
New York in July, but this soon reversed and the dollar became the world’s dominant cur-
rency, a status it has enjoyed ever since.
However, not all loans are callable, and financial institutions heavily exposed to cross-
border lending were vulnerable, as were those exposed to those exposed to cross-border
lending, and so on. In effect, everybody was exposed, even those who only did domestic
banking. This is an example of network effects in the financial system.
The supply of sterling dried up. The banks stopped loaning funds to the discount mar-
ket, which was a sort of interbank market, as they needed to hold everything in cash in
order to meet demand.
The collapse of the discount market, which was the single most important source of
sterling funds, meant that London acceptance houses, already facing losses from clients
unable to remit funds, were unable to take on more lending and, hence, stopped making
acceptances on 27 July.
One example of this slowdown is seen in Figure 1.1 which shows the 1914 volume of
new long-term foreign loans made in the City of London. In the second half of 1914 the
volume collapsed, along with activity on the financial markets.
Foreign countries and domestic firms sought to convert their stock holdings into cash,
but found no buyers on the London Stock Exchange. Finally, the stock exchange closed
shortly after 10 a.m. on 31 July.
40 million
30 million
20 million
10 million
0 million
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Figure 1.1 Amount of foreign loans made in the City of London in 1914 (in sterling)
Data source: Brown (1940)
3
A general moratorium was issued on banker acceptances, lasting for two months. The
Bank of England (BoE) made further advances of £200 million to the money markets, and
on 13 August obtained a government guarantee to discount bills in the market without
recourse, whilst also allowing for delays in repayments. The London Stock Exchange was
only reopened on 4 January 1915.
Analysis
It does not take all that much for a panic to happen when confidence evaporates. Even
rumours that a large institution might fail can cause panic. In 1914, it was clear that some
institutions were heavily exposed to cross-border lending and would fail in case of war.
That was enough to trigger the crisis. We did not need the actual event of a war to make
this happen; the expectation of a war was sufficient.
The authorities played a clear role, similar to what they did at the height of the crises
from 2007 but on a much larger scale. The BoE resorted to what now would be called
quantitative easing (QE), literally printing money. In the 10 days to 1 August, the Bank
made securitised loans, worth £31,700,000, to discount houses and the Stock Exchange,
most of which had to be paid back one year after the war. The bank rate jumped from 3%
to 10% in three days in order to prevent the gold bleeding to France. On 1 August, the Peel
Act was suspended and the Bank was allowed to issue notes in excess of its gold holdings.
The policy intervention was successful and the City of London did survive, even if not
unscathed. A systemic crisis was averted. By comparison, events during the crises that
started in 2007 are quite mild.
There are two main lessons to be taken from these events. First, in a crisis, the authori-
ties have no choice but to take extreme measures if they want to save their economies.
Secondly, because decisions have to be taken rapidly it is much better if the authorities
are prepared. In 1914 the BoE did follow the established practice of lending of last resort
(LOLR), developed half a century earlier.
1.2 The concept of systemic risk

The financial system is vulnerable to many different types of shocks, both coming from
outside the financial system and generated by the financial system itself. Some shocks
are idiosyncratic, affecting only a single institution or asset, whilst others are systematic,
impacting on the entire financial system and the real economy.
An example of an idiosyncratic shock is the failure of a single small or medium-sized
bank, perhaps due to internal fraud or mismanagement. Such failure is generally not a big
public concern. Banks fail all the time and the authorities have robust resolution mecha-
nisms in place for unwinding failed banks, ensuring that banking services are uninter-
rupted and contagion averted. The situation is different in the special case of ‘too big to
fail’ (TBTF) institutions. However, in some circumstances an idiosyncratic shock can lead
to systemic risk, usually because of in-built vulnerabilities amplifying a relatively small
event into a full-blown systemic crisis.
4
We need to make a distinction between the term systemic and systematic. Systematic
risk relates to non-diversifiable risk factors that affect everybody and is always present,
perhaps the stock market, whilst systemic risk pertains to the danger of the entire finan-
cial system collapsing.
The consequences of a systemic crisis in the financial sector are more devastating com-
pared to other economic sectors. Suppose a firm making chocolate goes bust: the share-
holders and employees suffer, but competing chocolate firms will benefit. There is no
significant damage to the economy, and such developments may even be positive, as
argued by Joseph Schumpeter (1942) with his notion of creative destruction. By contrast,
if a bank collapses, it may well lead to the seizure of the entire financial system, crippling
the real economy.
1.2.1 Notions of systemic risk

Systemic risk is a term used frequently, both in popular media and by specialists.
Unfortunately, most usage is imprecise and contradictory. Often one gets the impres-
sion that commentators are only talking about the last crisis event when they define
systemic risk.
The global authorities, in the form of the International Monetary Fund (IMF), the Bank
for International Settlements (BIS) and the Financial Stability Board (FSB) (2009), define
a systemic event as:
‘the disruption to the flow of financial services that is (i) caused by an impairment
of all or parts of the financial system; and (ii) has the potential to have serious nega-
tive consequences for the real economy.’
A more comprehensive definition is:
Definition 1.1 Systemic risk The risk that the entire financial system may fail,
causing a general economic collapse, as opposed to risk associated with an individual part
of the system.
Systemic risk arises from the interlinkages present in the financial system, where the failure
of an individual institution may cause spillovers and even cascading failures, amplified by
the inherent pro-cyclicality of banking and regulations.
The conditions for systemic risk tend to be created when all outward signs point to
stability and low risk.
There are few or no recorded instances of a systemic event according to this strict defi-
nition. We have seen events that got close in their severity, and other events that, if left to
fruition, might have ended up as a systemic event.
It is important to recognise that there is no single generally accepted definition of sys-
temic risk. Some commentators have a much looser definition than the one above, often
considering a severe crisis, and even a routine crisis, to be a systemic event. However, the
most common usage is similar to Definition 1.1.
5
1.2.2 Who is vulnerable?

A systemic event can be global in reach, or just affect a single country. Some countries
are more vulnerable to systemic risk than others, especially those that have based their
economies on finance and are exporters of financial services.
One way to identify who is susceptible is to consider the size of the banking system.
Within the European Union (EU), the European Central Bank (ECB) publishes statistics on
the size of the banking system, split into domestic and foreign parts. Figure 1.2 shows the
relative size of the banking system in EU member states, measured using the ratio of total
banking assets to GDP, both before the crisis in 2007 and also in June 2011.
The country with the smallest banking system, Romania, comes in at 64% whilst the
largest banking counties are Luxembourg, Malta, Ireland and Cyprus. The latter two have
been badly hit by the European sovereign debt crisis. This does not necessarily mean
that these large banking countries are more vulnerable to systemic risk than the rest. It
matters how much of the banking system is domestic owned and how much is foreign
owned. The countries with the smallest and largest banking systems have predominantly
Romania All
Poland Domestic
Bulgaria
Slovakia Upper column 2007
Lithuania
Lower column June 2011
Czech Republic
Hungary
Slovenia
Latvia
Finland
Italy
Greece
Estonia
Portugal
Germany
Spain
Sweden
France
Denmark
Austria
United Kingdom
Belgium
Netherlands
Cyprus
Ireland
Malta
Luxembourg
0% 500% 1000% 1500% 2000%
Figure 1.2 EU banking system, total assets/GDP ratio in 2007 and June 2011, ordered
by size of banking system in 2007
Data source: ECB and Eurostat
6
foreign banking, albeit for different reasons. For the small banking countries regular banks
are owned by foreigners, whilst for the large banking countries the foreign-owned banks
mostly service foreign clients. These countries export banking services, produced by
foreign-owned banks.
This reduces vulnerability. For example, even if the assets of the Luxembourg banks
exceed its GDP 20 times, Luxembourg is relatively insulated because the failure of a bank
would not directly involve taxpayers’ money in most cases. Of course, if a number of the
banks failed, the Luxembourg government might find itself short of funds from taxes, so it
is directly connected to the banks. This threat is, however, less direct than if Luxembourg
would be called on to bail out the banks. This suggests that in a country like Ireland or
Cyprus, even if their banks are relatively smaller than the banks in Luxembourg, bank fail-
ures pose a much bigger danger.
The large banking countries are also at risk of Dutch disease,1 where the banks become
the dominating part of the economy, crowding out other economic activity.
The United Kingdom

Figure 1.2 shows that the United Kingdom (UK) had the seventh largest banking system
in the EU in 2007, and the largest amongst the big economies. This means that banking is
more important for the UK than for most other countries. The importance of banking can
be seen in Figure 1.3, which shows the UK output index. Essentially, all economic growth
has come from finance, highlighting the growing importance of the City of London.
This suggests that countries with large financial systems need to pay special attention
to financial sector policies, such as financial regulations. The UK aims to be at the fore-
front of developing policies towards the financial system, developing the doctrine of the
‘hands-off approach’ before 2007. This was widely admired at the time. The UK now dem-
onstrates its frontier thinking by activities such as the Independent Commission on Banking
and the activities of the government’s Foresight group.
200 Finance
Whole economy
150 Manufacturing
100
1995 1997 1999 2001 2003 2005 2007 2009
Figure 1.3 UK output index, 1995 Q1 = 100

Data source: Thomson Datastream
1
Dutch disease arises when a new large profitable economic sector, traditionally exploiting natural resources,
causes a decline in the manufacturing sector. This happens because the new revenue from natural resources
strengthens the currency, making the manufacturing sector less competitive. It is used here in a more
general sense where one sector crowds out other sectors.
7
Banks, bank size and politics

The contribution of the financial sector to systemic risk also depends on the structure
of the financial sector. For example, suppose two countries have banking systems of the
same magnitude. The first country has one bank while the second has 10 equally sized
banks. In this case, the first country is much more vulnerable.
This happens for two main reasons. First, the country with the single bank is likely to
suffer significantly if its single bank fails, whilst the country with 10 banks would find
it easier to cope with the failure of several, but not all, of its banks. Second, the single
large bank is likely to have more political power than the 10 smaller banks combined.
The reason is that the single large bank is more important and therefore carries more
political weight, whilst the 10 banks would be unlikely to speak with a single voice and
maintain a uniform lobbying agenda, implying that the single big bank has more political
power.
The stronger political power would probably mean that the bank creates more sys-
temic risk because it could use its lobbying power to push back on regulations and could
appeal to the politicians if the supervisor is giving it difficulty.
1.2.3 Should we eliminate systemic risk?

This leaves the question of whether it is a desirable policy objective to prevent financial
crises altogether and reduce systemic risk to zero. While it is quite straightforward to do
so, the cost may not be acceptable.
The answer depends on the definition of systemic risk. While it may not be desirable to
prevent systemic risk altogether in its loosest definition, we are willing to do what it takes
to prevent a general economic collapse according to the strict definition.
Systemic risk is created by risk taking and the complex interactions within the financial
system, so all that is needed to prevent systemic crises is to drastically reduce the size of
the financial system. Countries with small and unsophisticated financial systems are not
very vulnerable to systemic risk. They may suffer, of course, if the outside world enters
into a systemic crisis. The question is whether a modern society can exist and grow with-
out a sophisticated risk-taking financial sector. The answer is no.
At the same time, we cannot simply ignore systemic risk; the consequences of a sys-
temic crisis are so severe that we would do almost anything to prevent it. Indeed, as we
see later in the book, governments have on occasion proclaimed they would not inter-
vene in the markets to contain systemic risk, but have then reneged once a crisis episode
is under way. Ignoring systemic risk as a matter of policy is not credible.
The correct balance lies between those two extremes. We need to encourage enough
risk taking by the financial industry so that economic growth is not hampered, whilst at
the same time have mechanisms in place that prevent risk taking from causing systemic
crises. This is a classical risk–return trade-off but is not a trivial undertaking since regula-
tions change the behaviour of economic agents and can by themselves lead to undesir-
able outcomes, like systemic risk. This means that one cannot look at individual policies
in isolation; policymakers instead need to consider the impact of government regulations
of the financial system in their entirety.
8
1.3 Who creates systemic risk?
The current financial crisis gives us a good example of the challenge facing the authori-
ties. The failure of individual institutions like Bear Stearns, Lehman Brothers and AIG
caused serious disruptions. Deciding on whether these institutions should have been
bailed out is a clear example of the risk–return trade-off we have discussed here.
That task is made easier by the observation that most crises are fundamentally similar,
even if the details differ. For example, the 1914 crisis and the ongoing crisis have strong
commonalities — financial institutions building up risk in good times, as if a crisis could
never happen, only to see confidence and liquidity evaporate suddenly, causing cascad-
ing failures. The authorities reacted similarly by providing liquidity injections. This sug-
gests that policymakers should be well prepared and know their history. Getting caught
short in 2007 did not reflect well on many policymaking institutions.
1.3 Who creates systemic risk?

While episodes of (near) systemic events seem to happen quite suddenly, in reality they
take a long time to build up, as noted by the former head of the BIS, Andrew Crockett,
in 2000:
‘The received wisdom is that risk increases in recessions and falls in booms. In con-
trast, it may be more helpful to think of risk as increasing during upswings, as finan-
cial imbalances build up, and materialising in recessions.’
In other words, when looking for the origins of systemic risk one should not focus on
current events; decisions made long ago are usually to blame.
This leaves open the question of who is responsible for a crisis episode. The public
media and affiliated pundits delight in blaming the target of the day, whether the greedy
bankers, incompetent politicians, pernicious academics or some other easy scapegoat.
The reality is more nuanced. It is the interplay between the various parties that creates
conditions for systemic risk, with no key players blameless. It is equally difficult to identify
what is the real problem when confronted by the news of the day.
1.3.1 Role of financial institutions

Financial institutions are in active competition with each other; they are ranked by size,
number of clients, profits, etc. Because of the relationship between risk and return, per-
haps coupled with a short-term outlook, profit-maximising behaviour can cause financial
institutions to take on considerable risk. We do not even need competition for this to hap-
pen; greed is sufficient.
Over time, risk taking of that nature can become destabilising, creating systemic risk.
One problem is that during boom times the risk is often hidden, so financial institutions
experience large profits at what seems like low risk. However, the risk builds up and can
materialise quite suddenly. This is an example of the theories of Minsky (1992) who
argued that economies have either stable or unstable financial regimes. Even if the econ-
omy starts out stable, continued prosperity paves the way for an unstable system. The
essence of Minsky’s financial instability hypothesis is that stability is destabilising because
9
financial institutions have a tendency to extrapolate stability into infinity, investing in ever
more risky debt structures that ultimately undermine stability. At some point, a disrup-
tive event occurs, and markets go through an abrupt correction — the further along in the
cycle, the more violent the repricing.
A clear example of such mechanisms occurred prior to 2007, when the financial indus-
try, and almost everybody else, were blind to the hidden risk being created. The pre-crisis
period was even labelled the ‘great moderation’.
This also demonstrates the double-edged nature of otherwise useful risk control meth-
ods such as mark-to-market accounting, which enable financial institutions to realise prof-
its up-front. This makes the banks even more profitable when things are good, but at the
expense of larger losses when the markets take a turn for the worse — a clear example of
pro-cyclicality.
1.3.2 Role of the government

Financial institutions do not operate in a vacuum. Their behaviour is shaped by govern-
ment policies and they also influence government decisions. Government policies can
create conditions for low or high systemic risk, because after all, financial institutions
react to the rules of the game. A direct example of this is moral hazard: if the governments
bail out banks, the banks will take more risk.
Governments can also be direct sources of systemic risk, where the most extreme
examples are wars. The main creator of systemic risk in the European sovereign debt crisis
is European governments, with the financial system in a supporting secondary role. Of
course, it didn’t help that the banking system was already in a vulnerable state.
The impacts of government policies on systemic risk are often indirect and counter-
intuitive. It may be desirable to implement policies to contain a particular type of ‘high
risk’ behaviour by the industry; unfortunately, any policy carries with it unforeseen
adverse consequences. For example, regulations preventing risk taking in a highly visible
part of the financial system can simply shift risk-taking activities into more opaque corners
of the system where the banks can continue as before, undetected, perhaps in shadow
banking. That outcome increases systemic risk. Competition makes this near-certain.
Consequently, it is necessary to consider the systemic impact of the entirety of govern-
ment banking policies together.
It is useful to draw an analogy to a different sector of the economy, using an exam-
ple provided by Kaufman (1996) who in discussing Merton (1995) points out that
governments often provide flood insurance and information about water levels. The
reason why the government provides flood insurance is that the private sector, know-
ing the risks, refuses to do so at a reasonable price. This means that homeowners have
an incentive to build on floodplains in the knowledge that they would be bailed out
by government flood insurance when the eventual flood comes. Flood insurance cre-
ates moral hazard that makes the eventual costs much higher than they otherwise
would be.
As a consequence, some commentators have made robust remarks on the origins of
systemic risk. Fisher Black (1995), of Black–Scholes fame, states:
10
1.4 Fundamental origins of systemic risk
‘When you hear the government talking about systemic risk, hold on to your wallet!
It means that they want you to pay more taxes for more regulations, which are likely
to create systemic risk by interfering with private contracting . . . In sum, when you
think about systemic risks, you’ll be close to the truth if you think of the govern-
ment as causing them rather than protecting us from them.’

Systemic risk arises because of inherent structural weaknesses in the financial system, for
example pro-cyclicality, information asymmetries, interdependence and perverse incentives.
These factors enable systemic risk to be created without much scrutiny, only to be realised
when it is too late. Therefore, it is necessary to address the underlying causal factors in
order to develop policies to mitigate systemic risk.
1.4.1 Fractional reserve banking and systemic risk

One avenue for systemic risk is the inherent vulnerabilities in the financial system because
of fractional reserve banking and the nature of money.
It may seem surprising that such a fundamental economic concept as money does not
have a clear and unambiguous real-world definition, but it does not. Anything that can
be freely used to make purchases falls within most definitions of money, but what about
highly liquid assets that can easily be converted into money, or even savings accounts
that will become money at some stage in the future? Economic agents also consider such
assets to be money.
Types of money
It is useful to classify money into several different categories, or monetary aggregates as
they are technically known, as illustrated by Figure 1.4.
M0, also known as monetary base, or high-powered money, consists of paper money in
circulation and reserves at the central bank. M1 is known as narrow money, and consists
10
M3−M2
8
USD trillions
M2−M1
6 M1−M0
4 M0
2
0
M0 M1 M2 M3
Figure 1.4 United States (US) monetary aggregates at end of 2005. Identifies unique
components of each aggregate, and labels the main component
Data source: Federal Reserve Board. Data ends 2005 since that is the last full year when M3 was published
11
of M0 plus current (checkable) accounts, in essence funds that are readily accessible for
spending — liquid funds. M2 is M1 plus savings accounts and represents money and close
substitutes for money. It is a key indicator for inflation forecasting. M3 is the broadest
measure of money, including M2 plus large time deposits, institutional money market
funds, short-term repurchase and other larger liquid assets. The higher forms of money
are created out of the lower forms, typically by means of fractional reserve banking.
Reserve requirements
Today, most banking systems use fractional reserve banking which is an arrangement
whereby most of the money is created by the banks; subject to a particular limitation —
the reserve requirement, d — banks are required to hold a certain fraction of deposits on
account with the central bank. Fractional reserve banking expands money supply beyond
what it would otherwise be, as seen by the following example:
Example 1.1 Fractional reserve banking
Person X deposits $100 (M0) into bank A. d is 10% so the bank lends $90 to per-
son Y who deposits $90 at bank B which keeps fraction d and lends out $81 and so
on. Hence, in the limit M1 = 100 + 90 + 81 + c = 1/d = 1000. The relationship
between M1 and the monetary base can be expressed as:
1
M1 = g * M0 = * M0
d
g is the money multiplier, which tells us how much the money supply changes for a given
change in the monetary base. If the reserve requirement is 10%, every dollar in the form
of deposits uses up only 10 cents of high-powered money, or each dollar of high-powered
money held as reserves can support $10 of deposits. Hence, the higher the required reserve
ratio, the lower the money multiplier. We see in Figure 1.5 how fractional reserve banking
aids the expansion of credit.
Fragilities
The fractional reserve banking system is inherently fragile and hence is a cause of systemic
risk. The reason is that when depositors put money into a bank, creating M1 from M0, the
bank then turns around and lends most of it out, keeping a small fraction as reserves. The
fragility arises because deposits generally are of short maturities, and some can be with-
drawn whenever the depositor wants — demand deposits — whilst the loans tend to be of
longer maturities. If a sufficiently large number of depositors want their money, the bank
will run out of cash, because it cannot similarly call on its own borrowers to repay their
loans. We call such an event a bank run. Bank runs are contagious and can spread quickly
throughout the financial system. The reason is that the banking system is built on trust,
so if depositors lose confidence in banks, they flock to withdraw their deposits as cash.
12
$750 10% 30% 50%
Money supply
$500
$250
2 4 6 8 10 12 14
Number of iterations
Figure 1.5 Credit expansion of $100 under fractional reserve banking and various
reserve ratios
1.4.2 Pro-cyclicality
A process that is positively correlated with the economic cycle is described as pro-cyclical.
Bank capital and leverage are two examples of pro-cyclical processes in which risks build
up during stable periods. Banks tend to have surplus capital when the economy is boom-
ing, whilst capital levels drop during recessions. Likewise, economic agents have a ten-
dency to borrow too much during good times, and borrow too little in downturns.
Pro-cyclicality is often created by the various amplification mechanisms built into the
financial system, and is encouraged by risk-weighted capital, mark-to-market accounting
and the fact that the strength of financial regulations tends to erode in boom times and
come back with a vengeance during crises.
Amplification mechanisms
Financial crises often seem to be triggered by relatively small events. This is not dissimilar
to what chaos theorists mean when they say that a butterfly in Hong Kong causes a hur-
ricane in the Caribbean. Many such butterfly effects exist in the financial system.
The main enabling factor is leverage, whereby a bank borrows money to make investments.
High degrees of leverage enable a bank to multiply profits when investments are success-
ful. For this reason, banks remain highly leveraged. Unfortunately, rapid de-leveraging
amplifies losses, perhaps resulting in bank failures and causing firesale externalities.
Firesale externalities
Definition 1.2 Externality Externality occurs where cost or benefit accrues to

someone who was not involved in the decision-making process that led to the cost or benefit.
Firesale externalities (see, e.g., Kashyap et al., 2008) arise when financial institutions need
liquidity and aim to convert risky assets into cash. At that time, there are many sellers and
few buyers of risky assets. That means prices collapse, making it even harder to raise cash
and forcing institutions to sell even more risky assets, in what has become a vicious cycle
(Figure 1.6). It is the individual self-preservation behaviour of each institution that causes
negative externalities for the rest of the financial system. Such behaviour can cause a rela-
tively innocuous shock to become a full-blown crisis.
13
External
shock
Financial dif ficulties
Prices Risk
Firesale
fall increases
Forced
sales
Figure 1.6 Firesales and acceleration effects
8%
4%
Asset growth
Asset growth
4%
2%
0%
0%
−4%
−1.0% 0.0% 1.0% −25% 0% 25%
Leverage growth Leverage growth
(a) Households (b) Commercial banks
Asset growth
6%
Asset growth
20%
4%
0%
2%
0% −20%
−1.5% 0.0% 1.5% −25% 0% 25%

Leverage growth Leverage growth
(c) Corporates (d) Broker–dealers
Figure 1.7 Asset growth and leverage growth, with a regression line
Data source: Adriam and Shin (2010)
Asset growth and leverage growth

Adrian and Shin (2010) document how the various categories of economic agents con-
tribute to systemic risk, focusing on the relationship between asset growth and lever-
age growth. We use their data in Figure 1.7 which shows the relationship between asset
growth and leverage growth for four categories of agents: households, commercial banks,
corporates and broker–dealers. The last is a term used in the US for an institution trading
securities for its own account or on behalf of its customers.
14
For households, the relationship is negative: the richer we become the less leverage we use.
If a household owns a house financed by a mortgage, leverage falls when the house price
increases, since the equity of the household is increasing at a much faster rate than assets.
Similarly, most people pay back their mortgages over time, reducing leverage. For corporations
and commercial banks there is little correlation, but for the broker–dealers it is highly positive.
For such firms, leverage is pro-cyclical, increasing when balance sheets are expanding
and decreasing when balance sheets are shrinking. The slope is close to 1 for broker–dealers,
suggesting that equity is increasing at a constant rate on average.
This result is counter-intuitive in light of standard theories in corporate finance, where
it should not make a difference how a firm is financed. Here, the willingness to use bor-
rowed funds to increase leverage is a key factor in firm growth.
From the point of view of systemic risk, this suggests that broker–dealer type institu-
tions have a tendency to continually increasing leverage. The owners of these institutions
might be rewarded by higher profits when things go well, but at the expense of increased
systemic risk.
1.4.3 Information asymmetry

Banks rely on the confidence of their depositors and counterparties to operate. Depositors
trust the bank to guard their money and counterparties need to be reassured that the bank
will honour its obligations. A loss of confidence can lead to bank failure. It does not matter
much whether the loss in confidence arises from unfounded rumours or from real negative
information. This leads to what is known as an agency problem between counterparties and
the bank, caused by information asymmetry — the bank knows more than the counterparties.
A confidence crisis at one bank can quickly spread if other institutions are perceived
to share the same weakness. A loss in confidence may result in a bank run — depositors
queuing up to withdraw their money. Depositors at other banks, observing problems at
the first bank and lacking information about the soundness of their banks, may decide to
pull out their deposits too, for fear of losing their deposits. If left unchecked, bank runs
may swiftly spread to the entire system, causing significant economic damage.
Similarly, counterparties may refuse to enter into new transactions or renew existing ones
if they suspect a bank is in trouble. For example, if bank A is allegedly holding toxic assets,
then other banks may stop trading with it, and if bank B is thought to be active in that asset
class then other banks may also stop trading with bank B. This can lead to markets collapsing.
In both these examples, the loss of confidence adversely affects liquidity. Financial
institutions often operate under the general assumption that liquidity is infinite, and the
evaporation of liquidity can cause acute distress for the financial system and, hence, is
often the first sign of a pending crisis.
1.4.4 Interdependence
The financial system consists of a network of interwoven obligations that during normal
times significantly increases the efficiency of financial markets. This means a financial
institution can have an indirect exposure to another financial institution without any
15
Bank B
xposed via B and C

D is e
Bank A Bank D
Bank C
Figure 1.8 D is indirectly exposed to A via B and C
direct dealings with it. This can be quite dangerous during crises, most obviously because
of interlocking exposures, which create the potential for one institution’s failure to have
‘knock-on’ effects on the financial health of other institutions. At any given time, a finan-
cial institution simultaneously owes money to other institutions and is owed money often
from the very same institutions. These linkages make the financial system fragile.
For example, consider the situation depicted in Figure 1.8, where we have four banks
A, B, C and D. A borrows and lends from B and C, which also borrow and lend from D.
Suppose A suffers a negative shock, and needing funds calls in a loan to B and C. They
now need funds, and in turn may call in their loans to the hitherto healthy bank D. The
difficulties facing A have now been transferred to D, even if A and D have no direct busi-
ness relationships. This is exactly the situation in the 1914 example discussed above.
Because banks have only a limited amount of liquid funds, even a relatively small but
immediate demand for cash, or an interruption in the flow of funds, can cause serious dif-
ficulties and even failure. This means that if financial institutions suspect others may be in dif-
ficulty, their natural instinct is to withdraw, spreading a crisis throughout the financial system.
We see a particularly damaging example of this in the interbank market during the crisis
from 2007. The proliferation of derivatives, in particular structured credit products such as
structured investment vehicles (SIVs), credit default swaps (CDSs) and collateralised debt
obligations (CDOs), has been increasing the interdependence between financial institu-
tions, in turn increasing the fragility of the financial system.
1.4.5 Perverse incentives

Because of the interconnectedness of the financial system and the very high cost of sys-
temic crises, the government will have no choice but to do anything it can to prevent such
an outcome. This often takes the form of bailouts of various types, creating moral hazard.
A small, prudently run and non-systematically important institution is less likely to get
support from the government than a very large, badly run, interconnected bank.
16
1.5 Summary
This can have the unfortunate outcome that a badly run bank actually has a lower
cost of funding than a well-managed bank because only one of them would be bailed
out. Ultimately, this means banks have incentives to become as big, interconnected and
dangerous as possible in order to maximise the chance of a bailout. A particularly inter-
esting example can be seen by the supposedly expressed desires by some hedge funds to
become ‘banks’ in order to enjoy low-cost government guarantees of funding.
There are many other types of perverse incentives. For example, lenders who ultimately
intend to securitise their loan books do not have proper incentives to monitor the quality
of their loans. In addition, the presence of financial instruments, such as CDSs, may create
incentives for some market participants to increase instability.
1.5 Summary
Systemic risk has always been a part of the financial system, and many crisis events of the
past are quite similar to modern crises, for example, that of 1914.
There are many definitions of systemic risk, but most emphasise the risk of widespread
failures in the financial system, caused by the interlinkages between financial institutions,
eventually resulting in a severe economic downturn.
Systemic risk is the inevitable result of having a market-based economy and is not eas-
ily eliminated or reduced significantly without unduly restricting risky activities, adversely
affecting the real economy. This means that the authorities need to balance the various
pros and cons in their approach to systemic risk policies, doing cost–benefit analysis.
Stated differently, the authorities have to find the appropriate risk–return combination,
similar to what investors do for their own portfolios.
Although any sector of the economy may be subject to systemic risk, it is especially relevant
for the financial sector, because it is uniquely dependent on the interplay between confidence
and network effects. The failure of a single institution quickly spreads to other banks, even if
they have been prudently run. The damage caused by even relatively small events in the mar-
kets can be amplified into systemic proportions because of the inherent pro-cyclicality in the
financial system, perhaps aided by the perverse incentives of market participants.
Questions for discussion

1 What is systemic risk?
2 How frequently can we expect systemic crises?
3 What does it mean when we say that systemic risk builds up out of sight, and what
could that mean for the government’s formulation of financial stability policies?
4 How does the European country with the largest banking system relative to GDP man-
age to be mostly unaffected by the banking crisis whilst other countries with relatively
smaller banking system suffer serious banking crises?
5 Do you think it would be prudent for the British government to reduce the size of its
banks in order to reduce systemic risk?
17
6 Can you think of some systemic events in the past, anywhere in the world, excluding
what has been happening since 2007?
7 What is pro-cyclicality, how is it amplified, and how could it be mitigated?
8 Explain the concept of firesale externality.
9 Do you think Europe is at the risk of a systemic event at the moment?
10 What are the main origins of systemic risk?
References
Adrian, T. and Shin, H. S. (2010). Liquidity and leverage. J. Finan. Intermed., 19: 418–37.
Bandt, D. and Hartmann, P. (2000). Systemic risk: a survey. Working paper no. 35, European
Central Bank.
Black, F. (1995). Hedging, speculation, and systemic risk. J. Derivatives, 2: 6–8.
Brown, W. A. (1940). The international gold standard reinterpreted, 1914–1934. Technical
report, National Bureau of Economic Research.
Crockett, A. (2000). Marrying the micro- and macro-prudential dimensions of financial stabil-
ity. The General Manager of the Bank for International Settlements, www.bis.org/review/
rr000921b.pdf.
Ferguson, N. (2008). The Ascent of Money. The Penguin Group.
International Monetary Fund, Bank for International Settlements, Financial Stability Board
(2009). Report to G20 finance ministers and governors. Guidance to assess the systemic
importance of financial institutions, markets and instruments: initial considerations.
Technical report.
Kashyap, A. K., Rajan, R. G., and Stein, J. C. (2008). Rethinking capital regulation. http://online.
wsj.com/public/resources/documents/ Fed-JacksonHole.pdf.
Kaufman, G. G. (1996). Bank failures, systemic risk, and bank regulation. Cato J., 16(1): 17–46.
Kindleberger, C. P. (1996). Manias, Panics, and Crashes: a History of Financial Crises, 3rd edition.
John Wiley & Sons.
Merton, R. (1995). A functional perspective of financial intermediation. Financial Management,
24: 23–41.
Minsky, H. (1992). The financial instability hypothesis. Working paper. Mimeo, Yale University.
Reinhart, C. M. and Rogoff, K. (2009). This Time Is Different: Eight Centuries of Financial Folly.
Princeton University Press.
Schumpeter, J. (1942). Capitalism, Socialism and Democracy. Harper, New York.
18
2 The Great Depression, 1929–1933
The Great Depression, 1929–1933, was the largest worldwide economic catastrophe
the world has ever seen. A Wall Street crash in 1929 was followed by the collapse of
financial institutions and an implosion of activity on financial markets, soon spilling
over to Main Street. This happened because of vicious feedback loops between con-
tracting economic activity, financial crises and government mistakes.
The Depression shaped economic policy long after it ended. While its impact had
been diminishing, that all changed with the crises from 2007, and over the past few
years, the lessons from the Depression have significantly impacted on policy. The
reason why events in the fall of 2008 did not lead to another depression is because
policymakers had learned the lesson of the Great Depression and acted correctly. The
continuing influence of the Great Depression on government policy make it a worth-
while subject in studies of financial stability in the twenty-first century.
The Depression came after one of the longest expansion periods in history, the
roaring twenties, with rapid economic growth and rampant stock market speculation,
when many investors were highly leveraged, buying stocks on the margin. Underneath
were significant frictions. The First World War (WWI) fundamentally affected the
international order, and universal suffrage – for both women and the poor – altered
political power structures as did the emergence of labour unions. Extreme politi-
cal ideologies, communism and fascism, gained a significant foothold, not least as a
result of the economic turmoil marking those years.
International trade patterns were altered considerably, as non-combatants prof-
ited and extended their global reach. Many countries attempted export-led growth,
typically via agricultural products, leading to overproduction and price deflation. And
Chapter 2 The Great Depression, 1929–1933
finally, disputes over war debt, and the determination in some countries to extract
reparations from the losing Central Powers, poisoned international relations and acted
to prevent effective crisis resolution once the Depression was underway.
The first sign of the problems leading to the Great Depression emerged in the agri-
cultural sector and industries dependent on agriculture such as banks and insurance
companies, spreading economic turmoil from the farm to the city. This was followed
by widespread bank collapses, contraction in money supplies, with surpluses, liquid-
ity crises, exchange rate depreciations and trade restrictions, all acting in a vicious
feedback loop, compounding the problems.
Note that it can be difficult to assess the costs of the Great Depression because only
limited data exists and it is often contradictory. This especially applies to national
accounts, as global standards for the calculation of GDP were only determined in the
late 1940s, so comparisons before that time are difficult.

Many of the concepts discussed in this chapter are addressed in much greater detail
elsewhere in the book, for example, in Chapter 4 (liquidity), Chapter 5 (the cen-
tral bank), Chapter 14 (bailouts), Chapter 8 (bank runs and deposit insurance) and
Chapter 17 (the ongoing crisis: 2007–2009 phase).
Key concepts
■ Deflation
■ Trade policy
■ Monetary policy
■ Bank runs
■ International coordination

The seminal work on the Great Depression is Kindleberger (1986) and much of the
material is drawn from him, as well as Eichengreen (1996), Friedman and Schwartz
(1963) and Bernanke (1995). More recent books and articles, for example Ferguson
(2008), Ahamed (2009) and Eichengreen and Irwin (2009), discuss the various
aspects of the Depression in a more modern context.
2.1 Build-up to a depression

Many factors contributed to the Great Depression and the underlying causes remain con-
troversial to this day. The single most important causal event is WWI and the significant
social and economic upheaval caused by the war. The countries that participated in the
war were in a much weaker position than before; some even ceased to exist, with new
countries formed. Those that stayed away from the war profited from selling to the com-
batants. This meant that after the war ended, the relative position of countries altered,
20
for example, the United Kingdom (UK) was no longer the world’s superpower, yielding
to the United States (US).
Following the war, what would now be called significant global imbalances built up,
with resentment and conflict resulting. Many countries accumulated large debts, with the
largest creditor being the US.
GDP rankings in 1929

It is interesting to note the evolution of the various economies from the start of the Great
Depression in 1929 until today. Figure 2.1 shows the GDP per capita of a selection of coun-
tries in 1929. The US was richest, followed by Switzerland, with the UK in fourth place.
Argentina was in the 11th place, wealthier than Germany in 13th place, with Norway
poorer than Venezuela, and Japan below average in 27th place.
Paying for the war

Perhaps the worst single source of tension was the question of war reparations. The coun-
tries on the losing side of WWI, the Central Powers (Germany, the Austro-Hungarian Empire,
Turkey and Bulgaria), were blamed for causing the war, whilst the countries suffering most
on the winning side (the Allies), especially Belgium and France, demanded compensation.
The question of war reparations remained a continuing source of friction in the interna-
tional agenda throughout the 1920s. The US refused to accept reparations from Germany
and only wanted to be repaid for war loans extended to the Allies. The UK was in favour
of cancelling war debts but, given the US position, had no choice but to collect debts, at
least up to the limit of the British debt to the US. France and Belgium, however, wanted
to collect reparations from Germany and suggested that Germany borrow from the US in
order to make reparations to them, so they could pay off their obligations to the US.
Japan 27
Norway 18
Venezuela 17
Uruguay 14
Germany 13
Sweden 12
Argentina 11
France 10
Denmark 7
Australia 5
United Kingdom 4
Netherlands 3
Switzerland 2
United States 1
0 1000 2000 3000 4000 5000 6000 7000
Figure 2.1 1929 GDP per capita in constant 1990 dollars and global rank out of
52 countries
Data source: www.ggdc.net/MADDISON/oriindex.htm
21
The policy of extracting very large reparations from the losers was quite popular right
after the war, but opinion soon changed as it was recognised that the amounts demanded
were simply too large. Some voices, most importantly Keynes (1920), warned against the
consequences of treating the losers that way. In order to pay reparations, Germany began
a credit-fuelled period of growth, ending in hyperinflation in 1923.
2.1.1 Bad monetary policy in the UK

The UK made one of its worst policy mistakes ever in 1925 when the Chancellor of the
Exchequer, Winston Churchill, decided to put the UK back onto the gold standard at
prewar parity. Churchill later admitted that going back on gold was one of his biggest mis-
takes. He was advised by the Bank of England (BoE) that it was the right decision, needed
to restore the credibility of the UK. Keynes argued strongly against the decision, maintain-
ing correctly that it would lead to deflation and general economic misery, having said the
year before that ‘In truth, the gold standard is already a barbarous relic’ (Keynes, 1924).
As the UK had experienced significant inflation during the war, going back on gold at
prewar parity meant that sterling was now significantly overvalued. Therefore, it had no
choice but to implement what is now called an internal devaluation, lowering various
factor costs. This led to significant economic upheaval, such as long-running strikes and
the haemorrhaging of gold, not least to France. The UK remained on gold, and with an
overvalued currency, for the next six years, in recession for the duration.
2.1.2 The United States

The war strengthened the competitive positions of American manufacturers and opened new
markets for its exporters. As a consequence, the US enjoyed a trade surplus, building up signif-
icant capital reserves. After the war, the US maintained a relatively accommodating monetary
policy with low interest rates that encouraged American gold to flow abroad. The monetary
policy also acted as to stimulate the capital markets, in particular the Wall Street bubble.
The main danger arising from the stock market boom was not the immediate impact on
prices or volume, but rather the interconnectedness with global credit markets. Significant
amounts of money flowed into New York from around the globe, causing other countries
to raise interest rates to prevent the loss of gold reserves to New York. This meant that
money was diverted from productive investment to stock market speculation, both in the
US and abroad, adversely affecting economic development and monetary policy.
The US government was increasingly concerned with the amount of resources being
diverted into the equity markets. Federal Reserve System (Fed) officials concluded that
Wall Street speculation was diverting funds from more productive uses, and began to
tighten monetary policy. In turn, that adversely affected debtor nations, which were forced
to adopt increasingly stringent monetary fiscal policies to maintain their exchange rate.
The New York Federal Reserve Bank (NYFed) raised its discount rate to 6% on 9 August
1929 in order to slow down the market, but to little effect. In September, the stock
exchange added more seats, with the price of a seat at an all-time high. The higher inter-
est rates in the US set off a round of interest rate increases in Europe.
22
These problems are not dissimilar to the problems currently affecting the euro zone
and the European exchange rate mechanism (ERM) before that. A monetary policy appro-
priate for a leading country in a boom is likely to be too strict for other countries facing
recessions. This leaves the weaker countries with only two options, either to implement
an internal devaluation or to devalue their currency, with neither choice palatable.
2.1.3 Agricultural depression

Up until the middle of the nineteenth century, the harvest was the main measure of busi-
ness conditions. A bumper crop lowered the price of bread and increased industrial out-
put and farm income. Crop failures led to depression. At some point around the 1860s,
business cycles in the industrialised countries became independent of agriculture but
it remained a big part of the economy for countries outside western Europe. Farming
accounted for a quarter of total employment in the US in 1929.
Major agricultural countries suffered significantly from overproduction in the 1920s,
causing a fall in prices and income, as can be seen in Figure 2.2. There was an initial boom
in prices right after the war, but from about 1926 prices were steadily falling, bottoming
in the main crisis year of 1933.
The reaction of many governments to overproduction and falling prices was to protect
their producers, subsidising exporters and accumulating stockpiles of agricultural prod-
ucts. In many countries, for example, the US and Canada, attempts were made to support
the price of wheat (which fell by 50% between December 1929 and 1930). Many other
countries had neither the necessary financial capacity nor the required storage facilities,
feeling obliged to export wherever they could, forcing prices downwards.
2.1.4 Competitive devaluations and deflation

In order to help their exporters, some countries resorted to currency devaluations.
From the point of view of an individual country this may seem a sensible policy. It
should help exporters and reduce imports, strengthening the economy. However, such
‘beggar-thy-neighbour’ policies encourage others to follow, leading to waves of com-
petitive devaluations, making everybody worse off. The sentiments are captured by the
comments of NYFed governor Benjamin Strong to the House Committee on Banking and
200
Wheat
Depression
150 Cotton
Great
100 Sugar
Rubber
50
0
1920 1925 1930 1935
Figure 2.2 Commodity prices
23
Currency in 1926: ‘It should not be overlooked that one of the greatest menaces to the
trade of this country is depreciating foreign exchange.’
Effectiveness
Surprisingly, the devaluations were not all that effective in helping the producers. When
a country devalues, two things happen: the local currency price of exports increases and
the foreign currency price falls. Elasticities determine which effect ends up dominating.
For a small country with no impact on world prices, local currency prices will increase.
For a major exporter facing an inelastic demand, prices abroad will fall. For example,
when Argentina devalued in the 1930s, local currency prices did not increase but foreign
prices fell, exacerbating that country’s difficulties. The experience of other major agricul-
tural exporters, like Australia, was similar.
Deflation
The falling agricultural prices hit sectors that depended heavily on credit especially hard,
because deflation erodes the price of commodities while increasing the real value of debt.
Throughout the 1920s farm debt was increasing, in the US from $3.3 billion in 1910 to
$9.4 billion in 1925.
One might expect the ever falling prices of commodities to stimulate global spend-
ing and prosperity. This did not happen because the response of countries facing the
improved terms of trade (lower import prices) is slow, as it takes time to realise that real
incomes increased, permitting an increase in spending. The adverse impact on producers
is more immediate.
This is similar to how financial crises reallocate wealth: the losers scream immediately,
but the winners are harder to identify and don’t really notice for a while, and hence delay
spending their newfound wealth.
The end result is a deflationary cycle, as seen in Figure 2.3: prices fall, causing difficul-
ties for debtors, who in turn curtail consumption. Creditors realise they will get more for
their money in the future and also delay spending. This causes demand to contract and
prices to fall further.
Prices
fall
Demand falls
Creditors Debtors in
prosper difficulty
Both delay
spending
Figure 2.3 Deflationary cycle
24
2.2 The Great Depression

While the Great Depression is often said to have started with the stock market collapse
in Wall Street, most analysis indicates that a serious downturn commenced earlier.
The economy started sliding from April 1929 in Germany, from July in the US and from
August in the UK. Only in France was production rising. The conditions were the worst in
Germany, with high levels of unemployment and business failures.
As the crisis picked up pace, a sharp deflationary cycle, like that in Figure 2.3, emerged.
International lending and investment dropped off. Bankruptcy, default, c urrency depre-
ciation and falling commodity and asset prices made many businesses and foreign
countries unattractive risks for loans. Spreads between high-quality and low-quality
bonds increased, reflecting a loss of confidence, apparent in the sharp decline of direct
investment.
Impact on the US economy

As the crisis in the US started, real-estate prices crashed and industrial production
collapsed. Figure 2.4(a) shows net investment from 1929 until 1939. In 1929 it was
$10.5 billion, but it fell by $15 billion during the crisis. The impact on GDP was even more
significant, as seen by Figure 2.4(b), whilst unemployment exceeded 25% at the height of
the crisis. As the crisis continued, inventories were reduced and durable goods wore out.
Eventually, this helped recovery by stimulating demand.
10.0 bn
7.5 bn
5.0 bn
2.5 bn
0.0 bn
−2.5 bn
−5.0 bn
1920 1925 1930 1935
(a) Net capital formation in constant prices
120 bn GNP
Unemployment
Unemployment 20%
100 bn 15%
GNP
80 bn 10%
60 bn 5%
1925 1930 1935 1940

(b) GNP in constant prices and unemployment
Figure 2.4 US macroeconomic data

Data source: NBER
25
2.2.1 Financial markets

Before the Great Depression, the stock markets, and especially the New York market,
enjoyed spectacular price increases. The DJIA (Dow Jones Industrial Average) increased
from 191 in early 1928 to a peak of 381 in September 1929, with both prices and volume
doubling in two years.
Stock market crash

The initial outward sign of the Great Depression was seen in the financial markets. September
1929 marked the peak of the New York market. It started slipping on 3 October and declined
throughout the week of 14 October. On 24 October, known as ‘Black Thursday’, panic
ensued and a record number of 12.9 million shares were traded. On Monday 28 October,
more investors exited the market, and the slide continued with a record loss in the Dow for
the day at 13%. The next day, about 16 million shares were traded. This volume was a record
not broken for nearly 40 years. The DJIA fell by another 12%. In total, the market lost $14 bil-
lion in value that day, bringing the loss for the week to $30 billion, 10 times more than the
annual budget of the federal government and far more than the US had spent in all of WWI.
The crash is shown graphically in Figure 2.5. The DJIA did not regain its 1929 peak until 1954.
Capital markets
The collapse of the stock market caused large losses to individual investors, and affected
firms that relied on capital markets for liquidity, conspiring to create a liquidity crisis
which soon spread to other sectors of the economy such as mortgages. At the time, mort-
gages were normally unamortised three-year obligations, rolled over every three years.
Because of the liquidity crisis, homeowners were unable to refinance, leading to wide-
spread foreclosures and a large drop in house prices. All these factors fed on themselves,
in an example of an endogenous risk feedback loop.
Schadenfreude
Initially, the crash led to widespread schadenfreude – the satisfaction or pleasure felt at
someone else’s misfortune – as people blamed the depression on greedy speculators. It
was felt that banks got what they deserved and should rightly suffer. In the words of the
treasury secretary Andrew Mellon:
‘It will purge the rottenness out of the system. High costs of living and high living
will come down. People will work harder, live a more moral life.’
350
250
150
50
1926 1928 1930 1932 1934 1936
Figure 2.5 Stock market crash 1929 – DJIA
26
While such views were common, others, such as Keynes, worried about the adverse
impact of falling markets, arguing that a more useful policy would be for the authorities to
prevent the extreme price drops.
2.2.2 Financial crisis of 1931

Even though the crisis had started around 1929, initially it seemed like a financial markets
crisis, with serious but not catastrophic real economy impacts. The consensus of most
post-Depression studies indicates that the various authorities around the world could
have prevented the recession from turning into the Great Depression, but failed to do so.
The financial crisis in 1931 became a turning point. As deflation continued, commod-
ity prices fell, businesses lost money and the stock markets collapsed, adversely affecting
banking. Eventually, this led to banking crises in several countries.
Austria
Austria was the first country to experience a general economic crisis. The balance sheet
of its largest bank, Credit-Anstalt, equalled total government expenditure. In May 1931,
bank officers informed the government that deteriorating loan performance had com-
pletely wiped out its capital. It was bankrupt. The authorities and some private financial
institutions attempted to mount a rescue, but were unable to secure the necessary loans
in time. In response to the crisis, Austria went off the gold standard and imposed capital
controls.
Credit-Anstalt had a controlling interest in Budapest’s largest bank, and as soon as the
crisis erupted in Vienna, foreign investors withdrew their deposits from this and other
Hungarian institutions. By 15 May, bank runs in Budapest were underway.
Germany
The crisis in Austria provoked immediate withdrawals in Berlin. There were worries that
if Austria could freeze foreign deposits so could Germany. In July a fully-fledged banking
panic erupted, starting with Danat Bank, the financier of a failed textile firm, Nordwolle.
As the German government was dependent on foreign creditors for its own financing,
the crisis caused a sudden stop in lending, triggering a run on the mark. The German cen-
tral bank, the Reichsbank, lost $250 million in gold and was forced to ration credit to the
banking system.
US president Herbert Hoover announced the so-called Hoover Moratorium in June,
which called for a one-year moratorium on debt arising from WWI. Whilst quite unpopu-
lar, especially in France, it got support from 15 major nations and Congress. The Hoover
Moratorium was not effective in engineering an international support operation, and
Germany was forced to abandon the gold standard.
The UK and the gold standard

The UK was already in serious economic difficulties as a result of its decision to go back
on the gold standard in 1925, and suffered persistent capital account and budget defi-
cits. The European banking crisis caused Austria, Hungary and Germany to freeze British
27
deposits, whilst earnings from services declined and capital income from abroad col-
lapsed. Unemployment exceeded 20% and foreign banks were unwilling to provide credit.
As the domestic currency denominated assets were converted into foreign exchange,
the BoE, committed to pegging the exchange rate, was haemorrhaging gold. To defend
the gold parity, the Bank raised interest rates and restricted credit, aggravating the bank-
ing crisis.
Eventually, this forced the BoE to abandon the gold standard in September 1931. In
retrospect, this was the right decision as the British economy quickly recovered, and the
impact of the Great Depression was not as big in the UK as in some other countries.
The decision to abandon the gold standard in 1931 resonates strongly in the ongoing
crisis. Because it was key to economic recovery, it has been used to justify subsequent
policies of devaluing sterling, in particular quantitative easing (QE). Similarly, many com-
mentators maintain that the gold standard has strong parallels with the euro, and if only
countries like Spain could devalue, their economies would recover.
2.2.3 The US crisis

More than one-third of all banks in the US failed in the Depression, but the government’s
reaction was for most parts inadequate in the early stages. A right move was made in
December 1931, when the Reconstruction Finance Corporation (RFC) was set up to pro-
vide finance for banks and firms in need of liquidity. However, the Speaker of the House
then insisted that the loans to banks were publicised. The immediate impact of imple-
menting this requirement in January 1933 was that banks receiving RFC loans were hit by
bank runs, and other banks in difficulty therefore became unwilling to seeking help from
the RFC.
With the benefit of hindsight, it is now well understood that any liquidity assistance
provided to banks needs to be done in secret. A modern equivalent is that the run on
Northern Rock only started when the BoE announced it was providing liquidity assistance
to it. Most liquidity support in the ongoing crisis has been secret, with some central banks
going so far as to force all banks to borrow from the central bank, even the healthy ones,
so as to hide the identity of those in difficulty.
While Federal Reserve (Fed) officials were fully aware of the pending banking crisis, the
governors of the various Federal Reserve Banks refused to act. A large number of the banks
in trouble were not members of the Federal Reserve system and as such the regional bank
governors did not feel any responsibility for these non-member banks, despite their obvi-
ous impact on the nation’s overall supply of credit. The feeling was that propping up fail-
ing banks would be throwing good money after bad and the regional governors made it a
principle to let them fail.
New president — Roosevelt

The US voted in a new president in November 1932, Franklin Delano Roosevelt, who took
office in March 1933. As one of his first acts, he declared a nationwide banking holiday, clos-
ing all banks for more than a week. Eventually, good banks were to be reopened and insol-
vent banks closed down, and those facing difficulties but not insolvent received government
28
2.3 Causes of the Great Depression
support to reopen. This was enough to restore confidence, and as the banks reopened, long
lines of depositors formed outside the reopened banks, waiting to put their money back in.
By the end of the week, a total of $1 billion had been redeposited in the banks.
Impact on financial regulation

The incoming administration saw excesses and abuses within the financial system as a
major cause of the Depression. There was little to no oversight of financial institutions,
individuals frequently used 10 times leverage when buying stocks on the margin, insider
trading was rampant, and lack of deposit insurance directly contributed to the domino-
style failures of the country’s banks.
As a consequence, widespread regulations were introduced, such as the Glass–Steagall
Act separating commercial and investment banking, the guaranteeing of bank deposits up
to $2,500 and the establishment of the Securities and Exchange Commission (SEC). These
regulations remained in place for decades, and many are still with us. Many commenta-
tors have blamed the loosening of the Great Depression regulations as a major factor in
the crisis starting in 2007.
Double dip
While the worst of the Great Depression occurred at the beginning of 1933 and the US
economy recovered significantly in the subsequent years, the crisis returned in 1937.
There are several reasons why that happened. The surviving banks had massive amounts
of cash on reserve with the Fed, both as insurance and also because of lack of demand.
The Fed worried about inflation and tightened monetary policy. Inevitably, this led to a
contraction in lending. At the same time, taxes were on the increase. All these factors left
the economy fragile and susceptible to a double dip in 1937.
In retrospect, a looser fiscal and monetary policy in the years immediately after the
main crisis year of 1933 might have prevented the double dip. That was of course not
known to the policymakers at the time. Eventually, it was the rapidly increased govern-
ment spending with the onset of the Second World War (WWII) that pulled the US out of
the recession.

The Great Depression has generated significant academic research. Whilst there are many
explanations for the Depression, and the various factors are hotly debated to this day,
most commentators attribute the main blame to trade policy and monetary policy. Other
factors include the lack of global leadership and a narrow focus on national interests,
exchange instability and allowing the banks to fail.
2.3.1 Trade and tariffs

Trade restrictions were a common tool in the 1920s and 1930s for stimulating domestic
economies. This was especially common among agricultural exporters. Trade restrictions
took various forms, ranging from import tariffs and outright restrictions on imports to
29
export subsidies and the manipulation of exchange rates. Policymakers at the time were
aware of pitfalls in using trade restrictions, and many attempts were made to discour-
age their use, for example in the World Economic Conference of 1927, but overall such
attempts were not effective.
Trade restrictions were seen much more favourably than today, and both mainstream
political parties and well-known economists like Keynes advocated their use. One reason
is that governments were prevented from employing other measures to stimulate their
economies because of the gold standard and the prevailing orthodoxy that government
should run balanced budgets, even in downturns.
Smoot–Hawley Act
The most infamous example of trade restrictions is the Smoot–Hawley Act of 1930 that
significantly raised US average tariffs. Even before the Act became law, France, Italy, India
and Australia increased their tariffs in response. Once it became law, it let loose a wave
of retaliation. Spain sharply increased its tariffs in July 1930, Switzerland boycotted US
exports, Canada raised tariffs three times, and there were similar tariff increases in many
other countries, such as Cuba, Mexico, France, New Zealand and Australia.
In his 1932 presidential campaign, Roosevelt attacked the Smoot–Hawley Act as a
cause of the Depression. He claimed that when other countries ran out of gold, they paid
their debts by sending more goods, which meant that tariffs had the effect of lowering
prices rather than raising them.
Collapse in trade
As a consequence of the Smoot–Hawley Act, and the various other similar initiatives in
other countries, global trade seized up, with countries experiencing serious difficulties
because they could not export, and frequently defaulting. The severity of the collapse
of trade during 1929–1933 can be seen in the list below and in Figure 2.6, the numbers
being taken from Kindleberger (1986). Globally, trade was reduced by more than two-
thirds. The impact on a sample of countries can be seen in the following list.
■ Over 80%: Chile

■ 70–80%: Cuba, China, Peru
■ 60–70%: Netherlands, Greece, Brazil, Spain, Estonia
■ 50–60%: Denmark, New Zealand, Finland, Columbia.
2.3.2 Role of monetary policy

One of the main causes of a recession in 1930 turning into the Great Depression in 1933
was the rapid contraction in money supply. This can be seen for the US in Figure 2.7. Note
how M0 increases rapidly whilst M2 contracts even more sharply. This indicates that people
were converting less liquid money to deposits. Running the calculation in Example 1.1 in
reverse shows that money available for productive uses contracted sharply. This is a form of
deleveraging and was a major contributor to the Great Depression, since when the money
supply is reduced, individuals and firms do not have sufficient access to capital to invest
and consume, which slows down the economy. This can become a form of vicious feedback
30
January
r 2998 Fe
be bru
em 2739 ary
c
De
1839
r
be
Ma
em
rch
v
1206
No
October 992
1932
1931
1930
1929
April
r
be
Ma
m
pte
y
Se
st Ju
gu ne
Au
July
Figure 2.6 Reduction in trade 1929–1933

Source: adapted from Kindleberger (1986) with data from the League of Nations
60
50 M0 M1 M2
USD billions
40
30
20
10
0
1926 1928 1930 1932 1934 1936 1938 1940
(a) Money
60%
45%
30%
M0/M2
15%
M1/M2
0%
1926 1928 1930 1932 1934 1936 1938 1940
(b) Ratios of M0 and M1 to M2
Figure 2.7 US money supply 1926–1941
31
cycle, with the reduction in money supply leading to deflation, further restricting the money
supply and contributing to a liquidity trap.
Post-Depression analysis is mostly unanimous in identifying the contraction of the money
supply in the US as a major cause of the Depression, in no little measure due to one of the
most influential economic studies ever made, by Friedman and Schwartz (1963) in their
book A Monetary History of the United States: 1867–1960, where they argued that it was the
Fed that was primarily responsible for turning the crisis of 1929 into the Great Depression.
Friedman and Schwartz critique

The mistakes of the Fed, as discussed by Friedman and Schwartz, are summarised as
follows:
1 The Fed did too little to counteract the credit contraction caused by failing banks. If
banks fail, money supply (credit) contracts, therefore the Fed should have increased
the money supply to counteract the reduction.
2 The Fed actually reduced credit between December 1930 and April 1931. This led to
more and more banks being forced to sell assets at firesale prices, and to deteriorating
liquidity conditions.
3 When the UK abandoned the gold standard in September 1931, the Fed raised dis-
count rates in anticipation of a rush to convert dollars into gold. This did stop the drain
of gold but drove yet more banks into insolvency. But the Fed was in no immediate
danger of running out of gold, since it held 40% of global gold reserves and had more
than enough to meet its legal requirements.
4 Only under enormous political pressure did the Fed start undertaking open market
operations in April 1932. Unfortunately, this was too little, too late, and failed to pre-
vent a wave of bank failures in the last quarter of 1932.
5 When rumours that the incoming Roosevelt administration would devalue the dol-
lar led to a renewed flight from dollars into gold, the Fed once again raised the
discount rate.
According to Friedman and Schwartz, the Fed should have aggressively injected liquid-
ity into the banking system from 1929 onwards, using open market operations on a large
scale, and encouraging lending through the discount window. Gold outflows should not
have been such a priority consideration.
It is still a matter of debate why the Fed reacted the way they did. Friedman and
Schwartz argued that it was the death in 1928 of Benjamin Strong, the Governor of the
NYFed, which caused the deterioration in Fed performance.
An alternative explanation by Epstein and Ferguson (1984) and Anderson et al. (1988)
maintains that this was a deliberate act by the Fed to protect the interests of member
commercial banks rather than the general economy, and that the Fed was content with
seeing non-member banks fail, so that the relative importance of the Fed would increase.
The conclusion of Friedman and Schwartz that the collapse in money supply was a major
factor in causing the Great Depression has significantly influenced policymaking in the ongo-
ing crisis, and is a key reason why central banks have provided significant amounts of QE.
32
The US goes off gold

The incoming president, Roosevelt, recognised that the Depression had been associ-
ated with a deflationary spiral, so recovery could only come about when prices began to
increase again. The question vexing him and his advisers was how to get prices to increase
without waiting for economic recovery. One of his economic advisers, George Warren,
noticed that when large gold discoveries came onto the world market and the supply of
gold outstripped demand, commodity prices tended to rise. Therefore, one way to raise
prices would be to devalue the dollar by going off gold.
This proposal was met with outrage given the Fed’s vast gold reserves, with pundits argu-
ing that confidence in the US would be dented. Indeed, almost all of the president’s eco-
nomic advisers opposed going off gold. Roosevelt was in favour nonetheless, and slipped
the necessary legislation into an agriculture bill so as to escape notice. On the evening of
18 April, he announced to his team: ‘Congratulate me. We are off the gold standard.’
In the days after, as the dollar fell against gold, the stock market soared by 15%. Going
off gold seemed to have reversed the psychology of deflation because of a renewed confi-
dence in banks, an activist Fed, and a government commitment to drive up prices. In the
following three months, wholesale prices increased by 45%, vehicle sales doubled, and
overall industrial production went up by 50%. This was not the end of the matter, how-
ever; by October 1933, though the dollar had fallen by more than 30%, commodity prices
began to sink again and it was time for a new initiative — buying gold in the open market.
In the following three months after October, every morning at 9 a.m., Roosevelt and
his economic advisors would determine the price of gold for that day, in a completely ran-
dom fashion — one day, Roosevelt priced an increase of 21 cents, and when asked why,
replied that it was a lucky number! Keynes dismissed this exercise as ‘the gold standard
on the booze’. In January 1934, Roosevelt agreed to stabilise gold at $35 an ounce. In
total, the dollar had been devalued by over 40%.
Impact of going off gold

The US was not alone in going off gold; every major country did the same during the
Great Depression. Argentina was one of the first to leave in 1929 and Switzerland the last
in 1936. We show GDP per capita and year of going off gold in Figure 2.8.
US France Germany
UK Argentina Japan
7000
5000 x
x
x x
x
3000
x
1000
1925 1930 1935 1940
Figure 2.8 GDP per capita in constant 1990 dollars, and year of going off gold
Data source: www.ggdc.net/MADDISON/oriindex.htm
33
Abandoning the gold standard was successful. Once prices started rising, the real value
of debt was reduced, making businesses more willing to borrow and consumers more
ready to spend.
This happened after a long period of little or no investment and sharply reduced
personal consumption, meaning there was significant pent-up demand for both indus-
trial goods and consumer goods. Once consumers and companies realised that there
was less benefit in holding onto savings because prices were rising sharply, they were
encouraged to spend, thereby stimulating the economy. Note the difference between
the global abandonment of the gold standard and the piecemeal approaches discussed
in Section 2.1.4. There, an individual country in difficulty devalued whilst the major
industrialised nations, and the main importers of agricultural goods, did not. In this
case, it was the industrialised countries that allowed their money supply to expand,
resulting in prices increasing across the board.
Implication for future policy

The post-Great Depression analysis has mostly concluded that going off gold was a turn-
ing point in fighting the Great Depression. The implications of this success had a signifi-
cant impact on future policy. Governments following WWII started to employ very loose
monetary policy as a routine technique to stimulate their economies, with the conse-
quence that monetary policy lost its effectiveness and inflationary expectations got built
in. Solving that problem was very costly in the 1980s.
2.3.3 Lack of global leadership and narrow focus on national interests

An important reason why the Great Depression happened was the lack of global lead-
ership and international coordination. At the time, international organisations, like the
League of Nations, were in their infancy and we did not know how to use them to exer-
cise effective diplomacy. International coordination and policymaking were not seen as
important. This meant that the leadership role of individual countries was more impor-
tant than today.
In the second part of the nineteenth century, right up until WWI, the UK assumed the
role of a leading nation, helping to quickly prevent pending crises from getting out of con-
trol. From that time the effective global leader was the US but it was not ready to assume
that mandate when the time came to formulate a global crisis response.
Leading countries at the time tended to focus on narrow short-term national inter-
ests, disregarding long-term consequences, the potential reaction of other countries and
the impact on the global economy. They had various long meetings, but strong disagree-
ments prevailed and no conclusion on policy response was reached.
One example of the failure to respond is provided by an anecdote from Ahamed
(2009), discussing an international government conference in Paris in April 1929. At the
time, the danger facing the global economy was becoming increasingly clear. Regardless,
France insisted on focusing the proceedings on war reparations, even though Germany
had no ability to pay. In the words of the report of the observer from the Swiss National
Bank (SNB), Felix Somary:
34
2.4 Implications for future policy
‘Almost all the great powers have been negotiating for months about how many
billions a year should be paid until 1966, and thereafter until 1988, by a country
that is not even in a position to pay its own civil servants’ salaries the next day.’
2.4 Implications for future policy

The Great Depression had a strong impact on policymaking and academic research.
Economists were blamed for providing bad advice, before and during the Depression, and
the field of macroeconomics was developed in response.
Since the Great Depression, the authorities have been well aware of which pitfalls
to avoid if threatened by another crisis episode: avoid protectionism, employ fiscal and
monetary stimulus, coordinate crisis response internationally, and do not let banks fail.
Stimulus in one country benefits its neighbours as well, by encouraging international
trade, leaving no reason to employ protectionist measures. This leaves a problem of free
riding since stimulus packages create positive externalities. However, if countries focus on
the common objective of preventing another Depression, free riding will not be a signifi-
cant problem.
In the early 1940s, the Allied leaders, excluding the Soviet Union, discussed how the
world economic order should be set up, and ended up establishing a new range of inter-
national government bodies and coordinating mechanisms. Of those, the most important
was the establishment of the Bretton Woods system, especially the International Monetary
Fund (IMF) and the World Bank. Policymakers also decided to implement specific mecha-
nisms to prevent the use of trade restrictions. The most prominent example of this is
the General Agreement on Tariffs and Trade (GATT) in 1947, eventually replaced by the
World Trade Organisation (WTO) in 1995. While free trade has remained a controversial
subject, and many countries employ trade restrictions in various forms, the scale of such
restrictions is much lower than in the 1920s and 1930s. Finally, as a direct response to
the Depression, many countries have set up formal mechanisms for safeguarding finan-
cial stability, such as rules for lending of last resort (LOLR), deposit insurance and bank
regulations. Most of these arrangements are fundamentally with us to this day, even if the
specific form has changed.
2.4.1 Parallels with the crisis of 2007–2009

In the decades after the Great Depression ended, financial and economic crises were
common but were always confined to individual countries, or at worst a handful of
countries. This finally changed in 2007 when a global financial and economic crisis
started. Below, we directly draw parallels between the Great Depression and the crises
from 2007.
The crisis response this time around is directly influenced by the Great Depression. It
has often seemed as if authorities had made the list of Great Depression mistakes, with
the intention of not repeating them. This policy was broadly successful for the first phase
of the crisis from 2007 to 2009, but less so for the European sovereign debt phase.
35
Comparisons
Great Depression Crisis from 2007
Roaring twenties Great moderation
The 1920s enjoyed rapid economic growth, with The boom prior to 2007 was based on new
new technology, silent films and the Model T technology like the Internet, and the renewed
Ford, and the beginnings of radio and prominence of financial markets. The financial
commercial air services. Stock market speculation system used high levels of leverage, with strong
was rampant, leverage was high and capital capital flows to financial markets.
flowed to Wall Street.
Financial markets
Rapid development of financial markets, with Financial innovation, with new forms of real-
ordinary citizens for the first time having access estate financing, like subprime, emerged, and a
to financial services such as mortgages and stock rapid increase in homeownership and low-cost
market investments bought on the margin. A mortgages fuelled a real-estate boom. The market
stock market and real-estate boom ensued, boom led to the belief that prices could rise
creating the feeling that prices could rise indefinitely. New financial practices like structured
indefinitely, leading to reckless risk taking. credit and shadow banking enabled risk taking
out of the view of supervisors and the market.
US bailouts
RFC TARP
In 1931, President Hoover set up the RFC, making The main bailout programme was the Troubled
loans to financial institutions of all types, with Asset Relief Program (TARP), authorised at
$500 million outright, and the possibility of $700 billion, but total disbursements were only
issuing up to $1.5 billion in government debt (1% $431 billion, 4.8% and 3% of GDP respectively.
to 3% of GDP respectively). The RFC is seen as a
failure, not least because the bailouts were public.
Short selling
President Hoover compelled the New York Stock The Bush administration temporarily imposed
Exchange to curb short selling. restrictions on short selling of shares in some
financial institutions.
LOLR
The US did not do LOLR until Roosevelt took Various large-sized LOLR-type credit and
office in March 1933. liquidity facilities were provided.
Trade
The Smoot–Hawley Act and various other trade No restrictions on trade have yet been imposed.
restrictions reduced global trade by more than
two-thirds.
Money and liquidity
The US authorities allowed the money supply to Significant QE was provided and deflation was
shrink and deflation to set in. prevented.
36
2.5 Summary
Lesson learned
Generally, it seems that the policy response in the fall of 2008 was the right one, and
the authorities did not repeat the mistakes made in the Great Depression. Deflation
was avoided, widespread bank failures were prevented, trade remained free and global
authorities actively cooperated in their crisis response. These measures seemed successful
in preventing the crisis from turning into another Great Depression.
2.4.2 European sovereign debt crisis

There are fewer parallels between the Great Depression and the European sovereign debt
crisis. The latter is created by the unique circumstances of a common European currency
without the necessary conditions being met, with no comparable problem in the Great
Depression. There is one important commonality, the gold standard.
Impact of the gold standard

The presence of the gold standard meant that the money supply grew slower than the
real economy, resulting in deflation. This was especially problematic in countries like the
UK that opted to reenter the gold standard at an overvalued rate, as well as the various
agricultural exporters exposed to the commodity price deflation. The only way for these
countries to remain on the gold standard was by internal devaluation.
The gold standard frustrated efforts at containing the banking crises by limiting the
scope for individual central banks to do LOLR. When the central banks provided liquid-
ity to domestic banks, it signalled that the government attached a higher priority to the
stability of the banking system than to the defence of the gold standard. Inevitably this
encouraged depositors to get money out of the country in advance of the expected deval-
uation, in a self-fulfilling crisis scenario. Any liquidity injected into the banking system just
leaked back out as the inevitable balance of payments crisis loomed.
Impact of the common currency

Similar forces are at work with the common European currency. For some countries, such
as Germany, the Netherlands, Finland and Austria, European inflation rates and exchange
rates have been appropriate, whilst for others the euro has been significantly overvalued,
as is the case for Greece, Italy, Spain and Portugal. The latter countries, therefore, are
in a similar situation as were the gold standard countries during the Great Depression,
being forced by the constraints of the monetary system to make internal devaluations and
impose extreme austerity, rather than simply allowing their currency to devalue.
2.5 Summary
The biggest economic crisis the world has ever seen is the Great Depression. The causes of
the Depression are varied and remain controversial, though they include war reparations,
agricultural overproduction, inappropriate monetary policies (too loose and too strict),
competitive devaluations and trade restrictions.
When the crisis was underway, the policy response could not have been worse; coun-
tries focused on narrow national interests, trade collapsed by more than two-thirds,
37
liquidity (money supply) was allowed to contract and deflation to set in, and banking
systems were allowed to fail in some countries.
The causes of the Great Depression, as well as the needed policy response, have been
extensively studied. The world authorities facing a new global crisis in 2007 effectively
prevented a new Great Depression from emerging by explicitly avoiding the mistakes
made before and during the Great Depression.
The European authorities fighting the European sovereign debt crisis have not been as willing
to take on board Great Depression lessons, especially problems arising from the gold standard.

1 The Great Depression is the most severe global economic crisis we have ever seen. In
what way was it more damaging than the ongoing crisis in 2007?
2 Winston Churchill is said to have stated that going on gold in 1924 at prewar parity was
his biggest mistake? What were the implications of this decision of his?
3 What was the impact of competitive devaluations in the 1920s, and how has the mem-
ory of those events shaped current policy debate?
4 What role did tariffs play in worsening the Great Depression, and how has the memory
of those events shaped current policy debate?
5 Explain the role of monetary policy in the US as contributor to the Depression.
6 What are the main lessons of the Depression that policymakers seem to have learned
when fighting the crisis in 2008?
7 What is the main lesson from the Great Depression that has not been applied in the
European sovereign debt crisis? Why do you think that is the case?
References
Ahamed, L. (2009). Lords of Finance, The Bankers Who Broke the World. The Penguin Group.
Anderson, G. M., Shughart, W. F., II., and Tollison, R. D. (1988). A public choice theory of the
great contraction. Public Choice, 59: 3–23.
Bernanke, B. (1995). The macroeconomics of the Great Depression: a comparative approach.
J. Money Credit and Banking, 27(1): 1–28.
Eichengreen, B. (1996). Golden Fetters: The Gold Standard and the Great Depression, 1919–1939.
Oxford University Press.
Eichengreen, B. and Irwin, D. (2009). The slide to protectionism in the Great Depression: who
succumbed and why? NBER Working Paper 15142.
Epstein, G. and Ferguson, T. (1984). Monetary policy, loan liquidation, and industrial conflict:
the Federal Reserve and open market operations of 1932. J. Econ. Hist., 44: 957–83.
Friedman, M. and Schwartz, A. (1963). A Monetary History of the United States: 1867–1960.
Princeton University. Press.
Keynes, J. M. (1920). The Economic Consequences of the Peace. Harcourt Brace, New York.
Keynes, J. M. (1924). A Tract on Monetary Reform. Macmillan.
Kindleberger, C. (1986). The World in Depression, 1929–1939, 2nd edition. University of California Press.
38
3 Endogenous risk
The term endogenous describes an outcome or process having an internal cause or

origin. In the context of financial markets, endogenous risk refers to the view that
risk is created by the interplay between market participants, rather than the chance
that a shock might arrive to the financial markets from the outside — exogenous risk.
The classification of risk into endogenous and exogenous risk was first proposed by
Danielsson and Shin (2003).
Endogenous risk arises when individual economic agents react to their environment
and their actions in turn affect their environment to such a degree that an endog-
enous feedback cannot be ignored. Financial markets, where all market participants
are constantly competing against each other, trying to gain advantage by anticipating
each other’s moves, are a clear example of an environment creating endogenous risk.
Endogenous feedback between the behaviour of market participants can suddenly
create a vicious cycle, causing a crisis. This arises from the presence of mechanisms
creating accelerator effects, whereby a relatively small event can trigger a large out-
come. This is similar to the idea by chaos theorists of a butterfly in Hong Kong trigger-
ing a hurricane in the Caribbean. Many commentators, and the popular press, focus
on the trigger event — the butterfly — while it is the mechanism that enables the but-
terfly to trigger the hurricane that really matters.
Endogenous risk tends to be low most of the time, because economic agents usually
behave individually and have different objectives and information sets. This means
that in aggregate their behaviour resembles noise when viewed from the outside.
Under certain conditions, market participants start behaving much more harmoni-
ously than usual, amplifying price movements that result in asset price bubbles and
Chapter 3 Endogenous risk
finally market crashes. This creates risk that is hidden until it is too late, providing a
false sense of stability, while the market is really heading for the precipice.
Such spirals of coordinated selling need to be strong enough to overcome the usual
stabilising forces in markets such as arbitrageurs, hedge funds, sovereign wealth funds
and the like that could be expected to step in and buy the cheap assets, putting a
floor under prices.
Financial markets are subject to both exogenous and endogenous risk, but it is the
latter that is more damaging. It is behind some of the biggest financial crisis episodes
in history, while being much harder to model than exogenous risk. For a single bank,
most risk is exogenous. For the financial system, risk is almost entirely endogenous,
even if not to a high degree most of the time. As an approximation, those concerned
with financial risk might not be too far off by assuming risk is exogenous 99.9% of the
time. It is, however, the other 0.1% when endogenous risk becomes a serious concern.
That is when economic agents become harmonised in behaviour, and the conditions
are ripe for a crisis to emerge.
Endogenous risk has direct implications for financial models, since most such mod-
els assume risk is exogenous. This suggests that the models may work well 99.9% of
the time, but fail when needed the most — at times of extreme market turmoil. In
turn, this leads to direct policy conclusions on financial stability, the use of risk mod-
els by bank supervisors and the wisdom of prudential regulations.
The concept of endogenous risk directly relates to other economic theories, such
as the beauty contest in Keynes (1936). It also connects to Minsky (1992) who argued
that economies have either stable or unstable financial regimes; even if the economy
starts out stable, continued prosperity paves the way for an unstable system.
Similarly, endogenous risk impacts on policies towards financial stability. For
example, prudential banking regulations aim at ensuring that each and every finan-
cial institution behaves properly, avoiding excessive risk. Endogenous risk analysis
predicts that such prevention regulations are directly destabilising, because as banks
act prudently in a crisis, their very acts of self-preservation cause them to dispose of
the same risky assets and buy the same safe assets, amplifying price movements and
distress. This suggests that any financial stability policies explicitly need to consider
the potential for endogenous feedbacks.

This chapter focuses on the hidden mechanisms that create potential crashes. Many
of the concepts mentioned in this chapter are discussed in much more detail later, for
example in Chapter 4 (liquidity) and Chapter 9 (trading and speculation).
Key concepts
■ Endogenous risk
■ Trading strategies
■ Actual and perceived risk
■ Dual role of prices
40
3.1 Millennium Bridge

While the chapter is self-contained, the material is drawn from several of my papers,
all of which are available on my website (www.RiskResearch.org). The notion of
endogenous risk was first introduced by Danielsson and Shin (2003) and further
developed by Shin (2010). Two recent papers discuss endogenous risk from a policy
point of view, Danielsson et al. (2012a) and Danielsson et al. (2012b), with the former
focused on systemic risk and financial regulations while the latter addresses extreme
outcomes in financial markets. Finally, a more theoretical treatment is provided by
Danielsson et al. (2012c).
A Value of assets
C Cash
D Debt
d Change operator
E Equity
f Option price
L Leverage
n indicates iteration
P Prices
Q Quantity (units of assets held)
X Strike price
∆ Option delta from the Black–Scholes equation
E Random number
L Price impact factor
S Standard deviation
3.1 Millennium Bridge

Endogenous risk is not exclusively confined to financial markets, and we start with an
illuminating example from engineering. The first new bridge to span the River Thames for
100 years was the pedestrian Millennium Bridge, opened by Queen Elizabeth on 10 June
2000. On the opening day, thousands of people used it to cross the river. This should not
have been a problem, as the bridge was designed to cope easily with such large crowds.
Within moments of being opened to the public the bridge began to wobble violently,
and was soon closed to the great embarrassment of the bridge’s designers – Arup and Lord
Foster – and the authorities. In the process, it earned the nickname the wobbly bridge.1
The wobble came as a surprise, as no such outcome had been predicted by the exten-
sive computer modelling and human testing conducted by Arup. It is well known that
soldiers marching across bridges can cause them to collapse, which is why they are asked
1
See http://news.bbc.co.uk/hi/english/static/in_depth/uk/2000/millennium_bridge/default.stm and www.
youtube.com/watch? v=eAXVaXWZ8.
41
to break step when crossing bridges. This happens because marching soldiers generate
harmonised frequencies that can create a feedback between the internal frequency of a
bridge and the soldiers’ steps, leading to a collapse. However, the people crowded on the
Millennium Bridge on the opening day were not soldiers, but people of all walks of life.
The chance of them spontaneously marching was considered next to impossible.
Diagnosis
Every bridge is designed to move with the elements, and the Millennium Bridge was sup-
posed to sway gently in response to the Thames breeze. A gust of wind – an exogenous
shock – hit the bridge, causing it to move sideways and wobble. When this happens a
natural reaction is to adjust one’s stance to regain balance. By doing so, the bridge gets
pushed back, making it sway even more, causing people to adjust their stance yet again –
more and more at the same time – this time pushing the bridge in the opposite direction.
As an ever increasing number of pedestrians started to adjust their stance simultaneously,
the bridge moved more and soon almost all the pedestrians joined in. This created a
mutually reinforcing feedback loop between the synchronised adjustments of the pedes-
trians’ stance and the bridge’s wobble,2 shown in Figure 3.1.
We cannot identify the endogeneity of the Millennium Bridge by considering either the
properties of the bridge or the crowd in isolation; they have to be studied jointly. This is
what makes endogenous risk problems hard.
Wider lessons
The wobble of the bridge was created by people reacting to a gust of wind pushing the
bridge. The wobble continued and intensified even though the initial gust of wind had
long passed. This is a concise example of how a relatively small exogenous shock (the gust
of wind) causes a large endogenous response. It is a clear example of endogenous risk in
action, unlike, for example, an earthquake which is an exogenous shock. The ultimate les-
son from the Millennium Bridge is that it is not the shock that matters but the feedback
mechanism that allows a small shock to be amplified into a large event.
Pedestrians
adjust
stance
Bridge
Bridge
gets
moves
pushed
People further
adjust their
stance
Figure 3.1 Feedback loop of the Millennium Bridge
2
For more information, see www.arup.com/millenniumbridge/challenge/results.html.
42
3.2 Dual role of prices
The financial system is replete with analogous processes, whereby an innocuous shock
has the potential to trigger a systemic crisis. Financial markets are examples of environ-
ments whereby individuals not only react to events around them, but also by their actions
directly affect market outcomes, because of balance sheet adjustments. The pedestrians
on the Millennium Bridge were like traders reacting to price changes, and the movement
of the bridge was like price moves in markets. Consequently, the notion of endogenous
risk provides a clear mechanism for analysing feedbacks in financial systems.

Prices of financial assets play two important roles. The first is quite familiar.
Prices reflect the underlying fundamentals, and aggregate all available informa-
tion into a single number, the price. For example, prices might be the present discounted
value of dividends. This means that market prices provide the best indication of the cur-
rent value of an asset. However, it does not mean that current prices provide a reliable
guide to the future value or future risk of an asset, because prices also affect the future
behaviour of market participants.
Prices are also an imperative to action, not least because of constraints on market
participants, constraints above and beyond the basic objective of maximising profits. This
can include accounting rules, legal obligations, disclosures, risk constraints or marking
to market. A bank may not react to market outcomes in the way we expect, because the
constraints dictate a certain type of behaviour. If the bank is large, it will significantly
affect prices by its trading activities, which means that its constraints may directly affect
the statistical properties of market prices. One could say that the presence of constraints
undermines the integrity of the prices, taking them away from their fundamental values.
In extreme cases, prices can become so distorted that they lead to undesirable extreme
outcomes, like bubbles and crashes.
What this means is that while market prices are the best guide to the current valu-
ation of an asset, they might not provide a good indication of the future valuation or
future risk. The deterministic impact of constraints gets in the way. While this may result
in prices being predictable, traders cannot exploit this because of the constraints they are
under. This has particular implications for investment decisions and policy. In particular,
the practice of treating prices as exogenously determined by an efficient market might be
a poor policy guide in times of crisis.
3.2.1 Leverage constraints and upward-sloping demand

We illustrate the surprising impact of external constraints on trading decisions by a sim-
ple model of leverage constraints. Even in this basic example, a financial institution does
behave unexpectedly, buying when prices increase and selling when prices fall.
Standard microeconomic theory predicts that demand functions are downward slop-
ing. When the price of something falls we buy more, and when prices increase we buy
less. The only exception is a strange asset called a Giffen good. The leverage constraint
43
example below demonstrates that constraints can lead to upward-sloping demand func-
tions for regular assets.
Setup
Suppose a bank’s leverage (L) is restricted to 5. In other words, the ratio of assets (A) to
equity (E) has to be 5:
A
L = = 5
E
We set the initial price of the asset, P, to $10, where the bank holds Q = 100 units of the asset,
and has debt of D = +800. This means that the value of its portfolio is A = Q * P = +1000,
and its equity is E = A - D = +200. The initial balance sheet is therefore:
Assets Liabilities
A = 1000 E = 200
D = 800
and the bank meets its leverage constraint. Suppose the asset is hit by a negative exog-
enous shock, causing its price to fall to P = +9. This price drop directly affects equity and
the balance sheet becomes:
Assets Liabilities
A = 900 E = 100
D = 800
Leverage is now 9 = 900>100, and the bank needs to take action to meet the constraint.
The most direct way is to reduce debt by selling assets.
Consider two different cases. In the first case, the bank is a price taker, meaning that
no matter how much it buys or sells, prices remain unaffected. Prices are exogenous. In
the second case the bank exerts a significant market power, so that its buying or selling
decisions cause prices to move. Prices are endogenous. The former case corresponds to a
small individual trader, whilst the latter case reflects a large bank or fund.
Prices are exogenous

Suppose the bank is a price taker and as a response to the price drop sells enough of the
asset to keep leverage at 5. The change in debt, D1 - D0, will be the amount sold at cur-
rent market prices P:
D1 - D0 = P1Q1 - Q02
D1 = D0 + PQ1 - PQ0 (3.1)
The leverage constraint is then:
A1 PQ1
L = =
E1 PQ1 - D1
44
Substitute D1 by its value in (3.1):
PQ1
L =
PQ0 - D0
D0
Q1 = L a Q0 - b (3.2)
P
In our example, the bank should now hold 500/9 units of the asset. By selling 100 - 500>9
units, the target leverage ratio is reached and the bank needs to do nothing further. Its
balance sheet becomes:
Assets Liabilities
A = 500 E = 100
D = 400
This means that in order to meet the leverage constraint, the bank had to sell $400 worth
of the asset. This happened because prices dropped, so the demand function is upward
sloping. If the price shock had been positive, the bank would have bought the asset. We
show the demand function in Figure 3.2.
Prices are endogenous

Consider the case where the bank is a large market participant, and thus exerts a sig-
nificant price impact. When it buys, prices increase, and when it sells, prices fall. We
capture this phenomena by the so-called price impact function, showing how much prices
change. Because there are several steps in the calculation, we use n to indicate the step.
The change in price Pn - Pn - 1 due to a transaction worth Pn - 11Qn - 1 - Qn - 22 is
Pn - Pn - 1 = lPn - 11Qn - 1 - Qn - 22
l is known as the price impact factor, with Pn - 11Qn - 1 - Qn - 22 the amount the bank wants
to sell, in our case $400.
Because prices fall in response to the sale, the bank will not be meeting the leverage
constraint even after selling the same amount as in the exogenous case above. It will have
to sell more assets. That in turn triggers a further price drop, and so forth, until we get
convergence. The process to compute the variables is as follows:
1 Pn = Pn - 1 + lPn - 11Qn - 1 - Qn - 22
2 Qn = L1Qn - 1 - Dn - 1 >Pn2
3 An = PnQn
4 En = An >L
5 D n = A n - En
Let the price impact factor be l = 0.001. We then repeatedly apply (3.2) until we get
convergence, with the results of 10 rounds shown in Table 3.1.
45
Table 3.1 Ten selling iterations
Iteration Q P A
1 100.000 10.000 1000.000

2 55.556 9.000 500.000
f f f f
9 42.934 8.492 364.585
10 42.934 8.492 364.585
The initial sale was for 44.444 units of the asset, but as we keep on iterating we get
further sales of 12.6 units, causing prices to drop further to $8.492.
Supply and demand functions

We can demonstrate this more generally by considering the bank’s demand for various
magnitudes of initial price shocks. This is shown in Figure 3.2 for both the exogenous and
endogenous cases.
Exogenous
Endogenous
Final change
50
in quantity
−50
−1.0 −0.5 0.0 0.5 1.0

Initial change in price
Figure 3.2 Demand functions
The x-axis shows the initial price change, with our $1 price drop from above on the extreme
left. We then go all the way to a $1 price increase on the right. The y-axis shows the final
change in the bank’s holding of the asset. On the extreme left we get 55.55 for the exogenous
case and 42.93 in the endogenous case. This demonstrates the upward-sloping demand
functions and shows that when prices are endogenous, the impact is much stronger.
In this case, the price impact was quite small, 0.001, and we still manage to get a sizable
endogenous change in prices and quantity. This succinctly demonstrates how a small exoge-
nous shock can trigger a large outcome, where the constraints dictate a ‘sell cheap, buy dear’
strategy that generates precisely the kind of vicious feedback loops that destabilise markets.
3.3 Risk
There is no single way to define risk. A basic textbook definition is the volatility of returns,
but that is only a complete description if financial returns are normally distributed, with
constant unchanging volatility. Neither assumption is true in reality. This means that any
46
3.3 Risk
textbook way of making investment decisions, such as a mean–variance diagram or a capi-

tal asset pricing model – essential elements of modern portfolio theory (MPT) – will be
approximate because the underlying assumptions are violated.
One could argue that even if the underlying assumptions are problematic, MPT pro-
vides reasonably good approximations to the real world. We can distinguish low risk from
high risk, we can often assume markets are sufficiently liquid, and we can partly predict
volatility. For this reason, many, even most, practitioners make some use of MPT, perhaps
extending it to incorporate fat tails and volatility clustering. When considering longer
horizons, especially for assets that are illiquid or have very fat tails, other alternatives may
be preferred. In the words of Warren Buffett (2012):
‘The riskiness of an investment is not measured by beta [a Wall Street term encom-
passing volatility and often used in measuring risk] but rather by the probability –
the reasoned probability – of that investment causing its owner a loss of purchasing
power over his contemplated holding period. Assets can fluctuate greatly in price
and not be risky as long as they are reasonably certain to deliver increased purchas-
ing power over their holding period. And as we will see, a nonfluctuating asset can
be laden with risk.’
Financial regulations employ Value-at-Risk (VaR) (see the appendix to Chapter 13) as a
preferred way to quantify market risk. This is based on the risk of daily outcomes (linearly
scaled to 10 days) that happen with 1% probability, so once every five months. In most
cases, the VaR is calculated by volatilities. It therefore falls under Buffett’s criticism.
Risk-free
Some assets have such low risk that they are referred to as risk-free. This is typically a short-
term obligation of sovereigns considered extremely unlikely to default, like the United
States (US), United Kingdom (UK) and Switzerland. Even in that case, the investment is
not totally risk-free because it involves putting money into a local currency which is at risk
from inflation. Some governments offer inflation-linked bonds which come close to being
risk-free.
Gold is, however, often considered a good hedge against inflation, and one might be
tempted to invest in an asset like gold in order to avoid risk, but it is far from risk-free, as
can be seen in Figure 3.3.
2000
1500
1000
500
1970 1980 1990 2000 2010
Figure 3.3 Real USD spot price of an ounce of gold

Data source: World Gold Council and US Department of Labor
47
3.3.1 Dependence
Correlations between asset returns are often higher when the markets are falling, even
tending to 1 in a market crash. An increased level of correlations is often a sign of market
turmoil. While the theoretical analysis above focused on the interaction between market
and funding liquidity for a single security, it also has implications for how multiple asset
prices move together in times of stress.
Definition 3.1 Non-linear dependence and correlations A common way to

refer to how two random variables are related is by the concept of correlations. However,
correlations only properly describe the interrelationship when it is linear, and often variables
are related in a complicated non-linear way, not captured by correlations. Therefore, it is
more accurate to use terms like dependence or non-linear dependence.
Danielsson et al. (2012b) demonstrate that a chain of events like that described in
Figure 4.4 in Chapter 4 also affects the dependence between assets. In normal times,
dependence between assets is generally based on the inherent properties of the assets
themselves, with similar assets tending to move in similar ways.
In a crisis, the situation changes, and is much more affected by the liquidity positions
of the banks that trigger the crisis. This can happen because if a bank suffers from a lack of
funding liquidity, it may be forced to sell its assets across the board, adversely impacting
on the price of each and every one. This is manifested in higher correlations.
Forced selling in a firesale externality scenario tends to create new forms of depend-
ence between assets held by similar investors. This leads to the empirical prediction that
correlations increase sharply in times of market turmoil, and especially in crises.
3.4 Dynamic trading strategies

While the leverage constraint example above is somewhat artificial, we have seen many
concrete examples of endogenous risk causing extreme outcomes in financial markets.
One of the clearest examples is the stock market crash of 19 October, 1987 when global
stock markets fell by around 23%. We show the impact in Figure 3.4.
2600
2400
2200
2000
1800
Jan Mar May Jul Sep Nov
Figure 3.4 1987 Dow Jones Industrial Averge index values

Data source: finance.yahoo.com
48
The postmortem analysis by the US government, the Brady Commission (1988) report,
found that the main underlying cause was the use of an automatic trading strategy called
portfolio insurance.
Portfolio insurance
Portfolio insurance is a trading strategy whereby an investor wishes to hedge downside
risk by dynamically replicating a put option. In principle, the investor could just buy the
put option, but it may not be available for the particular asset, or be illiquid and/or too
expensive. Under the assumptions of the option pricing model of Black and Scholes
(1973b), there should be no difference between the price of an option bought outright
and a dynamically replicated option.
Delta and dynamic replication

An option gives the holder the right but not the obligation to buy (call) and sell (put) an
asset at a pre-agreed price (the strike price, X) at a fixed date in the future. While one can
use the model of Black–Scholes to get the option price, we can also dynamically replicate
the option by a combination of cash and the underlying asset.
The delta (∆) of a put option is the rate of change of its price, denoted by f, with
respect to the change in price of the underlying asset, P. We can obtain Δ from the deriva-
tive of the Black–Scholes equation with respect to prices.
df
∆ =
dP
The ∆ of a put option is negative. Graphically, ∆ is the slope of a curve representing the
option price against the price of the underlying asset, as shown in Figure 3.5.
A put can be replicated by holding ∆ units of the underlying asset, so a portfolio con-
sisting of
∆ of underlying asset - 1 put
is risk-free for small movements in the underlying price. When the price changes, the
gain or loss from holding the underlying asset is matched by an exactly offsetting loss or
gain from the change in the price of the put option.
Option price f
pa
yo
pa
ff a
yo
f fp
te
rio
xp
r to
ira
exp
i
t
rati
ion
∆ on
X Stock price P
Figure 3.5 Put option and delta
49
3.4.1 Numerical example

Let us examine the impact of dynamic delta hedging by means of numerical simulations.
This example is adopted from Shin (2010) where it is explained in more detail. We first
decide on the magnitude of the inputs, and set the strike price at $90, the risk-free rate
at 0%, the annual volatility at 25% and the time to maturity to 9 weeks. The underlying
stock is currently trading at $100. If we plug these values into the Black–Scholes equation,
we get the value of the put as $0.8012. We simulate the market in one-week increments,
and show the details of the simulation in Table 3.2.
We use a superscript * to denote the actual outcomes, so for example P refers to the
theoretical price and P* to the actual price. C refers to cash holdings.
Simulation
The realised return from day t - 1 to day t is
1 + et + 1At - At - 12 (3.3)
where et is the exogenous random return and At - At - 1 the purchase (or sale) of the stock.
First week. The trader starts with no funds. The initial stock price is 100. The trader
shorts the stock by ∆ units, and from the Black–Scholes equation we find that
∆ = -0.14. This gives the trader 14.3 in cash. We assume this is the same in both the
endogenous and exogenous price cases in order to start both off in the same place.
Exogenous prices
Second week. We start by drawing a random shock,
0.252
P ∼n a 0, b
52
resulting in -0.016 in this particular case, so the price falls to 98.4, and ∆ becomes
-0.17. As a consequence, we sell more of the stock, increasing the cash balance to 16.8.
Table 3.2 Dynamic replication strategy
T - t e P P* ∆ ∆* C C*
9/10 100.0 100.0 -0.14 -0.14 14.3 14.3

8/10 -0.016 98.4 98.4 -0.17 -0.17 16.8 16.8
7/10 0.022 100.6 98.1 -0.10 -0.16 10.3 16.1
6/10 0.004 101.0 99.3 -0.08 -0.12 7.9 11.5
5/10 -0.040 97.0 99.9 -0.16 -0.08 15.5 8.4
4/10 -0.062 91.0 96.8 -0.42 -0.14 39.6 13.6
3/10 -0.014 89.7 90.2 -0.51 -0.47 47.1 43.6
2/10 -0.085 82.1 52.6 -0.97 -1.00 84.9 71.5
1/10 -0.018 80.6 23.8 -1.00 -1.00 87.5 71.5
0/10 0.045 84.2 24.8 -1.00 -1.00 87.6 71.5
50
We continue repeating this until at expiration, when T - t = 0, the price of the stock
ends up in the money at 84.2, the price where we have to buy it back, while the final cash
position is 87.6. This means that the final payoff is
87.6 - 84.2 = 3.4
If we had instead bought a put option at the Black–Scholes price of $0.80, the payoff
would have been
90 - 84.24 - 0.80 = 4.96
So why the discrepancy? If we had delta hedged continuously, as assumed by the Black–
Scholes model, the outcomes would have been identical. However, with one week pass-
ing between re-hedging, the error becomes non-trivial.
Endogenous prices
Third week. We start by drawing a random shock,
0.252
P ∼n a 0, b
52
resulting in 0.022. In the exogenous case this would make the price rise to 100.6; how-
ever, we sold at the end of the previous week and (3.3) indicates that the price falls
further to 98.1.
This continues for a few weeks, but it is really by the eighth week that things heat up
and the downward spiral begins to gather momentum.
Eighth week. The random shock is -0.085, coming after a price drop of 6.6 over the
previous week, resulting in a large negative price movement, all the way to 52.6. Delta
is on its way to become -1, which is obvious from Figure 3.5 since we are so deep in
the money.
This continues on in the following week, slightly reversing in the last week. At that
time, the actual price has fallen to 24.8 and the cash balance is 71.5. This means the profit
from the strategy is
71.5 - 24.8 = 46.7
If we had instead purchased a put option, and the prices had followed the same pattern,
the profit would have been
90 - 24.8 - 0.80 = 64.4
Here the difference between the dynamic trading strategy and an outright purchase of the
option is much larger than in the exogenous case. This is not surprising, since it follows
from Figure 3.5 that if prices move a lot, the delta approximation becomes less accurate.
This is an illustration that the Black–Scholes formula is underpricing the put option when
there is feedback.
Analysis
We summarise these results in Figure 3.6 which shows the two price columns from Table 3.2.
The dynamic strategy amplifies the price movements, both going up and going down.
If the price shocks are small, it doesn’t make much difference, but by the middle of the
51
100
80 Theoretical price
Actual price
60
40
9/10 8/10 7/10 6/10 5/10 4/10 3/10 2/10 1/10 0/10
T−t
Figure 3.6 Dynamic replication strategy
Price falls
∆ falls, sell
Figure 3.7 Endogenous dynamics of delta hedging
period we get a few small but negative price shocks in a row. Initially, it does not matter
much, but all of a sudden the endogenous shock kicks in and prices fall sharply over the
span of two weeks. This is captured in Figure 3.7. This is exactly how many market crashes
happen, and demonstrates the importance of considering endogenous feedback.
3.4.2 Endogenous risk and the 1987 crash

Estimates in 1987 indicated that around $100 billion in funds were following formal port-
folio insurance programmes, representing around 3% of the pre-crash market capitalisa-
tion. However, this is almost certainly an underestimate of total selling pressure arising
from informal hedging techniques such as stop-loss orders.
In the period from Wednesday 14 October to Friday 16 October, the market declined
by around 10%, with sales dictated by dynamic hedging models amounting to around
$12 billion (in either cash or futures), but the actual sales were only around $4 billion.
This means that by Monday morning, there was a substantial amount of pent-up selling
pressure, causing the Dow Jones to fall by over 20% on Monday 19 October.
Trading rules
The stock market crash of 1987 is a classic example of the destabilising feedback effect
on market dynamics of concerted selling pressure arising from certain mechanical trading
rules, like the sell-on-loss considered here.
What is especially striking in this example is that the underlying destabilising behav-
iour is completely invisible so long as trading activity remains below some critical but
unknown threshold. It is only when this threshold is exceeded that the endogenous risk
52
3.5 Actual and perceived risk and bubbles
becomes apparent, causing a market crash. This clearly demonstrates the difference
between perceived risk and actual risk.

Endogenous risk is relevant when individuals not only observe and learn what is happen-
ing in the environment around them but also react to what they observe, and thus affect
their operating environment. By exerting a significant price impact, market participants
change their environment. This means that the financial system is not invariant under
observation.
By incorporating endogenous risk, we quickly arrive at situations whereby we cycle
between virtuous and vicious feedbacks. When things are good, we are optimistic and buy,
which endogenously increases prices, with a bubble feeding on itself. This eventually goes
into reverse, and negative news feeds on falling prices, with the markets spiralling downwards.
This manifests itself in the difference between the risk reported by most risk forecast
models – perceived risk – and the actual underlying risk that is hidden but ever present. As
a bubble is building up, perceived risk is low and falling, whilst actual risk is increasing. After
the bubble bursts, prices will not continue falling so actual risk falls, but because observed
volatility increases, so does perceived risk. We illustrate this phenomenon in Figure 3.8.
Consider first the evolution of prices, shown by the blue line. They start low and
increase at an ever more rapid rate, peaking around period 13, after which they collapse
and remain thereafter constant. This is typical of a bubble, where we say the prices go up
by the escalator and down by the elevator (lift). We see many examples of such price pat-
terns in this book, just one being the 1987 crisis shown in Figure 3.4.
Perceived risk
The two risk lines are even more interesting. The red line labelled perceived risk shows
how market participants view risk when using typical risk forecast models used by the
industry. As the prices increase, perceived risk falls.
When market participants observe increasing prices and falling risk they are encouraged
to continue buying, an example of a momentum strategy. In the short run, this becomes a
virtuous cycle of ever increasing prices and lower risk.
9 Prices
Perceived risk
7 Actual risk
1 3 5 7 9 11 13 15 17 19
Figure 3.8 Endogenous bubble
53
Eventually, the bubble bursts. It is as if somebody yells ‘the Emperor has no clothes’
with traders realising there is nothing fundamental behind these high prices. This prompts
them all to sell at the same time, causing prices to collapse. Because the fall in prices
makes perceived volatility increase, risk forecast models will report sharply increasing risk.
Actual risk
What happens to the actual risk? We show that in the green line. Actual risk builds up
before market prices shoot up, eventually indicating a constant high probability of a crash
in the near future. Eventually, as the market collapses, so does actual risk.
This means that perceived risk sends the wrong signals in all states of the world. Before
the crisis, it is biased downwards, giving a too optimistic view of the world, and after the
main crisis event, it becomes too high, making us too pessimistic. This is one manifesta-
tion of pro-cyclicality.
This is one of the main explanations for the phenomenon observed by Andrew Crockett
(2000), mentioned earlier:
‘The received wisdom is that risk increases in recessions and falls in booms. In
contrast, it may be more helpful to think of risk as increasing during upswings, as
financial imbalances build up, and materialising in recessions.’
Active risk management and empirical predictions

These results have a direct impact on how risk management is generally practised in finan-
cial institutions. We can expect banks to become increasingly good at managing day-to-
day risk, but at the expense of an increased probability of very large and uncommon
outcomes. Borrowing language from the statistical literature on risk modelling, we can say
that active risk management may cause volatility to decrease and the tails to thicken.
We see a depiction of this in Figure 3.9. The blue line shows what the distribution
of market outcomes could be if financial institutions were not actively trying to manage
With active Without active

risk management risk management
0.5
Risk level targeted
Risk of very large
risk management
and uncommon
outcomes
0.4
by active
Distribution of risk
0.3
0.2
0.1
0.0
−3 −2 −1 0 1 2
Market outcomes
Figure 3.9 Impact of active risk management
54
day-to-day risk, whilst the red line shows what happens when they do. The impact is to
reduce the probability of uncommon events, shown by the red line being below the blue
line in the sides of the distribution, but at the expense of the red line being higher in the
tails, signalling the higher probability of extreme outcomes. Keep in mind that the area
under both functions must be 1 because these are probability distributions.
3.5.1 The crisis from 2007

The causes of the crisis that started in 2007 show many elements of the distinction
between perceived and actual risk. The crisis was initially triggered by turmoil in the US
subprime market, but quickly spilled over to most other markets, causing a massive wipe-
out of wealth across the globe.
Up until 2007, US house prices had been rising relentlessly, as we can see in Figure 3.10,
fuelled by low interest rates and an explosion of money coming in because of securitisa-
tion. The popularity of the various types of structured credit products ultimately led to
a flood of cheap credit and allowed NINJA (‘no income, no job, no assets’) borrowers to
become home owners.
The first sign of an impending problem was noted in February 2007. Moody’s put 21
US subprime deals on ‘downgrade review’, indicating an imminent downgrade on these
bonds. This led to a deterioration of prices of mortgage-related products.
As mortgage-related structured credit products fell in price, lenders to the purchasers
of these products started to demand higher margins, causing difficulties for many highly
leveraged traders, forcing many to sell. This spread, with forced sales depressing prices,
leading to more margin calls, and so on. A vicious feedback loop was formed, causing the
bubble to burst. Endogenous risk was doing the harm.
One of the implications of a highly leveraged market going into reverse is that a moder-
ate fall in price is unlikely. Prices either remain unchanged or crash. This is because of the
mutually reinforcing effects of selling into a falling market, causing a firesale externality.
Figure 3.10 also provides a real-world depiction of the phenomena shown in
Figure 3.8, both for prices and especially for perceived risk. Perceived risk – delinquency
rates – was very low between 2005 and 2007; meanwhile significant problems were
building up in the real-estate market, clearly highlighting the distinction between per-
ceived risk and actual risk.
10%
200
8%
Case−Shiller
Case−Shiller
Delinquency
180
160 Delinquency 6%
rate
140 rate
4%
120
2%
100
2000 2002 2004 2006 2008 2010 2012
Figure 3.10 US real-estate prices (Case–Shiller index) and delinquency rates

Data source: Standard & Poor (S&P) and Board of Govermors of the Federal Reserve System
55
3.6 The LTCM crisis of 1998

One of the most turbulent times in financial markets prior to the recent crisis episodes
was the summer of 1998.3 For a lucid description of events see, for example, Lowenstein
(2000).
Long term capital management (LTCM) was the most celebrated hedge fund of its time.
It was founded in 1994 with illustrious partners such as Myron Scholes and Robert C.
Merton of Black-Scholes fame, who shared the 1997 Nobel Memorial Prize in Economic
Sciences.
LTCM started with over $1 billion in capital, and was very successful initially in spite
of charging exceptionally high fees. A typical hedge fund charges 20% of profits earned
plus 1% of an investor’s assets as fees. LTCM charged 25% and 2% and investors were
required to commit their funds for at least three years.
In the first two years investors received profits of 43% and 41% after fees. $10 million
invested in 1994 was worth $40 million four years later. In September 1997, LTCM’s net
capital was $6.7 billion, leveraged to $126.4 billion. They returned $2.7 billion to inves-
tors in December 1997, to focus on investing their own money.
An impression was created that the firm would make extraordinary profits from techni-
cal expertise unavailable to anyone else. Myron Scholes summed up the strategy thus:
‘LTCM would make money by being a vacuum sucking up nickels that no one else
could see.’
LTCM operated with high leverage, at around 30. It drove very hard bargains on financ-
ing and was able to get low rates and special deals because its prime brokers did not
want to get left out of LTCM business and were prepared to cut corners both on the rates
charged and on the amount of collateral demanded.
Trading strategies
The mainstay of LTCM’s trading strategies was convergence or relative value trades, in
which a long position in one asset is hedged by a matching short position. The under-
lying principle is that two fundamentally identical assets should have identical prices,
otherwise there is an arbitrage opportunity. Their VaR in 1998 indicated that it would
take a 10s event for it to lose all capital in a single year. Since it makes sense to analyse
sigma events only under the normal distribution, this implies a probability of default
of 7.6 * 10-23 . By comparison, the Earth is 4.5 * 109 years old and the Universe is
1.3 * 1010 years old.
LTCM’s very success attracted copycats. As more and more players with similar trading
strategies crowded into the market, spreads narrowed, eroding profit margins.
3
The events are well summarised in two official reports by the Bank for International Settlements (BIS)
(Committee on the Global Financial System, 1999) and the International Monetary Fund (IMF) (1998).
The first is at www.bis.org/publ/cgfs12.htm, and the second in Chapter III of www.imf.org/external/
pubs/ft/weo1298/index.htm.
56
3.6 The LTCM crisis of 1998
Central bank of volatility

By 1998, LTCM took large positions in areas such as merger arbitrage and short options
on the S&P index. In other words, they were net short of the S&P volatility, that is the VIX
index, seen in Figure 3.11.
The long-run average VIX until the middle of 1998 was 17%, and as we see in Fig
ure 3.11(b), the VIX exceeded that level throughout 1998.
Note that the statistical properties of volatility are different from those of some other
financial instruments, such as stock prices or exchange rates, because it is unlikely that
volatility will become very low or very high, at least not for any appreciable amount of
time. In other words, volatility is mean reverting.
With volatility much higher than the long-run mean, it was quite likely that one could
make money by betting on mean reversion, exactly what LTCM intended, becoming a
major supplier of S&P 500 vega.4 In order to make significant profits, LTCM needed very
high degrees of leverage.
A perfect storm
In May and June 1998 fund returns were -6.42% and -10.14% respectively, reducing LTCM’s
capital by $461 million. This was further aggravated by the Russian financial crises in August
and September 1998, when the Russian government defaulted on their government bonds
on 17 August. This triggered a panic: investors sold Japanese and European bonds to buy US
30
20 Long run mean = 17
10
Jan 1990 Jan 1992 Jan 1994 Jan 1996 Jan 1998
(a) 1990 to June 1998
40
30
20
Jan Mar May Jul Sep Nov

(b) 1998
Figure 3.11 VIX
4
Vega is the sensitivity of the Black–Scholes price of an option with respect to volatility, that is, the first
derivative of the Black–Scholes equation with respect to volatility.
57
Treasury bonds (flight to quality), and hoped-for profits from convergence trades became
large losses, as the value of the bonds diverged. Meanwhile, credit spreads widened, volatil-
ity shot up to 45%, and correlations went to 1, as usually happens in crises.
In the first three weeks of September, LTCM’s equity tumbled from $2.3 billion to $600
million, leading to a sharp increase in its already high leverage, further increasing its fra-
gility. Eventually, it had to unwind the convergence trades, sell the long positions, and
buy back the short positions. This caused an adverse price shock for all other traders with
similar positions and for some triggering margin calls. A vicious feedback loop was set in
motion in which adverse price moves led to liquidations, which further fed adverse price
moves. Schematically, this can be seen in Figure 3.12.
Realising the potentially disastrous consequences of letting LTCM fail, the New York
Federal Reserve Bank (NYFed) organised a bailout of $3.625 billion by the major credi-
tors of LTCM to avoid a wider collapse in the financial markets. In return, the participat-
ing banks got a 90% share in the fund and a promise that a supervisory board would be
established. After the bailout, the panic abated, and the positions formerly held by LTCM
were eventually liquidated at a small profit to the new owners.
Analysis
What befell LTCM was a classic example of endogenous risk. The owners of LTCM main-
tained that it was quite safe as perceived risk was low. Investors could expect losses in
excess of 20% in one year out of 50.
One LTCM partner said that ‘what happened to LTCM was a perfect storm – a
100-year flood’. But does that stand up to scrutiny? The endogenous risk analysis
above suggests otherwise: the unprecedented price movements in the summer of
1998 were not the results of extremely bad luck, in the same way that the Millennium
Bridge did not wobble because of bad luck. Instead, it was only a matter of time
before a small exogenous shock would hit LTCM, triggering its failure. Once the bub-
ble started collapsing, the internal dynamics of the feedback loop took hold with a
vengeance. Under the right conditions, the crisis was a near certainty, with endog-
enous risk doing the harm, because of the endogenous feedback loops between mar-
gin calls and distress.
Deleveraging
Margin calls
Adverse price move
Distress
Figure 3.12 The vicious circle of deleveraging and margin calls
58
3.7 Conclusion
Irrationality of markets?
LTCM invested in a mean reverting asset whose value was far away from the long-run
mean. We would expect that given time, the VIX would eventually fall, bringing signifi-
cant profits to LTCM. Indeed, profits were made, but only by those who bailed LTCM out.
So why did LTCM not profit from this? The explanation is provided by an observation
often attributed incorrectly to Keynes, who supposedly said ‘The market can stay irra-
tional longer than you can stay solvent.’ The very high levels of VIX were explicitly caused
by the uncertainty created by LTCM’s existence. That meant that a necessary condition for
the VIX to return to its long-run mean was the failure of LTCM.
3.7 Conclusion
Most risk models and pricing models used by the financial industry and financial supervisors
assume risk to be exogenous, that is, risk arrives to the markets from outside but is not endog-
enously generated within the financial system. While this assumption is clearly incorrect, it is
usually accepted as a necessary evil because modelling endogenous risk is very hard.
The concept of endogenous risk is not confined to finance. For example, the Millennium
Bridge wobbled when it was opened in 2000 because of endogenous feedbacks.
One reason why endogenous feedbacks are so prevalent within the financial system is
the constraints imposed on market participants, like leverage constraints. Endogenous risk
is also created when a substantial number of market participants mechanistically follow
the same trading rules, which is exactly what caused the biggest stock market crash in
history in 1987.
The most infamous hedge fund failure in history, that of LTCM, happened because
the managers of the fund did not understand the endogenous risk it created, and naively
assumed that in essence it could not fail. That led them to take ever more risk, ensuring
the fund’s eventual failure.

1 Define the terms endogenous and exogenous risk, and provide examples of each.
2 The former general manager of the BIS, Andrew Crockett stated in 2000:
How does his view on risk relate to endogenous risk, and what is the implication for
governments’ policies on financial stability?
3 It has been said that financial risk models are least reliable when needed the most.
Do you agree?
4 Discuss financial bubbles, considering both prices and risk, from the point of view of
endogenous and exogenous risk, using the terminology of actual and perceived risk.
59
5 To what extent can crisis events from 2007 be explained by the notion of endogenous
risk?
6 To what extent is the sovereign debt crisis an endogenous risk story, and to what extent
is it not?
7 Suppose a hedge fund operates with four times leverage (assets over equity), where its
equity is worth $10 and the current price of the asset it holds is $5.
(a) Suppose the fund is currently meeting its leverage constraint. How many units of
the asset does it told?
(b) Suppose the price of the asset increases to $6 and the fund is a price taker. How
does it respond?
(c) Suppose instead the fund exerts a price impact, whereby the price of the asset
moves by 0.1% of the amount the fund trades. How does it react to the price
increase?
References
Black, F. and Scholes, M. (1973b). The valuation of option contracts and a test of market effi-
ciency. J. Finance, 27: 399–418.
Brady Commission (1988). Report of the presidential task force on market mechanisms.
Technical report, Government Printing Office, Washington DC.
Buffett, W. (2012). Why stocks beat gold and bonds. Fortune, 27 February.
Committee on the Global Financial System (1999). A review of financial market events in
autumn 1998. Technical report, Bank for International Settlements.
rr000921b.pdf.
Danielsson, J. and Shin, H. S. (2003). Endogenous risk. In Modern Risk Management – a History.
Risk Books, London. www.riskresearch.org.
Danielsson, J., Shin, H. and Zigrand, J.-P. (2012a). Endogenous and systemic risk. In Haubrich,
J. G. and Lo, A. W., editors, Quantifying Systemic Risk. University of Chicago Press for NBER;
www.riskresearch.org.
Danielsson, J., Shin, H. and Zigrand, J.-P. (2012b). Endogenous extreme events and the dual
role of prices. Ann. Rev. Econ., 4(1): 111–29; www.riskresearch.org.
Danielsson, J., Shin, H. and Zigrand, J.-P. (2012c). Procyclical leverage and endogenous risk.
Mimeo, LSE, www.riskresearch.org.
International Monetary Fund (1998). World economic outlook and international capital mar-
kets: Interim assessment. Financial turbulence and the world economy. Technical report.
www.imf.org/external/pubs/ft/ weo/weo1298/ index.htm.
Keynes, J. M. (1936). The General Theory of Interest, Employment and Money. Macmillan.
Lowenstein, R. (2000). When Genius Failed – the Rise and Fall of Long-Term Capital Management.
Random House, New York.
Minsky, H. (1992). The financial instability hypothesis. Working paper. Mimeo, Yale University.
Shin, H. S. (2010). Risk and Liquidity. Clarendon Lectures in Finance. Oxford University Press.
60
4 Liquidity
Liquidity is essential for the functioning of the economy. It increases the efficiency
of markets, makes individual assets more attractive to investors, enables financial
institutions to withstand maturity mismatches and gives central banks the ability to
conduct monetary policy. Unfortunately, liquidity is transient and has a tendency to
evaporate suddenly.
It is quite easy to miss the importance of liquidity. When it is ample, which is usu-
ally the case, markets function well. Herein lies a hidden danger. When market par-
ticipants and governments take liquidity for granted, they assume more risk than is
prudent. This means that over time hidden liquidity risk builds up. Eventually, the risk
becomes so high that a crisis is inevitable. When it finally happens, liquidity becomes
of paramount importance. Unfortunately, when liquidity hits the newspaper head-
lines it is too late: a crisis is already under way.
It is easier to talk informally about the importance of liquidity than to formally
conceptualise it, making it hard to measure and manage. Many a disagreement has
arisen because different problems are analysed under the mistaken assumption that
they are the same. This is no wonder, since the notion of liquidity is elusive and poorly
understood both in the economic literature and by practitioners.

Several other chapters expand on the various aspects of liquidity in much more detail,
for example Chapter 3 (endogenous risk) and Chapter 14 (bailouts), whilst some
chapters draw directly on concepts introduced here, especially the various crisis-
focused chapters like Chapter 6 (the Asian crisis of 1997 and the IMF), Chapter 17 (the
Chapter 4 Liquidity
ongoing crisis: 2007–2009 phase) and Chapter 19 (sovereign debt crises). Liquidity
also impacts on the ongoing developments in financial regulation as described in
Chapter 18.
Key concepts
■ Market liquidity and funding liquidity
■ Limit order markets
■ Feedback loops
■ Cash-in-hand pricing

The most commonly used notions of liquidity are based on the Basel Committee on
Banking Supervision (2006). Two early influential papers on liquidity are Allen and
Gale (1994) and Kyle (1985), whilst recent analysis, taking into account the crisis from
2007 and endogenous risk, are Brunnermeier and Pedersen (2009) and Danielsson
et al. (2012b).

Q Quantity (units of assets held)
4.1 The liquidity crisis of 1998

By the mid-1990s the world’s main financial markets had shown considerable stability for
quite some time. There had been crises, but they mostly involved falling asset prices, such
as in 1987 or happened in developing countries, for example, in Asia in 1997. Then the
unthinkable happened in the autumn of 1998, when liquidity in the world’s most mature
markets suddenly collapsed. The then Chairman of the Federal Reserve System (Fed),
Alan Greenspan, stated that ‘we do not as yet fully understand the new system’s dynam-
ics’. We follow the story as told by Davis (1999).
Chain of events
From the mid-to late 1990s, the bond and equity markets of the United States (US) expe-
rienced a long bull run, growth was steady and inflation low. This was sometimes referred
to as ‘a new era of high productivity’, a phrase that was echoed a decade later in the ‘great
moderation’. Spreads were falling, and the issuance of debt securities saw strong growth,
reflecting a move away from more traditional bank lending.
The Asian crisis which started in July 1997 had initially little effect on OECD countries
until it become apparent in July 1998 that recovery would take longer than expected.
Banks with exposure to the Asian markets saw their share prices drop and corporate bond
spreads widened in the first half of August 1998 on concerns about US growth, a sign of
the impending flight to quality and liquidity.
The trigger for a major crisis was when Russia declared a moratorium on its sovereign
debt and devalued the rouble on 17 August. This accelerated the repricing of emerg-
ing market risk and led to large-scale deleveraging. A flight to quality caused yields on
62
4.1 The liquidity crisis of 1998
US government bonds to fall and spreads on less safe bonds to increase, reaching an
eight-year peak. Figure 4.1 reveals how sudden these events were; issuance of new
debt fell from $228 billion to negative $156 billion, signalling very severe and rapid
deleveraging.
It was then that LTCM ran into trouble. The uncertainty created by its pending col-
lapse, and the resulting impact on the markets, triggered a global collapse in liquidity
which was the motivation for its bailout, organised by the New York Federal Reserve Bank
(NYFed).
The stable pre-crisis environment resulted in low margin requirements and the increase
of leverage. It was not surprising, therefore, that highly leveraged investors suffered heavy
losses during the crisis. They had to meet margin calls, further aggravating the crisis, in an
example of an endogenous risk feedback loop.
A clear example of the liquidity problems can be seen Figure 4.2 which shows the rapid
increase in spreads as the crisis unfolded. This parallels the impact on volatility (VIX) we
saw in Figure 3.11. The crisis spread to even supposedly safe assets, like US Treasuries, as
investors cared only about investments that could be monetised quickly. This rejection of
even the safest of assets is a typical characteristic of liquidity crises.
200 bn
100 bn
0 bn
−100 bn
Q3/97 Q4/97 Q1/98 Q2/98 Q3/98 Q4/98
Figure 4.1 Net funds raised in US capital markets, USD billions

Data source: US Flow of Funds
150
Spread (bp)
100
50
0
Q3/97 Q4/97 Q1/98 Q2/98 Q3/98 Q4/98
On−off the run Treasuries CDs
Commercial paper Corporate bonds
Figure 4.2 Spreads of various debt instruments (bp = basic point, one-hundredth of a
percentage)
Data source: Bank for International Settlements (BIS)
63
Chapter 4 Liquidity
As the International Monetary Fund (IMF) (1998) noted, ‘... markets became one-sided
until prices declined enough to bring buyers back ...’. The Wall Street Journal reported that
market makers ‘cut back on the size of the trades, quoted wider bid–offer spreads or did
not quote at all’ (as noted by Davis, 1999), further contributing to the general illiquidity.
The crisis finally ended when LTCM collapsed and the Fed, along with other major cen-
tral banks, lowered interest rates, increasing liquidity. The pace of recovery did surprise
many commentators at the time. However, for a pure liquidity crisis, where no productive
resources are destroyed, that is to be expected.
Analysis
The 1998 liquidity crisis highlights the common features of all liquidity crises, and many
a banking crisis. Securities markets usually help investors and borrowers to engage in
maturity transformations and help in ensuring against idiosyncratic liquidity shocks. One
criterion in investors’ decisions is based on the convertibility of assets into cash, that is
liquidity. The more liquid an asset, the less it yields, other things being equal. If investors
fear a liquidity crisis, their rational response is to sell before anybody else. If everybody
thinks the same, the result is a rush to sell, ending in a crisis.
Before a crisis, financial institutions often seem to assume that liquidity is infinite, that
they can always execute their trading strategies and borrow what is needed. Unfortunately,
these assumptions hold only when things are good, at the time when increasing prices
create an illusion of credit-worthiness, resulting in seemingly infinite liquidity. However,
during crises, such virtuous feedback loops suddenly turn vicious, causing distress.
Why did the safeguards not work?

The 1998 liquidity crisis was typical of liquidity crises, just like those of 1914 and 1863 and
many others. These crises have the same underlying fundamental cause – liquidity – even
if the details differ. After every crisis there is a clamour for safeguards preventing future
crises, often ineffectively, as the next crisis takes new forms. This does not mean that the
safeguards do not work: we do not see the successes, only the failures. After 1998 new
extensive crisis prevention mechanisms were put in place, though it took only nine years
for the next liquidity crisis to occur.
Fighting crises is like fighting bacteria. We are able to develop a medicine that fights the
parasite in the short run, but eventually the bacteria evolve and the medicine becomes
ineffective. It is the same with policy: we can prevent the old crises from recurring, but the
next one will simply take a new form.
The reason is that regulations look backwards at the causes of previous crises, while the
next invariably takes a new form. Market participants seek to take more risk than is consid-
ered prudent from society’s point of view, and have a direct incentive to find loopholes in
the safeguards, taking excessive risk exactly where nobody is looking. Often, it takes a fresh
crisis to expose the risk, by which time it is too late to do anything about it. In a way, we
could say that designers of financial regulations face the same problems as successful gener-
als who only prepare to fight the last war, whilst the enemy changes the rules of the game.
In this, 1998 was no different. The previous crises facing Europe and the US had been
the S&L banking crisis in the US in the 1980s, the 1987 crash and the Scandinavian
64
4.2 What is liquidity?
banking crisis of the early 1990s. Regulators worked to prevent those from happening
again, for example, by further developing prudential regulations and in the US prompt cor-
rective action.
The 1998 crisis happened elsewhere. It was a conjunction of several events, most
importantly the Asian crisis, the Russian crisis and LTCM. Therefore, it is not surprising
that decision makers were caught off guard. Once the problem was identified, the NYFed
acted decisively and correctly to address the problem.
This left us with the view that the next crisis would involve hedge funds. Many safe-
guards to prevent hedge fund crises were put into place, and the role of the prime broker
(the hedge fund bankers) was beefed up. However, the crisis that started in 2007 hap-
pened elsewhere, in the most regulated part of the financial system, banking, which had
been thought to be kept safe thanks to the Basel Accords.

There is no single answer to this question. Those concerned with liquidity often focus
on one particular aspect of it, at the expense of the overall picture. This may happen
because liquidity takes many different forms and different pundits may use very different
language when talking about essentially the same concept, perhaps appearing to strongly
disagree while fundamentally having the same view.
4.2.1 Funding liquidity and market liquidity

Currently, most discussion of liquidity follows the Basel Committee for Banking Supervision
(BCBS) in 2006, which indirectly defined liquidity by means of liquidity risk, decomposed
into two interrelated parts.
Definition 4.1 Funding liquidity and market liquidity Funding liquidity

risk is the danger that a firm may be unable to obtain the cash it needs to maintain smooth
operation of its business activities.
Market liquidity risk describes the possibility of being unable to reverse or eliminate
a position in the securities markets without significantly and adversely affecting the price.
Funding liquidity refers to the ready availability of cash to finance day-to-day opera-
tions, and market liquidity to the ease of buying and selling securities at a fair price.
These two notions of liquidity are of course highly interdependent. When things are
going well, cash is ample and trading is easy, whilst during market turmoil and especially
crises, financial institutions simultaneously will find access to cash curtailed and trading
difficult. Therefore, market liquidity and funding liquidity impact on each other, either
virtuously or viciously.
Regardless, policy initiatives directed at liquidity often focus exclusively on fund-
ing liquidity, ignoring market liquidity. One reason is that risks associated with funding
65
Chapter 4 Liquidity
illiquidity have been acknowledged for several hundred years, but market liquidity is a
newer topic and perhaps harder to understand.
4.2.2 Distinction between illiquidity and insolvency

In the analysis of liquidity, it is important to identify the distinction between insolvency and
illiquidity. A firm that is insolvent is bankrupt. It owes more money than it has assets, assets
minus liabilities are negative and all capital has been exhausted. Generally, banks are closed
down long before this happens. If a bank gets close to hitting its regulatory 8% capital ade-
quacy ratio (CAR), warning bells are likely to ring with the supervisor who then takes action.
Illiquidity is where assets minus liabilities are positive, but the bank does not have
enough liquid assets on hand to pay off creditors. In the simple case of a bank collecting
demand deposits and making long-term loans, if enough depositors demand their money
at the same time and the bank is unable to monetise the assets quickly enough, the bank
faces an illiquidity crisis, which could trigger its failure, even if it is not technically bank-
rupt. The reason is that if a firm fails to meet a single payment, all of its obligations might
be considered in default, perhaps because of pari passu clauses.
Illiquidity means that banks cannot raise cash by selling assets, because they cannot get
acceptable prices in the short run. In a clear liquidity case, it should be possible to set a
timescale in weeks or months for the required sales to be implemented and the cash raised.
In a clear insolvency case, the market value of a bank’s aggregate asset holdings is below the
cash it needs, even before sales are attempted. Naturally, many cases fall in between.
In practice, it can be difficult to distinguish between illiquidity and insolvency. Suppose
a country is facing a financial crisis and one of its banks is in difficulty. Before the crisis,
this bank was considered very solid, with no visible problems. However, it has the same
fragility as all banks, such as significant maturity mismatches. During the crisis, it needs to
meet demands for cash, but because of the crisis it finds it is much harder to monetise its
assets, and the bank may only be able to get firesale prices for them. This could even apply
to assets that are otherwise safe, like government bonds. Is this bank insolvent or only
illiquid? It depends on whether the value of the assets is low because of special circum-
stances during the crisis or because they were just overvalued before and now are getting
to more realistic valuations. It is often impossible for the authorities to make a distinction
between insolvency and illiquidity in real time when fighting a crisis.
If the authorities believe that the only issue is one of liquidity, then supporting their
banking system is a natural decision; support can be given on a temporary basis, and the
money lent can be recouped later when pricing returns to normal. With insolvency, how-
ever, the costs will be greater because the difference between assets and liabilities will
not be recoverable.
4.2.3 Funding liquidity

Of the two notions of liquidity, the more traditional and straightforward is funding liquid-
ity. It relates to the availability of funds, both for individual financial institutions and also
for the financial system as a whole.
66
Funding liquidity and crises

Funding liquidity risk often arises because of maturity mismatches, for example, when
a bank borrows short term and lends long term. This could arise in traditional banking
operations where banks collect demand deposits and make long-term loans, but could
happen just as easily in most funding models.
An economy that suffers from a scarcity in funding liquidity is usually also experienc-
ing a sharp reduction in money supply – reduction in liquidity – arising from fractional
reserve banking and how M0 gets converted into M1 and M2. Because of the money
multiplier, if the financial system is deleveraging – converting fewer liquid assets such as
bonds into cash – the money supply contracts sharply, with adverse consequences for the
real economy.
During such crisis episodes, there are often calls for liquidity injections by central
banks, perhaps in the form of temporary increases in the money supply by means of open
market operations, interest rate reductions, direct lending to financial institutions in the
interbank market, lending of last resort (LOLR) or direct bailouts.
A similar problem arising from liquidity happened in the Asian crisis of 1997. Several
countries of East Asia had resorted to short-term borrowing from international capital
markets to finance longer-term domestic investment. Eventually, a lack of confidence
caused the creditors to refuse to roll over loans, triggering a crisis.
4.2.4 Market liquidity

The second notion of liquidity, market liquidity, is less well understood and refers broadly
to the ease of buying and selling financial instruments. If one is able to buy or sell infinite
amounts of an asset in financial markets at the same price at any time, that asset can be
said to be perfectly liquid. Of course, no such asset exists in the world, even though many
come close, at least enough for the needs of all but the very largest entities.
Increasing importance of market liquidity

Market liquidity has gained more attention over time. Banks are increasingly active in
securities markets, involved in not just underwriting and issuance, but also trading,
market making and providing backup credit lines to clients. Therefore, if the market
becomes illiquid, banks suffer losses on both their positions and market-making activi-
ties, and depositors may lose confidence and run the bank. The popularity of securiti-
sation has also led to a greater dependence on liquidity. One example is the funding
difficulties experienced by many banks in 2007. Non-financial companies may also
depend on markets for funding, providing a direct channel from liquidity crises to the
real economy.
Moreover, the rapid growth of the derivatives markets has also raised concerns because
of its particular susceptibility to liquidity risk. Over-the-counter (OTC) derivatives are illiq-
uid by definition, as they are bespoke instruments. More crucially, many banks use deriva-
tives such as swaps and options for risk management and hedging. A lack of liquidity not
only exposes banks to significant risk, but also increases the possibility of other counter-
parties refusing to trade, spreading the illiquidity further afield.
67
Chapter 4 Liquidity
4.2.5 Limit order markets

A key ingredient in the concept of market liquidity is price impact. If we sell an asset, its
price drops, and if we buy the asset, the price increases. This is clearly seen in markets using
limit order books, which includes most actively traded assets. A limit order market is a com-
puter system where traders can make two types of trades, limit orders and market orders.
Definition 4.2 Limit order A limit order is an order that specifies an offer to buy or to
sell the security, the volume to be bought or sold and a price at which to execute a transaction.
Definition 4.3 Market order A market order specifies buying or selling immediately
at the best price.
Market orders usually execute immediately upon entry, provided there are existing limit
orders in the system. By contrast, limit orders are executed only if they are hit by a market order.
There is not a single price for a security in limit orders markets, but two: one price at
which you can buy a security, called the ask or offer, and another, lower price at which
you can sell the security, called the bid.
Almost all exchanges now use limit order books. This includes assets such as foreign
exchange, equities and many of the more common forms of derivatives. A more traditional
form of exchanges uses a market maker, but this is becoming less common over time.
It is perhaps best to explain the mechanics of limit order books by means of an example.
We see from the example that a sufficiently large market order will exhaust the supply
of limit orders at that particular price, causing the price to move, an example of a price
impact. As is apparent from Figure 4.3, the price impact for an order of a given volume
Price
depth of
best asks
$4.07
$4.05
bid–ask spread
$4.04
$4.02 depth of
best bids
$4.00
0 50 100 Q 150 Depth

(’000s of shares)
Figure 4.3 Market depth in a limit-order book. The figure depicts the depth of the top
of the book for shares of Citigroup on the BATS Exchange, 08:10 EDT, 21 July 2010
68
Example 4.1 Citigroup

Figure 4.3 shows the limit order book for Citigroup at one point in time in 2010.
The available limit buy at the best bid totals 35,000 shares at $4.04, whilst sell
orders at best equal 41,800 shares at $4.05. More shares are available at inferior
prices.
The figure shows the amount of shares available at particular prices, and the
figure can be interpreted as showing the excess supply and demand curves for
Citigroup.
A market bid received should execute at $4.04 and the next market ask received
should have executed at $4.05. We could write ‘did’ instead of ‘should’ if we knew
that the next market ask was for a volume of 41,800 shares or fewer and the next mar-
ket bid was for 35,000 or fewer shares, and additionally that the next market orders
arrived before the composition of the book changed as traders cancelled their out-
standing limit orders.
For example, a large market ask order for Q = 120,000 shares would execute against
the best outstanding limit asks to a total of 120,000 shares. In the situation shown in
Figure 4.3, where Q is marked with a vertical green line, this would be 41,800 shares
at $4.05, then 66,500 shares at $4.06, and finally 11,700 shares at $4.07. After the
order executes, the best asks would be priced at $4.07.
The best bids would be unaffected, so the spread would have widened from 1 cent
to $4.07 - $4.04 = 3 cents. Similarly, a market bid order for Q = 120,000 shares would
drop the best bid to $4.03 and widen the spread to 2 cents.
is lower in markets with higher volumes of limit orders outstanding, and with larger vol-
umes at and near the current best bids and asks – that is, price impact is smaller in
deeper markets.
Market liquidity terminology

In order to better understand how liquidity relates to financial markets, many researchers
have classified the categories of market liquidity in limit order markets. Of those, the best
known are due to Kyle (1985) (first three in the following list) and Black (1971) (the last
in the list):
■ Tightness is the cost of turning around a trading position in a short period of time.
Tighter markets mean lower trading costs.
■ Depth refers to the size and continuity of the market, in terms of participants and vol-
ume of trading. Large orders can be absorbed by deep markets without affecting the
price much.
■ Resiliency refers to the speed with which the market price recovers after being driven
away from its intrinsic value by uninformative shocks.
■ Immediacy refers to the time it takes to find a counterparty, initiate, clear and settle a
trade.
69
Chapter 4 Liquidity
Taken together, these notions provide the formal way to understand the various fac-
ets of market liquidity in financial markets. For example, if applied to the quant crisis of
2007, both depth and tightness sharply fell, because of the large amount of orders on the
same side of the market, usually sell, which were unusually large compared to the depth.
Similarly, the investor strike that commenced in the fall of 2007 signified the near disap-
pearance of depth in asset backed securities (ABS) markets.
4.3 Liquidity models

A number of important theoretical models of liquidity have been proposed, helping
to clarify this elusive concept. Below, we discuss two of the most important liquid-
ity models: the interaction between funding and market liquidity, and cash-in-the-
market pricing.
4.3.1 Interdependencies between market and funding liquidity

If only a single financial institution is facing liquidity problems it is usually due to fund-
ing liquidity difficulties. The reason is that if the rest of the market is functioning as
normal, the institution in difficulty would not face any special difficulties in trading,
beyond its own idiosyncratic funding problems. This might be different if the institu-
tion was especially large or in a special situation like LTCM. However, when funding
liquidity problems are widespread, they invariably imply shortages of market liquidity
as well, and vice versa. Market and funding liquidity go hand-in-hand and cannot usu-
ally be separated.
Vicious feedback loops

There are several different ways one can capture this interplay between market liquidity
and funding liquidity. Just one example comes from Brunnermeier and Pedersen (2009),
and another from Danielsson et al. (2012b). In what follows, we draw on both strands of
the literature.
Consider a leveraged trader subject to margin requirements, perhaps a hedge fund or a
proprietary trading desk, making large bets on a stock. At the same time, many other traders
are actively trading large positions in the same stock. Individually, these traders might not
exert that much price impact, but in aggregate they do. In the terminology of endogenous
risk, if something unifies their beliefs and constraints, their behaviour becomes harmonised,
so in aggregate they will make similar decisions and hence exert a significant price impact.
So long as the market is doing well and prices are increasing, some of these traders will
increase their positions whilst others decrease them, because of heterogeneous beliefs,
objectives and constraints. Suppose, however, that the market is hit by an exogenous
shock. Prices fall, the traders are hit by margin calls, and they need to obtain funds to pay
the margin, where the only way to do that is by selling the stock. This happens exactly
when the price is falling, further amplifying the price collapse.
70
4.3 Liquidity models
In other words, the exogenous shock triggers a margin call – a funding liquidity prob-
lem. In order to meet the margin call the traders have to sell the stock into a falling
market – a market liquidity problem. This in turn leads to more margin calls, triggering
more selling, etc. A vicious feedback loop is formed between deleveraging and funding
problems. These feedback loops are further amplified by client redemptions, as investors
have a tendency to redeem from poorly performing funds. Such loops explain much of
the quant event of 2007. Figure 4.4 illustrates these mechanisms.
A fundamental result, shown mathematically by Brunnermeier and Pedersen (2009), is
that the cumulative effect of a margin spiral coupled with a loss spiral is greater than the
sum of their separate effects. This can be seen mathematically by using simple convexity
arguments, and intuitively from the flow diagram in Figure 4.4.
4.3.2 Liquidity and asset pricing

A different way to analyse how liquidity affects the price of assets is the mechanism of
cash-in-the-market pricing proposed by Allen and Gale in a series of papers over the past
couple of decades, for example, Allen and Gale (1994, 2005). In their models, markets
are incomplete and financial institutions may be forced to sell assets to obtain liquidity.
Because the supply of and demand for liquidity is likely to be inelastic in the short run,
small aggregate uncertainties may result in large fluctuations in asset prices, which might
become so severe that financial institutions are unable to meet their obligations, ending
in a crisis.
Initial losses
Reduced positions
Funding problems Prices move away

for speculators from fundamentals
Higher margins
Client
Losses on redemptions
existing positions
Figure 4.4 Vicious liquidity feedback loops

Source: based on Brunnermeier and Pederson (2009) and further modified
71
Chapter 4 Liquidity
Allen and Gale refer to the relationship between liquidity and asset prices as cash-in-
the-market pricing, whereby the price of a risky asset in equilibrium is equal to the lesser
of the present discounted value of future dividends and the amount of cash available
from buyers divided by the number of shares sold – the cash-in-the-market price. The lat-
ter case arises when there is a shortage of liquidity, resulting in assets being underpriced
and returns therefore excessive.
If funding liquidity is scarce, the price of cash gets bid up and the price of the risky
asset bid down. This could happen, for example, if banks hoard cash because of some
precautionary principle. It may even happen with longer-dated government bonds of an
AAA-rated government. We have seen this phenomenon in crises past, and it played a key
role in the events in the fall of 2008, and with some banks in the European sovereign debt
crisis.
One could say that the opportunity cost of holding cash goes down during such crisis
episodes. Not only are there fewer investment opportunities, but financial institutions are
under risk constraints and thus prevented from increasing their holdings of risky assets,
or even longer maturities of safe assets. This creates an endogenous risk feedback loop
between falling prices for illiquid assets, preference for cash and difficulties in holding
illiquid assets, demonstrated in Figure 4.5.
Taken to the extreme, if everybody wants liquidity (cash), the price of risky assets falls
to zero, which in turn creates incentives for some agents to buy those assets, which then
puts upward pressures on prices. If the prices increase too much, everybody prefers the
risky asset, but then no liquidity is provided, so its price gets bid up. Of course, rational
economic agents realise this and, as a consequence, there must be an equilibrium price
where the profit in the states of the world with high liquidity demand is sufficient to com-
pensate for all the other states where no profits are made, so that the agents simply bear
the opportunity cost of holding liquidity.
Initial
Prices falling away
shock
from fundamentals
Risky
assets
become
illiquid
Preference for cash
Risky assets sold
Difficulties in holding
illiquid risky assets
Figure 4.5 Cash-in-the-market pricing
72
4.4 Policy implications
In other words, asset prices are low in states where banks need more liquidity. But this
is exactly the wrong time from an efficiency point of view for there to be a transfer from
the banks needing liquidity to the providers of liquidity.
Liquidity premium
This suggests that assets should bear a liquidity premium, where liquidity, or the lack
thereof, is priced in by the markets. Investors require compensation for taking on liquidity
risk when they purchase securities that may later be difficult or expensive to sell. Thus,
illiquid securities must offer higher returns than liquid assets.
Several empirical studies have shown that liquidity effects are indeed priced. For exam-
ple, Acharya and Pedersen (2005) tested their liquidity adjusted CAPM model on portfo-
lios of stocks varying by firm size and found that illiquid stocks – defined as stocks with
high average transactions costs – are shown to have high liquidity risk. These results help
to explain the higher than expected returns on small-cap stocks and other illiquid invest-
ments such as private equity.
4.4 Policy implications

Theoretical analysis of liquidity suggests that small changes in funding conditions, or
liquidity demand, can lead to sharp reductions in liquidity, and even crisis. This has direct
policy implications.
Central banks’ liquidity injections

Central banks have a direct role in mitigating market liquidity problems by breaking the
vicious feedback loop between market and funding liquidity, preventing a relatively small
event from spiralling into a full-blown crisis. They can do that by boosting market par-
ticipants’ access to funds during a liquidity crisis, or by simply stating their intention to
provide extra funding during times of crisis, which in turn loosens margin requirements
and other constraints.
The liquidity provision needs to be substantial, because if the authorities can only
credibly promise to provide some finite amount of liquidity, that might by itself provide
enough comfort to market participants to ensure that they build up sufficiently large
imbalances so as to exceed the amount of liquidity provided by the authorities. In other
words, an insufficient liquidity guarantee may perversely create incentives for traders to
worsen a crisis. This suggests that the guarantees may have to be infinite.
However, we do not know what that means. The most obvious way a government can
provide infinite guarantees is by printing infinite amounts of money, leading to hyperinfla-
tion. Another way is to make the guarantee uncertain enough to prevent the changes in
behaviour from causing a self-fulfilling crisis event – often referred to as constructive ambi-
guity. In other words, a policy that is too explicit might be counterproductive.
The insufficiency of the IMF facilities in the Asian crisis, and the ineffectiveness of the
seemingly colossal funds promised in the European sovereign debt crisis’ demonstrates
the importance of magnitude and credibility of the liquidity provision.
73
Chapter 4 Liquidity
Moral hazard problems

Central bank liquidity provisions have to be made reluctantly and be of very short term,
otherwise market participants may take them for granted, creating moral hazard and mak-
ing the central bank the liquidity provider of first resort. If the financial industry starts to
rely on the central bank for day-to-day funding, it creates dependency problems that are
very difficult to break. We see a clear example of this in the European sovereign debt crisis,
when the European Central Bank (ECB) is increasingly the only liquidity provider to some
banks and even entire national banking systems.
4.4.1 Liquidity and the crisis from 2007

Prior to 2007, financial regulation focused on credit risk and market risk, with liquidity
risk taking a distant secondary role. The reason was that liquidity crises were considered
infrequent and financial markets efficient providers of liquidity, whilst the nebulous nature
of liquidity made it quite challenging to translate justifiably high-level concerns about
liquidity into hands-on regulations. Since liquidity was not a priority and the problem was
hard, it got neglected. After all, the architects of financial regulations are constantly being
pulled into many directions by the various stakeholders, and liquidity seemed relatively
unimportant.
With the crisis this view has changed significantly. Liquidity is of a first-order priority to
policymakers, and researchers have made significant contributions in understanding the
nature of liquidity.
Regulations
In the ongoing reform of financial regulations, especially Basel III, liquidity is for the first
time an integral part of international regulation, in the form of the net stable funding ratio
(NSFR) and the liquidity coverage ratio (LCR). These aim at bringing the funding needs of
banks more in line with their outflows, both with maturity matching liabilities and assets
and ensuring that banks have sufficient liquidity to withstand a short-term liquidity dry-up.
Macro-prudential policies
Liquidity is at the heart of what makes systemic crises different from more routine market
turmoil. This arises when all financial market participants – not only the banks – with-
draw all possible liquidity from the market. In today’s highly integrated economy, that is
extremely grave.
The challenges in mapping out the financial network mean that even though it is
clearly necessary to consider network effects when formulating macro-prudential policy,
it is less clear how the immense complexity of the interconnections can be handled in
practice. The challenge for the supervisor is that in order to adequately address concerns
about liquidity, not only do they have to collect more and better information about the
amount of risk taken by financial institutions individually, but more importantly they
need to consider the impact of aggregate positions across the financial system. They fur-
ther have to take into account factors that may lead to heavy selling as a response to a
small exogenous shock.
74
4.5 Summary
Many initiatives have been launched for the purpose of mapping out the network infra-
structure of the financial system, with all major supervisory agencies, various partnerships
between public and private entities, private institutions and academic researchers all work-
ing on this problem. This is not an easy task. Interbank exposures, followed by lender/bor-
rower relationships, are probably clearest and easiest to map out, though true exposures
are obscured by collateral and contingent liabilities. Supervisors are able to get access to
every financial transaction made, including quantity and prices and names of counterpar-
ties, and some preliminary steps have been made in making use of this information in
network analysis. It is, however, not sufficient to look at direct exposures, because market
participants have to make decisions about the creditworthiness of other financial institu-
tions, even those to which they have no direct link. These decisions may influence how
reputation is perceived by third parties. This is the problem that was analysed in Figure 1.8.
4.5 Summary
Liquidity is essential for the functioning of financial markets. Because the financial system
works best when liquidity is ample, it is easy to take it for granted, only to miss it during
crises. One reason is that liquidity is a complicated and multifaceted concept, with no
single understanding of what it is.
The most common high-level notions of liquidity were provided by the BCBS, identi-
fying two interrelated facets: funding liquidity and market liquidity. The former relates
to the availability of cash to the markets and the second to the ability to trade. These
are highly interrelated and their disappearance during crisis is often caused by feedback
effects from one to the other. That gives rise to policy actions cutting the feedback loops.
Liquidity directly affects the pricing of financial assets, which can be demonstrated by
cash-in-the-market impacts.
Liquidity has a direct impact on government policy, for example, LOLR provision by
central banks, but carries with it problems of moral hazard. Liquidity has become a prior-
ity since 2007, and many initiatives have been launched trying to understand it and incor-
porate it into macro-prudential policies.

1 Why is it so difficult to formally conceptualise the notion of liquidity?
2 Explain what it means when we say that ‘liquidity is priced’.
3 Compare and contrast the two notions of liquidity proposed by the Basel committee.
4 What is a limit order book, and how can we think about liquidity from the point of view
of limit order books?
5 Describe a mechanism whereby margins and fire selling viciously interact with each
other, creating a crisis.
6 How can the supply and demand for an asset move its price away from the fundamental
price?
75
Chapter 4 Liquidity
References
Acharya, V. and Pedersen, L. H. (2005). Asset pricing with liquidity risk. J. Finan. Econ., 77:
375–410.
Allen, F. and Gale, D. (1994). Limited market participation and volatility of asset prices. Amer.
Econ. Rev., 84: 933–55.
Allen, F. and Gale, D. (2005). From cash-in-the-market pricing to financial fragility. J. Eur. Econ.
Assoc., 3: 535–46.
Basel Committee on Banking Supervision (2006). The management of liquidity risk in financial
groups. Technical report, pp. 1–25.
Black, F. (1971). Toward a fully automated exchange, part I. Financial Analysts J., 27: 28–35.
Brunnermeier, M. and Pedersen, L. H. (2009). Market liquidity and funding liquidity. Rev. Finan.
Stud., 22(6): 2201–38.
Danielsson, J., Shin, H. and Zigrand, J.-P. (2012b). Endogenous extreme events and the dual
role of prices. Ann. Rev. Econ., 4(1): 111–29; download from www.riskresearch.org.
Davis, E. P. (1999). A reappraisal of market liquidity risk in the light of the Russian/LTCM global
securities markets crisis. Bank of England.
International Monetary Fund (1998). World economic outlook and international capital mar-
kets: interim assessment. Financial turbulence and the world economy. Technical report.
www.imf.org/external/pubs/ft/weo/weo1298/index.htm.
Kyle, A. (1985). Continuous auctions and insider trading. Econometrica, 53(6): 1315–36.
76
5 The central bank
The most important single institution in the financial system is the central bank, also
known as a reserve bank. Central banks have a monopoly on the creation of money
and hence play a key role in ensuring price stability, as well as ensuring stable mac-
roeconomic conditions and the soundness of the financial system. Even though these
objectives are often in conflict with each other, they also lead to the same ultimate
objective, succinctly expressed by William McChesney Martin Jr, former head of the
United States (US) Federal Reserve System (Fed) who said that the Fed’s most impor-
tant job is ‘to take away the punch bowl just as the party gets going’.
Central banks were initially created with a narrow purpose in mind, to help com-
merce or, as in the case of the United Kingdom (UK), to provide stable war funding.
Over time, the roles of the central banks have expanded significantly, they acquired
a monopoly on the printing of banknotes, and they have significant autonomy in the
setting of interest rates. Most central banks started out as private institutions, but all
are now directly controlled by the central government, and most are owned by it.
With the expansion of objectives and power of the central banks, they often find
it difficult to reconcile the various tools and objectives. For example, the exercise of
financial stability may require significant injections of liquidity into the economy,
undermining price stability. Over time, as the nature of the challenges facing the
economy has changed, their priorities have shifted. From the second part of the nine-
teenth century, until 1914, financial stability was the main objective of most central
banks as monetary stability was taken care of by the gold standard. During the Bretton
Woods era, and until the 1980s, these roles switched and inflation became the main
Chapter 5 The central bank
problem facing central banks, with financial stability less important because heavy
regulations limited the scope for financial crises.
Eventually, this led to the neglect of financial stability, contributing to the build-up
of systemic risk, and the crisis starting in 2007. This, in turn, has made financial stabil-
ity the main objective of central banks. However, massive injections of liquidity into
the financial system in recent years are likely to make monetary policy yet again the
main objective of central banks.

This chapter directly relates to Chapter 2 (the Great Depression, 1929–1933),
Chapter 7 (banking crises) and Chapter 14 (bailouts).
Key concepts
■ Quantitative easing
■ Open market operations
■ Central bank interest rate
■ Objectives of central banks
■ Independence of central banks

For a good overview of monetary policy, see Lewis and Mizen (2000), and for banking
supervision Goodhart and Schoenmaker (1995) and Goodhart (2002). For a discus-
sion of central banks, see www.centralbanksguide.com.

ap, ag Parameters
it Target short-term nominal interest rate
r*t Real interest rate
yt Logarithm of real GDP
yt Logarithm of potential output
G Money multiplier
D Reserve requirement
pt Inflation rate (GDP deflator)
p*t Desired rate of inflation
5.1 The origins of central banks

The formal objective of the various central banks is specified in their governing law or
internal regulations. The first central bank, the Swedish Riksbank established in 1668,
engaged in collecting deposits, lending and facilitating trade. It pioneered the practice of
fractional reserve banking, which meant it was susceptible to bank runs. Indeed, it was cre-
ated out of Stockholms Banco which was the first European bank to print banknotes and
collapsed because it printed banknotes on a seemingly unlimited scale.
78
5.1 The origins of central banks
The Bank of England

The second central bank was the Bank of England (BoE), created in 1694, sometimes
referred to in Britain as the Bank. It was initially a private bank set up to assist the gov-
ernment with war finance and, in order to facilitate that function, it was endowed with
certain privileges such as permission to issue banknotes.
The Bank obtained a monopoly on issuing banknotes in England and Wales in 1844
with the Peel Act, which fixed the amount of notes that could be issued for a given
amount of gold, thus preventing an excessive expansion of the money supply. Existing
provincial banks could still issue money, with the last bank in England doing so in 1921.
This did not apply to banks in Scotland and Northern Ireland where private banks still
issue money.
The government retained the power to suspend the Peel Act in case of a financial crisis,
which happened a few times, including during the 1914 crisis and the 1866 Overend and
Gurney (O&G) crisis. After the 1866 crisis, the Bank’s role in lending of last resort (LOLR)
was formalised. This can be considered the first modern statutory financial stability func-
tion of a central bank.
The BoE was nationalised in 1946 and it now has two main objectives – monetary sta-
bility and financial stability. The first entails meeting an inflation target set by the govern-
ment. Subsidiary to that, the bank is to achieve ‘high economic growth in a low-inflation
environment’. The UK inflation target is 2% as measured by the consumer prices index
(CPI), but has been significantly above the target in recent times.
The Federal Reserve System

Amongst the major economies, the US was unique in the nineteenth century for not hav-
ing a central bank. While that may not have mattered much most of the time, it was
increasingly felt that a central bank was needed during financial crises. Ultimately, it was a
severe crisis episode in 1907 that convinced lawmakers of the necessity to establish a cen-
tral bank. During that crisis, the private sector provided LOLR, in particular, J. P. Morgan.
The authorities at the time felt that it would be better for a government agency to have
that role rather than a private individual.
As a consequence, the United States established the Fed in 1913. It could not be
called a central bank for political reasons. The system is made up of 12 federal reserve
banks, with each responsible for member banks located in its area. The best known is
the New York Federal Reserve Bank (NYFed). Over time, the Fed’s function has evolved
and it is now directly under the control of the federal government whilst retaining
significant autonomy. It is responsible for monetary policy and financial stability, and
is the supervisor for some banks. The formal objective of the Fed is ‘to promote effec-
tively the goals of maximum employment, stable prices, and moderate long-term inter-
est rates’.
The ownership structure of the Fed is somewhat convoluted. All nationally chartered
banks hold stock in one of the federal reserve banks. These are not regular stocks, since
they cannot be sold or traded nor used to exercise control, but provide a dividend of 6%.
The remainder of the Fed’s profits go to the government. In 2009 over 98% of the divi-
dends went to the government.
79
5.2 Banking supervision

Central banks are sometimes responsible for the supervision of financial institutions, that
is the enforcement of financial regulations. Sometimes it is felt that supervision is an inte-
gral part of the central bank’s core function; at other times it is thought that it should be
left to a separate institution. The discussion below draws on Goodhart and Schoenmaker
(1995) and Goodhart (2002).
Arguments for separation

The main argument for separating bank supervision from monetary policy is that the com-
bination of the two functions may lead to a conflict of interest between the monetary
authority and the supervisory authority. The central bank might desire higher interest
rates to fight inflation, while the supervisor is worried about the adverse impact of high
interest rates on solvency and the profitability of the financial sector. Similarly, a central
bank may want to close a failing bank because of systemic stability concerns, while bank-
ing supervisors may want to rescue it to protect depositors. Clearly, the decision on how
these conflicts should be reconciled is political and will depend on recent experience.
The cyclical effects of financial stability and monetary policy tend to be in conflict
because monetary policy is usually countercyclical, while the effects of regulation and
supervision tend to be procyclical.
If the central bank were the sole supervisor, it would need to be in charge of activities
quite distinct from its core mission, such as consumer protection. Having the same gov-
ernment agency in charge of many different functions is not efficient and is likely to lead
to some being favoured and others neglected. Central banks are already burdened by a
multiplicity of conflicting objectives.
Finally, in the situation where a central bank is in charge of supervision, it will also be
subject to reputation risk, since when banks fail, the supervisor takes the blame, and the
credibility of the central bank is dented. This happened to the BoE following the failure of
BCCI in 1991. Such reputation risk is inevitable for any supervisor, and since supervisors
get blamed for failures but do not tend to get credit for crises prevented, it can make them
excessively risk averse. For a central bank that depends on its reputation for competence
when it comes to monetary policy, such reputation risk is something it may prefer to do
without, and hence the central bank may want to keep supervision in a separate agency.
Maintaining central bank credibility might not be compatible with being in charge of
supervision.
Arguments against separation

If the central bank is only concerned with monetary stability, there is no need for supervi-
sion to be part of the central bank. However, in reality, central banks will have to engage
with financial stability, whether they want to or not, and may be called upon to provide
liquidity support to banks.
In order for the central bank to be responsible for financial stability, it needs to have
detailed knowledge of the financial institutions within its domain, exactly the type of
information provided to supervisors. It might be more efficient, therefore, for the central
bank to be in charge of supervision.
80
5.2 Banking supervision
As universal banking becomes more common, the same bank may be engaged in very
different activities and, therefore, fall under the supervision of several different regulatory
agencies. Having a multiplicity of separate supervisors, all crawling over parts of the self
same institution, is neither efficient nor cost effective.
Example 5.1 The UK and the tripartite regulatory system
In 1998, the incoming government in the UK transferred responsibility for the prudential
supervision of commercial banks from the BoE to the newly created Financial Services
Authority (FSA). This meant that the UK had a ‘tripartite’ regulatory system, which included
the BoE, the FSA and the Treasury. However, when the crises started in 2007, the tripartite
arrangement was found lacking. It failed to identify the build-up of problems in the finan-
cial system prior to the crisis and did not deal adequately with the early stages of the crisis.
A major reason for these failures was that the tripartite arrangement lacked a leader-
ship structure. It gave the Treasury responsibility for maintaining the overall legal and
institutional framework, with the Bank responsible for financial stability but without the
tools and information needed to carry this out adequately; the tools lay instead with the
FSA, which did not bear the responsibility. This meant that nobody was really in charge.
The government partially rectified this in 2009, giving the Bank the statutory objective of
contributing to the maintenance of financial stability. The Bank took over the FSA’s macro-
prudential role in early 2013.
Example 5.2 Supervision in Europe
The situation is more muddled in the euro zone where the European Central Bank
(ECB) has responsibility for monetary policy but supervision is primarily under national
control. Even though it is generally considered desirable for a central authority to be
responsible for supervision within the European Union (EU), or at least the Euro zone,
this is difficult given the current political structure of the Union.
The reason is that when a supervisory authority takes action, it often has substantial
national implications, potentially reducing revenues (and tax income) or committing
large amounts of taxpayers’ money. The government agency in charge of the taxpayers’
funds is the treasury or the ministry of finance. Under the current EU structure, each
country is directly responsible for its own finances, and therefore supervision has to be
the purview of the nation state.
This means that the ECB is in the rather curious situation of being in charge of mon-
etary policy, as well as financial stability, but without any supervisory powers. The EU has
attempted to partially solve this problem by hosting the new European Systemic Risk Board
(ESRB) within the ECB, in principle, providing the necessary information for the ECB to
efficiently exercise its financial stability function. However, with supervision in the hands of
the nation states, its powers are necessarily more limited than those of other central banks.
The EU is currently aiming to set up a common supervisor in Europe, and eventually
a banking union, starting with the single supervisory (SSM).
81
5.3 Monetary policy

The main day-to-day function of central banks is monetary policy, the control of the sup-
ply of money. Monetary policy is either expansionary, where a central bank increases the
total supply of money in the economy, contractionary, when it decreases the money sup-
ply, or neutral.
The most commonly used tool for monetary policy is interest rates, but central banks
may also use open-market operations. A traditional method is reserve requirements, but this
is now more common in emerging markets. A more recent tool is quantitative easing (QE).
Often described as unorthodox monetary policy, it was adopted by the Bank of Japan (BoJ)
in response to their problem with deflation after the crisis in the early 1990s. Similar meas-
ures have since been adopted by many major banks in response to the crisis from 2007.
5.3.1 Central bank interest rate

The most visible demonstration of monetary policy is the setting of interest rates. A num-
ber of different terminologies are used when referring to central bank interest rates and
often these can be confusing and even contradictory. Common names include the target
rate, the short rate, the risk-free rate and the Fed funds rate. We discuss these in detail in
the appendix to this chapter.
The central bank rates determine the overnight risk-free market rate, because it would
not be profitable for banks to conduct business at other rates because of the presence of
such a large counterparty. The central bank rates affect money supply directly. By increas-
ing the interest rate, banks are more likely to deposit money with the central bank, taking
money out of circulation, causing borrowing rates to increase throughout the financial
system, and reducing demand and hence money creation.
Interest rates can be raised without limit and so provide an effective contractionary
tool under inflationary conditions. The contrary is not true, because deflation would
require negative interest rates, something not possible except in special circumstances.
That means different tools are needed to combat deflation.
Central banks exercise only limited control over longer maturities, which are based on
supply and demand in the bond markets, with inflationary expectations an important
ingredient. The central banks’ control of interest rates is sometimes compared to hold-
ing a string on one end. The central bank can control one endpoint, but the other end
of the string does what it wants and intermediate maturities obey only a smoothness
requirement.
Taylor rule
Having a monetary policy objective, perhaps a formal inflation target, does leave open
the question of how the central bank should meet the objective. One approach is the
Taylor rule, proposed by Taylor (1993), whereby the central bank sets the nominal
interest rates based on changes in inflation, output and possibly other economic vari-
ables. Under the rule, the central bank should increase nominal interest rates by more
than 1% in response to a 1% increase in inflation. By having a formal rule, a central
82
5.3 Monetary policy
bank may avoid inefficiencies induced by a discretionary policy. Mathematically, we

can state the Taylor rule as:
it = pt + r*t + ap1pt - p*t 2 + ay1yt - yt2
where it is the target short-term nominal interest rate, pt is the inflation rate (the GDP
deflator), p*t is the desired rate of inflation, rt* is the equilibrium real interest rate, yt is
an estimate of the logarithm of real GDP, and yt is the logarithm of potential output,
obtained by a linear trend. yt - yt is the output gap. The parameters are restricted to be
positive, ap, ay 7 0, and Taylor (1993) proposed setting them at 0.5.
In general terms, the Taylor rule seeks to apply negative feedback to the economy,
increasing rates when either capacity is stretched or inflation is above target and reducing
them when the opposite applies. This clearly matches central banks’ objectives in qualita-
tive terms, but an important practical problem is the dependence of the rule on quanti-
ties that can only be approximated. In particular, GDP is only known with a considerable
lag, and is subject to frequent revisions. Even in the long run, GDP is only an approximate
measure of the economy.
Many central banks, explicitly or otherwise, use a form of the Taylor rule to set inter-
est rates. It is, however, most suited for very large currency areas, such as the US, because
it disregards the impact of interest rates on exchange rates. For much smaller countries,
higher interest rates may lead to inflows of hot money and carry trading.
5.3.2 Open-market operations

Central banks can directly control the supply of money by open-market operations. This
entails buying or selling securities, normally the debt obligations of the central bank’s
own government, in the open market. Typically, the counterparties are major banks.
When a central bank buys securities, it pays by increasing the reserve account (a bank’s
account with the central bank) of the seller’s bank. It is not a transfer into the account;
rather the central bank simply increases the account balance by some number by fiat.
Doing so increases the total volume of reserves (money) held collectively by the banking
system. This is a modern version of printing money. Similarly, when the central bank sells
securities, it deducts the proceeds from the reserve accounts of the buyers, which reduces
the total volume of reserves, and hence money.
Expanding or shrinking the total volume of reserves in this way matters because banks
can trade reserves among one another or exchange them for other assets. Because the
central bank pays only a low rate of interest (often zero) on these balances, any bank
that has more reserves than it needs typically will try to exchange them for some interest-
bearing asset.
Expansionary open-market operations, when the central bank buys short-dated
securities, create a downward pressure on short-term interest rates via two main routes.
A direct impact arises because this removes an instrument from the market, increasing
their prices and lowering yields. An indirect effect arises because the bank now has
cash instead of a security, and hence has a greater capacity to lend, also lowering inter-
est rates.
83
5.3.3 Reserve requirements

Reserve requirements give the central bank a degree of control over the money supply.
Recall Example 1.1:
1
M1 = g * M0 = M0
d
Changes in the reserve requirements, d, lead to changes in the money multiplier, g,

and the volume of M1 given an amount of M0. Lowering the reserve requirement has a
similar effect as an expansionary open-market operation, provided that banks are con-
strained by reserve requirements. Altering the reserve requirement used to be relatively
common, but nowadays most central banks rely on other methods. The main exceptions
are in less developed economies, for example Brazil, China, India, Russia and Uruguay. For
an example of the Chinese use of reserve requirements, see Figure 5.1.
20%
15%
10%
5%
0%
2004 2005 2006 2007 2008 2009 2010 2011
Figure 5.1 Chinese reserve requirments

Data source: The People’s Bank of China
5.3.4 Quantitative easing

The monetary policy measures discussed above can be expected to be successful when
an economy is not in a recession and inflation is comfortably above zero. If the economy
is close to deflation, traditional monetary policy tools may not be effective, because
the central bank interest rate cannot be negative, and deflation provides an incentive
to banks to hold on to funds and not to lend them out, making open-market operations
ineffective. This can be viewed as a form of liquidity trap, as illustrated in the following
example.
Example 5.3 Liquidity traps
John Maynard Keynes (1936) identified the pathological case of liquidity traps, illus-
trated in Figure 5.2. The supply of money intersects the demand for money (D) on the
perfectly elastic part of the demand curve (the flat part to the right), and an increase
in money supply from S1 to S2 does not therefore change the interest rate. That means
a conventional monetary policy is unable to stimulate an economy.
84
5.3 Monetary policy
Interest rate
S1 S2
D
Money balances
Figure 5.2 Liquidity trap
In order to prevent the economy from sliding into deflation and to provide an economic
stimulus, some central banks have resorted to a more direct control of the quantity of
money, called quantitative easing (QE), to be used when other measures are not successful.
Conceptually, open-market operations and QE may seem to be the same, because in both
cases the central bank is purchasing assets from banks using money (M0 or M1) it has cre-
ated ex nihilo (out of nothing). In practice, the difference between these two operations is
significant, in scale, frequency, asset composition, maturities and motivation. We might say
that while open-market operations are a scalpel, QE is more like a sledgehammer.
In QE, the central bank may buy exactly the same assets as in open-market opera-
tions, typically short-dated government bonds, but under QE a much broader range of
assets may be purchased, including longer-dated securities and even non-government
assets, such as corporate bonds. Open-market operations tend to be more frequent and
to involve much smaller amounts and shorter maturity assets than QE.
The motivation is also different for open-market operations and for QE. In the former,
the explicit objective is to fine-tune the quantity of money, while QE is also presented as
a way to stimulate the economy and a means for the central bank to directly support the
government financially. The stimulus happens for several reasons, for example, because
the increased money supply encourages demand whilst putting downward pressure on
exchange rates, helping exporters and the trade balance. The direct support of govern-
ments happens because the central banks have become significant purchasers of govern-
ment bonds in some countries, helping to keep yields and government debt down. Even
though central bank holdings of government debt count as much as any other holding of
government debt, because the central bank is owned by the government, on a net basis
government debt falls when the central bank buys it.
Example 5.4 Friedman’s helicopter
Milton Friedman proposed an innovative solution to the problem of liquidity traps

whereby the central bank bypasses banks and gives money directly to consumers and
businesses.
➨
85
‘Let us suppose now that one day a helicopter flies over this community and drops
an additional $1000 in bills from the sky, . . . Let us suppose further that everyone is
convinced that this is a unique event which will never be repeated . . .’
Milton Friedman (1969)
The BoE’s QE operations amount to £325 billion at the time of writing, or £5,242 for
each of the 62 million people that live in the UK. Instead of buying government bonds,
the Bank’s QE operations would probably have been much more effective if it had sim-
ply sent an envelope with £5,242 in notes to each person in the UK.
5.4 Financial stability

One of the key functions of central banks is financial stability, even if it is not usually one
of their statutory obligations. One reason is that while financial stability has been an
integral part of central banking from the beginning, it played a secondary role during the
highly regulated Bretton Woods era. It is only with the crises from 2007 that financial sta-
bility has again become a core function of central banks. The term itself is not of a recent
vintage, originating from the convention establishing the OECD in 1960.
Definition 5.1 Financial stability Financial stability refers to policies aiming to

moderate the extremes of financial volatility, prevent crises, contain systemic risk, keep
financial markets functioning efficiently and resolve financial crises.
Financial stability ends up being a responsibility of central banks in part by default.

Governments need to implement financial stability policies and the only government
organisation capable of doing so is the central bank because it is the only entity c apable
of printing money at will. This also means it is often difficult to separate out the mon-
etary policy and financial stability functions of the central banks. In particular, they
often use monetary instruments to implement financial stability and in some cases the
particular implementation may be in conflict between either the financial stability or
the monetary policy objective.
Implementing financial stability

While the concept of financial stability is broad and touches on many different aspects
of policymaking, we can delineate it into three different components: passive prevention,
active prevention and resolution.
Passive prevention refers to the prevention of adverse outcomes ex ante, by setting up
the rules of the game in a way intended to enhance financial stability. This can take many
different forms: just to name a few, loan to value ratios, central counterparties (CCPs)
and limits on how much a bank can lend to a single counterparty. Some of the clearest
examples of passive prevention policies were implemented in the US following the Great
Depression and include regulations on buying equities on the margin, the Glass–Steagall
86
5.5 Bailing out Governments
Act keeping the commercial banking separate from investment banking, and rules aiming
at minimising liquidity risk in mortgages.
Active prevention relates to policies designed to smooth outcomes in financial mar-
kets in real time, making sure that regular activity does not get out of hand. Active preven-
tion is intended to be reactive to the situation at hand rather than being an essentially
static framework like passive prevention. Examples include rules specifying minimum
bank capital and reserves, amount of liquidity or maximum leverage. These are generally
based on a dynamic assessment of bank activities, and nowadays may involve sophisti-
cated financial models.
Finally, resolution refers to dealing with a crisis already underway – ex post. Such poli-
cies were first formalised after the O&G failure in 1866 which led to the establishment of
LOLR, aiming to provide liquidity to solvent but illiquid banks. Resolution also relates to
having a special resolution regime for failed banks, such as prompt corrective action in the
US. Policies like living wills and bail-ins also fall under this category.
Resolution and passive prevention share in the advantage that the key decisions are
made during previous crises, with policies less exposed to shifting priorities than in active
prevention.
The introduction of new financial stability tools often follows when the authorities do
a postmortem analysis on a crisis. Besides the examples mentioned above, a prominent
case is establishment of the Bretton Woods system, especially the World Bank and the
International Monetary Fund (IMF).

One of the core functions of central banks is printing money and it is not surprising that
governments, upon finding themselves in a spot of financial trouble, resort to money
printing as a revenue source – the central banks bail out the government.
The obvious way to do this is for the central bank to print money and purchase gov-
ernment bonds. As an emergency device, this is a core function of central banks, but if
abused, it leads to inflation and economic difficulties.
A second way for central banks to bail out governments is by creating unexpected infla-
tion and reducing the real value of past borrowing. Of course, for this to be fully effective,
the central banks must first invest many years of effort in establishing the strength of their
commitment to fighting inflation.
While any government with its own central bank enjoys the benefits of seigniorage –
revenue from printing money – explicitly using the central bank to finance the government
has long been considered a taboo amongst developed economies, that is until the recent
crises when several governments, such as those of the US and the UK, have resorted to
extensive QE to prop up their economies.
Figure 5.3 shows relative holdings of the Fed and the BoE of government bonds.
The ECB has engaged in similar activities but on a much smaller scale, around 2% of
GDP. However, it has lent massively to the banking sector and, hence, indirectly to
governments.
87
16% UK
12% USA
8%
4%
0%
2005 2006 2007 2008 2009 2010 2011
Figure 5.3 Central bank holdings of government bonds relative to GDP, end of year
Data source: Federal Reserve Board, UK Debt Management Office and the World Bank
Pros and cons

Printing money to finance a government is not recommended except in exceptional cir-
cumstances. When done as a routine response to an economic downturn, it locks in
inflationary expectations and increases government financing costs, ending in instability.
When inflation eventually becomes a priority for the central bank, it will find it very costly
to fight.
There are special cases, however, where a central bank bailout of governments is justi-
fied. If inflation is low during deep recessions and the economy is way below its output
potential, printing money to finance the government carries with it two significant ben-
efits: first, it relieves the pressure on the government’s budget, and secondly it reverses a
contracting money supply.
This should only be done when the economy is below its output potential, because
newly minted money will increase demand; if the economy is close to its output poten-
tial, the demand will pass directly through to increased prices, whilst if the economy is
operating below its output potential, such price increases are much less likely.
5.5.1 The European Central Bank

Bailout of member governments
Printing money is in effect a tax on assets denominated in that currency because of infla-
tion. The assets lose purchasing power and a smaller amount of wealth is transferred to
the central bank, the rest being made up of deadweight loss. Since the majority of assets
are held by domestic agents, printing money can be seen as just a tax like any other,
with its particular distributional effects, but where most of the impact is confined to the
country. When a central bank buys the debt of its own government with freshly printed
money, we can therefore view that in the same light as any other tax.
The situation is more complex in the euro zone because the ECB is the central bank
for the 17 countries that are members of the zone. If the ECB only buys the debt of some
member countries, it is indirectly subsidising those at the expense of the others. By doing
so, the EU takes an important step towards becoming a transfer union. That is a major
political decision that the Union has been unwilling to make.
88
That means the ECB has limited legitimacy in its efforts to to fulfil one of the funda-
mental objectives of a central bank – bailing out its government in times of crisis. This is
the reason why the ECB’s purchases of European government bonds have been so limited
compared to the purchases of the BoE and Fed. Instead, the ECB tries to do its job indi-
rectly, by supplying liquidity to financial institutions, in the hope that it will relieve the
pressure on the sovereign. This is not as efficient as using the central bank to bail out the
sovereign directly. The ECB signalled change in policy in September 2012, announcing its
intention to buy the bonds of distressed European sovereigns. This has run into significant
opposition, not least from Germany, and at the time of writing it is unclear what the out-
come will be.
Private sector bailouts

If a national central bank bails out financial institutions in its home country, it is under-
stood that this is directly paid for by the people of that country, via either taxes or cur-
rency debasement. It is up to the authorities of the country to weigh the pros and cons of
such operations.
The situation is different if the central bank is owned by many countries. Bank bail-
outs by the ECB imply that taxpayers of one country have to support a financial institu-
tion in another country. If you feel your banks are sound and well supervised, you might
feel aggravated if your money is being used to bail out banks in countries with unsound
banks because they are imprudently run and supervised. This means that bailouts by
the ECB cannot have as much political support as bailouts by the national central banks.
Furthermore, this leaves open the question of who bears the cost. In the extreme case of a
central bank going bust because of ill-advised bailouts, who bails the central bank out? In
the eurozone the cost would likely be split amongst member governments.
Ownership and operational independence

A national central bank has direct connections to the national government. The governor
and board members are generally residents of that country and are appointed by the gov-
ernment. This ties the national central bank close into the power structure of the country.
The ties between a multinational central bank and the political superstructure are not
as strong. This means the ECB is weaker than national central banks, which can make it
easier for the political leaders to use, or abuse, an international central bank to provide
bailouts. While this can be prevented by strict rules, it reduces the flexibility of the central
bank.
This means that a mutually owned central bank like the ECB might be less effective
in its role as a guarantor of financial stability than a national central bank. However, the
lack of accountability and independence of the ECB might actually make it more active
than a national central bank. With national banks there is always a threat of the govern-
ment taking control, while with the ECB this is one step removed because treaties make
fundamental change a matter of international unanimity, not national vote. This can
then make the ECB become a law unto itself. However, it might also make it highly rule
and bureaucracy driven, and the same requirement for unanimity restricts and limits its
scope for action.
89
5.6 Challenges for central banking

Central banks often find it difficult to reconcile the various conflicting objectives they
have to meet. Traditionally, the main disputes have involved those in charge of the mac-
roeconomic objectives and monetary policy. The former demanded low interest rates to
stimulate the economy and the latter high interest rates to prevent inflation. During the
1980s, when the defeat of inflation was a key goal, the debate was settled in favour of
monetary policy, which central banks made independent to prevent political interfer-
ence. More recently, financial stability has become a priority, threatening not only the
primacy of monetary policy but also the independence of central banks.
5.6.1 Conflict between financial stability and monetary policy

In the early years of central banking, during the gold standard, monetary policy mostly
took care of itself, leaving financial stability as the most important function of central
banks. By contrast, in the decades after the second World War (WWII), the financial
system was heavily regulated, limiting systemic risk and hence the importance of the
financial stability function. Many governments came to rely on a rather unfortunate inter-
pretation of the theories of Keynes, whereby they used monetary policy to stimulate the
economy on a more or less permanent basis. The eventual outcome was persistently high
inflation and economic stagnation, termed stagflation. Conquering the inflation in the
1980s was very painful, leading to the establishment of policies aimed at preventing the
re-emergence of inflation. A key element was the doctrine of central bank independence,
preventing politicians from manipulating interest rates for election purposes.
This meant that in the 1980s and 1990s monetary policy became the main, and some
would say the only, objective of central banks, which generally were in the comfort-
able position of being able to ignore issues of financial stability. The central banks were
content to be in that situation because financial stability operations often require the
injection of liquidity or the lowering of interest rates, upsetting the monetary policy
objective.
Of course, these objectives may coincide. Economic agents may deleverage during a cri-
sis, converting illiquid funds into liquid cash, resulting in a sharp reduction in the money
supply, the end result being deflation. In this case, an increase in liquidity is needed for
both the monetary policy objective of meeting a target inflation rate and the financial
stability objective of providing liquidity to an economy in distress.
Generally, however, the financial stability objective and the monetary policy objec-
tive are quite different. Central bank employees are likely to be specialised in one field or
the other, but not both, leading to the danger of silos within central banks, one with the
monetary policy staff and the other with the financial stability employees, who may not
talk much to each other even if both functions reside within the same institution.
One reason is that we are trying to do two different things with one tool – interest rates.
With interest rates fundamental to monetary policy, it would be better for the financial
stability policymakers to have access to independent tools, if these can be developed, in
order to make it more practicable to pursue both sets of objectives at the same time.
90
5.6 Challenges for central banking
One of the clearest manifestations of how the emphasis is shifted from monetary
policy to financial stability and macroeconomic objectives is UK inflation, which has
frequently exceeded its annual target rate since the crises from 2007. The risk is the crea-
tion of inflationary expectations, which would be costly to eventually fight. It is far from
obvious that interest rates could currently be raised without a recurrence of the crisis.
With this key lever unavailable it is not clear that a contractionary monetary policy could
be effective.
5.6.2 Central bank independence

The financial stability objective of central banks conflicts with the monetary policy objec-
tive when it comes to central bank independence. Politicians take great interest in the
setting of rates. Low rates stimulate the economy in the short run, and therefore help the
prospects of unpopular governments facing election. For this reason, central banks are
often under considerable pressure to keep interest rates low. While tempting politically,
this is bad economic policy, except in special cases, since any temporary well-being is
outweighed by the costs of long-term inflation. This might be prevented by central bank
independence, and most countries have made significant steps in that direction over the
past decades.
It is not as obvious that central banks should be independently in charge of financial
stability. When the authorities are called upon to fight financial crises, this necessitates
deep involvement in the structure of the financial system and the use of public money
to bail out private sector institutions. The central banks should not have the authority to
implement such policies on their own. The ultimate guardian of the public purse is the
treasury, or the ministry of finance, which should assume the pivotal role when it comes
to using significant amounts of public money to fight a financial crisis. This places the
treasury directly above the central bank, which then cannot be considered independent.
The supervisor, the central bank and the treasury all need to cooperate in implementing
financial stability, and the treasury has to have the ultimate power. This does not mean
the treasury is better at exercising financial stability than the central bank. The opposite is
more likely to be true. The central bank is more likely to have the necessary expertise and
be less sensitive to cronyism and corruption.
This is a recipe for conflict. It is essential for central banks to be independent to
effectively implement monetary policy, but they cannot be fully independent when
it comes to implementing financial stability. Some compromise is called for, perhaps
leaving the central banks in charge most of the time but yielding to the treasury when
needed. This seems to be the direction taken by the governments of the UK and US and
arguably Europe.
5.6.3 Losing control of money

Many governments have used massive amounts of money creation as part of the resolution
of the 2007 crisis, obeying the imperative need to prevent a systemic collapse, stimulate
their economies, and pay for the large amounts of debt assumed by the government. In
91
the short run, printing money has been a success and has prevented the financial system
from collapsing, the economy from sliding into a full-blown depression, and governments
from defaulting.
In the longer term, there is a price to pay. Until the crises from 2007, such policies
were generate generally dismissed out of hand, because of their inevitable inflationary
effects and the cost imposed on the economy because of increased uncertainty. The only
legitimate use of monetary policy was to maintain low and stable inflation, a point made
explicit in many central bank charters.
Many of the governments now enthusiastically embracing money creation have in the
past condemned others for responding to domestic crises in exactly the same way. For the
next decade or two this will make it very hard for them to credibly forswear the further
use of monetary policy for such purposes: financial entities will simply assume that if the
need is sufficient, governments will do what it takes.
This has serious implications. At present, contractionary influences remain strong and
there is little evidence that inflationary expectations are building up, but the threat and
the associated costs are clear. It took many years to defeat the 1970s inflation, with cen-
tral bank independence and credibility playing a vital role in that fight. These have now
been impaired, suggesting that the coming battle against inflation will be similarly long
and painful.
Hyperinflation
Governments sometimes completely lose control of money. The reason might be that
as inflation increases, government revenue decreases, so the government needs to
print money to finance itself. That, however, leads to more inflation, and a vicious
cycle is formed. This process was modelled by Cagan (1956) who showed that it is
necessary to increase the money supply at a double exponential rate for the govern-
ment’s revenue to keep up. The empirical evidence is consistent with his model, as
seen in Figure 5.4 which shows the growth of money supply during the hyperinflation
in Germany and Zimbabwe.
Even though the y-axis in the figure is on a logarithmic scale, the growth is exponential,
so inflation is growing at a double exponential rate. Besides Zimbabwe and Germany,
many countries have experienced hyperinflation, as seen in Table 5.1.
Jan/1922 Jul/1922 Jan/1923 Jul/1923

1014 1010
Zimbabwe
Zimbabwe
Germany
1010 Germany 107
106 104
101
102
2001 2002 2003 2004 2005 2006 2007
Figure 5.4 Hyperinflation in Germany and Zimbabwe
92
5.7 Summary
Table 5.1 Highest monthly inflation rates in history
Month with Highest Equivalent Time required

highest monthly daily for prices
Country inflation rate inflation rate inflation rate to double
Hungary July 1946 4.19 * 1016 % 207% 15.0 hours

Zimbabwe November 2008 7.96 * 10 % 10 98.00% 24.7 hours
Yugoslavia January 1994 3.13 * 10 % 8 64.60% 1.4 days
Germany October 1923 29,500% 20.90% 3.7 days
Greece October 1944 13,800% 17.90% 4.3 days
China May 1949 2,178% 11.00% 6.7 days
Notes. The Reserve Bank of Zimbabwe reported inflation rates for March 2007–July 2008. The authors (Hanke and Kwok, 2009, Table 2)
calculated rates for August 2008–14 November 2008
Source: On the measurement of Zimbabwe’s hyperinflation, Cato Journal, Vol. 29, No. 2, 356 (Hanke, S. H. and Kwok, A. K. F. Spring/
Summer 2009). Copyright © Cato Institute. All rights reserved.
The general reason why a country ends up with hyperinflation is that the government
surrenders control of money creation because it is under some imperative to raise revenue at
all costs. In the end, hyperinflation is extremely costly. As Vladimir Lenin said: ‘The best way
to destroy the capitalist system is to debauch the currency’, as quoted by Keynes (1920).
5.7 Summary
The central bank is the most important institution in the financial system because
it has a monopoly on creating money. It generally has five objectives: price stability,
macroeconomic performance, financial stability, banking supervision and bailing out
its government. There are differences between countries. The first two objectives are
likely to be legal objectives, but financial stability has risen in prominence during the
ongoing crisis. The supervisory function remains controversial with good arguments
for and against the central bank taking on the responsibility. The bailout function is
more like a dirty secret, not something the central banks or governments want anyone
to notice.
Central banks control the supply of money either directly or indirectly to achieve price
stability. Central banks have resorted to unconventional methods, such as QE, since
2007 to stimulate the economy and finance the government. Their long-term impacts are
unclear but include some risk of inflation.
The priorities of the central bank objectives have changed over time, from financial
stability to price stability and back to financial stability. These objectives are quite
different and often in conflict with each other. For example, financial stability may
require money creation, undermining price stability, whilst prudential banking super-
vision could lead to the creation of hidden risks. Another contentious issue is central
bank independence, considered desirable for monetary policy but not entirely com-
patible with financial stability. In a worst case scenario, the central banks may lose
control of money, resulting in hyperinflation.
93
Appendix: Central bank interest rate

The most visible demonstration of monetary policy is the setting of interest rates. Several
different terminologies are used when referring to central bank interest rates and these
can often be confusing and even contradictory. Common names are the target rate, the
short rate, the risk-free rate and in the US the Fed funds rate.
The interest rate set by a central bank is typically a short-term rate, usually overnight
rates for secured or unsecured large institutional borrowers and depositors with the cen-
tral bank. It is often called the target rate because it is the interest target for the central
bank, the short rate because it refers to short maturities, the risk-free rate because the
government is risk-free, the discount rate for historical reasons, and the Fed funds rate in
the US because the Fed stands for the federal reserve which sets the rates.
Confusingly, the short rate could also refer to the interest rate on short maturities
bonds issued by commercial entities, while the risk-free rate could also refer to the inter-
est rate paid by the government on any of its borrowings, regardless of maturity.
The term discount rate has multiple meanings. It originates from an old type of loan
where a borrower would sell an obligation at a discount, promising to buy it at the full
price in the future, which might be called a repo today. The effective interest rate is the
discount rate. This term is most commonly used nowadays for short-term borrowing from
the discount window of the Fed. However, it could just as easily be used for the interest
rate used in regular present value calculations.
The term prime rate also has multiple meanings. Traditionally it referred to the interest
rate charged by banks to their best clients, hence the word prime. It now means generally
interest rates paid by somebody who is very low risk, but we could easily see interest rates
below the prime rate. Furthermore, in different countries it can refer to particular institu-
tional setups. The US prime rate is not a single standardised rate. Each bank has its own
prime rate, and the one that is quoted most frequently comes from the Wall Street Journal,
which polls the 30 largest banks in the country.
Benchmark interest rates

We outline the terminologies of the three major central banks here. The Fed sets a target
federal funds rate. This is the unsecured rate that banks charge each other in the interbank
market for borrowing reserves held at the Fed (known as federal funds), usually overnight.
The federal funds rate is determined in the market, but the Fed can influence this rate
through open-market operations, reserve requirements and the discount rate. The dis-
count rate is the interest rate charged to banks when they borrow overnight directly from
the Fed, through the Fed’s discount window.
The main interest rate for the ECB is known as the main refinancing operations fixed
rate. This is essentially a one-week repo rate where banks put up acceptable collateral
with the ECB and get a loan in return. Refinancing operations are conducted via auctions,
where the ECB specifies the rate at which it is willing to lend money and the amount
of liquidity available, and the banks then express their interest. The ECB also sets the
‘marginal lending facility’, which is a secured overnight rate. Banks can use this facility to
94
References
borrow overnight from their national central banks after providing acceptable collateral.
This rate provides a ceiling for the overnight market interest rate, and is similar in nature
to the discount rate used by the Fed. There is also a deposit facility where banks can make
overnight deposits with their national central banks. The interest rate on this facility simi-
larly provides a floor for the overnight market interest rate.
The benchmark interest rate for the BoE is known as the BoE base rate, which is the
rate that the BoE charges banks for secured overnight borrowing. It is usually transacted
as an overnight repo against high-quality collateral. This is more comparable to the Fed’s
discount rate rather than to the federal funds rate.

1 What are the main functions of a central bank?
2 What is more important as a central bank function: financial stability or monetary policy?
3 How have the common views of the relative importance of financial stability versus
monetary policy changed over time?
4 What is the danger of deflation, and what policy remedies for it do you recommend?
5 Should the central bank be in charge of supervision?
6 What is quantitative easing, and how does it compare to more conventional monetary
policy operations?
7 Why is quantitative easing discussed more widely today than in the past?
8 What is the long-term danger of quantitative easing?
9 Who are the main losers from quantitative easing?
10 What are some of the unique challenges facing the ECB?
11 How can a central bank bail out its government? Why is it harder for the ECB to fulfil
this function than central banks in other countries?
12 What are the main arguments for central bank independence?
13 What are the main arguments against central bank independence?
References
Cagan, P. (1956). The monetary dynamics of hyperinflation. In Friedman, M., editor, Studies in
the Quantity Theory of Money. University of Chicago Press.
Friedman, M. (1969). The optimum quantity of money. In Friedman, M., editor, The Optimum
Quantity of Money and Other Essays, chapter 1, pp. 1–50. Adline Publishing Company,
Chicago.
Goodhart, C. (2002). The organizational structure of banking supervision. Economic Notes,
31(1): 1–32.
Goodhart, C. and Schoenmaker, D. (1995). Should the functions of monetary policy and bank-
ing supervision be separated? Oxford Economic Papers, 47: 539–60.
95
Hanke, S. H. and Kwok, A. K. F. (2009). On the measurement of Zimbabwe’s hyperinflation.

Cato., 29(2): 353–64.
Keynes, J. M. (1920). The Economic Consequences of the Peace. Harcourt Brace, New York.
Keynes, J. M. (1936). The General Theory of Interest, Employment and Money. Macmillan.
Lewis, M. K. and Mizen, P. D. (2000). Monetary Economics. Oxford University Press.
Taylor, J. B. (1993). Discretion versus policy rules in practice. Carnegie-Rochester Conference
Series on Public Policy, 39: 195–214.
96
6 The Asian crisis of 1997
and the IMF
In 1997, several countries in East Asia experienced one of the most severe currency
and financial crises since the Great Depression. Exchange rates, asset prices and eco-
nomic activity collapsed with financial and corporate insolvency widespread.
The crisis caught everybody by surprise. The East Asian countries seemed to have
done everything right, avoiding mistakes causing crises elsewhere, for example, in
Latin America and Europe. The macroeconomics appeared sound, and the countries
had balanced budgets and tight monetary policy.
Instead, the underlying weaknesses were in the private sector, with excessive for-
eign currency borrowing that led to bad investment decisions. This was fuelled by the
apparent belief that the governments would ensure stable exchange rates, and even
bail out borrowers. The macroeconomic fundamentals were sound, the financial fun-
damentals not.
There is not complete unanimity on which countries were affected. Four are always
included: Indonesia, South Korea, Malaysia and Thailand, as all experienced spec-
tacular growth in the 1990s, followed by the 1997 crash and respectable recovery.
All but Malaysia were forced to call in the International Monetary Fund (IMF), so
we mostly omit Malaysia from the analysis. Other Asian countries affected include
Hong Kong, Singapore and Taiwan, but these avoided the worst, so we also exclude
them here.
Detailed analysis of the Asian crisis, with the benefit of hindsight, indicates that
foreign currency speculation and a sudden stop in foreign lending were a significant
immediate cause of the crisis, fuelled by weak fundamentals. Given the weaknesses in
Chapter 6 The Asian crisis of 1997 and the IMF
the financial fundamentals, a crisis in all of the countries was probably inevitable; if it
had not happened in 1997, it would have occurred within a few years.
The main policy conclusion drawn by the crisis countries themselves was that they
could only rely on themselves for liquidity support and, in response, have built up
vast currency reserves.
These reserves did not protect them from significant exchange rate fluctuations
once the global crisis was underway in 2007, and South Korea had to enter into cur-
rency swap contracts with the Federal Reserve System (Fed), helping to stabilise its
exchange rate. This clearly demonstrates the limited protection afforded by vast
reserves, and it is quite possible that the reserves only serve to prevent small crises,
making the eventual one large. In particular, large reserves do not imply that a coun-
try should avoid addressing underlying structural weaknesses.
The IMF, or the Fund as it is sometimes referred to, played an important and con-
troversial role in the crisis. It is a part of the United Nations and was originally set up
as a part of the Bretton Woods system, to manage the fixed exchange rate regime at
the time. Over time, the role of the Fund has shifted; for the past few decades it has
been the main international mechanism for responding to financial and economic
crises, as well as the enforcer of foreign claims on governments. It can provide sig-
nificant amounts of money as emergency assistance – a package – but does in return
demand what is known in the Fund’s jargon as structural adjustments. This includes
macroeconomic policies like fiscal and monetary tightening and improving the capi-
tal accounts, but also some microeconomic adjustments, for example, labour market
reform, cuts in subsidies and more recently anti-corruption measures. These policies
are a part of the Washington consensus.
At its core, the Asian crisis was not very different from many other crises throughout
history, a clear example of the observation that all crises are fundamentally the same,
only the details differ. It has clear parallels with the ongoing global crisis; many of the
underlying fundamental weaknesses are the same, and some policymakers involved
in the European sovereign debt crisis have drawn direct lessons from the Asian crisis.
Perhaps the most interesting parallel between the Asian crisis and the crisis from 2007
is how different policies are when they have to be implemented domestically or abroad.

This chapter draws on the theoretical concepts discussed in Chapter 4 (liquidity) and
Chapter 14 (bailouts), and relates to the policy analysis in Chapter 11 (currency mar-
kets) and Chapter 19 (sovereign debt crises).
Key concepts
■ East Asia crisis of 1997
■ IMF
■ Maturity and currency mismatches
■ Challenges in fighting crises
■ Washington consensus
■ Structural adjustment
98
6.1 Building up to a crisis

A large number of studies have been produced on the Asian crisis. We drew on the
early works of Goldstein (1998), Radelet and Sachs (2000) and Alba et al. (1998). We
also used two more recent papers: Weisbrot (2007) focusing on the role of the IMF,
and Ito (2007) providing a summary of recent analysis.

Large capital flows have long been an integral part of the global economy. In the era
of the first globalism, 1873–1914, much of the world’s infrastructure was financed by
London. Capital flows were lower between the two world wars and during the Bretton
Woods era, but took off significantly in the early 1970s, with newly wealthy investors in
the Middle East and liberalised financial institutions working together to build up global
capital markets.
Investors and borrowers alike often prefer short-term loans, those with maturity of less
than one year. Investors want to minimise risk, whilst borrowers prefer the lower cost of
the short-term loans. Therefore, maturities on foreign loans to developing countries are
often quite short. The ultimate use of the funds is generally longer-term, perhaps the
construction of a factory that will only start production in a few years. Therefore, interna-
tional capital flows often result in significant maturity mismatches.
In the 1970s, the preferred destination for excess capital was the Latin America, but it
soon became apparent that the continent made poor use of the money and a series of sov-
ereign debt crises ensued, followed by difficulties for many of the lenders. Consequently,
lenders were looking for new borrowers.
Lending to East Asia

In the 1990s, capital flowed to East Asian countries, helped by their adherence to the
Washington consensus, which is a term used to capture the common view of the major
economies, the World Bank and the IMF, and the European Union (EU) that emphasises
free capital mobility.
Definition 6.1 Washington consensus

1 Fiscal discipline (eliminate deficits)
2 Broaden the tax base, keep taxes low
3 Market interest rates
4 Raise spending on health and education
5 Secure property rights
6 Privatisation
7 Deregulation
8 Free trade
9 Competitive/sensible exchange rates
10 Free capital flows (remove barriers to foreign direct investment).
99
The East Asian countries opened up their capital accounts and liberalised the financial
sector, all while maintaining sound macroeconomics. Sovereign debt was low and falling:
see Figure 6.1. GDP growth was quite spectacular, ranging from 7.9% to 9.5% across the
region, as can be seen in Table 6.1. Other standard indicators, like inflation and unem-
ployment, all pointed to macroeconomic stability.
80%
1990
60% 1996
40%
20%
0%
Indonesia South Korea Malaysia Thailand
Figure 6.1 Debt to GDP ratio 1990–1996

Data source: World Bank and IMF. Indonesian starts 1991. No data for Thailand 1990
Table 6.1 Average GDP per capita growth
Annual
average Indonesia South Korea Malaysia Thailand
1990–1996 6.3% 6.8% 6.7% 7.5%

1997 -14.3% -7.5% -9.6% -11.6%
1999–2007 3.8% 4.7% 3.4% 4.0%
Data source: World Bank, http://data.worldbank.org.
Money and investment

Respectable macroeconomic performance does not mean much if based on transient fac-
tors. The success of the East Asian countries was founded on extraordinarily high invest-
ment rates, as can be seen in Figure 6.2, ranging from 28% to 40% of GDP. By comparison,
investment rates are around 18% in Europe.
40% 1990−1996 1999−2007
30%
20%
10%
0%
Figure 6.2 Investment rates (gross fixed capital formation, % of GDP)

Data source: World Bank
100
If investment is based on domestic savings, and especially if loans are maturity matched
to the ultimate investments, high investment rates are not a big concern for financial sta-
bility, even if one might question whether the funds can be sensibly invested. If, however,
the origin of the funds is short-term foreign currency borrowing from international capital
markets, the picture is different.
Initially, this was not the case for the East Asian countries; financing from international
markets was less than 1% of GDP in 1990. International borrowing was growing rapidly,
however, reaching 4% to 6% in 1996, as seen in Figure 6.3. Because these funds were
predominantly short-term, this rapidly growing exposure to international capital markets
signalled the building up of vulnerabilities to liquidity crises.
The high economic growth was partially reflected in the stock market (Figure 6.4). The
best year was 1993, caused by foreign inflows, both from direct purchases of equities
by foreigners and also from the indirect effect of domestic companies borrowing from
abroad. From 1994, the stock markets were doing poorly.
Perhaps the stock market anticipated the pending crisis better than foreign creditors, the
government or the IMF. After all, many investors in the stock market are well-connected
and well-informed local investors who would know of the build-up of vulnerabilities bet-
ter than most. Alternatively, those relatively poor returns may represent divestment by
foreigners, belatedly realising that with poor minority protection they had very limited
real ownership. The lenders felt themselves safe from the corporate governance prob-
lems, protected by convertibility, short duration, government guarantees and good mac-
roeconomic policies, and so continued to lend.
6%
1990 1996
4%
2%
0%
Figure 6.3 Financing via international capital markets (gross inflows, % of GDP)
Data source: World Bank and IMF
1990 1991 1992 1993 1994 1995 1996

90%
60%
30%
0%
−30%
Figure 6.4 Annual stock market performance 1990–1996

Data source: Global Financial Data
101
6.2 The crisis in individual countries

The main crisis event started with a speculative attack on the Thai baht in 1997, and soon
spread throughout the region. Within weeks practically every currency in East Asia came
under pressure, and subsequently the currencies of South Korea, Indonesia, Malaysia,
Taiwan and the Philippines were devalued. Singapore and Hong Kong managed to main-
tain their exchange rates.
As investors pulled out of the region, asset prices plunged and GDP growth turned
sharply negative. The crisis severely undermined public finances in a number of countries
and the IMF was called in to help.
We see the impact on the stock market and exchange rates for the second part of 1997
in Figure 6.5. The markets collapsed, with the stock market in Malaysia and the foreign
exchange market in Indonesia worst affected.
Stock market Exchange rates

0%
−20%
−40%
−60%
Figure 6.5 Drop in stock markets (local currency) and exchange rates in the second part
of 1997
6.2.1 Thailand
The first country to be hit by the crisis was Thailand. It had experienced spectacular
growth in the 1990s, fuelled by rapidly growing exports, and had maintained a relatively
fixed exchange rate of roughly 25 baht per USD since 1984.
A particular feature of the economy of Thailand was the prominence of special finance
companies, which were similar to banks and provided about 20% of all the credit in the
economy, on easy terms. The finance companies mainly financed themselves by short-
term dollar denominated loans from foreign investors.
The Thai economy started to run into difficulties towards the end of 1996, when inter-
est rates rose, and the economy slowed down due to sagging exports. One reason was
that the dollar appreciated in 1995, strengthening the baht, reducing the attractiveness of
exports and worsening the current account position.
As domestic companies ran into problems, Thai banks and finance companies reported
large increases in their non-performing loans. To keep the financial system stable, the
Central Bank of Thailand secretly lent about $20 billion to finance companies and weaker
banks at below market interest rates. These loans do not seem to have been a part of a
102
lending of last resort (LOLR) operation, but direct bailouts. Ultimately, this meant that the
Central Bank had lower reserves than expected, reducing confidence when that became
known.
The currency came under pressure towards the end of 1996, with the Central Bank
intervening in the currency market, building up increasingly large forward positions, fur-
ther reducing the effective amounts of reserves. Finally, a massive speculative attack took
place on 14 and 15 May 1997. Figure 6.6 shows the evolution of the currency. The Central
Bank of Thailand lost $10 billion defending the baht without success. On 15 May, the
Central Bank ordered all local and foreign banks to stop lending baht to anyone outside
the country, which eased the selling pressure for about a month.
Fresh concerns about the Thai economy surfaced on 19 June, prompting Thai
banks and corporations to sell baht in large numbers. The foreign creditors, which
used to roll over short-term debts, were suddenly demanding immediate repayment
as their loans matured. Few Thai borrowers had hedged against the possibility of a
collapse of the baht, and it would have been silly to do so, since the borrowing costs
when the frictional costs are included are higher than just borrowing in baht. On 2
July, the authorities concluded that the fixed rate system could not survive and let
the baht float.
This turned out to be the trigger for a fully fledged crisis in East Asia. The baht devalued
swiftly and lost more than half of its value, reaching its lowest point of 56 bahts to the
USD in January 1998. The Thai stock market dropped by 75% in baht terms and Finance
One, the largest Thai finance company at that time, collapsed.
Thailand got an IMF programme on 20 August 1997, with a package of $17.2 billion. The
markets were not impressed by the size of the package, nor the disclosure of $23.4 billion
in forward commitments by the Central Bank of Thailand, especially considering that
short-term private sector debts exceeded $30 billion. The baht continued falling.
The IMF recommended the closure of weaker financial institutions, but problems in
distinguishing between liquidity problems and solvency problems frustrated this process,
partly because financial institutions would rationally pay almost limitless bribes to be
placed on the list of liquid institutions. The Fund attached further conditions to its aid,
requiring fiscal and monetary tightening, the raising of taxes and reduction in expendi-
tures. This is a recipe it continued to follow in the other crisis countries.
40
IMF
worries
Baht/USD
attack
initial
35
float
30
25
May Jun Jul Aug Sep Oct Nov Dec
Figure 6.6 Thai baht in 1997

103
6.2.2 Indonesia
Indonesia under President Suharto embarked on an economic reform agenda in the early
1980s, restraining government spending, opening the economy to foreign investments
and easing regulations. The economy grew rapidly, whilst enjoying a current account sur-
plus, with foreign exchange reserves substantially increasing.
Similar to Thailand, Indonesian corporations were increasingly using short-term dollar
denominated debt to finance themselves. This worked well when the currency appreci-
ated during the 1980s and early 1990s, setting up a virtuous feedback loop between cur-
rency inflows, reduced cost of debt service and currency appreciation.
Behind it all was a strong culture of what was called ‘KKN’: corruption, collusion and
nepotism. Any foreign firm investing in an Indonesian company had to hand over partner-
ship rights to a presidential relative to ensure smooth approval. Individuals with personal
relationships to the Suharto regime made some of the largest fortunes in Indonesia.
Throughout the 1990s, difficulties were building up in the banking system whilst the
number of banks grew rapidly. The government heavily influenced lending decisions so it
could channel funds to favoured businesses, with banks made to lend to borrowers who
already had stopped servicing old loans – that is, in default. By the mid-1990s, non-per-
forming loans exceeded 25% of total loans made. The Indonesian currency came under
increasing pressure after the baht floated in July 1997, adversely affecting corporations
with unhedged dollar liabilities.
Indonesia asked for IMF assistance in October. The IMF General Manager, Michel
Camdessus, responded by saying that ‘The IMF strongly supports the approach that has been
followed by Indonesia, which sees this as an occasion to strengthen its economic policies even
if fundamentals are basically sound’ (IMF, 1997). This statement sounds naive: a country with
‘basically sound’ fundamentals should not suffer a crisis. In its analysis, the Fund seems to
have focused on macroeconomic fundamentals and neglected the financial fundamentals.
The IMF provided a package of $40 billion, attaching conditions to its programme,
most controversially that 16 banks be closed immediately, with limited protection to
depositors. The particular 16 banks seemed to have been chosen quite arbitrarily, and it
was unclear whether other banks would also be closed. Not surprisingly, the end result
was a run on the banking system and the country. While closing the 16 banks might
have been designed to demonstrate strength, within a few weeks a son of the president
reopened his bank under a different name, demonstrating a lack of commitment by the
Indonesian government, thus undermining the programme.
Similar to Thailand, the Fund required monetary and fiscal tightening, which the
Indonesian government found difficult to implement. For example, riots ensued when
subsidies were reduced. Eventually, the policies were adjusted and the Indonesian gov-
ernment agreed on another letter of intent with the Fund in January 1998.
6.2.3 South Korea

The situation in South Korea was quite different. It emerged as one of the poorest coun-
tries in the world from its civil war in the 1950s, and was governed by a series of mili-
tary regimes that dictated financial system lending decisions. South Korea pursued an
104
industrial policy to build up world-leading corporations called Chaebol. The Chaebol

responded by investing rapidly, with investments rising by nearly 40% a year in the early
1990s, and by then South Korea had become the second most industrialised economy in
Asia after Japan.
The Chaebol were financed with extremely high levels of debt, exceeding 400% of
shareholder equity in some cases. The largest became in effect too big to fail (TBTF), as
their default would threaten systemic crisis. Very high domestic savings rates were not suf-
ficient to meet industry demand for funding, and liberalised South Korean banks increas-
ingly resorted to borrowing dollars short-term internationally and lending long-term to
the Chaebol. Behind it all was an understanding that the government would bail out any
Chaebol that got into trouble.
Not all Chaebol used the seemingly unlimited amount of funds wisely and by early
1997 some were unable to meet their obligations, straining the banking system. The gov-
ernment decided to bankrupt the Daewoo Group and the Hanbo Group in mid-1997, and
prosecuted their owners.
The currency depreciated in the beginning of 1997, and after holding steady for a few
months, the depreciation accelerated in August. This further increased the difficulties for
the corporate sector, with its large dollar denominated liabilities, and put a strain on the
foreign currency reserves of the Central Bank.
There was little immediate fear about the solvency of the South Korean state – recall
that its debt was quite low – or the bulk of its financial and non-financial corporations.
However, if the Central Bank ran out of its limited foreign currency reserves, a crisis would
likely ensue.
South Korea’s foreign currency reserves started to fall at an accelerated rate from
October, and by November a rumour that usable reserves were far smaller than the
official statistics spooked the market. The Central Bank started to intervene in the
currency market, further depleting its reserves. According to official statistics, foreign
reserves declined from $30 billion at the end of October to $24 billion at the end
of November, and to $20 billion at the end of December. However, usable foreign
reserves were exhausted by November. Realising this, foreign investors stopped roll-
ing over South Korean loans by November. A self-fulfilling crisis was created: a vicious
feedback between solvency problems, falling exchange rates and the depletion of cur-
rency reserves.
Faced with massive defaults in the financial sector, the government requested IMF
assistance, getting a package of $57 billion. The IMF demanded significant conditions,
including its customary fiscal and monetary tightening, along with trade and capital
account liberalisation, and labour market and financial sector reforms.
The package was not effective, and the currency continued falling. There are several
reasons for this, not least that the IMF had lost credibility by this time as neither the Thai
nor the Indonesian packages had proven effective, whilst the South Korean package was
considered insufficient. As a response, the IMF accelerated the payout of funds and along
with the G7 applied what was known as ‘jawboning’, strongly leaning on creditor banks
to roll over South Korean exposures. This succeeded in halting the currency decline and
prevented more corporate defaults.
105
6.3 Reasons for the crisis

One of the most striking features of the Asian crisis was how surprising it was. Strong
growth rates concealed many of the weaknesses building up. Over time, we have gone
through several waves of explanations of the crisis. Initially, it was explained by weak fun-
damentals and moral hazard, but as the crisis developed, such explanations were found
lacking. Eventually, after the crisis countries recovered sharply, most commentators found
that liquidity and sudden stop had been a central element.
6.3.1 Weak fundamentals and moral hazard

Initial explanations for the crisis focused on rampant moral hazard problems and weaken-
ing macroeconomic conditions. Politics and corruption were key ingredients in the allo-
cation of capital and guaranteeing of loans, with Indonesia the worst offender. The end
result was not surprising: substantial over-investment in risky and poorly performing pro-
jects and asset price bubbles.
When the crisis erupted, most outsiders, in particular the IMF, focused on weak funda-
mentals and moral hazard. In that case, the optimal policy response is to directly address
the problem of corruption and moral hazard and to improve macroeconomic fundamen-
tals. This view is reflected in the IMF conditions.
Subsequent analysis has played down the pivotal importance of those two factors. Many
other countries that are not afflicted by crises have similar problems, and other countries
with better macroeconomic fundamentals and less moral hazard get hit by crises.
6.3.2 Panic and contagion

Another category of explanations emphasises panic in financial markets that leads to contagion,
whereby the initial devaluation of the Thai baht caused a panic, which spread to other countries
in the region. Many statistical studies have focused on such explanations, and it is quite straight-
forward to document the presence of interconnectedness and causality amongst the affected
economies. The problem is that any such analysis is conditional on an actual crisis happening
and simply provides a statistical description of how a crisis moves from one country to the next.
A panic does not happen by itself; there must be an underlying reason. Sophisticated
investors providing short-term funds to the East Asian countries were interested in profit
maximisation and panicking is not compatible with that objective. Deeper reasons are
needed to understand their motives.
Statistical descriptions focusing on panic and contagion may help in mapping how
a crisis evolves, but do not help very much in understanding the underlying causes nor
more importantly in formulating future policy.
6.3.3 Market factors

A different class of explanations focuses more directly on financial markets. All the
countries liberalised their financial markets quite rapidly but had little experience in
supervising sophisticated open financial markets. Related to this was the problem of
106
6.3 Reasons for the crisis
maturity and currency mismatches. Much funding was short-term foreign currency while
lending or investments were long-term in domestic currency. This creates a classical mis-
match problem for borrowers. What was special, but not unique, in the Asian crisis was
how prevalent such mismatches were.
The capital markets in East Asia were relatively underdeveloped, so that most financial
intermediation occurred through the banking system. If a country does not have sufficient
reserves to cover a sudden stop in the rollover of loans, it is vulnerable to a liquidity crisis.
Ito (2007) reports that the ratio of short-term bank borrowing to foreign reserves at the
end of June 1997 was 2.1 for South Korea, 1.6 for Indonesia and 1.4 for Thailand, signal-
ling significant fragility. In later analysis, the so-called Guidotti–Greenspan rule argues that
the ratio should not exceed 1.
A major ingredient was overconfidence. A history of currency stability encouraged for-
eign banks to lend, for the simple reason that their whole experience had always been
profitable. Central banks tolerated this for the same reason. Faith in convertibility and
the IMF made lenders comfortable, which is a necessary condition for future crises to
develop.
6.3.4 Performance before and after the crisis

With the benefit of hindsight, we can use post-crisis data to analyse what sort of a crisis
hit the East Asian countries. The latest year is 2007 in order to avoid being affected by the
ongoing crisis since then.
If the crisis had been driven by fundamental weaknesses, recovery would have taken many
years, with a successful implementation of the reform programmes demanded by the IMF a
necessary condition for sustainable recovery. This is not what happened. Instead, the main
crisis period was quite brief, and partial recovery swift, as seen in Figure 6.7.
Such crises are often called ‘V-shaped’ because they happen quickly and are over
quickly. Economic growth has been lower post-crisis than before, but is more sustainable
and based on a more sensible model of investment. Evidence of a V-shape is not as clear
if we look at the exchange rates, as seen by Figure 6.8. All four countries suffered a signifi-
cant fall in the exchange rates in the crisis, and the recovery was only partial. This applies
especially to Indonesia which suffered worst from the crisis and had very high inflation
rates in 1997 and 1998.
6%
Indonesia
0%
South Korea
−6% Malaysia
−12% Thailand
1995 1996 1997 1998 1999 2000
Figure 6.7 GDP per capita growth

107
1990−1996 Minimum 1999−2007

100
80
60
40
20
0
Figure 6.8 Average and minimum (in crisis) exchange rates. USD/local currency,
1990 = 100
Table 6.2 Relative GDP per capita ranking in percentiles
Year Indonesia South Korea Malaysia Thailand
1960 88% 36% 48% 71%

1990 74% 26% 44% 58%
1996 66% 23% 37% 47%
1999 70% 23% 39% 51%
2007 72% 20% 39% 49%
Data source: World Bank, http://data.worldbank.org.
Instead, a V-shaped crisis is more indicative of a liquidity crisis, where a sudden stop
causes significant temporary difficulties. This causes the currency to collapse, making local
corporations more competitive than before, often with much lower debt, creating condi-
tions for rapid export-led growth. This is exactly what happened in the East Asian crisis.
These effects were most pronounced in South Korea, which has performed best since
the crisis. Table 6.2 shows the relative per capita GDP ranking of the four countries at key
points, expressed in percentiles. Notably, Indonesia was essentially in the same place in
1990 and 2007, consistent with other empirical results presented. Malaysia has shown a
small improvement, but Thailand has made the largest move in the ranking, but is still
below its 1996 place.
6.3.5 Liquidity crisis – sudden stop

Recent analysis emphasises the role of liquidity in the crisis. Liquidity is a multifaceted
concept, and difficult to define concisely, but the use of the term here is more related to
funding liquidity.
The importance of liquidity is highlighted by Ito (2007), who finds the crisis to a signifi-
cant extent to be caused by foreign exchange speculation, rather than fundamental weak-
nesses, moral hazard or market panic. In other words, the crisis was caused by a sudden
stop, and reversal of capital flows. Net inflows were $93 billion in 1996 and net outflows
108
6.4 Policy options for the crisis countries
$12 billion in 1997. The observation that the crisis was broadly V-shaped does support
the view that it was more the sudden stop, rather than structural weaknesses, that caused
the crisis.
Of the four countries, South Korea exhibited the strongest signs of a pure liquid-
ity crisis. Events in the autumn of 1997 indicate that the possibility of the Central
Bank running out of reserves scared off investors and encouraged holders of the South
Korean currency to exchange it for dollars. This in turn become a self-fulfilling chain
of events, only ending after domestic assets became very cheap for holders of foreign
currency, and the IMF helped to provide liquidity. The fact that South Korea has the
smallest difference between pre- and post-crisis GDP further supports this observa-
tion. For the other three countries, the picture is more muddled. There was a clear
liquidity crisis but the relatively large shortfall in growth indicates that the very high
pre-crisis growth was caused as much by inflows of money as by the build-up of eco-
nomic strength.
Why did the sudden stop happen? If there had been no worries about the soundness
of the economies, there would have been no reasons for creditors to deny credit all of a
sudden. Furthermore, the fact that the economies performed well after the crisis does not
by itself justify the conclusion that everything was fine: after all, the affected economies
underwent significant structural changes as a consequence of the crisis. This means that
a sudden stop or liquidity explanations for the crisis are inadequate and only a manifesta-
tion of more fundamental problems.

The affected countries themselves have come to the conclusion that the crisis was caused
by a sudden stop, and hence was a liquidity crisis. This view leads to particular policy
conclusions.
6.4.1 Fighting a crisis

Consider the problem of policy response from the point of view of one of the Asian coun-
tries under attack. The attackers are motivated by the belief that if they go short on local
currency and long on dollars they will later be able to cover their short at a rate that pays
for the interest rate differential.
A first response might be to say that the country will never devalue. The speculators
will not believe that.
Next, the country can raise interest rates, pushing up the cost of maintaining the
attack for those short local rates; with luck the attackers will give up and close their posi-
tion, and the problem is over. This works provided the high local rates are not damaging
the economy, but if speculators perceive that the government is not prepared to take the
pain for as long as they are, they will wait the government out. If this does not work, the
country is in deep trouble. It can buy some time by closing off speculators’ access to cur-
rency, but eventually the speculators will do a deal with a local entity that will do the
109
transaction on their behalf. There is a myriad of ways to implement a speculative attack,

and the only way to plug all the holes would be to implement such draconian measures
that all foreign trade grinds to a halt.
The next step is to approach the IMF for a loan. The benefits come in two parts:
■ Soft. A humiliating set of anti-corruption and austerity measures should ratchet up

credibility with lenders by a notch or two and lend weight to statements about not
devaluing. Certainly such measures are likely to be popular with lenders, and other
things being equal should reduce local currency borrowing rates by reducing credit risk
and inflation expectations. Provided all goes well, this approach should be less damag-
ing to the economy than the rise in nominal rates rejected earlier.
Unfortunately, this optimistic outlook relies critically on the steps being credible.
Even if the government is entirely behind the process, there is plenty of potential for
disputes, ranging from votes of no confidence to mass riots. If these suggest that the
government cannot take the pain for long, then the overall effect may well be negative.
■ Hard. The country takes delivery of a large loan in dollars that has to be repaid in
dollars, with interest. This signals strong commitment because if the country devalues
it loses money on the deal. This is still problematic because if the excess foreign bor-
rowings were part of the problem to begin with, that problem just became worse, and
might offset all other benefits.
6.4.2 Long-term policy response

The affected countries drew the conclusion that a sovereign state, ultimately, can only
depend on itself if it is facing a liquidity crisis. While the IMF might be willing to help,
its aid may not prove particularly effective, it comes at an unpalatable political cost, and
it benefits mainly pre-existing lenders rather than the borrowing country, so a country
might prefer to rely on itself. According to this view, it is only prudent to take steps to pre-
vent such a crisis from happening in the first place. There are two alternatives.
First, a country can avoid having significant currency and maturity mismatches. This
may mean depending on local savings, matching the maturity of foreign loans to the
maturity of the ultimate investments and encouraging foreigners to make direct invest-
ments rather than just lending funds. The country may want to take steps to minimise
carry trading, and have an independent and credible central bank and a credible commit-
ment to convertibility. Of course, these may conflict.
Alternatively, a country might want to keep reserves higher than needed in any
plausible crisis. So long as it has more reserves than it has short-term foreign currency
borrowings, a liquidity crisis of the type discussed here should be forestalled. This
is exactly what the four crisis countries have done since the crisis. Before 1997, their
reserves were quite small but they have been building up rapidly since, as can be seen
in Figure 6.9.
However, reserves that are never meant to be used are useless, and having outsized
reserves might result in crises being less frequent than otherwise but being larger when
they do happen. Large reserves also come with a significant cost. A developing country
110
263
1990
225
2007
150
102 87
75 57
5 9 7 9
0
(a) Foreign reserves in 1990 and 2007, in billions of 2007 USD
60% Indonesia Malaysia

South Korea Thailand
45%
30%
15%
0%
1990 1992 1994 1996 1998 2000 2002 2004 2006
(b) Foreign reserves to GDP
Figure 6.9 Foreign reserves

with rapid growth provides much better investment opportunities than a wealthy devel-
oped country. It would seem more natural for a wealthy country to export its capital to
a developing country rather than the other way around. Indeed, that had been the norm
until recently; during the first globalism, between 1873 and 1914, the United Kingdom
(UK) was the greatest exporter of capital, and since then the United States (US), until
recently. The lesson drawn by the Asian Tigers, however, is that while the benefits are
shared, the risks of this arrangement fall almost entirely on the borrower.
By accumulating foreign currency reserves, the crisis countries, in effect, are subsidising
the wealthiest countries, in particular, the US as the owner of the world’s reserve currency.
This keeps interest rates in the US low and its currency relatively strong. This adversely
affects the exports of the high-reserve countries.
The policy adopted after the crisis was the sacrificing of wealth for greater financial
security and policy independence. It is an open question whether that is an advisable
policy. While this does subsidise US consumers, it also encourages them to buy the very
products exported by these countries, helping domestic industry.
There are also very conflicting opinions on the sensibility of maintaining substantial
holdings of US government bonds. Some might say that by becoming a large creditor, it
gives these countries power. Others highlight the potential for the US to effectively ‘tax’
the capital exporting countries simply by debasing its currency, perhaps by quantitative
easing (QE).
Some other developing countries have emulated the foreign exchange (FX) and reserve
policy of the East Asian countries. While that does not indicate the policies are sensible, it
does signal they have widespread support.
111
6.5 Role of the IMF

The IMF was severely criticised for its performance in the Asian crisis, condemned by almost
all commentators. Amongst the many critics are Ito (2007) and even more scathingly Weisbrot
(2007), who, with the benefit of hindsight, point out that the IMF misjudged the nature of the
crisis throughout. However, such criticisms might be founded on the view that the Fund has to
fix crises at any cost. That is not true: the Fund has many other priorities, including protecting
lenders. Furthermore, the Fund’s reaction to the Asian crisis, however misguided in subse-
quent analysis, needs to be considered in the context of the time of the decision.
The previous round of crises was in Latin America a decade earlier, where the underly-
ing factors were weak fundamentals. This suggests that the Fund’s approach to the Asian
crisis was based on recent experience, with the Fund a victim of the successful general’s
problem: fighting the last war.
In its defence, the Fund’s performance also reflected the many limitations of the envi-
ronment it operated in, the inability to do LOLR and the need to seek widespread consen-
sus along with slow decision making.
The Fund can be partly blamed for the build-up of the maturity and foreign exchange
mismatches. The rapid opening of capital accounts and liberalisation of financial markets
prior to the crisis were a part of the Washington consensus and of the IMF’s favoured
policy prescriptions.
Initial focus on moral hazard

The earliest reports of the emerging crisis focused on moral hazard and corruption as a
significant, and even the only, causing factor. The IMF embraced this view. The financial
systems certainly could do with more discipline, but should that be pursued to the mid-
dle of a crisis, when it is hard to separate insolvency from liquidity problems? Demanding
more capital and seemingly arbitrarily closing down financial institutions just deepens
the crisis, but the problem for the Fund is that if it does not pursue reform policies when
it has the lever, it will never achieve them.
For example, there is no reason to believe that President Suharto of Indonesia had any
intention of asking the IMF for advice on how to clean up. The Fund’s sole opportunity
was when he wanted help, to which it could attach strings.
The worst example seems to have been the closure of the 16 financial institutions in
Indonesia. Not only did those 16 seem to be chosen arbitrarily, but the closure signalled
that anybody having dealings with the financial sector in Indonesia might be facing losses
because of policy decisions. Not surprisingly, runs on the banking sector and the coun-
try followed, causing significant damage. This could have been prevented by a more
nuanced policy response.
Initial focus on structural adjustments

The second plank of the IMF programme was structural adjustment. This included macro-
economic policies such as fiscal and monetary tightening and improving capital accounts,
but also microeconomic adjustments like labour market reform, cuts in subsidies and
anti-corruption policies.
112
6.5 Role of the IMF
The fiscal and monetary tightening was questionable. None of the countries had dem-
onstrated a lack of discipline on that front prior to the crisis, and a sensible crisis response
was loosening, not tightening. The policies demanded by the Fund also were out of touch
with the underlying realities; for example, it set an inflation target for South Korea of 5.2%
for 1998, an impossible task given the sharp currency depreciation.
To achieve this, interest rates in South Korea were pushed above 20%, hitting the
industrial base, already reeling from the crisis, and further deepening the recession.
The demand for high interest rates disregarded the accumulated experience of fighting
crises over the preceding century or more, and is contrary to the policies used subse-
quently, for example in the ongoing crisis. The Fund has demanded very high inter-
est rates in some countries to which it has provided assistance in the ongoing crisis,
suggesting a deep-seated attachment to what most analysts consider an inappropriate
policy response.
Other reforms, such as cuts to subsidies and labour market reforms, might have been
laudable long-term goals but are hard to implement in the middle of a crisis. In Indonesia,
the resulting cuts in food and energy subsidies led to riots, and the end result in South
Korea was mass layoffs, leading to strikes and riots.
The problem is that the IMF can only exert influence when a country wants a
cheque. While riots are not palatable, avoiding them does not guarantee good policy.
The IMF’s policy was unpopular and disruptive, but one could argue that its structural
adjustment programmes laid the groundwork for the strong sustainable post-crisis
growth.
Heavy-handedness
The IMF was seen as acting heavy-handedly by the crisis countries, with the signature
manifestation of this being the signing ceremony of the agreement with Indonesia in
2008. In order to demonstrate commitment, both President Suharto and Managing
Director Michel Camdessus attended the ceremony. The photo taken at the occasion –
see Figure 6.10 – became the visible face of the IMF’s power. Of course, the photo says
nothing about the actual relationship between the IMF and Indonesia, but it did resonate
with popular perceptions of events.
Strong resentment against the Fund lingered, and the countries did their best to repay
their loans ahead of schedule: South Korea in 2001, Thailand in 2003 and Indonesia in
October 2006. The resentment was fuelled by politicians deflecting blame from them-
selves, by the real economic disruption in the crisis and by the dislike anybody feels when
beholden. Of course, the ability to repay early can also be seen as a testament to the easy
terms of the loans and the effectiveness of the Fund’s policy demands.
6.5.1 IMF as a lender of last resort

If the crisis was essentially a liquidity crisis, the optimal real-time response was a suffi-
ciently large provision of liquidity – LOLR.
The IMF, however, is not a suitable LOLR. LOLR operations can be quite large and any
LOLR needs to have ready large amounts of liquid money. Similarly, any provider of LOLR
113
Figure 6.10 IMF Managing Director, Michel Camdessus and President Suharto, in 1998
Source: Agus Lolong/Getty Images
needs to be able to act decisively, providing funds in a very short amount of time. That
is exactly the advantage the central banks have in the provision of liquidity during crises.
As an international lender of last resort (ILOLR), the IMF does not have the ability to
print money on demand, and needs a significant amount of time to get support for lend-
ing operations. Therefore, the Fund is not in a position to be a LOLR, except perhaps in
special cases. It is consequently unable to deal directly with an international liquidity
crisis. If it is not the IMF in such circumstances, it is nobody.
Furthermore, the loans from the IMF are in dollars. If the dollars are spent defend-
ing the local currency, which then devalues, the government’s local currency liability is
increased and this will worsen whatever the market is worried about — unsustainability
of finances, excess foreign borrowings, etc. Hence, the dollar loans would do little to
change market perception. To be effective, the loans would have to be fully in local
currency and the IMF would need to put its money where its mouth is, as one expects
of a LOLR.
This means that comparing the scale of the IMF loans to the scale of changes in reserves is
not correct. IMF loans enable countries to take bigger bets, but the loans do not remove the
risk, and the country is presumably under attack because the risks were already perceived as
excessive.
That said, the IMF does represent member governments and hence has consider-
able powers of persuasion beyond simply providing funds. In January 1998, it helped by
114
6.6 Wider lessons
jawboning, backed up by realistic carrots and sticks. Still, the IMF packages were in hind-
sight clearly insufficient, even though that was not clear at the time.
What the IMF could have done

What was primarily needed in the summer of 1997 were not structural reforms nor hec-
toring over moral hazard but LOLR at the onset of the crisis. If foreign investors had not
feared significant losses if they rolled over loans, the sudden stop would not have hap-
pened, the speculative attacks would not have occurred and most of the crisis would have
been averted. The countries would then have had time to improve their capital accounts
and address the most direct vulnerabilities. It is an open question whether the countries
would have taken advantage of this breathing space.
The IMF did not see things this way in 1997, and even successfully prevented other
efforts to provide emergency liquidity assistance, most importantly a Japanese-led ini-
tiative in September 1997 for an ‘Asian Monetary Fund’, amounting to $100 billion,
to provide rapid liquidity assistance without many conditions. Besides Japan, several
countries including China, Taiwan, Hong Kong and Singapore signed up. This met
strong opposition from the US Treasury and the IMF, indicating that the Fund does not
want competition.
Eventually, the Fund adjusted its position, moving away from immediate structural
reforms and especially fiscal macroeconomic tightening, towards helping the countries
cope with the liquidity crisis.
6.6 Wider lessons

The Asian crisis was a typical crisis, and any student of financial crises from the nineteenth
century would have recognised events and mistakes made both by the countries them-
selves and by the international community.
The underlying causes of the crisis were familiar. The East Asian countries had a desire
for rapid economic growth, and therefore needed high investment rates. Since domestic
savings were not enough, they had to resort to international capital markets. Being over-
confident of their ability to repay, they were prepared to accept terms that exposed them
to risks that proved to be excessive. Similar mechanisms have been at the heart of many
previous crises.
The authorities could have attempted to minimise the risk by keeping mismatches low,
but instead the balance was swung towards excessive risk by the opening of the capital
account and financial liberalisation, without appropriate safeguards. Left to their own
devices, the overconfident tend to go for the shortest-term funding because that is the
cheapest.
There is nothing unique about the Asian crises and, perhaps with benefit of hindsight,
observers should have seen the vulnerabilities build up. The fact that nobody was con-
cerned at the time is a cause for worry. Perhaps one reason is that crisis watchers were
focusing on the last round of crises, in Latin America. It is a reminder of how much easier
it is to diagnose problems with hindsight than when they are building up.
115
We should continue to question the extent to which the lessons that have been learned
from the last crisis will be relevant to the next. In the Asian crisis, as is so often the case, the
most damaging decisions were not those made before the crisis, but those during the crisis.
A longer-term view, focusing beyond the immediate newspaper headlines, might have dem-
onstrated that the crisis was as much about liquidity as structural deficiencies. Provision of
liquidity, backed up by a credible reform programme, would have prevented the worst.
Thinking back, it now seems that the countries’ governments have not sufficiently
taken on board the necessary lessons, perhaps with the exception of South Korea. The
objective of the structural adjustment programmes was to help the affected countries,
and their failure to grow sufficiently, demonstrated by their failure to improve their global
ranking post-crisis, suggests that the programmes did not take hold.
This is partly because the issues were misdiagnosed at the time, but also because
many of the regimes reverted to type as soon as the pressure came off. A thorough imple-
mentation of many of the Fund’s programmes would probably have helped the countries
perform better. For example, corruption remains a serious problem in all of the coun-
tries, except perhaps South Korea, and the IMF’s anti-corruption stance is laudable.
The crisis marks the beginning of a reduction in GDP per capita growth, but correlation
does not imply causation. The very high pre-crisis growth was in part based on unsustain-
able policies, such as the dependence on short-term foreign capital. Therefore, it is just as
likely that the post-crisis slower growth is more sustainable and less prone to crises, and
therefore to be preferred.
Ultimately, this means that Asian crisis countries remain vulnerable. Their large cur-
rency reserves might provide protection, but in the absence of sensible reforms, helping
to move them up the relative wealth ladder, a future crisis remains a likely eventuality.
6.6.1 Relation to the ongoing crisis

The Asian crisis has many connections with the ongoing crisis. It is a direct causal factor,
because the crisis response at the time – the low exchange rates, accumulation of foreign
reserves and export-led growth – all contributed to low interest rates and low inflation in
developed countries. This fuelled the growth and excessive optimism, leading to the 2007
crisis.
There are also direct parallels between the Asian crisis and the ongoing crisis. The
authorities in the US and Europe were quick to identify liquidity as a main cause, providing
vast amounts of funds to the financial system. In spite of considerable evidence that risk
taking had been excessive prior to the crisis, concerns about moral hazard and reduction
in risk, with the sole exception of the decision to allow the collapse of Lehman, have very
much taken the backstage in the crisis-fighting efforts.
Furthermore, there has been little serious discussion of structural reforms in the main
industrialised countries suffering from the crisis; reform is focused on those receiving bailouts.
The current phase of the global crisis has perhaps more parallels with the Asian crisis.
Here, the main crisis countries – Ireland, Greece, Spain, Portugal and Cyprus – play the
role of the Asian countries, and Germany and other northern European countries act as
the international creditors, with the IMF in a similar role.
116
6.7 Summary
Before the European crisis countries get assistance, they have to implement extreme
austerity programmes and structural reforms, just like the East Asian countries had to do
in 1997. Also, while some liquidity assistance is provided, it is far from enough to assure
the markets. Meanwhile, no end to the European sovereign debt crisis is in sight.
6.7 Summary
Some countries in East Asia suffered a severe financial and economic crisis in 2007, in
particular South Korea, Malaysia, Indonesia and Thailand. Of these, all but Malaysia
requested assistance from the IMF.
Prior to the crisis, the fundamentals appeared sound, and the crisis came as a surprise,
but underneath were significant private sector problems, in particular, extensive short-
term foreign currency borrowing.
Initial explanations for the crisis focused on weak fundamentals and moral hazard,
with subsequent analysis highlighting the importance of liquidity.
The IMF came under strong criticism for its performance in the crisis, but it was ham-
pered by institutional difficulties beyond its control. Many of the mistakes made by the
Fund are visible only with the benefit of hindsight.
Some of the lessons from the Asian crisis have been taken on board in the current crisis, espe-
cially the early stages. Unfortunately, similar mistakes now appear to be repeated in Europe.

1 How does the Asian crisis differ from the previous major regional crisis, in Latin America?
2 What does it mean when we say that the economic fundamentals were sound but not
the financial fundamentals?
3 Compare and contrast the causes of the crisis, the resolution of the crisis and the post-
crisis reactions across the three countries.
4 Some commentators have argued that panic and contagion played an important role in
the crisis. What is your view on this analysis?
5 Initial explanations for the crisis focused on moral hazard and related issues. These
were eventually dismissed as a main cause. Why?
6 Later analysis, especially in the countries affected, focuses on liquidity. Explain the basis
of the argument; discuss why it might be correct and what might be missing.
7 What was the role of the IMF? Did it give necessary help and support to the countries
concerned or did it make matters worse?
8 Is the IMF a suitable provider of liquidity, or even a LOLR?
9 Are there parallels between the Asian crisis and the crisis dating from 2007?
117
References
Alba, P., Bhattacharya, A., Claessens, S., Ghosh, S. and Hernandez, L. (1998). Volatility and con-
tagion in a financially integrated world: lessons from East Asia’s recent experience. Paper pre-
sented at the PAFTAD 24 Conference, ‘Asia Pacific Financial Liberalization and Reform’, 1998.
Goldstein, M. (1998). The Asian Financial Crisis: Causes, Cures, and Systemic Implications.
Institute for International Economics, Washington DC.
International Monetary Fund (1997). Camdessus announces IMF support for Indonesia’s eco-
nomic program. News brief no. 97/19, 8, October 1997, www.imf.org/external/np/sec/
nb/1997/nb9719.htm.
Ito, T. (2007). Asian currency crisis and the International Monetary Fund, 10 years later:
Overview. Asian Econ. Policy Rev., 2: 16–49.
Radelet, S. and Sachs, J. (2000). The onset of the East Asian financial crisis. In Krugman, P., edi-
tor, Currency Crises. University of Chicago Press. www.nber–org/chapters/c8691.
Weisbrot, M. (2007). Ten years after: the lasting impact of the Asian financial crisis. Technical
report. www.cepr.net/documents/publications/asia_crisis_2007_08.pdf.
118
7 Banking crises
Policymakers formulating banking policy and fighting banking crises need to consider
several conflicting objectives. We want banks to actively finance economic activity and
hence take on risk, but at the same time we want to curtail excessive risk-taking. Once a
crisis is under way, we need to balance moral hazard considerations against the benefit of
robust recapitalisations of banks, all whilst trying to minimise the costs to the taxpayers.
Banking crises cannot easily be avoided, as they are an inevitable consequence of
having a vibrant banking system and economy. The only way to completely prevent
banking crises is to restrict the banking system so severely that banks are unable to
fulfil their socially useful role of financing risky activities, seriously holding back eco-
nomic growth. What we can do is to reduce the incidence of banking crises and to
minimise the costs of fighting them once they are underway.
Banking crises tend to have the same underlying causes. When times are good, banks
enjoy access to significant amounts of credit. The banks, flush with money, look for bor-
rowers, but traditional companies can only absorb so much money, not enough for all the
liquid funds available. Therefore, as the banks search for borrowers, those they find are
increasingly of low quality, often involved in real-estate speculation. Initially, the result-
ing weaknesses in the banks’ loan books are not very visible since the price of real estate
and equities increases sharply, creating a virtuous cycle between bank lending and prices.
High prices create the illusion of a good collateral which further stimulates lending.
Eventually, valuations of assets are increasingly out of tune with the underlying eco-
nomic fundamentals, and as in the tale by Hans Christian Andersen, it is as if some-
body cries out ‘the Emperor has no clothes’ and everything reverses at warp speed,
prices collapse and credit is withdrawn. This is typical of endogenous risk.
Chapter 7 Banking crises
The reason why such booms and busts in asset prices cause banking crises is the
inherent fragility of banks that comes from fractional reserve banking. Under this
arrangement, when depositors put their money into a bank, the bank then turns
around and lends most of it out, keeping a small fraction as reserves. The fragility
arises because deposits generally are of short maturity, and some can be withdrawn
whenever the depositor wants – demand deposits – whilst the loans tend to be of
longer maturity. If a sufficiently large number of depositors want their money, the
bank will run out of cash, because it cannot similarly call on its own borrowers to
repay their loans. We call such an event a bank run. Bank runs are contagious, and can
spread quickly throughout the financial system. The reason is that the banking system
is built on trust, so if depositors lose confidence in banks they flee to cash.
Bank failures matter because they create negative externalities, adversely impacting
on the economy. Banks provide essential services, and without banks companies can-
not do business, nor can individuals go about their lives in their usual way. The failure
of an individual bank might not be that costly because the authorities have in place
robust mechanisms for preventing collateral damage from bank failures, but if we expe-
rience a wave of bank failures – a banking crisis – the impact on society will be serious.
The incidence and seriousness of banking crises in the developed world has fallen
significantly since the Great Depression because the authorities in most countries
have in place effective regulations preventing bank failures and robust mechanisms
for coping with the failure of individual banks, preventing failures from spreading to
the entire banking system.
Historically, this was not the case, and bank crises have caused significant economic
damage. For example, Kupiec and Ramirez (2009) studied the United States (US) from
1900 until 1930 and found that bank failures significantly reduced economic growth.
A one-standard-deviation shock to the share of liabilities in failed banks was found to
cause a 17% decline in industrial production and a 4% decline in GDP.

This chapter draws on theoretical concepts discussed in Chapter 1 (systemic risk),
Chapter 3 (endogenous risk), Chapter 4 (liquidity), Chapter 8 (bank runs and deposit
insurance), Chapter 14 (bailouts) and Chapter 13 (financial regulations).
Key concepts
■ Banking crises
■ Moral hazard
■ Good bank – bad bank
■ Causes of banking crises
■ Why it is so hard to prevent banking crises

No specific readings are required for this chapter as the material is self-contained, but
for a good survey see Kaminsky and Reinhart (1999), Banking Committee on Banking
Supervision (2004), Caprio and Honohan (2009) and Reinhart and Rogoff (2009). For
120
7.1 Money and early banking
a good account of early banking history and further discussions on fractional reserve
banking, de Soto (2009) is an excellent book. Ferguson (2008) documents the his-
tory of money and Graeber (2011) provides a comprehensive survey of the history of
money and credit.
7.1 Money and early banking

Money is defined by its function as a means of payment in exchange: in a monetary
economy, goods and services are bought and sold in exchange for money. Historically,
various goods have served as money, anything from seashells to copper. For example, the
Swedish government established a copper standard in 1625, which turned out to be an
effective way to prevent theft, but transaction costs were somewhat high.
Over time, the basic unit of money has converged to precious metals, a portable com-
modity which derives its value from its scarcity, with silver and gold most common. By the
early nineteenth century, the world was divided into three metallic blocs: gold, silver and
both (bimetal). Over time gold became dominant, lasting until 1914, due to the politi-
cal and military dominance of the gold bloc. Various forms of the gold standard played
an important role in the various monetary arrangements between the two world wars,
and gold was the cornerstone of the Bretton Woods system. Since the US left the Bretton
Woods system, and hence the gold standard, in 1971, almost all countries have used fiat
money, which is money that has value because of government regulation or law but is
without intrinsic value. Consequently, fiat money necessarily has to be legal tender. Fiat
money is not a new invention; an early example is its introduction in the twelfth century
in China as documented by Selgin (2003), leading to the first recorded nationwide infla-
tion. The first European bank to print banknotes (fiat money) was Stockholms Banco in
1656. It eventually printed too much money and went bust.
A major problem with most forms of money is the temptation of governments to debase
it. There are examples going back to the Roman Empire of governments reissuing coins in
the same denomination, but increasing the relative content of cheap metals at the expense
of gold and silver – the original form of debasement. The more recent practice of printing
too much money and creating inflation is simply a modern form of currency debasement.
Early banking
It is unclear who was first engaged in fractional reserve banking. Initially, such banking
was illegal, not least because of the Catholic Church’s ban on the charging of interest –
usury. Fractional reserve banking began to gain legal recognition as governments realised
they could also benefit by borrowing large sums of money from banks, often to finance
wars. Inevitably, banks often failed because of sovereign defaults.
Below we discuss two examples of early banks, the famous Medici bank of Florence,
which functioned as a fractional reserve bank, and the Amsterdamsche Wisselbank in the
Netherlands, which operated as a full reserve bank for most of its existence. We will see
how they innovated and made their money, what risks they faced and how they eventu-
ally failed.
121
Medici bank (1397–1494)

The Medici bank, 1397–1494, was the largest and most respected bank in Europe
during the fifteenth century, making the Medici family perhaps the wealthiest family in
Europe. The Medici bank was highly diversified, being involved in silk and cloth manufac-
turing as well as facilitating trade, for example, by lending to English sheep farmers and
wool merchants in return for buying their goods for low prices, thus bypassing usury laws.
The Medici bank made most of its money by selling bills of exchange, which were
invented to circumvent usury laws. When a depositor paid money into a branch of a
Medici bank, the general manager of the branch issued a bill of exchange to the deposi-
tor, instructing another Medici branch in a different country to pay the money back in
local currency upon the depositor presenting the bill, but at a pre-agreed exchange rate.
Thus, the interest rate charge could be hidden in the exchange rate.
The way the bank facilitated trade can be seen in Figure 7.1. If a merchant was travelling
from Florence to London, he could buy a bill of exchange for 15 florins from the Florence
branch, agreeing at the time of purchase that the London branch would cash the bill at
an exchange rate of 3 florins per pound. Three months later when the merchant arrived in
London, if the florin had appreciated against the pound, the merchant would suffer a loss.
This instrument meant that the Medici bank faced a significant exchange rate risk, and the
heads of different Medici branches often wrote to each other to obtain information and
forecast the direction of exchange rates. This risk could be hedged by offsetting trades, for
example, facilitating the business of a London-based merchant travelling to Florence.
Bills of exchange were freely traded, and were effective in acting as a medium of
exchange. The Medici bank has also been credited with the invention of double entry
accounting; all in all, it was a hugely profitable business, making around a 32% annual
return on equity from 1397 to 1420.
The Medici bank eventually failed for a variety of reasons, one of them too much lend-
ing to high-risk borrowers, like sovereigns. Just one example is the failure of the London
branch which lent too much to the English King Edward IV who could not repay due to
15 florins
Medici
Merchant
Florence
Bill of exchange agreed at
3 florins per pound
(a) Initial deal
Bill of exchange
Medici
Merchant
London
5 pounds
Gains if florin Gains if florin
depreciates appreciates
(b) Three months later
Figure 7.1 An example of a bill of exchange transaction
122
7.2 Moral hazard
a civil war. Failure of the Bruges branch followed, due to poor management, fraud and,
again, excessive lending.
Amsterdamsche Wisselbank (1609–1795)

In the late 1500s, Amsterdam was the up and coming financial centre. Its merchants pio-
neered new methods and the city’s wealth multiplied. The growth in its prosperity was
held back by the large number of foreign coins in circulation. The burghers of Amsterdam
solved that problem by creating a new form of bank, in a way an early form of a central
bank, called the Amsterdamsche Wisselbank, backed by the city of Amsterdam.
It was fundamentally based on a type of money, new to Europe but used earlier in
the middle east, known as bank money. The bank received foreign and local coins at
their intrinsic value, deducted a management fee, and then credited the client’s account
with bank money for the remainder value. The bank was helped by a law stating that
all bills drawn in Amsterdam and worth more than 600 guilders had to be paid in this
bank money, in effect, granting the bank a monopoly on the issue of currency. By allow-
ing merchants to set up accounts, the bank pioneered the system of cheques and debits.
Bank money was secure from accidents and theft, and essentially free from debasement
worries. Often bank money was worth more than its nominal value.
Most of the banks’ capital was composed of gold and silver bullion deposits, which
clients deposited for safekeeping. The client would receive a receipt for an amount that
was 5% less than the mint price, and could withdraw their bullion upon presenting the
receipt. Like the bills of exchange, receipts were freely traded; if a merchant was in need
of coins, he could sell his receipt, and if he wanted bullion, he would purchase a receipt.
The bank allowed no withdrawal of bullion except by means of receipt, and it maintained
that it did not lend out the deposited bullion.
This meant that the amount of deposits was almost the same as the money supply
and there was no credit expansion. In other words, Amsterdamsche Wisselbank practised
full-reserve banking. It is a practice in which the full amounts of each depositor’s funds are
available in reserve, as cash or other highly liquid assets. This made the bank very secure
but not very profitable. The bank made most of its profits from various fees. The bank
survived for almost 200 years, and its eventual demise began when the city’s government
ordered it to hand over its deposits to finance the Anglo-Dutch war; the bank also lent out
funds secretly to other provincial states of the Netherlands and the East India Company.
Its reserve ratio declined from 100% to 25% and it stopped exchanging receipts for bul-
lion when a liquidity crisis hit in 1790–1791, destroying its reputation.
7.2 Moral hazard

Moral hazard is fundamental to any study of financial crises.
Definition 7.1 Moral hazard Moral hazard is what happens when those taking risk
do not have to face the full consequences of failure but get to enjoy all the benefits of success.
The consequence of moral hazard is that those fortunate enough to be in that situation are
encouraged to take on more risk than they otherwise would do.
123
Moral hazard is pervasive throughout the economy, such as in standard insurance con-
tracts. Moral hazard is not something that should be eliminated, since that would sharply
curtail socially optimal risk taking. For example, an important development in modern
economies was the creation of limited liability corporations, firms where the owners do
not face the full consequences of failure but get all the benefits. Even though limited
liability corporations create moral hazard, few would argue they should be banned or
that the underlying moral hazard should be substantially reduced, since the benefits out-
weigh the costs.
Moral hazard is usually not much of a problem when those subject to moral hazard get
compensated for doing so. That compensation is why insurance companies are content to
write insurance contracts in the full knowledge that such contracts induce the insured to
take more risk. To contain such risk taking, insurance companies write into the contracts
clauses designed to prevent excessively risky behaviour, and try to price risk properly.
Unfortunately, this is often not the case in the financial system. Banks receive explicit
or implicit guarantees from the government but such insurance is almost never priced.
Governments can try to prevent excessive risk taking, but this is also frequently ineffective.
Addressing this problem is a main motivation for financial regulations. The problem arises
because the externalities from the failure of an important financial institution can out-
weigh any possible moral hazard considerations. This leaves the authorities with a typical
risk–return trade-off between the benefits of preventing panic now against the costs of
inducing riskier activity later.
Financial institutions, of course, know this and are incentivised to take on too much risk
when things are good in the full expectation of a bailout when things are bad. The bigger,
the more dangerous, the worse-run and more interconnected a financial institution is, the
more likely it is to be bailed out.
Taken together, the moral hazard problem in the financial industry is particularly hard,
with no obvious solution. In order to contain the problem, governments regulate the finan-
cial industry, but it can be difficult to the point of impossible to effectively contain excessive
risk taking. The powerful tools at the government’s disposal tend to be very blunt and could
easily curtail risk taking excessively, whilst the more surgical tools may not be effective.
7.3 Costs of banking crises

The World Bank and the International Monetary Fund (IMF) have developed and made
public a very comprehensive database on financial crises. We made use of the 2008 ver-
sion of the data, due to Laeven and Valencia (2008), although they are only the most
recent maintainers of the database and many others have preceded them. All data men-
tioned in this section are from this database unless otherwise mentioned. The latest ver-
sion of the data can be downloaded.1
Before the crises from 2007, the frequency of banking crises had been slowing down,
as indicated by Figure 7.2, with the worst year being 1995 when 13 countries suffered
1
www.imf.org/external/pubs/cat/longres.aspx?sk=26015.0.
124
7.3 Costs of banking crises
12
0
1978 1981 1984 1987 1990 1993 1996 2000 2003
Figure 7.2 Frequency of banking crises

Data source: Laeven and Valencia (2008)
banking crises. The relative calm before the 2007 crisis led to the misleading conclusion
that the world was becoming safer; instead, this period was more like the calm before
the storm.
It is remarkably hard to make sustained changes in levels of risk, because improvements
addressing existing fears have the side-effect of improving confidence, leading to greater
risk taking.
Costs
Banking crises impose costs on the economy in two main ways. The first is from the direct
fiscal costs of resolving a crisis, like the provision of deposit insurance, liquidity support,
recapitalisation and the like. This cost can be more or less accurately measured.
The second cost is the indirect effect on the economy, for example, the impact on
asset markets, reduced consumption and investment, the impaired channelling of funds
from savers to ultimate users, the contraction of government revenues, sharply expanding
fiscal expenditures and the shortfall in economic growth. This is much harder to meas-
ure, and the available studies inevitably have to rely on fairly crude methods to do this.
Consequently, we do need to take such measurements with reservation. Laeven and
Valencia (2008) estimate output losses by extrapolating from the trend of real GDP prior
to a crisis, and calculating the aggregate difference between the trend prediction and
actual outcomes during and after a crisis. There are many other ways of doing this analysis.
This calculation will often overstate the output losses. For example, suppose a country
is on an unsustainable growth path fuelled by inflows of foreign money. In this case, the
GDP growth in the years before the crisis will be artificially high, and the output loss,
therefore, will be overstated. Since most of the necessary macroeconomic data are in
publicly available databases, it would be straightforward to use alternative methodolo-
gies to calculate the costs of crises, and many authors have done so.
A small sample of the results is shown in Table 7.1. We see that Argentina in its 1980
crisis used more than half of its GDP to resolve the crisis, but the impact on GDP was rela-
tively small at 11%. This calculation is of course complicated by the fact that Argentina
experienced hyperinflation at the time.
The Scandinavian crisis of the early 1990s imposed significant direct fiscal costs on
Finland, but less so on Sweden. However, the output shortfall is significant. Note that it is
likely to be overstated because the GDP was artificially high at the top of the boom, and
125
Table 7.1 Country systemic banking crises. Fiscal cost and output loss
are represented as percentages of GDP
Fiscal cost Output loss
Argentina 1980 55.1% 10.8%

Finland 1991 12.8% 59.1%
Indonesia 1997 56.8% 67.9%
South Korea 1997 31.2% 50.1%
Mexico 1994 19.3% 4.2%
Sweden 1991 3.6% 30.6%
Turkey 2000 32% 5.4%
United States 1988 3.7% 4.1%
Venezuela 1994 15% 9.6%
Source: Systemic Banking Crises: a New Database, IMF Working Paper, pp. 34–51 (Laeven, L. and
Valencia, F. 2008), International Monetary Fund
these numbers do not quite correspond with those in Section 7.5.3 below, produced by
one of the central banks involved.
The size of the country matters: while the cost of the S&L crisis in 1988 in the US was
quite significant in monetary terms, relative to GDP the cost was not very high, in part
because only a small part of the financial sector was directly affected.
Costs of resolution
The IMF database was originally developed by the World Bank and some of the devel-
opers of the database published an interesting paper (Honohan and Klingebiel, 2003)
focusing on the fiscal costs of banking crises and how the various government responses
contribute to this cost. They found that if countries do not extend some policies of unlim-
ited deposit guarantees, open-ended liquidity support, repeated recapitalisations, debtor
bailouts and regulatory forbearance, the fiscal costs of resolution will be around 1% of
GDP on average, one-tenth of the actual costs. If, however, governments employ all of the
approaches above, the fiscal costs will be six times larger than in actuality.
It is important to note, however, that quantifying the costs of crisis and government
action, or inaction, is quite challenging. There are strong causal effects, where a bad crisis
might lead to bad outcomes, so that some of the policy measures taken might be associ-
ated with the severity of the outcomes.
It is also hard to estimate the cost of inaction. For example, the decision to allow
Lehman to fail was followed by a reversal of policy and effectively unlimited support to
many other institutions. Many commentators argue that the cost of continued inaction
would have been much higher than the cost of this support, but this remains controversial.
7.4 Causes of banking crises

The causes of banking crises are usually fundamentally the same: financial institutions
over-expand during good times, artificially inflating asset prices, creating positive feed-
back loops between bank lending, market prices and firm profitability. This is often
126
7.4 Causes of banking crises
coupled with inflows of hot money from abroad, and/or an accommodating fiscal and
monetary policy.
However, such booms are artificially created by money, not fundamentals, and even-
tually prices become so misaligned with the underlying economy that a small shock can
trigger a rapid reversal. This means that banking crises are often directly linked with mac-
roeconomic crises.
There are often country-specific differences. A report from the Basel Committee for
Banking Supervision (BCBS) in 2004 discusses the origins of banking crises in G10 coun-
tries and identifies several common factors – the regional economy, asset prices, financial
liberalisation and poor regulation in addition to bank specifics. We report a subset of
these results in Table 7.2. What is interesting is that bank-specific factors are least com-
mon and financial liberalisation is always present.
Financial liberalisation
Many governments faced with a heavily regulated financial sector and anaemic growth
opt for financial liberalisation in the belief that it may promote growth. If a government
opts for this policy, the execution does have to be right. A common mistake is to reduce
oversight and activity restrictions but maintain implicit or explicit government guaran-
tees such as deposit insurance. This creates a nasty moral hazard problem because it can
enable financial institutions to borrow cheaply and use the money for high-risk activities.
Deregulation has been at the core of many banking crises, such as the S&L crisis in the
US in the 1980s and the Scandinavian crisis in the late 1980s and early 1990s.
Corruption
A frequent cause of individual bank failures is corruption and mismanagement. We dis-
cuss one example in detail below, that of BCCI. Caprio and Honohan (2009) describe two
examples of what they call bad banking and bad policies in Latin American countries,
in Venezuela in 1994 and the Dominican Republic in 2003. In both cases, it involved
banks that were so large as to be systematically important. The banks did not seem to
Table 7.2 Banking crises in G10 economies
Macroeconomic Banking
factors system
Real Asset Financial Poor Bank-specific

Shock economy prices liberalisation regulation factors
Switzerland (1991–1996) ✓ ✓ ✓ : :
Spain (1978–1983) ✓ ✓ ✓ ✓ :
UK (1991) ✓ : ✓ : ✓
Norway (1988–1993) ✓ ✓ ✓ ✓ :
Sweden (1991–1994) ✓ ✓ ✓ ✓ :
Japan (1994–2002) ✓ ✓ ✓ ✓ :
US (1982–1995) ✓ ✓ ✓ ✓ ✓
Source: Table 6 in BCBS Bank Failures in Mature Economies, Basel Committee, Working Paper 13, p. 67
(Ms Natalja v. Westernhagen et al. 2004), Bank for International Settlements
127
record deposits as liabilities, enabling their insiders to loot the banks’ assets from inside,
that is, steal the banks’ assets. Because of the banks’ systematic importance, the central
bank of each country felt it necessary to make depositors whole, destabilising the macro-
economy. Furthermore, the rogue bank in Venezuela paid high deposit rates, forcing other
competing banks to do the same.
Zombie banks
If financial institutions, in effect, are insolvent but able to continue operating because
of government support, they are often referred to as zombie banks. The term was first
used by Kane (1989) when referring to the S&L crisis, but it became especially com-
mon when referring to the Japanese banking sector following that country’s crisis in
the early 1990s. See BCBS (2004) for more details on the Japanese crisis. In the 1980s,
Japanese banks made large loans that eventually turned out to be bad, with many
involving real estate. Instead of resolving the bank failures, closing down the failed
banks or recapitalising them, the government opted to keep them on life support,
often using a process called evergreening, whereby a bank is allowed to keep a non-
performing loan on its books as if it were performing. That often involved lending
money to failed borrowers, just so that they could repay old loans, steadily increasing
loan losses to the bank.
Eventually, this meant that a large proportion of the Japanese banking system con-
sisted of zombie banks, significantly associated with the Japanese financial and economic
malaise. While the zombies were not the cause of the crisis, their presence adversely
affected recovery.
The Japanese lesson is that policymakers should prevent the emergence of zombie
banks at all costs, and hence aim to resolve, restructure or shut down failed banks as
quickly as possible. That was the approach taken by the Scandinavian governments in
resolving their banking crisis in the early 1990s. In the ongoing crisis, the governments of
the US, Switzerland and the UK have actively tried to prevent the emergence of zombie
banks. However, this has not been the case in many of the European countries affected
by the European sovereign debt crisis, and there are significant fears that zombie banks
may be emerging in Europe. The fact that the European Central Bank (ECB) has become
the sole provider of liquidity for certain banks, and even entire national banking systems,
whilst losses have not been properly recognised, supports the view that some European
governments prefer to let their banks turn into zombies to make the necessary effort to
resolve their banking crises.
7.5 Bank and banking system failures

7.5.1 Individual bank failures
Among the many individual bank failures throughout history, several had an especially
important impact on future policy and financial regulations. We discuss three of the most
important cases below.
128
Herstatt
Herstatt was a large German bank that was forced into liquidation by the German author-
ities on 26 June 1974. On that day a number of banks had released payments of German
marks to Herstatt in Frankfurt in exchange for USD that were to be delivered in New York.
Because of time-zone differences, Herstatt ceased operations between the times of the
respective payments. Herstatt had taken receipt of payment from European banks (in
marks) but not yet made the corresponding dollar payments. Consequently, mostly US
creditors were left holding unsecured claims. This is an example of settlement risk. The
failure of Bank Herstatt was one factor that led to the creation of the continuous linked set-
tlement platform, which was launched almost 30 years later in 2002. For more details on
the Herstatt case, see BCBS (2004).
Banco Ambrosiano
Banco Ambrosiano was an Italian bank which collapsed in 1982 with important conse-
quences for future regulations. See, for example, Stoler and Kalb (1982) for a description
of events, and Herring and Litan (1995) for the regulatory implications. It was the largest
private banking group in Italy in its time, with operations in 15 countries. At the centre
of the bank’s failure was its chairman, Roberto Calvi, called by the Italian press ‘God’s
Banker’ due to his close association with the Holy See.
Calvi was determined to transform his bank into a major international financial insti-
tution from a relatively small regional bank with strong religious overtones. One of his
initial steps was to form a Luxembourg holding company not subject to Italy’s banking
regulations.
Calvi’s problems began in 1978, when the Bank of Italy conducted an extensive audit
of his financial empire, noting unorthodox operations involving $1.2 billion in unsecured
USD borrowings. Calvi was buying up Ambrosiano stock, using money borrowed on inter-
national financial markets. The bank collapsed because the Italian lira fell relative to the
dollar.
Calvi was sentenced to four years in jail, but released pending appeal. He fled Italy but
was found hanged under Blackfriars Bridge in London in 1982. The failure of Ambrosiano
left more than 200 international financial institutions with large losses, threatening the
stability of the entire international banking system. Ultimately, this brought changes in the
way the world’s major banks do business. The BCBS responded to the Banco Ambrosiano
collapse in 1983 with the first major revision of the Concordat.
BCCI
BCCI was a bank registered in Luxembourg with head offices in Karachi and London.
In July 1991, BCCI, one of the 300 or so branches and subsidiaries of foreign banks
operating in London, failed because of widespread fraud. BCCI’s complex structure
consisted of a holding company, incorporated in Luxembourg, and two main sub-
sidiaries incorporated in the Cayman Islands and Luxembourg. See Truell and Gurwin
(1992) for more information. BCCI had branches in over 70 countries with the UK
offices being branches of the Luxembourg subsidiary. Its principal shareholders were
in Abu Dhabi.
129
It is now believed that BCCI’s financial statements had been falsified from its establish-
ment in 1972. A scheme of deception was developed to support BCCI’s rapid growth and
to conceal lending losses. To achieve this, BCCI failed to record deposit liabilities and cre-
ated fictitious loans that generated substantial profits. Fraud also took place within BCCI’s
treasury operations. BCCI used depositors’ money to fund their own proprietary trading
activities, and covered up the resulting losses with more fictitious loans.
Prior to these problems surfacing in 1990, supervisors and commercial bankers were
wary of BCCI because of its rapid growth and opaque corporate structure. However, while
BCCI was sometimes mentioned in the press ‘chiefly for the mystery that surrounded
it’, financial market participants generally saw BCCI as a bank that made losses through
incompetence rather than fraud.
From the spring of 1990, concerns about the evidence of fraud within BCCI led to
ongoing discussions between BCCI’s auditors (Price Waterhouse), banking supervisors
and BCCI’s shareholders. In 1991, Price Waterhouse became increasingly convinced that
the fraud within BCCI was endemic, with published financial statements grossly inaccu-
rate, and informed the Bank of England (BoE) of its findings.
The liquidators, Deloitte & Touche, filed a lawsuit against the bank’s auditors, Price
Waterhouse and Ernst & Young. This was settled for $175 million in 1998. BCCI creditors
also attempted to sue the BoE as BCCI’s regulators. This case demonstrated the reputa-
tional risk for central banks that also oversaw banking supervision, and was one motiva-
tion for the separation of banking supervision from the BoE in 1997.
7.5.2 Savings and loans in the US in the 1980s

The US suffered a banking crisis in the 1980s in its S&L (savings and loans) sector. S&Ls
used to be a fairly quiet part of the US financial landscape, a category of financial institu-
tions common in many countries, often under the name of savings banks or something
similar.
The S&L industry suffered from the high interest rates and inflation of the late 1970s,
and as a consequence the authorities deregulated the industry with the view that the
S&Ls could grow their way out of trouble. At the time, deregulation took place in many
parts of the US economy and the S&L deregulation was a part of that process. The inten-
tion was to allow the S&L sector to expand into more parts of banking services previously
closed to them, because the objective of the S&Ls had been to promote housing and
homeownership. Other changes in oversight included authorising the use of more lenient
accounting rules to report their financial conditions, and the elimination of restrictions
on the minimum numbers of S&L stockholders. Such policies, combined with an overall
decline in regulatory oversight, contributed to the risk-taking in the sector.
The US has a fragmented regulatory structure and had a specific regulator for the S&L,
the Federal Home Loan Bank System. The government continued to provide deposit insur-
ance but did not increase oversight of the industry. There were clear signs of regulatory
capture, whereby the regulated have undue influence on the supervisor. This is a classical
problem in deregulation. The final cost of resolving failed S&Ls is estimated at just over
$160 billion, including $132 billion from federal taxpayers.
130
Lessons from the S&L disaster

The S&L crisis has particular lessons for the resolution of banking crises, as identified by
Moysich (2000):
■ First and foremost the need for strong and effective supervision of insured depository
institutions, particularly if they are given new or expanded powers or are experiencing
rapid growth.
■ Second, this can be accomplished only if the industry does not have too much influ-
ence over its regulators and if the regulators have the ability to hire, train and retain
qualified staff. In this regard, the bank regulatory agencies need to remain politically
independent. The S&L supervisor was too close to the industry it regulated during the
early years of the crisis, and its policies significantly contributed to the problem.
■ Third, the supervisors need adequate financial resources. The S&L supervisor was given
insufficient resources for the supervision of the newly deregulated industry, with many
new activities.
■ Fourth, the S&L crisis highlights the importance of promptly closing insolvent, insured
financial institutions in order to minimise potential losses to the deposit insurance fund
and to ensure a more efficient financial marketplace. The failure to promptly close fail-
ing institutions creates uncertainty, adversely affecting both the financial system and
the economy at large.
■ Finally, resolution of failing financial institutions requires that the deposit insurance
fund be strongly capitalised with real reserves, not just government guarantees. The
reason is that a deposit insurance fund is allocated to its intended purpose, ensuring
efficiency and rapid disbursement. It takes much longer to raise funds from the govern-
ment, because a political decision is needed.
The S&L crisis, and other bank failures in the US at the same time, had a significant
impact on how the US authorities have approached financial regulation. In response, they
have significantly improved the quality of supervision, absorbing the lessons of the S&L
crisis, for example, by developing one of the best resolution regimes in the world, prompt
corrective action.
7.5.3 The Scandinavian crisis in the 1990s

Three Scandinavian countries, Sweden, Norway and Finland, suffered a severe banking
crisis in the early 1990s caused by a huge lending boom in the late 1980s followed by
severe deleveraging in the 1990s. Finland was especially badly affected because it had the
extra factor of its important trading partner, the Soviet Union, collapsing. The crisis has
been documented in BCBS (2004).
The governments of these countries liberalised their financial markets and imple-
mented pro-cyclical macroeconomic policies in the late 1980s. This included removing
caps on lending and interest rates and encouraging more competition within the banking
sectors, whilst reducing the level of oversight.
The banks were not used to operating in such a free environment and neither were the
supervisors. The banks did not develop the necessary risk management systems, so at the
131
time when risk-taking was on the increase by inexperienced banks, government oversight
was decreasing. This is typical of the problems of deregulation as discussed above.
This led to a rapid increase in the balance sheets of the banks, as well as asset prices
generally, with increased leverage the main cause. For example, nominal non-financial
private sector debt increased by 52% in Norway and by 87% in Sweden, causing asset
price bubbles. Housing prices in the five years before the peak increased in real terms by
80% in Finland, 44% in Sweden and 38% in Norway. Similarly, equity prices tripled in
nominal terms in Sweden and Finland and doubled in Norway.
Moe et al. (2004) argue that the fiscal cost of the banking recapitalisation was 8.9% of
GDP in Finland, 3.9% in Sweden, and 2% in Norway. Table 7.2 gives 12.8% for Finland,
highlighting the challenge of estimating the cost of banking crises.
7.5.4 Bad bank – good bank

A common way for governments to resolve a banking crisis is the method of good banks –
bad banks. This was the method used to good effect by the Scandinavian governments in
resolving their crises. This has now become the ‘gold standard’ in banking crisis manage-
ment, and is the model many look to for resolving the European sovereign debt crisis.
The government splits up a failing bank, hiving the dodgy assets into one institution –
the bad bank – whilst keeping most of the bank’s operations and solid assets in the good
bank. In effect, the bad bank becomes an asset management firm. Over time, the govern-
ment aims to sell the good bank but will often hold onto the bad assets, like corporate
loans, until they expire or can be sold individually.
If the assets are valued at firesale prices at the time of doing this, the government has
the potential to make significant profits, an argument often used to justify this approach
to taxpayers. However, if the original bank was insolvent and the good bank is solvent,
then the bad bank must, by definition, have a negative value, so a profit for the govern-
ment is not the expected outcome.
Realistically, taxpayers should expect to lose money, but hopefully the efficiency gains
from having a well-functioning bank replacing a failing bank outweigh the expected loss.
7.5.5 Iceland, 2008

Iceland was the first country to be severely affected by the crises from 2007. The narration in
this chapter follows Benediktsdottir et al. (2011), and all the statistical data (except where
otherwise indicated) come from the parliamentary Special Investigation Commission
(SIC) whose report was published in 2009. One of the authors of Benediktsdottir et al.
(2011) sat on the commission.
The crisis started when the banking system of Iceland, mostly composed of three
banks, failed in early October 2008. A decade earlier, all three either had been in gov-
ernment hands or were highly regulated. The government banks were sold to politically
connected buyers for what was considered below fair market prices. The banks enjoyed
an enviable asset growth of 54% a year, and in the end, their asset size exceeded 800% of
GDP, placing Iceland close to the top among European banking nations, if compared to
the countries presented in Figure 1.2.
132
300 Iceland UK
Ireland US
200 Spain
100
0
2000 2002 2004 2006 2008
Figure 7.3 Stock market prices, 2000 = 100
A distinguishing feature of the Icelandic crisis was that the real-estate bubble was
smaller than in other bubble economies, with the bubble manifesting itself more clearly
in the stock market, as seen in Figure 7.3. While there are several reasons for this, the most
important is that the stock market was more easily manipulated and more useful in pro-
viding collateral in the alleged fraud committed.
Although the banks were highly risky, this was not evident from publicly available num-
bers, as they were amongst the most highly capitalised banks in Europe according to the
Basel capital regulations – see Figure 16.4. One reason is that the bank capital numbers
were illusory, since the banks funded purchases of their own equity issues. After that was
prevented by the supervisor, banks simply funded equity issues in other Icelandic banks,
hedging the exposure by using a contract for difference. This served the dual purpose of
increasing tier 1 capital, without the dilution causing prices to fall. All of this helped to
create the appearance of high demand for the stocks, fuelling stock-price increases early
in the bubble, and preventing prices from falling late in the bubble, after the professional
investors caught on to what was happening and started to sell the stocks.
Another way to inflate bank stock prices was by goodwill which was continually increas-
ing, sanctioned by the banks’ accountants, the main international accounting firms, with-
out much apparent concern as to value. This contributed to the accounts overstating the
value of assets.
Taken together, the causes of the banking crisis were over-ambitious goals without the
competence to pull it through, supported by aggressive accounting treatment and per-
mitted by poor regulatory oversight. As the banks expanded, this developed into corrup-
tion, incompetence and looting the banks from inside. The privatisation of the banking
system delivered the banks in the hands of politically connected groups, who continued
enjoying strong connections to the political classes. When the banks started to expand,
the government did not correspondingly expand the supervisory structure, so over time
the supervisors became less and less effective. The supervisor took an excessively legal-
istic approach, making it easy for the banks to find loopholes in the regulations. In other
words, the enforcement of the regulations targeted the letter of the law, not the spirit of
the law, and lost touch with the objectives for which the rules were created.
The banks enjoyed seemingly unlimited access to funds until late in the bubble, and
used some of those funds to manipulate their own capital and share prices, all in a posi-
tive feedback loop. Analysis by financial professionals did not seem to provide much
133
B3 2000
Rating 1000
CDS spread
Ba3
CDS spread 500
Rating
Baa3
200
A3
100
Aa3 50
Aaa
2006 2007 2008 2009
Figure 7.4 Moody’s credit ratings for Kaupthing and CDS spreads
Data source: Moody’s, Markit and Bloomberg
discipline. We can see this from Figure 7.4, showing the credit ratings and credit default
swap (CDS) spreads for the largest bank, Kaupthing. The ratings, in particular, were sur-
prisingly good right up until the collapse, whilst the CDS spreads were steadily increasing
from the middle of 2007 when the global crisis started.
Lessons
Countries with large financial systems have spent many decades or centuries develop-
ing banking structures and the related supervisory structures. Creating such a setup from
scratch in the span of a few years is quite challenging and needs significant political com-
mitment. This demonstrates the importance of institutions in ensuring that markets func-
tion efficiently.
This relates to peculiarities caused by operating within Europe.2 Because of the com-
mon European market, European banks were able to operate across the EU, with minimal
scrutiny from the host country, since supervision was mostly in the hands of the home
supervisor. The absence of pan-European supervision was exploited by the Icelandic
banks, enabling their rapid expansion.
7.6 Summary
Banking crises are an unfortunate consequence of modern economies, and while we can sig-
nificantly mitigate their incidence and consequences, they cannot be prevented altogether
in a cost-effective way. The main culprit is the inherent vulnerabilities in banking, such as the
fractional reserve system. Another factor is the importance of trust and interdependence in
banking, and the overconfidence and tendency to conceal weakness that result.
Banking crises, while well understood, are still quite frequent, because it can be dif-
ficult to identify the underlying problems until it is too late, and banks have incentives to
ignore and hide the problems.
The authorities are in a difficult position when it comes to resolving banking crises, hav-
ing to balance a robust response against moral hazard considerations.
2
Iceland is not a member of the European Union (EU) but is a member of the European Economic Area
(EEA), giving it full access to European markets.
134
References
Several bank crises have had a profound impact on financial regulations, such as those
of Herstatt, Ambrosiano and BCCI. Doubtless the current crisis will also be influential.
We discussed three banking crises in some detail: the S&L in the US in the 1980s, the
Scandinavian crisis of the early 1990s and the recent Icelandic crisis.

1 What are the main causes of banking crises?
2 Banking crisis are frequent and well understood. We know why they happen, the best
way to deal with them and we have a good understanding of how to prevent them.
Why then are they so frequent?
3 Do you think financial liberalisation is a recommended course of action for a country

aiming to stimulate economic growth?
4 What are the ingredients that enable some countries to remain financial centres for
centuries, while other countries do not seem to manage it?
5 What are the main lessons from the the savings and loan crisis?
6 The Swedish resolution of its banking crisis in the early 1990s is usually considered the
‘gold standard’ in crisis resolution. Why might that be the case?
7 What is the good bank – bad bank model?
References
Basel Committee on Banking Supervision (2004). Bank failures in mature economies. Working
paper no. 13, Bank of International Settlements.
Benediktsdottir, S., Danielsson, J. and Zoega, G. (2011). Lessons from a collapse of a financial
system. Economic Policy, 26: 183–231.
Caprio, G. and Honohan, P. (2009). Banking crises. In Berger, A., Molyneux, P. and Wilson, J.,
editors, The Oxford Handbook of Banking. Oxford University Press.
de Soto, J. H. (2009). Money, Bank Credit, and Economic Cycles. Ludwig von Mises Institute,
Auburn, AL.
Graeber, D. (2011). Debt: The First 5,000 Years. Melville House Publishing, New York.
Herring, R. J. and Litan, R. E. (1995). Financial Regulation in the Global Economy. The Brookings
Institution, Washington DC.
Honohan, P. and Klingebiel, D. (2003). The fiscal cost implications of an accommodating
approach to banking crises. J. Banking Finance, 27(8): 1539–60.
Kaminsky, G. L. and Reinhart, C. M. (1999). The twin crises: the causes of banking and balance-
of-payments problems. Amer. Econ. Rev., 89: 473–500.
Kane, E. J. (1989). The S&L insurance mess: how did it happen? Technical report, Urban
Institute Press.
Kupiec, P. H. and Ramirez, C. D. (2009). Bank failures and the cost of systemic risk: evidence
from 1900–1930. Technical report, FDIC.
135
Laeven, L. and Valencia, F. (2008). Systemic banking crises: a new database. Technical report,
IMF. IMF Working Paper.
Moe, T. G., Solheim, J. A. and Vale, B. (2004). The Norwegian banking crisis. Technical report,
Norges Bank. Occasional Paper no. 33.
Moysich, A. (2000). The savings and loan crisis and its relationship to banking. In Federal Deposit
Insurance Corporation, History of the 80s — Lessons for the Future. Volume I: an Examination of
the Banking Crises of the 1980s and Early 1990s, Chapter 4. FDIC, Washington DC.
Selgin, G. (2003). Adaptive learning and the transition to fiat money. Econ. J., 113(484): 147–65.
Stoler, P. and Kalb, J. B. (1982). The great Vatican bank mystery. http://time.com/time/
magazine/article/0,9171,951806,00.html.
Truell, P. and Gurwin, L. (1992). False Profits: the Inside Story of BCCI, the World’s Most Corrupt
Financial Empire. Houghton Mifflin, Boston, MA.
136
8 Bank runs and deposit insurance
During the Great Depression of 1929 to 1933, the United States (US) lost over one-
third of its banks to bankruptcies and people preferred to store their money under
their mattresses to keeping it in banks. At the same time, a loss of confidence in Austria
led to a run on its banks and government, spilling over to Hungary and Germany.
These bank failures were a major contributor to the Depression.
While the Great Depression is the worst example of contagious bank runs in his-
tory, waves of bank runs were quite common in the nineteenth and early twentieth
centuries. Since then, they have been much less common, especially in developed
economies, not least because policymakers have much clearer understanding of the
damage caused by bank runs and have developed much better tools for prevent-
ing them. The run on Northern Rock in 2007 is unique. This does not mean we have
become successful in preventing bank runs, rather they have changed form and are
more likely to happen in the wholesale markets.
Bank runs happen because of a fundamental weakness in fractional reserve bank-
ing systems. A basic model of a bank is an institution that collects demand deposits
and makes long-term loans. If a sufficient number of depositors want to get their
money back, the bank fails because most of its assets are tied up in long-term
loans. For this reason, if depositors are worried that a bank’s loans are of a low
quality they will demand their deposits back, and cause a bank run. A bank run
can happen even if there is nothing wrong with the bank; all we need is for deposi-
tors to get worried, and then a bank run becomes a self-fulfilling prophecy. A bank
run can lead to cascading failures within the banking system because depositors
Chapter 8 Bank runs and deposit insurance
might view the failure of a single bank as a symptom of system-wide difficulties.

Depositors have limited information about the quality of banks’ assets and may
feel that if hidden problems have been allowed to develop in one bank, the same
may have happened in other banks. Another explanation for cascading failures is
through cross-held assets, a failure in one bank directly impacting on other linked
banks, and on banks exposed to those banks, etc. This was a key reason behind the
1914 crisis.
Deposit insurance is often successful in preventing bank runs. Such schemes
became common after the Great Depression, and are now used in some form by
most major economies. The fundamental problem solved by deposit insurance
was eloquently modelled by Diamond and Dybvig (1983). The fundamental con-
clusion of their model is that so long as depositors believe they will be protected,
they will not run the bank and, in consequence, a deposit insurance scheme will
never be used.

This chapter directly relates to Chapter 2 (the Great Depression, 1929–1933), Chapter 4
(liquidity), Chapter 7 (banking crises) and Chapter 13 (financial regulations).
Key concepts
■ Bank runs
■ Deposit insurance
■ Retail and wholesale funding
■ Moral hazard
■ Diamond and Dybvig model

The main theoretical analysis in this chapter comes from the model of Diamond and
Dybvig (1983), but more broadly it fits in with the more general analysis of prudential
banking regulation such as Dewatripont and Tirole (1994).
More specific analysis is provided by Santos (2000). Miller (1996) studied the
Argentinian experiment of abolishing deposit insurance in the early 1990s, whilst
Shin (2008) is concerned with the run on Northern Rock. Finally, Borio (2009) focuses
on the intersection between more general liquidity provision and wholesale markets
with deposit insurance for retail depositors.

c1, c2 Consumption at t = 1, 2
N Number of depositors
R Gross return on deposits if withdrawn at t = 2
t = 0, 1, 2 A particular observation, e.g. a day
U( # ) Utility function
L Fraction (probability) of early consumers
P Transfer from late to early agents
138
8.1 Bank runs and crises

8.1.1 The United States in the Great Depression
The nineteenth-century economy of the US was characterised by frequent bank panics,
with corresponding economic downturns and unemployment. After a particularly severe
panic of 1893, legislators sought to arrange better security for bank deposits.
The resulting combined state–federal system failed to prevent bank panics during the
Great Depression which saw 4,004 banks close (see Figure 8.1). The panic suffered by the
US banking system in 1931 originated with the Bank of United States in New York, which
made a big bet on the value of New York real estate. In early 1931, as rumours that the
Bank of United States might be in trouble circulated, the New York Federal Reserve Bank
(NYFed) tried to engineer a merger and arrange a rescue package. But the Bank of United
States was clearly insolvent and in the end was allowed to fail.
Its failure marked a profound change in public sentiment towards banks. A real-estate
bubble in Chicago collapsed soon after in May 1931, and 30 Chicago banks defaulted.
Depositors, who were unable to tell whether a bank was good or not, began pulling their
cash indiscriminately out of all banks, good and bad, triggering multiple bank runs. Such
withdrawals had a negative multiplier effect on the supply of money, since in order to
maintain their own liquidity, banks had to call in three or four dollars of loans for each
dollar in cash withdrawn and to liquidate assets in a falling market at firesale prices. In
a fractional reserve system, this leads to a drastic fall in the supply of money. Moreover,
as their loans were called, borrowers in turn withdrew their deposits from other banks,
further fuelling the contagion.
In this climate, all banks felt the need to protect themselves, with bank failures snow-
balling as desperate bankers called in loans which the borrowers did not have time or
money to repay. With future profits looking poor, capital investment and construction
slowed or completely ceased. In the face of bad loans and worsening future prospects,
the surviving banks became even more conservative in their lending. The effect was to
speed up the scramble for liquidity right across the system. Banks built up their cash
reserves and made fewer loans, which intensified deflationary pressures. Servicing debts
became harder, because prices and incomes fell significantly but the debts remained at
the same dollar amount.
20%
10%
0%
1920 1923 1926 1929 1932 1935 1938
Figure 8.1 Bank failure rate 1920–1939 in the US. Main years of Great Depression
highlighted
139
The government reaction to the bank failures was inadequate. A potentially useful
move was made in December 1931, when the Reconstruction Finance Corporation (RFC)
was founded to provide finance for banks and firms in need of liquidity. However, then
the Speaker of the House insisted that RFC loans to banks be publicised and once this
requirement was enacted, in January 1933, banks in trouble stopped seeking RFC aid
because to be seen receiving RFC loans simply triggered runs.
The Federal Reserve System (Fed) was not prepared to sign off on a federal guarantee
for bank deposits in March 1933, nor did it feel there were any additional measures it
should recommend. Finally on 4 March, President Roosevelt closed all the banks in the
country for more than a week – bank holiday – and under the federal government’s super-
vision, weaker banks were merged into stronger banks. Depositors received compensation
for roughly 85% of their former deposits.
In response to this crisis, the Glass–Steagall Deposit Insurance Act was passed in
June 1933 establishing deposit insurance coverage in the amount of $2,500, and the
Federal Deposit Insurance Corporation (FDIC) as the authority to regulate and super-
vise state non-member banks. We can see from Figure 8.1 that these measures were
successful in stopping the cascading failures, since the US experienced virtually no
bank failures for the remainder of the decade, even if the Great Depression had a
double dip in 1936.
8.1.2 Northern Rock

The first bank run in the United Kingdom (UK) since Overend & Gurney in 1866 was the
run on Northern Rock on 14 September 2007 (see Figure 8.2). For detailed analysis, see
Borio (2009) and Shin (2008). The immediate bank run seemed to have been triggered by
an announcement by the Bank of England (BoE) that it was providing emergency liquidity
support for Northern Rock.
The underlying cause of its demise was its funding structure: borrowing on the asset
backed securities (ABS) markets to fund mortgages which then would be securitised and
sold off on the markets, with the proceedings used to pay back the ABSs. When that mar-
ket froze in the summer of 2007, the fate of Northern Rock was sealed. This meant that the
bank run shown on TV screens in September 2007 was only the endgame in a bank run
that started a month earlier.
Role of deposit insurance

The UK had a partial deposit insurance scheme in place at the time, and one might argue
that it was an invitation to a run. After the first £2,000, legislation protected only 90% of
savings up to £33,000 – guaranteeing a maximum payout of £31,700 – in the event of a
bank collapse, the so-called co-insurance. It took several months to get the insurance pay-
out. Co-insurance was intended to help provide incentives for depositors to monitor the
banks, thus putting prudential pressure on the banks. In the event, the Financial Services
Authority (FSA) missed the pending problems at Northern Rock, and it would be quite
surprising if ordinary depositors with less information could do better. The only sensible
strategy for depositors was to run the bank. As Sir Callum McCarthy, the former chairman
140
Figure 8.2 Northern Rock run

Source: Cate Gillon/Getty Images
of the FSA, said, the UK could have avoided the run on deposits at Northern Rock had
there been a depositor protection scheme such as exists in the US.
Role of wholesale markets

Northern Rock was vulnerable because of its unusually heavy reliance on wholesale, that
is non-retail, borrowings for 77% of its funding. This made it uniquely vulnerable when
the asset backed commercial paper (ABCP) market collapsed in early August 2007, in the
opening shots of the global crisis. Northern Rock informed the FSA and the BoE of its fund-
ing problems by the middle of August, after which the authorities tried to resolve the crisis
behind the scenes. The wholesale depositors’ run on Northern Rock demonstrated that the
financial markets had a much better understanding of the bank’s problems than the super-
visor or the general public.
Summary
Until its failure, the funding model of Northern Rock was considered sound by the bank
itself, its supervisors and the capital markets. With the benefit of hindsight, we recognise
that all of these parties were misguided. Northern Rock remained solvent only so long as
liquidity was in effect infinite. Because liquidity was thought infinite, Northern Rock and
most other banks in the world behaved in a way that ensured liquidity would dry up.
Therefore, the failure of Northern Rock was almost certain, given time.
If the deposit insurance scheme of the UK had been effective, the actual run on Northern
Rock probably would have been prevented, sparing the authorities the embarrassment of
141
seeing it play out on the TV screens. Deposit insurance would not have prevented the fail-
ure of the bank, considering that only 23% of its funding came from deposits.
As a response to the Northern Rock failure, the authorities were forced to announce
unlimited deposit insurance in order to reassure bank clients and prevent contagious runs
across the UK banking sector. This demonstrates the problem created when the authori-
ties have inadequate policies in place and then are confronted by a crisis. The authorities
are forced to overreact in order to prevent more damage. In other words, the UK went
from having a deposit insurance scheme that was too weak to one that was too strong.
8.2 Modelling deposit insurance

One of the most influential models in economics is the deposit insurance model pro-
posed by Diamond and Dybvig (1983). The model captures the fundamental uncertainty
leading to bank runs and answers how deposit insurance can prevent bank runs.
In their model, banks are viewed as pools of liquidity providing consumers with insur-
ance against their own idiosyncratic shocks. The banks transform illiquid assets into liquid
liabilities, and because the idiosyncratic shocks are not highly correlated, banks need to
hold only a small fraction of their total assets in liquid form. This is the basis of the frac-
tional reserve banking system.
The paper by Diamond and Dybvig demonstrates three important points. First, banks
providing demand deposits can improve on a competitive market by providing better
risk-sharing among people who need to consume at different random times. Second, a
demand deposit contract can prevent an undesirable equilibrium (a bank run). Finally,
deposit insurance provided by governments can prevent bank runs.
Background
In the simple version of the Diamond–Dybvig model presented here, the banking system
is illiquid, meaning that not all banking system obligations can be met if all holders of
those obligations simultaneously claim what they have been promised.
There are two main ingredients in the model:
1 Depositors are initially individually uncertain about their profile of consumption, that
is whether they prefer to consume early or late.
2 The banking system deals with depositors on a first-come, first-served basis.
Setup
The economy has three periods, t = 0, 1 and 2, the economy starts at date 0 and all
agents have an endowment of $1. There is an intertemporal technology (reflecting the
illiquidity of assets) which can be represented by the following:
t = 0 t = 1 t = 2
-1 1 R
142
So $1 deposited at t = 0 will yield $1 if withdrawn at t = 1 and R 7 1 if withdrawn at

t = 2. This captures the cost of early liquidation of assets.
Each agent can be either early or late. Early agents only care about consumption at
t = 1, with utility U1c12, where U( # ) is their utility function, and late agents only care
about consumption at t = 2, with utility U1c22, where we implicitly assume the discount
factor is 1, without loss of generality. The agent does not know if she is early or late at
t = 0 but learns it at t = 1. Realisations are independent across agents. Prior to learning
their type, all agents have identical preferences. A fraction (probability) l of agents are
early, taken to be constant.
Autarky
Autarky is the condition of economic self-sufficiency, without trade. Under autarky, there
are no means to shift consumption and agents solve the optimisation problem individu-
ally. Ex-ante 1t = 02, all agents are identical. The expected utility for each agent is given by
E1U2 = lU1c12 + 11 - l2U1c22

= lU112 + 11 - l2U1R2
Assuming the utility function is increasing and concave – standard form with non-satiation
and risk aversion – the late agent will have higher eventual utility than the early agent.
Utility under autarky

Given the presence of risk aversion, there is an optimal social insurance contract that
would allow agents to insure against the unlucky outcome of being early. This assumes
that types of agents are publicly observable as of period 1 (Figure 8.3(a)).
Suppose there are two agents; one is late, the other is early, so l = 0.5. They make the
following contract at t = 0: at t = 1 the late agent will pay the early agent some amount
c2 c2
2R
autarky
autarky
R R
op
tim
al
c1 c1
0 1 0 1 2
(a) Autarky (b) Optimal social insurance
Figure 8.3 Autarky and optimal social insurance
143
p, thus the early will have consumption ∼ c 1 = 1 + p and the late ∼

c2 = R11 - p2. If p is
chosen correctly, there will be improvement on the autarkic outcome of c1 = 1 and c2 = R.
The intertemporal budget constraint is
∼
c2 = R12 - ∼
c 12
The economic problem is then
∼ 2 + U1c
max E1U2 = U1c ∼2
1 2
∼
c1
∼ 2 + U1R12 - ∼
= U1c c 122
1
Optimisation, differentiating with respect to ∼c 1, and setting the result to zero, gives the
standard result (assuming an interior solution) that the marginal rate of substitution equals
the marginal rate of transformation:
∼2
U′1c 1
∼2 = R (8.1)
U′1c 2
The optimal consumption allocation satisfies the first-order condition and the budget
constraint. At the optimum, we do not want to give consumption to those who do not
value it, therefore, early agents’ date – 2 consumption is equal to late agents’ date – 1
consumption, which is zero.
Utility under optimal social insurance

Since by assumption, R 7 1, it follows that optimal consumption levels satisfy ∼ c1 7 1
and ∼c 2 6 R. The optimal indifference curve is higher than the indifference curve under
autarky (Figure 8.3(b)), so there are gains from trade. The optimal outcome devotes more
resources on average to date – 1 consumption and less to date – 2 consumption than the
autarky outcome.
Diamond and Dybvig show that under certain circumstances it is possible to achieve
the above optimal insurance contract given there is a bank in the economy. Banks can
underwrite this insurance by providing liquidity (allow for better risk sharing), guaran-
teeing a reasonable return when the investor cashes in before maturity. This provides an
explanation of how banks subject to runs can still attract deposits.
8.2.1 A fractional reserve banking system and equilibrium decisions

The optimal allocation characterised above can be implemented by a fractional reserve bank-
ing system in which banks collect consumers’ endowments and invest a fraction of them in
long-term investments while offering depositors the possibility of withdrawal on demand.
Suppose there is a large number of agents. Diamond and Dybvig show that the same
solution is obtained if a financial institution (a bank) creates a bank account or a tra-
ditional demand deposit contract that pays the optimal amounts 1 + p in t = 1 and
R11 - p2 in t = 2. Then it is an equilibrium for early agents to withdraw at t = 1 and for
late agents to wait.
144
This good equilibrium achieves optimal risk sharing and demonstrates the role of finan-
cial intermediation in increasing welfare. A crucial question regarding this fractional reserve
system is whether the bank will be able to fulfil its contractual obligations. This depends
on investors’ anticipation about the safety of the bank.
Consider first the case that a late investor anticipates that the bank will be able to fulfil
its obligations. If late investors trust their bank, they will prefer to withdraw at t = 2, thus
the proportion of time t = 1 withdrawals will be l. So in order to avoid premature liqui-
dation, the bank should have a liquid reserve of at least lc*1.
Multiple equilibria
There is a potential for multiple equilibria in this model: a good equilibrium of no run
described above and a bad equilibrium where a bank run takes place. The good equilib-
rium is better than the bad in the sense of Pareto efficiency. In the bank run equilibrium,
everyone prefers to receive a risky return with a mean of 1 even if holding the endowment
until t = 2 provides a return exceeding 1 if there is no run.
Bank runs are a type of self-fulfilling prophecy, caused by a shift in expectations. This
shift could happen for almost any reason. Each depositor’s incentive to withdraw her
funds depends on what she expects other depositors to do. If enough depositors antici-
pate that other depositors will withdraw their funds, they all have an incentive to rush to
be the first in line to withdraw their funds, since the face value of deposits is larger than
the liquidation value of the bank’s assets.
Bank runs
To illustrate a bank run, suppose there are N depositors, each endowed with $1, so the
amount the bank has on hand at t = 1 is $N. Given the optimal contract, the total value
of deposits is $N11 + p2, hence, the bank does not have enough cash to pay off all
depositors at t = 1.
The first person to demand her money at t = 1 will get the full amount 1 + p and this
applies to all the agents in the queue up to the fraction 1>11 + p2 when the bank runs
out of assets. The bank’s payoff to any agent depends only on the agent’s place in line
and not on future information about agents behind her in the queue. Therefore, the last
N11 - 1>11 + p22 get nothing, and all agents will want to be the first in the queue.
Note that for runs to be an equilibrium, demand deposits must yield more than 1 if
withdrawn at t = 1. If deposits yield 1 when withdrawn at t = 1, the outcome is not an
improvement on the autarkic outcome, with the bank simply mimicking each agent hold-
ing their endowment, but then this shows that a demand deposit which is not subject to
runs provides no liquidity services. It is precisely the transformation of illiquid assets into
liquid assets that is responsible both for the liquidity service provided by banks and for
their susceptibility to runs.
Decisions and utility

The inferiority of bank runs seems to rule out runs, since no one would deposit in the bank
if they anticipate a run. However, all agents will choose to deposit at least some of their
wealth in the bank even if they anticipate a positive probability of a run, provided that the
145
probability is small enough, as the good equilibrium dominates autarky. Since in a run, the
fraction 1>11 + p2 get ∼c 1 at t = 1, while the rest get nothing, the expected utility is
∼2
U1c 1
E1U2 =
1 + p
We therefore get that the run utility is inferior:

∼ 2 + 11 - l2U1c
No run = E1U2 = lU1c ∼2
1 2
∼2
U1c
Run E1U2 =
1 ∼ 2 + 11 - l2U1c
6 lU1c ∼2
1 2
1 + p
Analysis
The instability of fractional reserve banking systems comes mainly from the coordination
failure among investors. Once they have made their deposits, anything that causes them
to anticipate a run will lead to a run. In this framework, banks with pure demand deposits
will be very concerned about maintaining confidence because they realise that the good
equilibrium is very fragile.
8.2.2 Regulatory response: deposit insurance

Deposit insurance provided by the government can result in an allocation that domi-
nates the best that can be offered without insurance and never do any worse, resulting
in a dominant strategies equilibrium. The government makes the first agents pay a tax of
p, which is enough to compensate the unlucky ones late to the queue. The important
feature here is that the government imposes the tax on the agent after she withdraws the
money, an option not open to the bank.
The effect of this policy is to guarantee that every agent can get $1 at t = 1. So regard-
less of whether there is a run or not, agents know they will get their initial deposits back.
As long as the probability of a run is not 100%, late agents are better off not running, since
they get ∼c2 7 1 at t = 2.
Deposit insurance prevents bank runs because, for all possible anticipated withdrawal
policies of other agents, it never pays to participate in a bank run. As a result, no strategic
issues of confidence arise. This in turn makes the good equilibrium unique, so there will be
no run. Without runs, the insurance is never claimed, and hence is costless in equilibrium.
The presence of deposit insurance increases the efficiency of the banking sector,
because the bank does not have to worry about the potential for runs and, therefore,
can follow an optimal asset allocation policy. Otherwise, the bank might have to allocate
resources to reassure depositors, reducing its efficiency.
Analysis
Within the stylistic framework of the model, a government deposit insurance scheme has
a natural advantage over a privately funded scheme. The reason is that the government
has the power to tax depositors to finance insurance, therefore effectively preventing a
146
8.3 Pros and cons of deposit insurance
run. A private scheme cannot rely on such powers, and instead would need to have suffi-
cient reserves to make the insurance credible. Since holding such reserves would be costly,
a government provision of deposit insurance is more efficient than private provision.
In the real world, this is not as clear-cut. For example, it would be politically very dif-
ficult to actually tax those withdrawing early, and as a consequence the government
might have to maintain a funded deposit insurance fund, or provide guarantees. This
might remove the efficiency advantages of the government.
One example of difficulties in taxing agents withdrawing early can be seen from the
process of resolving the failure of Bernie Madoff. The authorities are trying to claw back
funds from those who successfully took the money out of the fund, but this has turned
out to be a very long and very expensive process, with uncertain outcomes.
8.3 Pros and cons of deposit insurance

In the simple Diamond–Dybvig model above, deposit insurance is effective in ensur-
ing socially optimal outcomes. However, in the real world more issues are at work, and
deposit insurance has often been controversial. The opposition to deposit insurance was
succinctly stated by Kovacevich (1996), President and CEO of Wells Fargo:
‘The deposit insurance system is indeed a monster and it is a monster that threatens
to devour the very system it is intended to protect.’
Deposit insurance is often criticised for creating moral hazard and incentives for
excessive risk-taking by banks. By guaranteeing deposits, market incentives to monitor
banks and to demand an interest payment commensurate with the risk of the bank are
diminished.
Moral hazard
Moreover, it can be difficult for insurance premiums to properly internalise the cost of
risk, because of problems of asymmetric information. The bank knows more about its
operations than outsiders, which therefore gives the bank incentives to take more hidden
risk – moral hazard. Furthermore, risk-sensitive deposit insurance gives rise to problems
of pro-cyclicality.
The depositors themselves also contribute to moral hazard. If they know they are
protected regardless of what the bank does, their incentive is to put their money into
the bank with the highest interest rates. This is very similar to the problem analysed in
the Krugman model of governments guaranteeing borrowers, and the problem in the
S&L crisis. In addition, it creates competitive problems for banks that are prudently run,
because they cannot attract deposits. This was the case in the Venezuelan banking crisis.
Impact of wholesale markets

The discussion so far has focused on bank runs by retail banking clients, the traditional
source of funding for banks. In recent years, banks have increasingly been relying on the
wholesale markets for funds. Borio (2009) argues that in this case, bank runs come in
two waves, first from sophisticated institutional investors and then by unsophisticated
147
retail depositors. Provided the institutional investors run the bank, deposit insurance is
not sufficient to prevent the retail depositors from running the bank. This was the case
with Northern Rock.
This picture is somewhat muddled by the fact that post-Lehman, it is often assumed
that governments insure all providers of liquidity to the banks, both institutional and
retail. This is supported by the fact that in the crisis from 2007, credit losses have been
very small. For example, by the middle of 2012 only two European countries had imposed
losses on bank creditors, the UK and Denmark, a minuscule amount in the former case
and small in the latter. In spite of the massive amounts of public money provided to the
banks, the authorities have not wanted to hit private creditors.
More generally, changes in the structure and functioning of financial markets have
been reducing the significance of deposit insurance schemes as devices to deal with sys-
temic risks of runs. This reflects the greater importance of wholesale financial markets
and funding in the system, as well as the increasing systemic relevance of institutions that
either do not have deposit insurance protection, such as money market mutual funds and
hedge funds, or are perceived as too big to fail (TBTF).
8.3.1 Argentina, 1991–1994

The case of Argentina’s short-lived effort to repeal its programme of deposit insurance
is instructive, and we follow this story as presented by Miller (1996). Before 1991, the
Argentine banking system operated under a regime of optional, explicit deposit insur-
ance, coupled with an extensive implicit deposit insurance in the form of central bank
assistance to failing banks. In 1991 and 1992, Argentina reversed this policy by repealing
the country’s deposit insurance programme, intending to convince financial markets that
it would not under any circumstances rescue a failing bank.
The government decision to forswear providing deposit insurance had several reasons.
One was the product of the unusual political and economic circumstances of 1989–1991.
Plagued by the disaster of hyperinflation and exhausted by years of political instability,
the Argentine public reached a consensus that fundamental reforms were necessary, with
the population willing to engage in a radical experiment to rectify a deteriorating situa-
tion. The decision to abolish deposit insurance did not appear to impose extreme risks,
compared to the dangers and costs of rampant hyperinflation. The government utilised a
variety of pre-commitment devices to assure the markets that it would not bail out depos-
itors. These included the newly established independence of the central bank, statutory
prohibitions on deposit insurance and stringent limitations on the central bank’s ability to
loan money to governments or to shaky banks.
Some private market mechanisms did also develop to respond to the withdrawal of the
government safety net. Although these private responses did not turn out to be adequate
to cope with the systemic crisis induced by the Mexican devaluation, they showed signs of
promise before being overwhelmed by events.
In 1995, in the face of a forthcoming election and a severe economic crisis sparked by
the Mexican peso devaluation of December 1994, the Argentine government reinstituted
a form of deposit insurance in an effort to stave off an all-out bank panic.
148
8.4 Summary
The financial crisis happened at a time of maximum political pressure, in the months
before elections. This forced the government to reinstitute deposit insurance for its banks,
even if the government and the central bank president had said strongly and often that
they would never do so. Even though the authorities at the time held the view that deposit
insurance creates moral hazard and free-rider problems, and therefore should be limited,
this view proved unsustainable.
This experience suggests that deposit insurance is a practical necessity for any indus-
trialised nation, and that governments cannot credibly commit to not providing deposit
insurance, because the political and economic pressures to do so in times of crisis become
overwhelming. Consequently, governments are better off recognising the eventuality and
maintaining a deposit insurance scheme.
8.4 Summary
The fractional reserve banking system is inherently fragile and subject to bank runs. The
experience from the Great Depression demonstrated how devastating contagious bank
runs can be, and how effective deposit insurance is in preventing bank runs. This is not as
clear-cut with modern banks depending significantly on wholesale funding.
The fragility of banks and the importance of deposit insurance are succinctly modelled
by Diamond and Dybvig who clearly demonstrate the positive contribution of financial
intermediation and simultaneously how it can increase the fragility of the financial sys-
tem. This fragility can be addressed by the government by providing deposit insurance,
perhaps by taxing those who aim to run banks and by compensating those who are at the
end of the queue and otherwise would not get their funds back. Within the simple con-
fines of the model, this is sufficient to ensure bank runs will not happen.
At the same time, full deposit insurance does lead to moral hazard. If depositors do not
care what a bank does with its money, a badly run or even fraudulent bank will simply
offer more interest than its competitors, expand rapidly, and make dodgy loans.
The alternative of having no deposit insurance scheme is not politically credible, as
the example of Argentina demonstrated. In case of a bank run, depositors will exert enor-
mous political pressure to fix the problem and make them whole. A bank run is a highly
visible event, demonstrating that the authorities have failed to adequately supervise the
financial system. This is highly embarrassing, as we saw in the case of Northern Rock. The
authorities are left to scramble to implement deposit insurance, and since the worst time
to make policy decisions is usually during a crisis, it is much better to be prepared and
have a well thought-out deposit insurance scheme in place.

1 The Northern Rock bank run embarrassed the British authorities in 2007. The previous
British bank run was in 1866, and the general view was that bank runs just did not hap-
pen in advanced economies anymore.
149
(a) What were the underlying risk factors that ultimately caused the the failure of
Northern Rock?
(b) How did the reaction of the British authorities to the crisis at Northern Rock
directly contribute to the bank run?
(c) Compare and contrast bank runs in wholesale markets and retail markets from
the point of view of Northern Rock.
2 Recall the model of Diamond and Dybvig (1983).

The model has three periods, t = 0, 1 and 2, with 1$ deposited in t = 0, yielding 1 if
withdrawn at t = 1 and yielding r 7 1 if withdrawn at t = 2. Agents are identical and
have a wealth of $1 in t = 0. There are 2 types of agents:
Early Prefer to consume c1 in t = 1, getting U(c1)

Late Prefer to consume c2 in t = 2, getting U(c2)
An agent does not know if she is early or late at t = 0, but learns it at t = 1. Fraction l
are early, and 1 - l late.
(a) Suppose the utility function takes the form
U1c2 = 1 - 1>C
Derive the utility levels of both early and late agents under autarchy.
(b) Design an optimal insurance contract for the two agents.
3 Before the crisis from 2007, the general trend was for deposit insurance being reduced
or even abolished. Very quickly, this sentiment changed, for example, in the case of
Northern Rock. Why was that, and do you think extensive deposit insurance is here to
stay?
References
Borio, C. (2009). Ten propositions about liquidity crisis. Working paper no. 293, Bank of
International Settlements.
Dewatripont, M. and Tirole, J. (1994). The Prudential Regulation of Banks. MIT Press.
Diamond, D. and Dybvig, P. (1983). Bank runs, deposit insurance, and liquidity. J. Polit. Econ.,
91: 401–19.
Kovacevich, R. M. (1996). Deposit insurance: It’s time to cage the monster. www.minneapolisfed.
org/publications_papers/pub_display.cfm? id=2682.
Miller, G. P. (1996). Is deposit insurance inevitable? Lessons from Argentina. Int. Rev. Law Econ.,
16: 211–232.
Santos, J. A. C. (2000). Bank capital regulation in contemporary banking theory: a review of
the literature. Technical Report 90, Bank for International Settlements. www.bis.org/publ/
work90.pdf?noframes–1.
Shin, H. S. (2008). Reflections on modern bank runs: a case study of Northern Rock. Mimeo,
Princeton University.
150
9 Trading and speculation
Trading is an essential part of economic life, facilitating the needs of economic agents
to acquire or dispose of tradable assets. For example, the main transaction for most
people involves real estate, and that is trading. Intimately connected to trading is
speculation. Few economic terms evoke such strong and different reactions as spec-
ulation. For some, speculation is the root of instability, undermining sensible gov-
ernment policy, while for others it brings rationality to a world where governments
constantly manipulate the economy for political reasons.
The term speculation has multiple meanings in common parlance. In a classic text,
Security Analysis, by Graham and Dodd (1934), speculation is the opposite of invest-
ment, borrowing money to buy assets without proper fundamental research. A more
common usage suggests that speculation is the making of risky investments for the
purpose of short-term profit. Most people tend to distinguish between speculation
and investing based on the amount of risk. A trader in a bank using 1 million to buy
an asset with a 50% chance of delivering nothing and a 50% probability of doubling
the money might be said to be speculating, whilst allocating the funds to a long-term
position in an AA corporate bond is investment.
There is a fine line between what might be considered legitimate and illegitimate
speculative activity. Speculators play a useful role in society, providing useful risk-
sharing and hedge facilitation, since for anybody wishing to hedge away risk, some-
one needs to be willing to take on that risk – to be the counterparty to the trade. That
counterparty is often a speculator.
For example, farmers need to make decisions about their crops early in the year,
but can sell them only much later, when they harvest. If crop prices collapse on the
market, the farmer may be wiped out. It is not surprising that the first hedging markets
Chapter 9 Trading and speculation
in the world developed in agriculture. Aristotle describes the use of a futures contract
for olive oil, but the first documented futures exchange is the Dōjima rice exchange in
Japan in the 1730s, set up to help the Samurai, who were paid in rice, needing to get
stable conversion to coin. The first exchange-traded commodities futures contracts
were listed on the Chicago Board of Trade in 1864, a model soon replicated around
the world, for example in India in 1875, covering cotton futures.
Other speculative activity is less salubrious and it might be tempting to impose
rules to prevent such behaviour. Large financial institutions are able to make such
large speculative bets that they not only risk bankruptcy but also threaten financial
stability, causing systemic risk. One of the most infamous examples is the failure of
LTCM in 1998. Unfortunately, it is not easy to effectively regulate speculation, because
it can be quite difficult to distinguish between illegitimate and legitimate speculative
activities, and if one is concerned about speculation, it may be better to target the
tools employed by speculators, rather than the motives of those trading.

This chapter directly relates to Chapter 1 (systemic risk), Chapter 3 (endogenous risk),
Chapter 13 (financial regulations) and Chapter 18 (ongoing developments in finan-
cial regulation).
Key concepts
■ Speculation
■ Moral hazard
■ Proprietary trading
■ Financial transaction tax
■ Trading strategies

There are many books that address the topics in this chapter, and we have found the
following to be especially useful: Kohn (2003), Madura (2010), Mishkin and Eakins
(2011) and Hull (2011).
9.1 Trading scandals and abuse

Illegal activities of various types are quite common in the financial system. In simple cases
an individual may be stealing from his or her employer or clients, and in sophisticated
cases a large financial institution may be subtly manipulating the system in an illegal way
for profit. Below, we discuss some well-known scandals.
Example 9.1 JP Morgan Chase and the ‘London whale’

One of the largest banks in the world, JP Morgan Chase, disclosed in May 2012 that
it had lost over $5.8 billion from proprietary trading in a bank unit called the chief
investment office (CIO), as reported in the Financial Times (2012a, 2012b). The losses
152
9.1 Trading scandals and abuse
arose from the activities of Bruno Iksil, a well-known player in the credit default swap
(CDS) market, who accumulated such large positions that he got nicknamed the
‘London whale’, supposedly moving prices with his trading.
It seems that the bank had little understanding of what he was up to, finding his
strategy ‘flawed, complex, poorly reviewed, poorly executed, and poorly monitored.
The portfolio has proven to be riskier, more volatile and less effective as a hedge than
we thought’, according to senior bank management. The trading scandal will not
much affect the financial fortunes of JP Morgan Chase; its first-quarter profits in 2012
dropped to $4.9 billion, from $5.6 billion a year earlier. The scandal may, however,
have a significant indirect impact on the bank because of how it affects the debate
on the regulation of financial institutions and whether banks should be allowed to
engage in proprietary trading under the Volcker rule.
While the ‘London whale’ reflected the activities of few individuals over a relatively short
period of time, a much larger scandal involving LIBOR price-fixing was first reported by the
Wall Street Journal in 2008, with Barclays Bank fined for its participation in the scandal in 2012.
Example 9.2 LIBOR-fixing scandal

In June 2012, it emerged that Barclays Bank was fined £290 million for its part in a
LIBOR price-fixing scandal: see BBC (2012b) for a timeline of events. According to the
Financial Services Authority (2012), derivatives traders at Barclays made 257 requests
to fix LIBOR and EURIBOR rates between 2005 and 2009, conspiring with other banks.
There are several reasons why the bank might want to do this. If derivative trad-
ers are able to manipulate LIBOR, they profit because they know its value before it is
made public. Even more importantly, the banks might have been short LIBOR so low
values fed directly into profits; after all, LIBOR is the rate ‘at which London banks can
borrow’. Similarly, at the height of the crisis in 2008, Barclays seemed to have had
higher funding costs than comparable institutions, signalling vulnerability, which it
could hide by reporting artificially low funding costs.
The manipulation of LIBOR significantly affects other market participants: many
mortgages are linked to LIBOR, and it is used to set rates in $800 trillion worth of
derivatives and borrowings.
Barclays cooperated with a number of other banks in fixing LIBOR; at the time of
writing, the Royal Bank of Scotland group and UBS have also been fined for this.
The LIBOR price-fixing scandal is quite serious, as it undermines the integrity of a key
interest rate and the banking system as a whole. The scandal will embolden efforts to
strictly regulate the financial system.
Ponzi schemes
A classical, and common, fraudulent investment operation is a Ponzi scheme. A fund pays
investors profit from their own money, or money paid in by new investors, rather than
153
profits made by investments. Operators of Ponzi schemes entice unsuspecting investors

by promises of large profits. However, the scheme can go on only so long as new money
comes in, at ever increasing rates. Eventually, all such schemes are doomed to failure.
Ponzi schemes are well known in history, but are named after Charles Ponzi who created
a large scheme in 1920, taking money from all over the United States.
The best-known recent Ponzi scheme was run by Bernie Madoff, who might have caused
paper losses to investors of perhaps $18 billion and consequently was sentenced to prison
for 150 years. The scheme is notable because he was a well-known establishment figure, well
connected on Wall Street and highly respected, attracting supposedly sophisticated investors.
Rogue traders
Employees of banks are not always willing to accept the restrictions imposed by risk poli-
cies, giving rise to the phenomenon of rogue trader. A recent example is Jérôme Kerviel’s
liberal interpretation of ‘low-risk arbitrage’ causing a €4.91 billion loss to his employer,
Société Générale, possibly costing him three years in prison. This may be the largest loss
attributed to a rogue trader.
Example 9.3 Nick Leeson and Barings

Perhaps the most infamous rogue trader is Nick Leeson, whose unchecked risk-taking
and manipulation of his risk controls caused the collapse of Barings Bank in 1995,
landing him 6.5 years in a Singapore prison. We follow events as described by himself
in Leeson (1999).
Leeson started working for Barings Bank in 1989, and three years later became the
general manager of its futures market operation in Singapore. His trading activity caused
significant losses for the bank, but he hid those losses by manipulating the accounts, put-
ting losses into a special account numbered 88888, the number 8 being considered lucky
in Chinese numerology. Eventually, the losses reached £827 million, bankrupting Barings.
Leeson fled from Singapore but was captured in Germany and sent to prison. Since then,
he has gained celebrity status, helped by his book and a popular film about his activities.
9.2 Trading and risk

The most publicly visible part of the financial market is the equity market, often in the form
of a stock market index like the FT100 index in the United Kingdom, S&P-500 in the US and
DAX in Germany. The fixed income markets are even larger in volume, and a number of other
types of assets are traded, including foreign exchange and commodities, as well as many
types of derivative assets. We present the basic terminology and institutions relating to trad-
ing in the appendix to this chapter, and will refer to those terms throughout the book.
9.2.1 Market participants

Within the financial system, a number of different types of financial institutions are
engaged in trading, and we use the term market participants as a catch-all term to refer
to them. Most trading takes place in banks, which dominate the debt markets and are
154
9.2 Trading and risk
actively involved in the market for most other financial assets. Banks are engaged in trad-
ing on behalf of their clients, but also often trade on their own account, termed propri-
etary trading, or prop trading in short. This is done with the explicit purpose of making
profit by taking on risk – that is, speculating.
While some market participants are engaged in proprietary trading, most act instead on
behalf of investors like pension funds, insurance companies, mutual funds and sovereign
wealth funds. Investors often choose to subcontract investment decisions to those market
participants, not exerting direct control over investment decisions but usually monitoring
investment performance by various means, such as comparison with a benchmark, which
might be a stock market index. This is often referred to as the ‘institutionalisation’ of the
market. A large number of different investment schemes are on offer, and relatively unso-
phisticated investors can find it rather hard to evaluate the different options. This means
there is an obvious risk of fraud or misrepresentation, or even Ponzi schemes. As a result,
mutual funds are highly regulated in most jurisdictions.
Hedge funds
An important category of financial institution is hedge funds which exist in some markets
influenced by a relatively liberal US approach to financial services. Hedge funds are less
regulated than other financial institutions, provided they sell their services only to rela-
tively sophisticated and wealthy investors, often called accredited investors.
It is sometimes said that hedge funds are unregulated, but that is not strictly true. They
are subject to securities law, their manager is typically regulated, and all their counter-
parties such as brokers and lenders will also be regulated, so they are at least indirectly
subject to regulation.
Beyond their regulatory status and types of clients, it is very difficult to classify hedge
funds. They are extremely diverse, engaging in a large number of different trading activities,
often using very complicated trading strategies. Because of the lack of regulation, hedge funds
are able to take on risk and investment strategies not open to other market participants, per-
haps their defining characteristic. Hedge funds are designed to provide good performance to
investors, where performance fees supposedly improve the alignment of interests between
clients and manager, though how much of this gain is captured by clients is less clear.
Before 2007, most commentators on financial stability thought the next crisis would
originate in the hedge fund sector, due to the impact of the failure of LTCM in 1998.
However, the crises from 2007 instead took place in the most regulated part of the finan-
cial system, and as a consequence, concerns about hedge funds do not attract a high
priority from policymakers, except in Europe where there has recently been a wave of
regulatory initiatives aimed at the sector.
9.2.2 Financial innovation

Market participants have always had the tendency to create new and different types of
financial instruments, a practice falling under the general heading of financial innovation.
The question of whether financial innovation is positive for society remains controversial.
Before the crisis, it was generally viewed favourably, perhaps because only the advocates
understood the nature of the newly created instruments and the adverse effects had not
155
materialised. However, as it turned out, many of the new financial instruments proved to be
quite damaging, enabling financial institutions to amass significant amounts of hidden risk.
As a consequence, some commentators have argued that financial innovation provides
no value to society whilst destabilising the financial system. Perhaps most succinctly, this
is stated as: ‘the most important financial innovation that I have seen in the past 20 years
is the automatic teller machine’ (Paul Volcker in the Wall Street Journal, 2009b). Volcker
further notes that financial innovation ‘moves around the rents in the financial system’,
benefiting the inventor, not the clients.
9.3 Trading activities

Trading and speculation typically involve the use of specialist trading activities, which is
how traders go about implementing their price forecasts. This may imply buying or selling
stocks or bonds, trading derivatives or any number of different activities. Several of these
have come under particular scrutiny by politicians and the popular press, in particular
short selling and carry trading.
Traders of financial assets use what is called a trading strategy to make decisions, which
can be considered as a set of rules that the trader follows when deciding what to buy
and sell. The extent to which trading rules are formalised varies greatly. At one extreme,
some investors employ largely automated strategies with formal rules. Perhaps more fre-
quent are less formal systems, for example, the common investor preference for ‘low risk’
or ‘safe’ investments, with the exact nature of safety left undefined. Rules may even be
unconscious; for example, if an investor prefers to repeat investments that have been suc-
cessful in the recent past, they may unconsciously engage in trend following.
Whether conscious or not, investor behaviour can be analysed in terms of trading
strategies, and because asset prices are determined by these strategies, their analysis is as
important as the types of instruments employed.
Value investing
A classical approach to investing is value investing, a strategy proposed by Graham and
Dodd (1934). This involves finding companies that are trading below their inherent worth,
by finding stocks with strong fundamentals, like earnings, dividends, book value and cash
flow. The most successful investor of all times, Warren Buffett, follows value investing.
Value is an example of a mean reversion trade, where assets are bought when prices have
fallen and sold when prices have risen. As such, under normal conditions, when prices move
within their typical ranges, value investing has a stabilising effect on prices, increasing demand
when they are low and reducing it when high. In extreme conditions, however, this may not
be the case, because value investors are likely to take severe losses and may be forced to liqui-
date portfolios, in a firesale. Value investors stabilise prices only when they are making money.
Seeking yield is maximising risk

Warren Buffett has argued that if one is seeking yield, one is maximising risk. It is impor-
tant to keep in mind that this risk is usually obscure, because management or supervisors
spot the obvious risk. This suggests that traders are actively seeking to take risk in an area
156
that is subject to the least amount of scrutiny. We will see many examples later in the
book of how this might done in practice and the consequences it has had.
9.3.1 Carry trades

A trading activity that has come under considerable criticism is carry trading, a form of value
investing in bonds or currencies based on exploiting yield differences. A higher-yield asset
is bought and a lower-yield asset sold, with the hope of profiting from the spread in yields.
This will be successful provided that capital losses are not taken on the higher-yield asset.
The most controversial form of carry trading is the foreign exchange carry, borrowing
money in a country with low interest rates, exchanging the money for the currency of a
country with high interest rates and lending it out at the higher rate. This creates profits
from both the interest rate differential and also the price impact on the exchange rates.
Example 9.4 Yen–dollar, 1998

Over two memorable days in October 1998, the 7th and 8th, the dollar fell from 131
yen to 112 yen by lunchtime in London on Thursday the 8th, bouncing back sharply
by the end of New York trading to 119 yen. The precipitous drop is shown in Figure 9.1.
The fall in the dollar was unexpected given its strength throughout the spring and sum-
mer of 1998, reaching a high of 147.26 yen on 11 August. The US enjoyed strong growth dur-
ing the period, Japan was suffering from a long recession and the conventional wisdom was
that the yen was set to continue its fall, with only the speed and the magnitude uncertain.
The combination of an appreciating dollar and the large interest rate differential,
4.9% between Japan and the US, gave rise to the singularly profitable trading oppor-
tunity of borrowing yen, buying dollar assets, and gaining on both the appreciation
of the dollar and the interest rate differential. This carry trade was widespread among
hedge funds, prop desks and Japanese commercial banks.
145
135
125
115
Feb Mar Apr May Jun Jul Aug Sep Oct
Figure 9.1 Yen/USD in 1998
Feedback effects
When carry traders are borrowing the low interest rate currency and buying the high inter-
est rate currency, it is likely to lead to the appreciation of the latter. Initially the effects
may not be particularly apparent, as positions are likely to increase gradually, but the
price impact becomes stronger with time.
157
However, it is quite likely at some point that traders will realise the currency has appreci-
ated too much, causing them to rapidly unwind their positions. The rapidity is self-generating
because unwinding increases losses, leading to a currency crash. This results in a weakening
of the high-interest currency, which in itself is an incentive for other traders to unwind their
positions. If everyone tries to unwind at the same time, the high-interest currency is sold into
a falling market, creating an adverse endogenous risk feedback loop, shown in Figure 9.2. It
is sometimes said that the prices go up by the escalator and down by the elevator (lift).
These effects were formally modelled by Morris and Shin (1998, 1999) in their global
games model, showing that speculators have incentives to wait until the last minute to
exit the trade, but no longer. If a trader exits too early, she will lose money, and also if she
exits too late. This leads to herd behaviour among speculators, caused by strategic com-
plementarities. We present their model in detail in Chapter 12, Section 12.5.
Carry trades as a source of contagion

Carry trades can be a source of contagion. Suppose a hedge fund is engaged in carry trades
with multiple countries, using a high degree of leverage. If a crisis hits one of the countries,
the hedge fund will face margin calls and will have to liquidate some of its positions to
meet those margins. It is likely that it will unwind some trades with countries not in crisis,
putting downward pressure on the exchange rate of those countries, causing them to fall,
and thus transmitting the crisis. See Figure 9.3 for a graphical description of this.
Dollar
weakens
Unwind
carry trade
Figure 9.2 Endogenous risk feedback in unwinding
Contagion
Hungary New Zealand
Ma
ZD
rg
ll N
in
c
Se
all
Crisis
Hedge fund
Figure 9.3 Carry trades as a source of contagion
158
This has happened several times, for example, in February 2006 when a mini-crisis in
Iceland adversely affected the exchange rates of several countries such as Hungary, New
Zealand and South Africa.
9.3.2 Technical trading

Technical trading is a broad category where a trader attempts to forecast future prices
based on statistical analysis of historical data, combining behavioural finance and quanti-
tative methods. These methods remain quite controversial and it is hard to verify whether
they are successful or not, because those developing successful technical trading rules
keep them strictly proprietary.
At high and very high observation frequencies, technical trading takes the form of statis-
tical arbitrage and HFT, where it can be highly successful, as information in prices arrives at
a much greater rate than that from any other sources. However, on human decision-making
timescales of hours or days the effectiveness of technical trading is very much in doubt.
There is a long list of studies proclaiming that it is possible to forecast prices with statistical
methods, and it is not hard to apply common methods to historical data and find significant
forecasting ability in-sample. However, as argued by Sullivan et al. (1999), such statistical
relationships tend to be spurious. If we test 20 variables to see if they help forecast prices,
one of them can be expected to be statistically significant at the 5% level, even if none is
related to prices. If we correctly take into account the impact of searching in the calculation
of statistical significance, common technical trading rules become statistically insignificant.
Momentum – following trend

A common technical trading strategy is momentum trading, buying assets that have seen
recent price increases and selling those that have fallen in price. If a sufficient number of
traders follow the momentum, their trading activities will directly impact on prices, in
the short run reinforcing the profitability of the trading strategy. Over long time periods,
momentum trading may cause asset price bubbles, followed by market crashes.
Momentum may be followed as a conscious trading strategy, but there are many ways
in which institutional decision-making can subconsciously lead to momentum trading.
For example, if investors prefer not to repeat losing decisions, or are prepared to take
more risk on positions that have in the past been successful, their trading will have a
momentum component. Similarly, it is widely surmised that investors who subcontract
management to external managers tend to award contracts to recently successful manag-
ers, and cancel contracts with unsuccessful managers, leading to buying assets that have
risen and selling those that have fallen. This suggests that a large proportion of momen-
tum trading is unconscious, perhaps taking the form of not repeating losing decisions
rather than actively seeking to repeat winning decisions.
9.3.3 High-frequency trading

Market participants have always employed various techniques to speed up their access to
information in order to gain an advantage on their competitors. Nathan Rothschild set up a
network of pigeons in Europe so he could send information between his operations faster than
159
anybody else. As legend has it, he was able to profit from the defeat of Napoleon at Waterloo
in 1815 because his pigeons brought the first news of the defeat to the City. Other technolo-
gies, such as post riders, semaphore and the telegraph, have served the same purpose.
Since the 1970s, high-speed computers, data networks and algorithmic trading have
become the main tools for gaining trading advantages. Today, computer algorithms
directly interface with trading platforms and place orders without immediate human
intervention, and are now the biggest generator of trading volume. As a consequence,
policymakers have become quite worried about the potential for systemic risk arising from
computer-based trading (CBT), especially after the Flash Crash of 2010.
Example 9.5 Flash Crash, 6 May 2010

The US stock market experienced a very sudden crash on 6 May 2010 when the market
went down by almost 10% in a span of a few minutes and recovered just as quickly.
This has been called the Flash Crash. Figure 9.4 shows the impact on the DJIA whilst
Figure 9.5 demonstrates the impact on two stocks, Accenture and Sotheby’s. The for-
mer saw its stock price drop from $40 to one penny at 14:48, whilst the latter jumped
from about $42 to $100,000 nine minutes later, when the market was already recov-
ering. This happened because the limit order books had been exhausted, and those
desiring to buy or sell found no counterparties.
‘These trades occurred as a result of so-called stub quotes, which are quotes gen-
erated by market makers . . . at levels far away from the current market in order
to fulfil continuous two-sided quoting obligations even when a market maker has
withdrawn from active trading.’
(Securities and Exchange Commission (SEC) and
Commodity Futures Trading Commission (CFTC) report)
The causes of the flash crash remain controversial. A joint report by the SEC and
the CFTC in September 2010 blames a fragmented and fragile market, where a sin-
gle institution selling a large number of so-called E-Mini S&Ps, that is a stock market
index futures contract on the S&P-500 index, exhausted available buyers.
‘Still lacking sufficient demand from fundamental buyers or cross-market arbi-

trageurs, HFTs began to quickly buy and then resell contracts to each other –
generating a “hot-potato” volume effect as the same positions were rapidly passed
back and forth. Between 2:45:13 and 2:45:27, HFTs traded over 27,000 contracts,
which accounted for about 49 percent of the total trading volume, while buying
only about 200 additional contracts net. At this time, buy-side market depth in the
E-Mini fell to about $58 million, less than 1% of its depth from that morning’s level.
As liquidity vanished, the price of the E-Mini dropped by an additional 1.7% in just
these 15 seconds.’
(SEC and CFTC report)
The crisis quickly spilled over from the futures markets into equities markets. As
trading systems detected the large volumes and price drops, many HFTs then exited
160
10800
10600
10400
10200
10000
10:00 11:00 12:00 13:00 14:00 15:00 16:00
Figure 9.4 Flash Crash, 6 May 2010, DJIA
40 105
30 104
20
103
10
102
0
14:45 14:50 14:55 15:00 14:45 14:50 14:55 15:00
(a) Accenture transaction prices (b) Sotheby’s transaction prices
1400
1200
1000
800
600
400
200
0
10:00 11:00 12:00 13:00 14:00 15:00 16:00
(c) Accenture number of trades
1000
800
600
400
200
0
10:00 11:00 12:00 13:00 14:00 15:00 16:00
(d) Sotheby’s number of trades
Figure 9.5 Flash Crash, 6 May 2010, individual stocks

Data source: NYSE, TAQ and Tick Data
from the market. The official report finds that the crash eventually stopped when the
Chicago Mercantile Exchange implemented automatic stabilisers, pausing trading.
That enabled market participants to react, verify prices and the system, leading the
market to recover quickly. This chain of events is typical of endogenous risk.
The official conclusion remains controversial, and many commentators have disa-
greed with both the assumptions and the conclusions.
161
Analysis
The speed of trading may make it easier for anomalous outcomes and asynchronicities
between different markets to self-correct, compared to human-only trading. Therefore,
events like the 1987 crash might have been less severe if trading had been faster. The
fact that the Flash Crash played out within a single day, and high-frequency extreme
outcomes do not generally result in extreme daily outcomes, indicates that problems in
high-frequency trading (HFT) self-heal. In other words, HFT might reduce the incidence
of market crashes. However, this issue has not been well studied, and it is not obvious
whether the effects will always be benign.
The critical difference between human and automated trading systems is common
sense. Faced with a new and unexpected situation we know that traders make some deci-
sions that look strange with hindsight, but their decisions will be expressed in shades of
grey and have some guiding principles behind them. By contrast, fully automatic systems
may fail catastrophically, and because of this will often shut down entirely when market
parameters exceed those of normal conditions. The price for liquidity always increases
under extreme circumstances, but where systems are highly automated and much of the
liquidity depends on them, this tendency becomes much stronger and the price may
approach infinity for short periods, as illustrated by Figure 9.4.
In the event, the Flash Crash did not cause much damage, because the markets recov-
ered quite quickly and the crash happened early in the trading session. If, however, the
Flash Crash had happened at the end of the trading day, and the markets closed around
the lowest point, the damage could have been substantial. It would have affected any-
body using daily marking, and many traders would have been hit by margin calls. It is
possible that an endogenous risk vicious feedback loop could have set in, causing prices
to continue spiralling downwards.
9.3.4 Short selling

Short selling, or shorting, is selling by market participants who do not own the asset sold.
Typically, they might borrow a stock or bond from somebody, promising to return it at
some future date. The trader then sells the asset. Until they repurchase and return the
asset, they are short. Typical reasons to short are to hedge some other position, or specu-
late on the possibility that the asset can be bought back for a lower price in the future.
Short selling is quite controversial, and has often been a source of abuse, for example,
in times past when management shorted their own stock, clearly putting their interests
at odds with those of the shareholders. Short selling is often perceived as profiting from
the misfortunes of others and contributing to price drops and even crises. As a result, it is
quite often banned in times of crisis.
Shorting has many legitimate uses, for example, in hedging where somebody making a
loan to a risky counterparty can hedge that risk by selling short the stock of the counter-
party. More generally, short selling increases liquidity, helps in preventing overvaluation
and provides incentives to produce negative research.
The problem with short selling is separating out economic from political or moral argu-
ments. While a short seller might be seen as attempting to benefit from the misfortunes of
others, in terms of market impact there is no difference between a short sale and a regular
162
9.4 Policy issues
sale from an asset. Both cause prices to fall. There is no reason to assign a pernicious
motive for short selling, any more than it is malicious to profit from prices increasing.
Naked short selling

Usually, most criticism is directed at naked short selling, which refers to two different prac-
tices, either a short speculative position rather than a hedge against another security, or
more recently the practice of short-selling an asset without first borrowing the asset.
It is often claimed that naked short selling in a crisis makes the crisis worse, and is
hence morally wrong. However, it is not clear why short sales are any worse than sales by
long holders or hedging with derivatives, with the latter practices more voluminous and
hence affecting markets much more than shorting.
Restrictions and bans

Short selling remains banned in many jurisdictions, and because of its connections to
profiting from and exacerbating crises, has often been temporarily banned once a crisis
is underway. For example, after the Great Depression started in 1929, one of the first
steps taken by President Hoover to regulate the activity of the markets was to compel the
New York Stock Exchange to curb short selling. Similarly, the SEC temporarily banned in
2008 what it called ‘abusive naked short selling’ in major banks in order to prevent their
share prices falling more than they were already. In 2011, France, Italy, Spain, Greece and
Belgium banned short selling of the shares of banks and other financial companies.
The available empirical evidence is unclear on whether such bans achieve their
intended purpose. They may even be counterproductive, as some studies suggest greater
falls for shares covered by a short selling ban, perhaps because this provides a powerful
signal that others would sell if permitted. Banning short selling signals panic on behalf of
the authorities which equally contributes to a crisis.
9.4 Policy issues

Before 2007, the prevailing view amongst policymakers was that most concerns about
trading and speculation were best left to the market. That did not mean that these activi-
ties were unregulated, far from it, and the world’s securities regulators, such as the SEC
in the US, have imposed a number of controls and checks on the trading process. These,
however, were more focused on improving the efficiency of markets and preventing fraud
and abuse, rather than improving financial stability.
There are many exceptions to this. Many of these policies date back to the Great
Depression. Just one example is US regulations on margins and counterparties. These are
explicitly aimed at addressing systemic risk concerns arising from the Depression. The
problem with these rules was that they were only partially effective when needed and did
not adequately address new areas of risk.
There was a general consensus that lightly regulated private enterprises lead to greater
efficiency than state enterprises in the provision of services ranging from telecommunica-
tions to finance, and that the state’s role should be peripheral – essentially to provide a
level playing field and restrict monopolies.
163
The financial crises starting in 2007 demonstrated the complacency of such attitudes.
Certain trading activities directly contributed to the crises, and policymakers ever since have
sought to develop rules that would regulate trading in a way that increases financial stability.
9.4.1 CCPs and financial stability

One problem that quickly became apparent when the crisis was underway was the myr-
iad of over-the-counter (OTC) derivative instruments. In the boom times before the crisis,
this did not seem such a big problem because the main risk factors, asymmetric informa-
tion, complexity and counterparty credit risk, were not apparent.
The problem of asymmetric information is demonstrated in Figure 9.6. In Figure 9.6(a)
we have four banks, representing the banking system, where all have exposures to each
other. The exposures are net zero, that is, if we add them all up, and hence net out all the
positions, the end result is zero. This is not how the individual bank sees it. For example,
bank A owes money to B, C and D, and is owed money by the same. If it worries about
the solvency of, say, B, it also fears not being able to make payments to C and D. At the
same time, it also expects C and D to be exposed to B. This means that if it becomes wor-
ried about the stability of the system, a prudent course of action would be to reduce its
exposures to everybody, even to the extent of causing a run on B.
This is what happened with Bear Stearns and Lehman in 2008. We call this problem
information asymmetry, since each participant does not know the full picture. If there was
a way to net out the positions across the system, the problem of information asymmetry
Bank B Bank C
Bank A Bank D
(a) Bilateral
Bank B Bank C
CCP
Bank A Bank D
(b) CCP
Figure 9.6 Role of CCPs
164
9.4 Policy issues
would be reduced. This is exactly what a central counterparty (CCP) does, because it acts
as a counterparty to each and every trade, and therefore knows all the exposures and
can net them out, as seen in Figure 9.6(b). This also solves the problem of counterparty
credit risk, provided there are no worries about the solvency of the CCP. As a side benefit,
it reduces complexity in the financial system because the CCP will prefer only a relatively
small number of standardised instruments.
Contrast the situation in which most contracts are bilateral. In this case the contracts
are likely to be customised, making it difficult to calculate aggregate positions for the
banks involved. For a third party such as a regulator it would be impossible.
How many CCPs?

Ideally, there should be only one CCP in the world, so that we could net all exposures
globally. That is impossible politically because major countries would want the CCP
to be within their own jurisdiction. At the very least, we would see one CCP in the US,
another in Europe and the third in Asia, and quite possibly more. Within the European
Union (EU), there are strong disagreements among member states as to who should
host the CCP, and similar debates rage in Asia.
If it is not possible to have a single global CCP, a particular problem is created as noted
by Duffie and Zhu (2011). If an institution has to use more than one CCP, its net position
across all CCPs might be zero, but this may not be visible to each CCP individually, which
sees only its own positions. The problem may be partially solved by different CCPs having
cross-margin arrangements and information sharing. This would somewhat reduce the
benefits of the CCP, particularly if CCPs do not standardise contracts between themselves.
However, so long as the number of CCPs is much smaller than the number of market par-
ticipants, significant benefits still accrue.
Safety of CCPs
CCPs eliminate most, but not all, counterparty credit risk, because market participants
are still exposed to the credit risk of the CCP itself. The failure of a CCP would significantly
contribute to systemic risk, since a substantial part of the financial system will be directly
exposed to it. For these reasons, the CCPs need to be safest of institutions, and they usu-
ally do enjoy explicit or implicit government guarantees. CCPs have been bailed out in the
past, for example, in Hong Kong in 1987. This, however, means that the CCP is an institu-
tion that is too big to fail (TBTF), with its own attendant problems.
The need for CCP safety creates another problem, as noted by Zigrand (2010).
During a crisis, the CCP might feel compelled to sharply increase margins, in order to
protect itself, thereby transmitting problems to members. Of course, the same might
happen regardless because any counterparty in bilateral transactions might behave in
the same way. Counting the arrows in Figure 9.6, it is clear that a CCP market with N
participants has 2N margined relationships, while without a CCP there are N(N – 1)
relationships. Assuming perfect netting between partners, exactly half of these will
require margin.
In the special case where position sizes are the same in each market, and if we make the
assumption that the CCP would respond to the crisis by increasing margins by an amount
165
similar to that of any other participant, the total increase in required margins will be
roughly 2/N of the size. Even if there are multiple CCPs, the pro-cyclical effect of margin
increases could be reduced by the presence of CCPs. In practice, gains will be more mod-
est, because in the CCP market the individual positions will typically be larger, and the
institution might have offsetting positions, recognised in the bilateral market but not the
CCP market, but the benefits of netting can still be significant.
Conclusion
The move towards CCPs is broadly positive, and to the extent that the various political dif-
ficulties can be overcome, would lead to a more stable financial system. However, CCPs
contribute to the commoditisation of the finance industry and will reduce profits; conse-
quently, industry opposition will be entrenched. Within the US most trading will move to
CCPs because of the Dodd–Frank Act.
Clearing houses follow recommendations made by the Bank for International
Settlements (BIS) in 2004, and the ongoing crisis has triggered a review of those stand-
ards. It is likely that CCPs will be required to further boost their financial safeguards, espe-
cially given their renewed importance.
9.4.2 Bonuses
A particular issue that has become quite controversial recently is bank bonuses. Bank
employees are amongst the most highly compensated individuals in the world, enjoy-
ing bonuses that in theory are connected to their individual profitability, their division
or bank.
In most cases, those engaged in trading for somebody else are incentivised by perfor-
mance bonuses, so that when they achieve high returns they share in the profits, but if
profits are low, or losses incurred, the trader receives only regular income. While this does
incentivise good performance and risk-taking, it does create certain problems, related
to the issue of incentives of guaranteed intermediaries discussed later in Section 14.7, as
both payoff structures have a similar option-like structure.
Perhaps the main problem is that the trader shares in the profits but not the losses. In
other words, the risk-return profile is asymmetric. This means that an individual trader
might be quite content taking significant amounts of risk; if the bet turns out well, she
will get large bonuses, but otherwise her employer fails. While that might not be the
type of risk profile her superiors, shareholders, supervisors and other stakeholders desire,
she is the one making the trading decisions. See Macrae and Watkins (1998) for more on
these issues.
Various stakeholders might want to prevent such excessive risk-taking, employing many
layers of controls intended to restrain risk. Management typically restricts traders in the
types of securities they may trade, the level of risk anticipated, and the amount of capital
employed. Senior management will impose similar restrictions on divisions, and regula-
tors on banks. Despite all these many-layered restrictions, in practice, it can be difficult
to contain risk-taking. Trading strategies are often very complicated and it may be impos-
sible for any outsider to make sense of what is happening. In addition, the trader has an
166
9.4 Policy issues
incentive to make the trading activities as complicated as possible, to prevent outside

scrutiny, and to take risk in areas that receive the least amount of scrutiny.
Does limiting bonuses reduce systemic risk?

Arguments against high bank bonuses often cloak unhappiness with high compensation
levels with arguments that bonuses increase risk. However, little evidence exists either
way, despite the strong a priori argument that incentives should influence behaviour, and
one can treat arguments on either side only as assertions rather than as evidence-based
conclusions. These issues are discussed by Danielsson and Keating (2009), for example.
Looking at crises past, there seems to be a recurring thread of overconfidence and herd-
ing. This may happen because competitive pressures cause market participants to emu-
late the successful initiatives of their competitors, even if it may look imprudent to outside
observers. Such pressures would be present regardless of any bonuses paid.
What is clear is that historically we have seen excessive risk playing a central role in
financial crises without a bonus in sight. Perhaps the best example is the Japanese excesses
and subsequent crisis in the 1980s and early 1990s. Similarly, the largest loss to a rogue
trader in history, that of Jérôme Kerviel, was caused by an individual paid mainly via a
fixed salary, with no prospects of receiving a bonus commensurate with his vast positions.
Similarly, strict punishments for failure have not prevented banks from collapsing. The
largest nineteenth-century bank that failed, Overend & Gurney, was a partnership, so that
its failure presumably meant poverty for the partners. Still, it did not dissuade them from
plunging headlong into the junk bond market at the time. Similarly, Francesch Castello
knew the punishment for bank failure was execution, and still took too much risk, failed
and was executed.
However, the presence of asymmetric risk-return profiles can only increase risk, and the
payment of large bonuses for successful risk-taking will therefore encourage socially unde-
sirable levels of risk-taking and the manipulation of the financial system. For example,
the LIBOR price-fixing would have been less likely if derivatives traders in Barclays did not
directly gain from fixing prices.
How easy is it to limit bonuses?

Policymakers have recently become concerned about bonuses, as expressed by the
Financial Services Authority (2009a), and several initiatives aiming at regulating compen-
sation have been launched. Even if the case for limiting bonuses seems fairly clear-cut, it is
not obvious how that could be done in practice, at least in a manner that clearly improves
financial stability.
If financial institutions attempt unilaterally not to pay bonuses, they are at the risk of
star employees – sometimes called rainmakers – leaving the firm. The banks maintain that
they cannot cut bonuses unilaterally for competitive reasons, but such arguments are of
course self-serving and not much support exists for or against that assertion. National reg-
ulators face similar threats from their banks if they aim to regulate compensation. Head
offices are mobile, a point made clear in negotiations with regulators.
For this reason, it is sometimes argued that the only way to prevent excessive bonuses
is for governments to coordinate in regulating compensation. Within the EU, bankers’
167
bonuses are to be limited by law. A new rule, from the European Parliament, applies
to senior employees of EU-based banks anywhere in the world as well as to EU-based
staff of non-European banks. This covers all forms of bonuses, where the ratio of bonuses
to salary will be limited to 1:1. This would have several different implications: we
could expect base salaries to increase, non-European banks to move trading activities
away from the Union, and traders to resign from European banks and move out of the
Union.
If we could assume that funding for their activities (and hence the risk) would actually
move with them, it could then make banking safer within the EU, provided that all core
banking activities actually remained onshore and European governments did not sim-
ply find themselves obliged to bail out foreign banks. The extensive support provided by
the Federal Reserve System (Fed) to foreign banks, like Royal Bank of Scotland (RBS) and
HSBC, discussed later in Section 14.5.1 is an interesting precedent.
Such initiatives also raise employees’ incentives to gain and retain a bank job, while
reducing the optionality that incentivises risk-taking. However, other distortions are
introduced. For example, there are indications that basic salaries in banking are sharply
increasing from their already high levels. This improves the industry’s position in the com-
petition for new graduates, to the detriment of other employers like regulators and the
non-financial industry.
9.4.3 Universal and narrow banking

Many observers have become worried about the fact that the same commercial banks
that are essential for the functioning of the economy are also engaged in significant risky
investments and speculation, where large losses might cause an otherwise solvent com-
mercial bank to fail, seriously disrupting economic activity.
For this reason, following the Great Depression, the US passed the Glass–Steagall Act,
which sought to separate risky investment banking from presumably safer commercial
banking. Over time, the Glass–Steagall Act became progressively watered down, and was
eventually repealed in 1999. It is ironic that by the end of 2008, all of the remaining
investment banks in the US became commercial banks as a result of Fed efforts to prop
up the financial system. This was, in effect, a recognition of the Lehman lesson that the
investment banks were systemically vital and that, as with commercial banks, regulators
could not, in practice, stand back and allow them to fail.
The end of the Glass–Steagall Act, and the lack of a similar law in Europe, is frequently
blamed for contributing to the excessive risk-taking in the banking system and, hence, the
ongoing crisis.
Initiatives
It is therefore not surprising that several initiatives aiming at preventing excessive risk-
taking in banks have been launched by various policymakers. For example, the US, as a
part of the Dodd–Frank Act, has proposed the so-called Volcker Rule, named after a former
governor of the Fed, Paul Volcker, aimed at restricting commercial banks from engaging in
proprietary trading.
168
9.4 Policy issues
A similar proposal was made by the UK Independent Commission on Banking1 in 2011,

in the Vickers report, named after the chairman of the commission. It states that British
banks must separate out high street banking activities from investment banking opera-
tions. This means ring-fencing all retail and small business deposits, with investment
banking activities, such as derivatives, debt and equity underwriting, and investing in pro-
prietary trading, falling outside. This protects the retail and payments operations of banks
so that a bank failure would not result in failure of the more economically important divi-
sions of the bank.
The effectiveness of these proposals is untested, and it is also unclear whether they
will in fact be adopted. One of the largest banks in the UK, HSBC, has warned that it may
move its headquarters if the government were to break up big banking groups (Financial
Times, 2010b). As well as the predictable opposition from UK banks, the measures also
face hostility from other European countries, which would prefer all European banks to be
regulated on an equal basis, in this case a less stringent one.
After the US and UK initiatives, the EU commissioned a study from Erkki Liikanen,
governor of Finland’s central bank (Liikanen, 2012), to look into possible reforms of the
European banking sector. The Liikanen report reaches similar conclusions in some aspects
as the other two reports, addressing ring-fencing and trading restrictions. It also aims to
limit bonuses and reduce risk in real estate. It remains unclear whether the EU will adopt
the report’s conclusions.
9.4.4 Financial transaction tax

One of the oldest types of tax is a financial transaction tax (FTT), often referred to as a
stamp duty or something similar. This does not tax profits or income but rather specific
transactions. It is common in real-estate transactions, and many countries apply it to rou-
tine financial transactions. For example, the UK imposes a 0.5% tax on share transactions.
While FTT was used historically to raise revenue for the government, Keynes in 1936
proposed a different objective, that is to curb speculation. This works because FTT is by
its nature proportional to turnover, and so creates a disincentive for rapid trading, thus
reducing speculation.
Tobin tax
A specific example of an anti-speculative FTT is the Tobin tax, as proposed by the econo-
mist James Tobin in 1972, which is a uniform worldwide tax on spot currency transactions.
He argued that, designed correctly, the tax would cushion exchange rate fluctuations
by penalising short-term financial capital flows and ‘throw some sand in the wheels of
super-efficient financial markets and create room for differences in domestic interest
rates, thus enabling national monetary policy to respond to domestic macroeconomic
needs’ (Tobin, 1996). He further claimed that such a tax could strengthen the weight of
fundamentals on exchange rates, diminish excess volatility and expand the autonomy of
national monetary policies.
1
bankingcommission.independent.gov.uk.
169
Given the large trading volumes in foreign exchange markets, estimated by the BIS to
be $4 trillion per day in April 2010, even a very small tax rate might be expected to bring
in significant revenue. That is not true, however, because trading volumes would probably
collapse as a consequence of the tax.
The main argument against the Tobin tax is that for it to be effective, it needs to be
introduced in every financial centre, otherwise foreign exchange trading would move to
the areas where the tax has not been implemented. It is hard to know how serious this
problem would be, because there are significant costs in relocating and such offshore
activities would still need to interact with the taxed onshore sector.
It might be harder to prevent the migration of trading to derivative contracts, beyond
the control of the authorities. For example, if a bank sells beef for pounds and buys a
matching quantity of beef for dollars, it has entered into two livestock contracts, not for-
eign exchange (FX) contracts, yet has still contrived to exchange dollars for pounds. There
are myriad similar ways to mask the true nature of financial transactions.
The FTT debate

While some proponents of the FTT argue that it would bring in substantial tax revenue,
the main arguments in favour of the tax focus on financial stability, like the fact that an
FTT would be an excellent anti-complexity device. Each time a financial asset is traded,
yet another link is created between financial institutions, whilst the distance between
the owner of the asset and the originator becomes longer. Because an FTT will reduce
the number of transactions, complexity is automatically reduced as a consequence. More
generally, as argued by Tobin, according to the general principle of control theory, the
introduction of friction into systems tends to increase stability.
The opponents of FTT, many from the finance industry, argue that the tax revenues are
overestimated, that it would not be effective in increasing financial stability, that domes-
tic banks would lose ground relative to foreign banks and hence jobs would be lost, that
trading would migrate to offshore locations or derivative contracts, and finally that this
is the wrong time to increase taxes on the financial sector, in the middle of a crisis, espe-
cially as the sector is already receiving significant public assistance.
While there are arguments on both sides of the discussion, little evidence exists either
way, and the debate is more often than not based on general attitudes to finance and
speculation, rather than the underlying facts. The question of FTT is under considerable
scrutiny by researchers but no clear answer has yet emerged.
FTT in Europe
As a part of its post-crisis regulatory reform agenda, the EU proposed a FTT in 2011 to
take effect in 2014. This would affect financial transactions between financial institu-
tions, taxed at 0.1% for shares and bonds and 0.01% for derivative contracts. The FTT
has met significant opposition, especially by the UK, the Netherlands and Sweden, and
there have been proposals to apply it only to the euro zone. At the time of writing, it is
unclear what the final outcome will be. Some states, including Germany and especially
France, are implementing a unilateral FTT regime, which will provide interesting experi-
mental data.
170
Appendix: Basic terminology of trading
9.5 Summary
Speculation and trading is an essential part of the financial system. They provide key ser-
vices to economic agents, but also have the potential for destabilising financial markets
and even causing crises.
Trading scandals of various types are common, such as the case of JP Morgan Chase
and the London whale, the LIBOR price-fixing scandal, Ponzi schemes like the Madoff
scandal, and rogue traders.
There are many categories of market participants, for example proprietary traders who
aim to generate trading profits for their employer. Most market participants act on behalf
of investors, a part of the institutionalisation of dishonesty markets. Perhaps the most
controversial market participants are hedge funds.
Financial institutions have been very active in financial innovation, the creation
of various types of new financial instruments, but this remains quite contentious,
and some respected commentators claim that financial innovation benefits only the
inventors.
Trading involves various types of activities like value investing, technical trading
and momentum trading, but also carry trading which has frequently been blamed for
undermining government policy. Perhaps the most controversial trading strategy is
short selling.
Following the crisis from 2007, there have been various initiatives aiming at curbing
perceived abuses in the financial system. These include protecting us from casino bank-
ing, limiting bonuses and taxing trading.
Appendix: Basic terminology of trading

We refer to each trading entity as a party, and when referring to the other party we use the
term counterparty. The trading process has three distinct stages.
1 In the first, traders submit orders to buy or sell a security. Buy orders are matched with
sell orders in the marketplace, often an exchange, and the parties trading are notified of a
match and a successful trade.
2 This is followed by a process called clearing that involves the exchange of information,
regulatory reporting and various other activities that must take place before a transaction
can be completed.
3 The last stage, the actual exchange of money, securities or goods, is called settlement.
Definition 9.1 Exchange An exchange is an organised venue for trading financial

assets, usually nowadays a computer system.
While some trading takes place in a formal venue called an exchange, most trad-
ing bypasses exchanges, and takes place over-the-counter in a bilateral deal between
counterparties.
171
Definition 9.2 Market maker A market maker is a financial institution or an

individual that quotes both a buy and sell price of a financial instrument, hoping to profit
from the spread between the buying and selling prices.
Definition 9.3 Broker A broker is an individual or institution that arranges transactions

between buyers and sellers, charging a commission when a deal is executed.
For example, on organised exchanges, market makers are often permitted to acquire
stock for the purpose of later selling it to meet anticipated customer demand.
Clearing and settlement involve procedures meant to reduce the risk to each party that
the other party will fail to complete the transaction as agreed, that is counterparty risk.
These activities are often concentrated in an institution called a clearing house.
Definition 9.4 Clearing house A clearing house is a firm that sits in between buyers
and sellers, acting as a counterparty for each, and also clearing trades, that is taking care
of the movements of assets and money. They net trades among their members, manage
collateral before facilitating payment and delivery at the settlement date, and take care of
the various accounting details. See Figure 9.7 for the various steps.
Definition 9.5 Netting Netting means that we consider the net amounts owed by
two counterparties, not the gross amounts. For example, if party A owes B $10 million, and
B owes A $11 million, the gross amount is $21 and the net amount $1 million.
Definition 9.6 Central counterparty (CCP) A CCP is the legal counterparty on

each side of every trade in the market, with the result that all market participants have
contractual relationships only with the CCP rather than each other.
A clearing house may act as a CCP in a particular market. In order to be included in a CCP,
it is necessary for a security to be standardised.
Assets Assets
Clearing
Seller Buyer
house
Money Money
Collateral
management, move
money and assets,
accounting,
regulations, netting,
reporting
Figure 9.7 Clearing house
172
References

1 Do you think scandals like the JP Morgan ‘whale’ indicate that something is fundamen-
tally wrong in the financial system?
2 What is the real damage caused by the LIBOR scandal?
3 Is Nick Leeson a hero, villain or neither?
4 Carry trades
(a) What are carry trades?
(b) Are speculative activities such as carry trades beneficial or not to the economy?
(c) How can carry trades be vehicles for contagion?
5 Is financial innovation a force for good or evil?
6 What are the main arguments in favour and against curbing high-frequency trading?
7 What is the main contribution of CCPs, and how could they increase financial instability?
8 Is naked short selling bad?
9 Is narrow banking a good idea?
10 Is the EU right in its pursuit of a financial transaction tax?
References
BBC (2012b). Barclays’ widening Libor-fixing scandal. www.bbc.co.uk/news/business-18671255.
Danielsson, J. and Keating, C. (2009). Bonus incensed. Vox EU, 25 May. http://voxeu.org/index.
php?q=node/3602.
Duffie, D. and Zhu, H. (2011). Does a central clearing counterparty reduce counterparty risk?
Stanford University Working Paper.
Financial Services Authority (2009a). Reforming remuneration practices in financial services.
Technical report.
Financial Services Authority (2012). Final notice. Technical report www.fsa.gov.uk/static/pubs/
final/barclays-jun12.pdf.
Financial Times (2010b). HSBC in clearest warning over relocation. Financial Times, 3 September.
www.ft.com/cms/s/0/5e0ba186-b6bd-11df-b3dd-00144feabdc0.html#axzz1zsbIzMf.
Financial Times (2012a). DoJ probes $2bn JPMorgan trading loss. Financial Times, 15 May. www.
ft.com/cms/s/0/39133792-9e9f-11e1-9cc8-00144feabdc0.html#axzz1x8DGs4zP.
Financial Times (2012b). JPMorgan whale harpooned. Financial Times, 11 May. ftalphaville.
ft.com/blog/2012/ 05/10/995211/jpmorgan-whale-harpooned/.
Graham, B. and Dodd, D. (1934). Security Analysis. Whittlesey House (McGraw-Hill).
Hull, J. C. (2011). Options, Futures, and Other Derivatives, 8th edition. Prentice Hall.
Kohn, M. (2003). Financial Institutions and Markets, 2nd edition. Oxford University Press.
Leeson, N. (1999). Rogue Trader. Sphere, London.
Liikanen, E. (2012). High-level expert group on reforming the structure of the EU banking sec-
tor. Technical report, European Union.
173
Macrae, R. and Watkins, C. (1998). A disaster waiting to happen. Mimeo, www.cs.rhul.ac.uk/

home/chrisw/Disaster.pdf.
Madura, J. (2010). Financial Institutions and Markets, 9th edition. South Western College,
Cincinnati, OH.
Mishkin, F. S. and Eakins, S. (2011). Financial Markets and Institutions, 6th edition. Pearson.
Morris, S. and Shin, H. S. (1998). Unique equilibrium in a model of self-fulfilling currency
attacks. Amer. Econ. Rev., 88: 587–97.
Morris, S. and Shin, H. S. (1999). Risk management with interdependent choice. Oxford Review of
Economic Policy, 15: 52–62, reprinted in Bank of England Financial Stability Review, 7, 141–50,
www.bankofengland.co.uk/fsr/fsr07art5.pdf.
SEC and CFTC (2010). Findings regarding the market events of May 6, 2010. Technical report.
Sullivan, R., Timmermann, A. and White, H. (1999). Dangers of data mining: the case of calen-
dar effects in stock returns. J. Econometrics, 105(1): 249–86.
Tobin, J. (1996). A currency transactions tax, why and how? Open Economies Review, 7: 493–9.
Wall Street Journal (2009b). Paul Volcker: think more boldly. Wall Street Journal, 14 December.
online.wsj.com/article/SB100014 24052748704825504574586330960597134.html.
Zigrand, J.-P. (2010). What do network theory and endogenous risk theory have to say about
the effects of CCPs on systemic stability? Financial Stability Review, 14, Banque de France.
174
10 Credit markets
The most important financial products are fixed income assets, composed of various
credit instruments, like bonds, structured credit products and credit derivatives. The
reason is that credit is an absolute necessity for the modern economy, allowing peo-
ple to save for retirement and get mortgages, governments to finance their operations
and stimulate the economy in recessions, and private companies to finance opera-
tions and expansions.
Of the fixed income assets, the most important are bonds. Financial institutions,
corporations and governments depend on bonds for financing, and any problems in
accessing the bond markets can quickly cause significant difficulties. That means the
bond markets exert significant power over borrowers. Perhaps the best expression of
their power was by a Bill Clinton strategist, James Carville, who said that he wanted
to be reborn as the bond market because it is the most powerful force in the universe.
Currently, some European governments are facing the wrath of the bond markets.
Investors in fixed income markets are often faced with borrowers of an unknown
quality and need some mechanisms to ascertain the credit quality, and monitor bor-
rower performance. The most common way to achieve this is by using credit ratings
issued by credit rating agencies (CRAs). These are private companies whose business it
is to report on credit quality.
Trying to forecast credit risk is difficult in the best of times and the CRAs have been
under almost constant criticism for the quality of ratings and the impact downgrades
have on borrowers. As a result, there are frequent calls for reforms of the credit rating
business, but it is not obvious how that could be accomplished.
Chapter 10 Credit markets
We can also use various types of mathematical credit models to forecast credit risk.
Of these credit models, the so-called reduced form models are particularly useful
when it comes to mapping between the probability of default and credit spreads.

This chapter directly relates to Chapter 9 (trading and speculation) and Chapter 15
(dangerous financial instruments).
Key concepts
■ Fixed income
■ Margins and haircuts
■ Credit rating
■ Reduced form credit models
■ Spreads
■ Securitisation

There are many books covering the topics discussed in this chapter. Besides the spe-
cific references in the text, we have drawn on books such as those by Hull (2011,
2012). A comprehensive discussion of bond markets is in Fabozzi (2009). Murphy
(2009) discusses many of these issues in context of the crises from 2007.

c Coupon payment
P Price
pd Probability of default
r Risky interest rate
rf Risk-free rate
s Spread, perhaps r–rf
T Maturity (years)
z Recovery rate
10.1 Market for credit

The financial market that gets most attention in public discourse is the equity market, but
in importance and size it is dwarfed by the bond market.1 In 2010, the size of the bond
market was $95 trillion whilst the global equity market reached only $55 trillion.
This counts only regular bonds, sometimes referred to as plain vanilla bonds. If we
include other credit instruments up to and including bank loans, the amount of overall
credit is much larger.
1
See www.imf.org/external/data.htm, www.bis.org/statistics/secstats.htm, www.sifma.org/research/statistics
.aspx and www.world-exchanges.org/statistics.
176
10.1 Market for credit
Fixed income assets

Fixed income assets are assets providing payments on fixed schedules. They are debt instru-
ments with an issuer (borrower) and an investor (lender). The markets for fixed income
assets range from simple instruments such as zero coupon bonds to the magical world of
credit derivatives.
Firms and governments issue debt in the primary market, generally by engaging an
investment bank to act as an underwriter or primary dealer via auctions. Subsequent
trading takes place in the secondary market, which could be an organised exchange or a
more informal over-the-counter (OTC) market. Secondary markets provide two important
functions: they give the original purchasers of securities the ability to resell, thus creating
liquidity, and they provide a near-continuous reassessment of securities’ prices. Bonds are
typically traded in OTC markets rather than on formal exchanges.
10.1.1 Bond markets

The Bank for International Settlements (BIS) publishes detailed statistics on global bond
markets, segmenting debt securities into domestic and foreign. The former are generally
issued by domestic residents in domestic currencies, sold to resident investors, whilst
international securities are global issues, mostly sold to and bought by international resi-
dents and issued in foreign currency. Of course, it is often difficult to precisely delineate
between the two, especially with regard to the buyers. The BIS distinguishes between
three categories of issuers: governments, financial institutions and corporations.
We show the development of the bond market in Figure 10.1. Both the domestic and
international markets have grown rapidly, but the international much more so. Most of
the growth is derived from financial institutions. In the domestic market, the persistently
40
Government
30 Financial institutions
Corporate issuers
20
10
0
1989 1992 1995 1998 2001 2004 2007 2010
(a) Domestic
20
Government
15 Financial institutions
10 Corporate issuers
5
0
1993 1995 1997 1999 2001 2003 2005 2007 2009 2011
(b) International
Figure 10.1 Outstanding bond volumes, 2011 prices, USD trillions

Data source: BIS and World Bank
177
Table 10.1 Outstanding bond volumes, June 2011, USD trillions
Domestic International
Gov. FI Corp. Gov. FI Corp.

US 11.61 10.94 2.93 0.01 5.48 1.84
Japan 12.09 1.18 0.90 0.00 0.36 0.05
France 1.94 1.32 0.31 0.07 1.74 0.46
Italy 2.17 0.78 0.38 0.27 1.12 0.11
China 1.49 1.05 0.60 0.00 0.09 0.02
Germany 1.92 0.52 0.41 0.34 2.47 0.15
UK 1.38 0.31 0.02 0.12 2.83 0.32
Rest 8.46 5.72 1.48 1.86 7.88 1.02
Notes. Gov. is government, FI financial institutions and Corp. corporate issuers

Data source: Bank for International Settlements, http://www.bis.org
largest issuer is governments, followed by financial institutions and then corporations.

This market has also grown rapidly: it is more than three times larger, in real terms, than
it was 23 years ago, growing by more than 6% a year. The international market has grown
by more than 13% a year.
Table 10.1 constitutes a snapshot of the the global bond market in June 2011, showing
the seven countries with the largest domestic bond market. The world’s largest economy,
the United States (US), is the biggest issuer, followed by Japan, which has the highest gov-
ernment debt in the world. China is number five, driven mostly by government bonds. The
United Kingdom (UK) is seventh, but in the market for paper issued by financial institu-
tions it is the second, not surprising given the importance of City, the UK financial sector.
Example 10.1 Bond pricing

The price of a bond is inversely related to changes in interest rates. The reason is that
as interest rates increase, future payments are less valuable than before, so the price
of the bond falls. In other words, future payments are discounted to the present at a
higher rate.
This can be shown formally by looking at the pricing equation for a bond. Interest
rates at a given maturity t are rt, whilst the cash flow from the bond is5ct6Tt = 1, that is
the coupon payments, where the last payment, cT, also includes the principal and T is
maturity in years. The price of a bond, P, is given by the present value of the cash flow:
P = a
T ct
t=1
11 + rt2t
10.1.2 Credit risk

Credit risk is the main risk factor facing an investor. While one might see the term credit risk
being used only for probabilities of default, most commentators use it in a broader sense
as referring to the overall risk of losing money from fixed income instruments. This might
178
10.2 Credit rating agencies
happen because of a downgrade causing bond prices to fall in price, the failure of a bor-
rower to make payments, changes in the probability that payments will not be made or will
be delayed, changes in interest rates, or the amount of money recovered in case of default.
Default and recovery

Probability of default refers to the risk of borrowers failing to repay their debts as agreed, or
otherwise failing to meet their contractual obligations by, for example, missing scheduled
payments. The common pari passu clause in bond covenants means that debtors are con-
sidered in default on all their debt obligations as soon as they default on any particular one.
Loss given default quantifies how much a creditor can expect to get back after a default.
It often takes many years, and even decades, to get recovery. In addition, there are signifi-
cant national differences, so the failure of an otherwise identical company in two differ-
ent countries might result in widely different recovery rates.

The creditworthiness of borrowers is of key interest to anybody buying fixed income
assets. While the buyers may want to estimate the creditworthiness by themselves, it can
be a time-consuming and difficult task, especially for small and unsophisticated investors.
As a consequence, they often outsource the analysis of creditworthiness to specialised
companies called credit rating agencies (CRAs). The two largest are Standard & Poor (S&P)
and Moody’s, with Fitch not far behind.
The CRAs do not provide investment advice or analysis of how much money an investor
might lose; instead, in the words of S&P:
‘A credit rating is S&P’s opinion of the general creditworthiness of an obligor, or the

creditworthiness of an obligor with respect to a particular debt security or other
financial obligation, based on relevant risk factors.’
The two key terms to note are opinion and creditworthiness. CRAs do not provide advice,
only opinion. There is an important legal difference between those terms, because advice
implies an obligation to be right, while opinion does not. It is therefore easier to challenge
advice in a court of law than it is to challenge opinion.
The CRAs are generally considered to have considerable expertise in credit risk and are
regarded by many as unbiased evaluators of such risk. As a consequence, their ratings are
widely accepted by both market participants and government agencies, and many inves-
tors are restricted to hold only rated instruments.
Ratings
Ratings take the form of a letter grade, as shown in Table 10.2 where we show ratings from
both S&P and Moody’s. For S&P the highest rating is AAA, followed by AAA-, AA+ all the
way down to default. Moody’s has a similar arrangement, but note that ratings from two
different agencies are not directly comparable, even if we broadly expect them to provide
similar information.
179
Table 10.2 Ratings
S&P Moody’s Quality

AAA Aaa Prime
AA + Aa1 High grade
AA Aa2
AA - Aa3
A+ A1 Medium grade
A A2
A- A3
BBB + Baa1
BBB Baa2
BBB - Baa3
BB+ Ba1 Non-investment grade, ‘junk’
BB Ba2
BB – Ba3
B+ B1 Highly speculative
B B2
B- B3
CCC + Caa1
CCC Caa2
CCC - Caa3 Near default, limited recovery
CC Ca
C
C In default
D
Ratings can apply to either an issuer, such as a corporation or sovereign, or to a specific

debt issue. The rating process includes quantitative, qualitative and legal analysis. The
quantitative analysis is mainly financial analysis based on a firm’s financial reports like the
balance sheet, cash flow statement and profit and loss account. The qualitative analysis
is concerned with the quality of management, a firm’s competitiveness and the expected
growth of the industry it operates in, as well as its vulnerability to technological, regula-
tory and labour changes. The nature of this assessment is, of course, subjective. Ratings
are usually reviewed once a year.
Spreads
The yield curve is a curve showing several yields or interest rates across different maturi-
ties, that is, the time at which the payment is due. Each debt issue has its own yield curve,
determined by its credit rating. The difference in the yields for different ratings is called a
spread. The lower the credit rating, the higher the yield. We show in Figure 10.2 a typical
term structure for bonds of different rating categories.
180
5.3%
5.2%
Yield
5.1% C AA
5.0% B Risk free
5 10 15 20
Time to maturity in years
Figure 10.2 Yields, maturity, and credit spreads
10.2.1 Issues with rating agencies

Because of the fundamental role played by the CRAs, it is not surprising that they are
under constant criticism. Ratings are an inexact field and involve forecasting future com-
pany performance and implicitly the economy generally, inevitably meaning that ratings
are inaccurate. Ratings are used in the regulatory process, and therefore both are espe-
cially important and attract special scrutiny.
Critics have been quick to jump on mistakes, attacking ratings for missing individual
defaults, such as those of Enron, WorldCom and Parmalat, and country problems as in
the Asian crisis of 1997 and more recently problems in European sovereigns. The CRAs
have also come under criticism for their aggressive approach to collecting fees. This was
documented by the Washington Post (2004) and we discuss one case from that article,
that of Hannover Re and Moody’s.
Example 10.2 Hannover Re and Moody’s

When the CRAs made a big push into Europe, they allegedly used aggressive tactics to
collect fees. According to the Washington Post (2004), Moody’s informed the German
insurance company Hannover Re in the mid-1990s that it had decided to rate the
company at no charge, but was looking forward to the day Hannover Re was willing
to pay for the ratings. Hannover Re refused, and never paid Moody’s. Moody’s rated
Hannover Re anyway, starting with Aa2 in 1998 and downgrading three times, eventu-
ally to Baa1 (near junk) in 2003; while S&P, which did get paid by Hannover Re, has
rated it AA- from 2003 until 2012. Moody’s stopped rating Hannover Re in 2008.
Conflict of interest
Ratings are generally solicited and paid for by the issuer of debt and not the investor.
Good ratings enhance the marketability of debt issues, enabling issuers to place debt
more easily at lower interest rates. The prospect of bad ratings may cause an issuer not to
go ahead, thus depriving CRAs of substantial fees. This leads to the perception that the
181
CRAs are too willing to provide good ratings in order to generate fee income. The agencies
strongly reject the possibility of any conflict of interest.
However, in a world where issuers can go ratings shopping, that is, talk to several CRAs
before picking one to rate an issue, worries about conflict of interest are unlikely to disap-
pear. These problems have become especially pertinent following the crises from 2007
because of how banks were able to use the ratings models to structure credit products.
We discuss that topic in detail later in the book.
Evaluating ratings quality

The term ratings quality refers to how accurate a rating is in predicting future performance
of a debt issue. The CRAs publish historical records on ratings and defaults, enabling
researchers to ascertain the long-run quality of ratings. This is especially relevant in the
case of US corporate ratings since the CRAs have operated there since the 1920s, resulting
in almost a century’s worth of data. It is hard to access the ratings quality of sovereigns as
well as non-US corporations because the historical record is limited.
It is perhaps most difficult to determine and evaluate the highest ratings because the
failure rates are so low that adequate data may not be available. This means that the
potential for model risk is especially important for AAA ratings.
European sovereign debt crisis

The CRAs have been severely criticised for their performance in the European sovereign
debt crisis, both missing the pending problems in Cyprus, Portugal, Italy, Ireland, Greece
and Spain, and then exacerbating their problems with downgrades.
A typical example was the reaction of European leaders to the downgrades of Portugal,
in the summer of 2011. Viviane Reding, the European Union (EU) justice commissioner,
said: ‘We can’t have a situation where a cartel of three US enterprises decides the fates of
entire national economies and their citizens.’ Either their ‘cartel’ should be smashed or
‘independent’ European and Asian ratings agencies set up, she said. The implication of
this remark is that EU commissioners seem to believe that the reason for the sovereign
debt crisis is downgrades by the CRAs, and not the mismanagement of public finances
by some European states. If the EU were to set up its own ‘independent’ European CRA,
it is unlikely that the ratings would be taken seriously. The EU announced in 2012 that
the Union was to set technical standards and regulate CRAs under the auspices of the
European Securities and Markets Authority (ESMA).
The oligopoly of ratings agencies

In the early 1970s, the US securities regulator, the Securities and Exchange Commission
(SEC), started to mandate that certain investors could only buy debt instruments rated
by a recognised CRA. That amounted to the creation of an oligopoly of recognised ratings
agencies, and over time the requirements that issues be rated before purchase became
more common. The oligopoly of the CRAs was enhanced by their central role in the Basel
regulations. Throughout this time, the authorities had no objections to the merger of
CRAs, and by the start of the crisis in 2007, following a wave of mergers, two agencies had
80% of the market.
182
10.3 Credit models
More recently, the US and European authorities have been encouraging the creation of
new CRAs, and we can expect new CRAs in the future, including some that charge buyers,
not issuers.
Legal protection
Given the importance of CRAs, it is no surprise that they are frequently sued by aggra-
vated investors. However, in the US, they have consistently maintained that they are no
different from financial journalists offering opinions, thus getting protection from the First
Amendment to the US constitution – free speech. This would make them different from
consultants and auditors who have a contractual obligation to provide a correct opinion,
given the facts available to them at the time.
The CRAs have until now prevailed in US courts, but changes in law may alter that.
Furthermore, no such protection of free speech exists in Europe so they need different
legal protection there.
This leaves the question of whether the CRAs should be held legally accountable for
inaccurate ratings. While it might seem just that investors be able to sue them for dam-
ages, the end result would only be that the expected cost of damages be built into their
fee structure. Since any damages would probably be substantial, the cost of ratings would
increase significantly. It is not clear if that would be of benefit to investors.
Furthermore, in a world where two agencies have 80% of the market it would be dif-
ficult to punish either severely, because that might force it to leave the business, leaving
only one major CRA. The case where the accounting firm Arthur Andersen had to close
down following its mistakes with Enron serves as a cautionary tale.
Solution
While it is not feasible to do away with the ratings agencies altogether, nor would it be
sensible to substantially restructure the existing ones, there are several steps that could
be taken to minimise the problems created by ratings agencies. It would be of significant
benefit to encourage competition in the ratings business, to try to move away from a
model whereby the issuer pays for the ratings, and finally to reduce the impact of CRAs in
financial regulations. Some steps have been taken in these directions, but it remains to be
seen how effective they will be.
10.3 Credit models

There are several alternatives to using ratings for assessing credit risk, and sophisticated
investors may prefer more formal theoretical or statistical modelling of credit risk. There
are two main alternative approaches, structural models and reduced form models. The latter
is particularly useful in providing a mapping between spreads and probability of default.
The structural models derive from option pricing theory, such as the Black–Scholes–
Merton (Black and Scholes, 1973a; Merton, 1974) option pricing model, and consider
risky debt as a combination of safe debt and a short put option on a firm, struck at the
face value of a firm’s debt. Under this approach, a firm’s liabilities are considered to be
183
contingent claims against the firm’s assets, where a default occurs whenever the firm’s
asset value falls below the debt value. A key advantage of this approach is that the prob-
ability of default and recovery rates is directly related to firm-specific characteristics.
However, a disadvantage is that asset values are not directly observable and may be dif-
ficult to estimate.
10.3.1 From credit spreads to probability of default

Credit spreads indicate what the market thinks about the credit quality of a particular
bond compared to a benchmark, typically the bond of a risk-free government. This means
that if we know the credit spread, it is possible to infer the implied probability of default.
The spread, s, is the difference between the yield on risky bond r and the yield of a risk-
free but otherwise identical bond, rf . While differences between risky and risk-free rates
can arise from many factors, such as convexity, tax and liquidity, the main reason is due to
the probability of default. In what follows we assume that the probability of default is the
sole determinant of the credit spread.
Supposing the yield is continuously compounded and quoted on an annual basis. The
realised returns on a risky and risk-free bond are erT and erfT respectively, where T is the
maturity of the bond measured in years. If an investor is risk neutral, she will be indifferent
between owning the risk-free bond and the risky bond, weighted by the risk-neutral prob-
ability of default, pd adjusting for the recovery rate, z:
11 - pd2erT + pdzerT = erfT
We therefore get:
erT11 - pd + pdz2 = erfT
Note here that the pd is the cumulative risk-neutral probability of default: if T = 2, then
pd gives the probability of default in two years, not the conditional probability of default,
which is the probability of default occurring in year 2 given no default in year 1. Since the
credit spread is defined as s = r - rf , we can substitute to get:
1 - pd + pdz = e - sT
Therefore
1 - e - sT
pd =
1 - z
Example 10.3 EU defaults

Consider the yields on European sovereign bonds in December 2011, where German
bonds are risk-free whilst Italian bonds are risky. If we assume a recovery rate of 40%,
the table below shows the implied probabilities of default for Italy. The markets find
that Italy has an almost 10% probability of default every year.
184
10.4 Margins, haircuts and mark-to-market
Yield Probability of default
Maturity German Italian Cumulative Annual
1 0.079% 5.753% 9.19 % 9.19%

2 0.320% 5.956% 17.77% 8.57%
3 0.493% 6.254% 26.45% 8.69%
4 0.751% 6.436% 33.90% 7.45%
5 1.030% 6.665% 40.92% 7.02%

The two entities on either side of a fixed income transaction are known as counterparties,
and generally one of the two owes money to the other. The risk of one counterparty not
getting paid is called counterparty credit risk. In some cases, it is always the same party
owing money to the other, such as the seller of a bond, but in other cases it could be
either, like in a swap contract. Several mechanisms have been developed to address this
counterparty credit risk, for example margins, haircuts, and mark-to-market.
10.4.1 Margins, haircuts and leverage

Margins
Market participants are very often required to post a type of collateral called the initial
margin, and maintain a portion of that, called the maintenance margin, in case the market
goes against them. We sometimes use the terminology to be in red when we owe money
and to be in black when we are owed money. If the posted margin is not sufficient, the
party in red may receive a margin call, requiring funds within a short period of time, per-
haps the end of the business day. Unless the trader provides the funds, the underlying
position might be closed out and the trader declared bankrupt.
Margins do provide significant protection, both as a buffer against possible losses and
also because they incentivise those in red not to default on their positions. Margins are,
however, a double-edged sword. Consider a trader whose position is in red and facing a
margin call. They will have to come up with the funds almost immediately, which may
require them to liquidate some of their positions at short notice, perhaps at firesale prices,
suffering significant losses and perhaps transmitting a crisis across markets.
If only one trader is in difficulty, this is not a serious problem. If, however, the problems
are widespread, such forced sales and difficulties in meeting margins may lead to a vicious
spiral between selling, prices falling and more margin calls, perhaps giving rise to a firesale
externality. Such endogenous risk feedback loops are at the heart of many crisis episodes.
Haircuts
Haircuts are intimately related to initial margins. When securities are pledged as collateral
for a transaction, only a portion of the current market value counts as a pledge, with the
rest termed a haircut. Used this way, the haircut serves the same function as initial margins.
185
The term haircut has other related meanings, such as indicating losses to bondhold-
ers in credit restructuring. For example, when discussing the losses facing the owners of
Greek sovereign debt, the term used is a haircut.
Leverage
Haircuts and margins are directly related to leverage. In particular, the inverse of the haircut
or initial margin shows implied leverage. Suppose a trader enters into a financial transac-
tion, buying a security for $100 million, posting a 25% initial margin and borrowing the rest
from her broker. Therefore, the trader uses $25 million of her own money and $75 million
borrowed from her broker. The trader has therefore leveraged her $25 million three times.
10.4.2 The case of haircuts in the 2007–2008 crisis

We show typical haircuts on a range of fixed income assets in Table 10.3, along with
approximate implied leverage, inverse haircut, over two sample periods, right before and
at the height of the 2008 crisis. The table shows directly how the riskiness of the underly-
ing security is reflected in the haircut. Risk-free US government bonds or investment-grade
bonds have a haircut of almost zero, allowing almost infinite leverage, while high-risk
bonds and risky tranches of CDOs attract a much higher haircut. What is perhaps most
interesting is how rapidly the haircuts increase for the fixed income assets but not for the
equities. The biggest change is observed for the safest assets, the government bonds and
investment-grade bonds.
All classes of securities see a significant increase in haircuts, with haircuts on even sup-
posedly safe US government bonds increasing from 0.25% to 3%, with the various collat-
eralised debt obligation (CDO) tranches losing their collateral value.
Table 10.3 Typical haircut or initial margin for selected securities
Haircut (%) Leverage
Jan–May April Jan–May April
Securities held as collateral 2007 2008 2007 2008
US government bonds 0.25 3 399 32

Investment-grade bonds 0–3 8–12 ∞–32 12–7
High-yield bonds 10–15 25–40 9–6 3–2
Equities 15 20 6 4
Investment-grade CDS 1 5 99 19
Asset-backed CDOs rated:
AAA 2–4 15 49–24 6
AA 4–7 20 24–13 4
A 8–15 30–50 12–6 2–1
BBB 10–20 40–70 9–4 2–0
Mezzanine 50 100 1 0
Data source: Financial Stress and Deleveraging Macrofinancial Implications and Policy, Global Financial Stability
Report, p. 42 (Citigroup and IMF staff estimates 2008), International Monetary Fund
186
10.4.3 Mark-to-market, model or magic

A traditional approach in accounting is based on historical values and cash flows. Cash
flow in is profit and cash flow out is loss. While this may be an appropriate approach
in many situations, for many financial assets, whose value changes continually in finan-
cial markets with maturities that span multiple years, the historical value approach may
lead to valuations that are very far away from the most likely eventual profit or loss. For
this reason, accountants have developed the concept of marking-to-market, using cur-
rent prices to account for the value of assets or liabilities on a firm’s book. This means
that values on balance sheets change along with market conditions, unlike historical cost
accounting which uses the original value. Note that while the initial margin functions like
a haircut, the maintenance margin is more akin to marking-to-market.
Mark-to-model or magic
A requirement for marking-to-market can cause particular problems when assets are illiq-
uid because in that case there may not be a proper market price or, even if a market price
were available, it would change if the firm bought the asset. In these conditions, firms
may have to resort to marking-to-model, by using an internal pricing model to get valua-
tions. This became common practice in the crises from 2007 for many structured credit
products. The obvious problem is that no model is correct, and some are particularly bad.
As a result, it is sometimes said that marking-to-market really means marking-to-magic
because the dependence on untestable assumptions is so strong.
10.4.4 Marking and financial stability

While margins and haircuts generally provide useful protection to counterparties, they do
carry with them a danger to financial stability if used improperly. In practice, they tend to
be pro-cyclical, meaning that if the markets are booming, margins and haircuts are low,
encouraging further borrowing which further inflates asset prices; and if the markets are
heading for a crisis, margins and haircuts may increase, causing difficulty for market par-
ticipants who may be already in trouble. This can lead to a vicious feedback loop between
increasing margins and market distress. One manifestation of this was seen during the
crises from 2007, as seen in Table 10.3.
It is important to recognise how the problem created by marking-to-market arises. Any
investor would prefer to know the mark-to-market values and more transparency about
asset values is generally of a positive benefit. Lack of transparency might be helpful in
hiding risk or even mistakes, but in a crisis investors tend to assume the worst, so a lack
of transparency directly leads to the worst possible expectations. The financial stability
problem created by marking-to-market arises because of how marking is misused by rule-
based systems, whereby more certainty can lead to undesirable results. In other words,
the problem created by marking-to-market is not the fact that we report assets at their
market values, but rather that the reporting of market values mechanically affects other
rules and requirements, for example, forcing banks to cut positions or raise capital in a
pro-cyclical way.
187
10.5 Securitisation
New fixed income assets are often created out of other fixed income assets in a financial
practice called securitisation. Securitisation has long been a common practice in financial
markets, often resulting in useful products, and is not itself problematic. Such views have
changed following the crisis of 2007 because of the role played by securitisation in the
shadow banking system, which undermined financial regulations prior to the crisis.
Some of the clearest examples of this are the failure of Northern Rock in the UK in 2007
and the subprime mortgage crisis in the US.
Securitisation is where various types of credit assets are pooled together and sold in
various forms to creditors. Common examples of credit assets include credit card debt,
car loans and mortgages. Securitisation is often related to what is known as the originate
and distribute model of banking, whereby a bank specialises in lending and then sells off
a portfolio of loans on the secondary market.
Securitisation has long been a feature of financial markets and throughout his-
tory various forms of securitisation have been used. For example, a particular securiti-
sation instrument called a covered bond, or Pfandbriefe in German, was first created in
nineteenth-century Germany and has been in use ever since. A more recent example is
the US government creating a government-sponsored entity called Fannie Mae whose
purpose is to create a liquid secondary market for mortgages.
Example 10.4 Bowie Bonds

Bowie Bonds are bonds backed by the revenues from David Bowie’s 25 albums
recorded before 1990. (Bowie was a 1960s and 1970s rock star.) They were issued in
1997 with a face value of $55 million, paid an interest rate of 7.9% and had an aver-
age life of 10 years.
Mechanics of securitisation
While there are many different ways to design structured products, a common way is to
pool assets that pay cash regularly in a reasonably predictable manner, put them into a
special purpose vehicle (SPV) and sell rights to the cash flow.
Definition 10.1 Special purpose vehicle A special purpose vehicle (SPV), also
known as a special purpose entity, is a legal entity, usually a limited company or limited
partnership, set up for specific, narrow and often temporary purposes. They might, for
example, be used to isolate a firm from risk or to hide ownership. SPVs can be very useful, but
can also be used for more nefarious reasons, such as Enron’s use of them to commit fraud.
Definition 10.2 Sponsor The financial institution that sets up the SPV is known as a
sponsor. It chooses the assets and engages a manager, and may provides guarantees to it. It
is the owner of the SPV and often retains an equity position.
188
10.5 Securitisation
Fees
Rating
agencies
Interest &
Borrow Sell to Sells to principal
Mortgage Investment Special

Homeowners purpose (SPV) Investors
originators bank
vehicle
Credit line Price
Insurance
companies
Premiums
Figure 10.3 Securitisation chain
In securitisation, the sponsor sets up an SPV which often buys assets already held
by the sponsor, perhaps financing the purchase by issuing bonds to outside investors.
Typically, the value of the bonds is lower than the value of the assets, with the difference
being equity contributed by the sponsor. The bonds are known as asset backed securities
(ABSs) since the bonds are backed by the underlying assets. Additional protection may be
provided by equity and junior tranches.
Figure 10.3 shows a typical securitisation process, the players involved and the flow of
funds. The process starts by a homeowner borrowing from a mortgage originator. After the
mortgage originator has accumulated enough mortgages, it sells them to an investment
bank that intends to package them into a structured credit product.
The investment bank creates an SPV out of the mortgages it bought, and sells the rights
to the payment flow to investors. The bank then pays fees to CRAs to obtain ratings to
reassure the investors, and may pay premiums to insurance companies to protect some
parts of the SPV.
10.5.1 Advantages and disadvantages

Advantages
The main advantage of securitisation is that by setting up a separate company, the origi-
nator does not have to count the SPV’s debt amongst its liabilities, nor does it have to
count the SPV’s assets as its assets. If the originator is regulated it may not have to keep
capital against the assets, depending on regulations.
Securitisation allows the originating bank to remove business loans, credit card loans,
mortgages and other assets from its balance sheet, whilst retaining control of the assets.
It also allows for the transformation of illiquid assets into liquid securities. Rather than
holding the asset on a balance sheet financed with liquid deposits, securitisation trans-
forms the asset itself from an illiquid one (perhaps a pool of loans) into liquid securities
189
issued by the SPV. Furthermore, by creating a separate SPV, the firm isolates the cash flow
generating assets and collateral so that securities issued by the SPV are not a general claim
against the company, just against those assets.
The obvious reason to securitise assets is to transfer risk. In the originate and distribute
model of banking, securitisation frees up capital and funding, allowing the firm to lend
again. Securitisation allows banks to provide more and cheaper credit than they other-
wise could, lowering the total cost of financing loans.
Investors also benefit in that securitisation allows them to invest in otherwise inac-
cessible assets, which is desirable for diversification purposes. For example, if you want
to have some exposure to Japanese retail credit risk, without securitisation you would
need a presence in Japan, advertising, regulatory approvals and so on. But if a Japanese
bank securitises its credit card portfolio, then a Western investor can get access to this
class of investment without the costs associated with opening a credit company in
Japan.
Securitisation also allows for increasing specialisation in the financial system. One can
separate what used to be one process in traditional banking into multiple different steps.
For example, the origination of loans might be done by those specialising in local markets
but not rich in capital. Borrowers get better terms and access to lending than they would
get otherwise.
These loans are then sold on to investment banks with specialised skills in packaging
loans and the necessary capital to buy them off the originators. Finally, the packaged
loans can be split up into different risk categories, so that the risk can be precisely tar-
geted at those who most desire it and will pay the most to get it.
Drawbacks
While the seller of portfolios of mortgages or credit card receivables might benefit
from getting rid of the risk, the buyers of the loans face a ‘lemons’ problem2 because
the originators are likely to be better informed about the loans than the sellers and,
therefore, may be tempted to securitise the lowest-quality loans and hold on to the
better loans.
This directly relates to moral hazard because if a seller is not intending to hold the loan
for more than a few months, it may care less about credit quality than if it held the loan
until maturity. This was exactly the key problem in the market for subprime mortgages
prior to the crises from 2007: the lenders did not care much about mortgage quality, as
noted for example by Bitner (2008).
Securitisation also carries with it risk for the sellers. In the case of Northern Rock, the
bank was dependent on securitisation, and when it encountered difficulties in placing its
mortgage portfolios on the secondary markets, it collapsed.
2
The term lemon problem relates to a famous paper by Akerlof (1970) on asymmetric information. In Ameri-
can English a lemon is a dodgy used car.
190
References
10.6 Summary
The credit markets are fundamental to the functioning of modern economies, channel-
ling savings to productive investments. In order to lend money, investors like to know the
credit risk of their counterparties. The most common way to do that is by using ratings
from CRAs. These are controversial, with many critics claiming that ratings quality is low.
An alternative way to access credit quality is by using mathematical models, such as
structural models and reduced form models. The latter are especially useful in allowing
for a conversion from spreads to probabilities of default.
Fixed income assets are often used as inputs into new assets, generally called struc-
tured credit. This is a long-established practice in financial markets, providing useful ben-
efits to market participants, but can also be abused. Such products were at the heart of
the crisis from 2007.
Margins and haircuts provide useful information to investors, but can increase financial
instability when used in a mechanistic fashion.
Securitisation, whereby different fixed income instruments are packaged together and
sold as one, is an old financial technique, dating back to the nineteenth century. When
used properly, it can increase the efficiency of financial markets, but when misused may
cause moral hazard and even crises.

1 What are OTC assets and markets?
2 Why are interest rates and bond prices inversely related?
3 What are the main components of credit risk?
4 What is a rating agency?
5 What are the main limitations of ratings?
6 Why is the EU so cross with the rating agencies? Is this warranted?
7 Suppose the markets view a one-year bond issued by the US government at 1% to be

risk-free. If Argentina issues a one-year USD bond at 10%, what is the implied probabil-
ity of default?
8 Explain the relationship between initial margins/haircuts and leverage.
9 What does it mean when we say mark to magic?
10 What is the reason why margins may increase financial instability?
References
Akerlof, G. (1970). The market for ‘lemons’: quality uncertainty and the market mechanism.
Quart. J. Econ., 84(3): 488–500.
Bitner, R. (2008). Confessions of a Subprime Lender: An Insider’s Tale of Greed, Fraud, and
Ignorance. John Wiley & Sons.
191
Black, F. and Scholes, M. (1973a). The pricing of options and corporate liabilities. J. Polit. Econ.,
81(3): 637–54.
Fabozzi, F. J. (2009). Bond Markets, Analysis, and Strategies, 7th edition. Pearson.
Hull, J. C. (2011). Options, Futures, and other Derivatives, 8th edition. Prentice Hall.
Hull, J. C. (2012). Risk Management and Financial Institutions, 3rd edition. Prentice Hall.
International Monetary Fund (2008). Global financial stability report: containing systemic risks
and restoring financial soundness. April 2008, pp. 1–211.
Merton, R. C. (1974). On the pricing of corporate debt: the risk structure of interest rates.
J. Finance, 29: 449–70.
Murphy, D. (2009). Unravelling the Credit Crunch. CRC Press.
Washington Post (2004). Credit raters’ power leads to abuses, some borrowers say. Washington
Post, 24 November. www.washingtonpost.com/wp-dyn/articles/A8032-2004Nov23.html.
192
11 Currency markets
No financial market sees more government intervention than the foreign exchange (FX)
market. After all, every economy is sensitive to the exchange rate and policymakers
often want exchange rates to reflect the underlying economic fundamentals, because
misalignments and excessive exchange rate volatilities imply significant economic costs.
Every government needs to decide on what type of exchange rate regime it wants,
ranging from letting its currency float freely, to various ways it can tie the currency to
that of another country, all the way to joining a currency union. However, determin-
ing an exchange rate regime is not a trivial proposition. Leave it to the market, and the
result may be FX volatility. Leave it to the government, and the outcome may be mis-
alignment and currency crisis. Choosing a sensible exchange rate regime is hard, and
governments are sometimes guided by wishful thinking and blind pursuit of outdated
goals, perhaps only wanting to replace the existing arrangement.
A key reason why governments find it so hard to manage their exchange rates is
because of the size of the FX market, as it has more trading volume than any other
financial market – see Figure 11.1.
Volume in the FX market has increased rapidly, from $1.5 trillion per day in 1998
to $4 trillion in 2010, or around one quadrillion dollars per year. The most actively
traded currency is USD, involved in 42% of all trades. This is not surprising given the
reserve currency status of the dollar. Because the dollar is so liquid, most currency
pairs do not trade bilaterally; for example, going from the Malaysian ringgit to the
Hungarian forint would probably involve a detour to the dollar. The dollar is followed
by the euro, the yen and sterling. Surprisingly, even though China is the second larg-
est economic area in the world, its currency constitutes only 0.15% of trading. This is
Chapter 11 Currency markets
4
2007
3
2010
2
1
0
Total Spot Forwards Swaps Options ETD
(a) Daily volume by instrument, USD trillions. ETD = exchange-traded derivatives
HKD
CAD
SRF
AUS
GBP
Yen
Euro
USD
0% 10% 20% 30% 40%

(b) Relative trade volume by currency in 2010
40% 1998
30% 2010
20%
10%
0%
UK US Japan Singapore Switz. Hong Kong
(c) Relative volume of trade by country where trade takes place
Figure 11.1 Bank for International Settlements (BIS) survey of the FX market
Data source: BIS triannual survey of the FX market: www.bis.org/publ/rpfx10.htm
because the Chinese government actively discourages currency trading and employs
capital controls.
At least 92% of FX trading involves the currencies of large developed economies,1 with
the rest mostly smaller developed economies or larger stable developing economies.
The FX market has no single physical location, and instead takes place on computer sys-
tems. However, the individual traders and banks do have a physical location and the main
geographic centre of FX trading is London, making up 37% of the global total trading vol-
ume, a number that has been steadily increasing; in 1995 it was 29%. This is not that sur-
prising, since trading is concentrated in the largest financial centres and large volumes in
one geographical area attract traders, further increasing volume. There are many reasons
for this, for example, the supply of specialised human resources and technical support.
1
In the order of volume: USD, euro, Japanese yen, pound sterling, Australian dollar, Swiss franc, Canadian
dollar, Hong Kong dollar, Swedish krona, New Zealand dollar, Norwegian krone and Danish krone.
194
11.1 Fixed or floating
The daily FX turnover dwarfs the underlying economic activities. The International
Monetary Fund (IMF) found the world’s GDP in 2010 to be $63 trillion and the World
Trade Organization (WTO) global trade to be $18.5 trillion. In other words, the global
trade in goods and services is only around 1.85% of the global trade in FX. It follows
that the bulk of the activity in FX markets is due to speculation, the buying or selling
of currencies solely to profit from anticipated changes in exchange rates, by means
of high frequency trading (HFT), exploiting arbitrage between currencies (triangular
arbitrage) and prices on different trading platforms.

This chapter directly relates to Chapter 2 (the Great Depression, 1929–1933), Chapter 9
(trading and speculation) and Chapter 12 (currency crisis models).
Key concepts
■ Currency regimes
■ Capital controls
■ Bretton Woods
■ Gold standard
■ Reserve currency
■ Sterilisation

There are many excellent texts on currency markets, and in writing this chapter we
have drawn on McCallum (1996), Sarno and Taylor (2003), Krugman and Obstfeld
(2006) and Marsh (2010). Eichengreen (1996), Ferguson (2008) and Ahamed (2009)
contain some of the historical discussion.
B Net domestic currency bonds

Et - 1 ∆e t Expectation of change in exchange rate from period to t - 1 to t
et Nominal exchange rate at time t
it Interest rate at time t
M Domestic nominal money supply
P Price level
Q Index of the real value of expenditures
R Government foreign exchange reserves
V Velocity of money
* A star indicates a foreign variable
11.1 Fixed or floating

One of the most important decisions made by governments is what sort of currency
regime it chooses to have. On one extreme, it could fix the exchange rate to the currency
of another country or, alternatively, it could let it float freely on the open market.
195
Arguments in favour of fixed rates and interventions

Many governments opt to fix the exchange rate at a value the central bank finds best. This
is based on what is known as the ‘wrong-rate argument’. There is a ‘correct’ exchange rate,
fully reflecting the economic fundamentals. The market, however, does not arrive at it.
The reason floating exchange rates are the ‘wrong rates’ is that the market is inefficient:
it does not make use of available information and it is prone to destabilising speculation.
Markets may attach too high a probability on a devaluation or appreciation, not usually
justified by economic fundamentals. In some cases, speculators may deliberately manipu-
late the exchange rate away from the fundamentals, perhaps so they can profit from the
resulting exchange rate volatility.
One of the best manifestations of the inefficiency of market-determined exchange
rates is very high exchange rate volatility, inconsistent with slow-moving economic fun-
damentals. Many of the most important fundamentals are measured quite infrequently,
for example, GDP is known only quarterly, hence, there should be no reason for exchange
rates to fluctuate as much as they do.
Such FX volatility might happen because speculation is subject to a bandwagon effect:
it feeds on itself rather than fundamentals, perhaps because traders use simple forecast-
ing methods, like momentum trading, or other simple extrapolative techniques or subjec-
tive pattern recognition.
Since there are several reasons why the actual exchange rate may be driven away from
the rate implied by fundamentals, official intervention may be a useful instrument to sta-
bilise the exchange rate at or close to a ‘correct’ rate.
Countries with fixed exchange rates expect significant benefits because transaction
costs and risk are minimised, encouraging trade and investment. However, in order to
obtain those benefits, the exchange rate needs to be sensible and backed up by a credible
monetary policy. If a government chooses the exchange rate incorrectly, or undermines it
by unwise monetary policy, the end result would be a costly realignment. Fundamentally,
a fixed exchange rate regime depends on the ability of the experts who set it, and their
independence from political interference.
Arguments in favour of floating rates

Opponents of official interventions in FX markets might argue that it is far from clear that
the authorities are capable of identifying the correct exchange rate in practice. Even if they
do have the ability, when trying to do so, they might arrive at the wrong exchange rates
for political reasons, perhaps preferring an overvalued currency to make voters feel artifi-
cially wealthy, or an undervalued currency to help industry.
The costs of an incorrect exchange rate might be significant, with interventions to
maintain the official exchange rate distortionary. The exchange rate is likely to affect the
expectations of speculators, who may then trade in a way that undermines the inter-
ventions so that the government might only be able to maintain the official rate at very
high costs.
If the exchange rates deviate from the presumed fundamental rate, it can be hard to
identify whether this is a permanent or temporary phenomenon; consequently, it is dif-
ficult for the authorities to identify when an intervention is justified.
196
11.2 Foreign exchange interventions
Furthermore, floating exchange rates might not be any more unstable than fixed
exchange rates; one is more likely to see more gradual adjustment in floating rate regimes
rather than the sudden and sharp jumps resulting from devaluations in fixed regimes.
Floating regimes might have higher volatility but fixed regimes bigger jumps.
Perhaps the main argument in favour of a flexible exchange rate is that it allows for an
independent monetary policy, enabling the government to respond to economic con-
cerns more easily. As a consequence, floating rate regimes might be less crisis prone than
fixed rate regimes. Furthermore, since differences in national inflation rates and price lev-
els are more likely to be directly reflected in the exchange rate, floating rate countries are
less likely to have an under- or overvalued currency for longer periods.
Conclusion
Both sets of arguments have compelling elements. Over time, countries have alternated
from one arrangement to another, usually not satisfied with their choice. The ultimate
conclusion seems to be that no exchange rate regime is perfect; governments are quite
willing to experiment with a new arrangement, as the large number of regimes discussed
in the appendix bear witness to.
Most, if not all, governments, however, take the view that at least some interventions
in the FX market are necessary. Whilst academic economists are often vocal in their oppo-
sition to FX interventions, those actually in charge disagree.

Governments intervene more in the market for FX than in that of any other market, and
every country has to implement some FX policy. When a central bank decides to fix its
exchange rate, it has two choices: either it must implement capital controls or it must
always be willing and able to trade with market participants at the target rate.
Mechanics of interventions
Denote the exchange rate by e, which is the amount of domestic currency per unit of
foreign currency, and the interest rate by i. Denote foreign variables by *. It is common to
assume the foreign country is the United States, since it has the reserve currency. Suppose
a central bank decides to fix its exchange rate (peg) at level e. The FX market is in equi-
librium when the uncovered interest rate parity (UIP) condition holds, meaning that the
domestic interest rate equals the foreign interest rate plus the expected movement in the
exchange rate:
it = it* + Et - 1 ∆et
where ∆et = et - et - 1 and Et - 1 refers to expectation at time t - 1. When the exchange

rate is fixed at e and market participants expect it to remain fixed, then
Et - 1 ∆et = 0
Given that investors are not expecting any changes in the exchange rate because of
the peg, they are only willing to hold the foreign and domestic currency if they offer the
197
same interest rate. To ensure equilibrium in the FX market, the central bank must set
it = it*.
That imposes a particular restriction on the central bank. When it intervenes by print-
ing money, it would normally affect interest rates, but they must not change. Hence, the
bank needs to adjust the money supply in order to keep the interest rates constant. In
other words, it sterilises the intervention. This eliminates any independence or discretion
that the central bank has over the money supply and therefore over monetary policy,
leading to the ‘impossible trinity’ discussed in Section 11.4.3.
11.2.1 Sterilisation
When a central bank intervenes in the currency markets to prevent an appreciation of
the currency, the money supply usually increases; after all, the central bank is buying
foreign currency with freshly printed domestic money. This is, of course, inflationary, and
since the central bank may prefer to prevent the inflation, it often couples interventions
with sterilisation. Sterilisation means that the domestic money supply is held fixed, while
an unsterilised intervention is where the central bank intervenes without offsetting the
domestic money supply.
Mechanics of sterilisation
Consider the balance sheet of the central bank in Table 11.1, where it purchases $10 billion.
Step 1. The central bank purchases USD 10 billion and invests that amount in US Treas-
uries, so B* = $10 billion.
Step 2. Its liabilities (monetary base) increase by the same amount: M c by e * $10
billion.
Step 3. The central bank now sells domestic bonds by the same amount and, there-
fore, takes money out of circulation: M Te * $10 billion and B c e * $10 billion.
Step 4. After the intervention the net assets of the bank have not changed, but:
■ Its holdings of foreign currency bonds have increased by e * $10 billion.

■ Its holdings of domestic bonds have decreased by the same amount.
■ Overall the monetary base has not changed.
The changes in the holdings of foreign assets are offset one-for-one by the changes in
domestic asset holdings, and the purchase of foreign assets has no effect on the domestic
money supply. However, the exchange rate would probably fall because of the purchase
of foreign currency.
Table 11.1 Central bank balance sheet

Assets Liabilities
Net foreign currency bonds (B*t ) Monetary base (Mt)

Net domestic currency bonds (Bt) Net worth
198
Sterilisation in developed markets

Sterilisation might not be a very effective tool in advanced countries where assets are
almost perfectly substitutable. The reason is that a sterilised intervention might lead to
an increase in the interest rate because it implies the selling of domestic bonds. In turn,
that might make domestic assets more attractive to outside investors, causing an inflow
of currency and appreciation of the exchange rate. Ultimately, this would undermine the
sterilisation.
The developed economy that has been the most active user of interventions and steri-
lisations – Japan – has faced exactly this problem. Currency interventions might lead to a
temporary fall in the exchange rate but the yen soon appreciates again. As a consequence,
Japan has very sizable holdings of US government bonds, $956.8 billion in September
2011, according to the US Treasury.
Sterilisation in emerging markets

The situation is different in emerging markets, whose assets are not as substitutable for
foreign assets. The reason may be that the government employs capital controls, or the
markets lack transparency and liquidity, deterring foreign investors. In this case, a non-
sterilised intervention would lead to an increase in the monetary base, resulting in infla-
tionary pressures and even asset price bubbles. As a consequence, sterilisation is likely to
be more effective in preventing inflation than in developed economies.
Uses of sterilisation
The use of intervention to keep exchange rates low has become increasingly common in
recent years. Lavigne (2008) estimates that from the $1.3 trillion in accumulated reserves
between 2000 and 2006 in emerging markets in Asia, 75% have been sterilised. He further
finds that China has been the biggest user of sterilisation, sterilising 80% of its interven-
tion. This is in addition to using extensive capital controls, administrative controls and
regulations, and state ownership of large parts of the banking system as means to manage
the exchange rate.
Limitations
In principle, a government can intervene indefinitely to prevent appreciations; however,
its ability to sterilise the interventions depends on the central bank’s holdings of govern-
ment bonds. This is not a major problem because the government, as the owner of the
central bank, of course, can just issue new debt. This does not mean that sterilisation is
always a recommended policy when intervening.
Sterilisation is costly because the central bank acquires foreign bonds and sells domes-
tic bonds. Since the yield on foreign bonds is likely to be much lower than the yield on
domestic bonds, the difference is a cost to the central bank.
Sterilisation involves encouraging banks to accumulate government bonds and so
might be considered as a tax on banks. The banks are holding government bonds instead
of lending to private investors, potentially holding back economic growth. Finally, as the
central bank continues to intervene and sterilise, sovereign debt keeps on accumulating,
affecting the solvency of the state.
199
11.3 Capital controls

Governments wishing to control the flow of foreign currency may resort to capital controls.
This can take the form of taxes on transactions or outright restrictions on the buying/sell-
ing of a foreign currency in order to maintain fixed exchange rates. We term the latter case
strict or traditional capital controls, and the former capital controls 2.0.
11.3.1 Traditional capital controls

Traditional capital controls (1.0) relate to policies aimed at controlling inflows and out-
flows of currencies for the purpose of maintaining fixed exchange rates or the balance of
payments. While these controls vary in severity, they often involve a licensing regime.
Example 11.1 United States

During the 1950s, the US was battling a balance of payments deficit. As a response,
the government imposed a surcharge on the purchase of foreign stocks and bonds,
excluding Canada and the developing world.
Investors were able to evade the tax by trading via Canada and using bank loans
instead of securities. The government moved to close the loopholes, but eventually
faced the controls out after leaving the Bretton Woods system.
Developed countries’ move away from capital controls

While capital controls were widespread in the early years of the Bretton Woods system,
they were always controversial and became to be seen as increasingly damaging. They
were always quite leaky and encouraged corruption. As a consequence, most developed
countries abolished them early, see Table 11.2.
The move away from capital controls was connected to the trend towards free global
flows of capital. Amongst the developed economies, Germany was the biggest historical
Table 11.2 Abolition of capital controls in OECD countries

Country Year Country Year Country Year
Australia 1978 Greece 1980 Portugal 1992

Austria 1980 Iceland 1993 Spain 1985
Belgium Italy 1984 Sweden 1986
Canada Japan 1979 Switzerland 1979
Denmark 1983 Luxembourg Turkey 1985
Finland 1991 Netherlands United Kingdom 1971
France 1986 New Zealand United States 1974
Germany 1980 Norway 1989
Note. Belgium, Canada, Luxembourg, the Netherlands and New Zealand did not have capital controls
in the OECD’s time
Source: OECD (2002), Forty Years’ Experience with the OECD Code of Liberalisation of Capital Movements, OECD
Publishing. http://dx.doi.org/10.1787/9789264176188-en
200
11.3 Capital controls
proponent of free capital flows, with most major economies, such as that of the US, opposed.
Abdelal (2007) argues that Germany was in favour of free capital flows not least because
they saw it as a fundamental component in the establishment of a common European cur-
rency. When the French governments of the 1980s came to the German view, support for
free capital flows globally was pushed to the forefront of the Washington consensus agenda.
Capital controls in developing countries

Developing countries have been frequent users of capital controls, either as a part of their
permanent arrangement for managing currencies or as a temporary reaction to a crisis.
In recent times, one of the best examples of the latter is Malaysia during the Asian crisis.
Example 11.2 Malaysia

With the onset of the Asian crisis in 1997, in order to protect its currency and to avoid
asking the IMF for help, Malaysia banned transfers between domestic and foreign
accounts and between foreign accounts, eliminated credit facilities to offshore par-
ties and prevented repatriation of investment until September 1999, whilst fixing
the exchange rate to the dollar. Foreign exchange transactions were allowed only at
authorised institutions. In February 1999, a system of taxes replaced prohibition on
repatriation of capital. The net effect was to discourage short-term capital flows but
permit longer-term transactions. The aim of this policy was to enable the government
to lower interest rates without fearing devaluation.
While many developing countries have permanent capital controls in place, the best
known and most controversial use of capital controls is in China.
Example 11.3 China

China tightly controls its exchange rates by means of capital controls. The capital
account is closed for inflows and outflows, except for foreign direct investment.
External debt is strictly controlled. However, more recently, China has been relaxing
the controls. Portfolio investment has been slowly liberalised, with foreign investors
allowed to invest in special categories of shares and bonds.
Re-emergence of strict controls

With the crisis starting in 2007, traditional capital controls have re-emerged as a crisis-
fighting tool amongst developed economies. The most prominent example of this is
Iceland: see Arnason and Danielsson (2011) for more details.
Example 11.4 Iceland and the IMF

The Icelandic capital controls were imposed following the collapse of the coun-
try’s economy in 2008. The IMF made capital controls a requirement for provid-
ing aid.
➨
201
This was a U-turn in Fund policy, because up until that point the IMF had been
opposed to capital controls, at least from the 1970s. This is also the first case of a
country subject to European Union (EU) rules resorting to capital controls. While
Iceland is not an EU member, it is in the European Economic Area (EEA). In order to
comply with both EEA and IMF requirements, Iceland had to impose traditional strict
capital controls, requiring Central Bank authorisation for all purchases of foreign cur-
rency, and in practice preventing foreign direct investment.
11.3.2 ‘Hot money’ and capital controls 2.0

A different reason for capital controls is the danger posed by ‘hot money’. While the term
was first used by President Roosevelt in 1936, it has become more relevant in recent times
with free global capital flows. Hot money refers to the type of flows that come quickly into
a country, often as part of carry trades, and can leave just as quickly. It is not foreign direct
investment but rather portfolio investment.
Such money flows are controversial because speculators can move significant amounts of
funds into a small country, causing the exchange rate to appreciate sharply, resulting in sig-
nificant difficulties for exporters. At the first whiff of problems, the money then leaves, caus-
ing what is called a sudden stop, and the exchange rate collapses, perhaps ending in a crisis.
To prevent this problem, some authorities have resorted to a special type of capital
controls designed to discourage hot money. They can take the form of a special tax on
inflows that leave within a certain amount of time, or even on particular types of invest-
ments. Since controls on hot money are quite different from the traditional controls dis-
cussed above, we might label them capital controls 2.0.
Example 11.5 Chile

Chile experienced heavy capital inflows in the 1990s, causing both a build-up of short-
term debt and fears of a real currency appreciation. The government announced in
1991 that direct investment was to become subject to a three-year stay, which was
reduced to one year in 1993. Portfolio flows were restricted to mandatory non-interest
paying deposits at the Central Bank with early withdrawal penalties. The effect was to
reduce the return to the investor and, hence, discourage inflows.
The effectiveness of the controls was controversial. Some commentators argued
that they prevented a real appreciation of the peso and changed the composition of
the inflows but not the magnitude of the flows. Others found that they did change
the magnitude of the inflows, and had real adverse impacts.
Since the crisis starting in 2007, several countries have employed various forms of 2.0
capital controls. For example, Brazil introduced a tax on financial transactions in October
2010. It is a tax of 6% on non-resident equity and debt inflows and also requires a 6%
margin payment to trade derivatives. Thailand also imposed controls in October 2010,
in the form of a withholding tax on non-resident interest income and capital gains from
state bonds. South Korea has restricted the size of the FX derivative books of its banks as a
202
11.4 Exchange rate regimes
macro-prudential measure, and imposed a 14% withholding tax and a 20% capital gains
tax on foreign purchases of government bonds.
11.3.3 Pros and cons of capital controls

Capital controls are quite controversial and have been rejected by policymakers in devel-
oped countries. The abolition of capital controls is central to the Washington consensus:
see Definition 6.1 in Chapter 6.
Arguments against capital controls

The main argument against capital controls is that they prevent investments and savings from
being used in the most productive manner. Economic agents will attempt to evade the con-
trols, finding loopholes, with a cat-and-mouse game ensuing between the authorities and
market participants, with ever tightening of the controls leading to yet more evasion. Keeping
capital controls watertight is an impossible task. This often leads to corruption whereby
exports are under-invoiced and imports over-invoiced, enabling funds to remain abroad.
Licensing regimes create the potential for corruption of government officials. Authorities might
have to choose between either making them much stricter or abolishing them altogether.
Arguments in favour of capital controls

Those in favour of capital controls point out that they can reduce the risk of crisis and the asso-
ciated costs. As a consequence, they are a useful macro-prudential tool. Reinhart and Rogoff
(2009) suggest that they were to thank for the low incidence of banking crises during the
Bretton Woods era. Uncontrolled inflows of hot money followed by a sudden stop are quite
damaging to countries, first creating a boom whilst undermining exporters and then terminat-
ing in a crash. As a consequence, countries often prefer to restrict flows of hot money.
Targeted 2.0 capital controls can be successful in preventing excessive capital inflows,
at small cost. They are designed to be genuinely countercyclical, aimed at preventing
excesses when times are good.
A rethink on capital controls

Capital controls have had a bad reputation for a few decades, with the opposition to
them led by the IMF and some European countries. Those introducing capital controls in
recent decades, such as Malaysia, came under strong criticism.
This has now changed. The IMF has in recent years expressed support for capital con-
trols. They were an enthusiastic supporter of the Icelandic capital controls from November
2008, and have continued to support their use. The current support for capital controls
usually refers to the 2.0 variety, aiming to use them as a macro-prudential tool to prevent
hot money flows, rather than a means to maintain an unsustainable exchange rate.

A number of exchange rate regimes have been proposed and implemented. Generally,
the choices are between either fixing the exchange rate to some value or letting it float
freely. Alternatively, countries can adopt the currency of another country or even join
203
a currency union. In between those choices are the various alternatives discussed in the
appendix at the end of this chapter.
11.4.1 The gold standard

The gold standard is the most important historical example of a fixed exchange rate
regime. Countries fixed the value of their currency to the price of gold and committed
to its free convertibility to and from gold. The gold standard developed gradually, with
more and more countries adopting it in the nineteenth century, led by the world’s most
important economy at the time, that of the United Kingdom.
Under the gold standard, gold is money, so the supply of gold restricts the growth in
the money supply, thus preventing inflation and ensuring long-run price stability, unless
there are massive gold discoveries. The result was long-term stability and predictability
in the international monetary system never before or since accomplished, with a positive
effect on trade and economic development.
The mechanics of the gold standard

The mechanics of the gold standard were described by David Hume (1752), who may
have created the first monetary model in economics. In his basic model, only gold coins
circulate and no banks exist, prices are flexible and the supply of gold is fixed. This gives
the quantity theory of money:
M * V = P * Q
where M is the total amount of money in circulation, V the velocity of money, P the price
level and Q an index of the real value of expenditures. Whenever goods are exported,
payment is received in gold, which can be converted into domestic currency. To pay for
imports, the importer exchanges domestic currency for gold. If a country has a trade defi-
cit, it is losing gold and, hence, prices must fall, assuming V and Q are constant. This
encourages exports and discourages imports, helping to eliminate the trade deficit. The
system self-corrects and speculation is stabilising. See Figure 11.2 for how this works.
Why the gold standard survived

The absolute credibility of the commitment of the central bank to maintain gold con-
vertibility was the main reason why the gold standard system was stable and flourish-
ing between 1870 and 1914. But the system also benefited from other unique factors.
The UK’s leading economic position allowed sterling to anchor the international system.
The openness of markets and surge in trade supported the operation of the gold stand-
ard adjustment mechanisms. On the political side, the monetary authorities enjoyed an
extraordinary political insulation, partly because of the lack of universal suffrage, allowing
them to commit themselves to the maintenance of an unquestionable gold convertibility.
Problem of deflation
The main downside of the gold standard is deflation. The world money supply of gold is tied
to the production of gold, and when gold discoveries are rare, the normal situation, the world
supply of gold will not increase as fast as global production and real world income, especially
204
Trade deficit
Net gold outflows (M )
Pimports Pdomestic
imports exports
Trade imbalance
eliminated
Figure 11.2 Gold standard adjustment mechanism
in times of a booming world economy, having a strong deflationary effect. Accordingly, British
price levels fell by 18% between 1873 and 1879 and by a further 19% by 1896.
Deflation is more costly than inflation for several reasons, the most important of which
is that since interest rates cannot fall below zero, those owning money benefit from delay-
ing purchases because they know their money will be more valuable in the future.
Deflation affects agricultural producers badly. For example, a farmer with a fixed mort-
gage receives less and less money for his products and has to pay more interest in real
terms. The main agricultural exporters, like the US, Australia, Canada and Argentina, all
suffered from deflation, and were quite crisis prone as a consequence. By contrast, capital
exporters like the UK did well out of the deflation and the stability of the gold standard.
The benefits for the capital exporters were not uniformly distributed. The landed
classes who supplied the capital profited, but those working, and especially the working
class, lost out. Voting rights were restricted to property owners, so governments remained
firmly in favour of the gold standard.
Several factors contributed to the increasing instability of this system and led to its
breakdown after its reintroduction following the first World War (WWI). The introduction
of universal suffrage and the emergence of political parties representing the working class
made domestic considerations more important than maintaining convertibility.
11.4.2 Bretton Woods

The time period between the two world wars was marked by extreme uncertainty in the
world’s financial markets, culminating in the Great Depression. As a consequence, the
Allied leaders in the early 1940s decided to create a new world financial order, more sta-
ble and less prone to destabilising activities by the markets.
These efforts led to what is known as the Bretton Woods system, named after the hotel
in New Hampshire in the US where the final agreement was signed in 1944. The par-
ticipants wanted stable exchange rates but realised they could not go back on the gold
standard. Instead they replaced it with the US dollar, which became the reserve currency.
205
Member countries were to adopt an adjustable peg system, with capital controls,
fixing their currencies to the dollar. The dollar, in turn, was pegged to gold at a value of
$35 per ounce. Par values could be changed to correct a ‘fundamental equilibrium’, which
was to be decided by the newly created International Monetary Fund (IMF), which was
also responsible for monitoring national economic policies and could extend balance of
payment financing to countries facing problems.
The capital controls were somewhat effective, especially early on, but required extensive
regulation. When current account convertibility was reintroduced in 1959 it became easy to
avoid the capital controls.2 The adjustable peg meant, in theory, that trade deficits could be
eliminated without experiencing painful deflations. In reality, parity changes were rare as they
eroded the credibility of a central bank, and were often accompanied by a political crisis.
Why Bretton Woods failed

There are three main reasons why the Bretton Woods system failed:
1 Domestic priorities had become increasingly important and led to policies that were
inconsistent with the exchange rate system.
2 The main mechanisms were not really working. Capital controls were envisaged to
allow governments to follow their own monetary policies but were difficult to enforce
since it was easy to over-invoice and under-invoice trade.
3 The final problem area was the so-called Triffin dilemma, named after the economist
Robert Triffin, who observed that the tendency for the Bretton Woods system to meet
excess demand for reserves through the growth of foreign dollar balances made it
dynamically unstable. Accumulating dollars as reserves was attractive only as long as
there was no question about their convertibility into gold. As foreign dollar balances
grew relatively large compared to the gold reserves of the US, doubts started to grow
about its commitment to convertibility. If all countries sought to convert their reserves
into gold, the US would be forced to abandon the dollar peg to gold.
This happened in the end, with a run on the dollar, because of inadequate mone-
tary and fiscal discipline in the US. Faced with high expenditure for the war in Vietnam
and an expansion of welfare programmes, the US increased spending, disregarding
the consequences on the Bretton Woods system. In the spring of 1971, flows from the
dollar to the German mark surged and Germany finally allowed the mark to float
upwards. Other countries followed. As the flight from the dollar grew stronger, the
Nixon administration finally suspended convertibility, bringing about the end of the
Bretton Woods system.
Should we adopt Bretton Woods now?

The ongoing crisis has prompted some commentators to argue that the more ordered
Bretton Woods financial system would be preferable to the current arrangement. The
reason is that excessive activity in financial markets, speculation and hot money flows
2
Current and capital account convertibility refers to the freedom to convert the local assets into foreign
assets and vice versa at an exchange rate determined by the market.
206
are all destabilising, and direct government intervention is needed to bring the system
under control.
Such views often seem to be based on a reinterpretation of what Bretton Woods was
rather than a call for the adoption of the particular arrangement. The Bretton Woods
system failed for particular reasons. It is not possible to maintain a fixed exchange rate
regime in a world where countries have widely differing prospects and economic policies.
Even in the relatively small world of Europe, it is proving to be quite challenging. It would
be necessary to reimpose strict capital controls, and even in that case, countries would be
subject to damaging periodic realignments.
Consequently, returning to a Bretton Woods-type arrangement for the world’s financial
system would not be a sensible response to the ongoing crisis.
11.4.3 Impossible trinity

The choice of currency regimes directly affects a government’s ability to pursue other
policies. This is expressed by the impossible trinity, which states that it is impossible for a
country to pursue the following three goals simultaneously:
1 Fixed exchange rate

2 Free capital movements
3 Independent monetary policy.
At best, a country can achieve two out of the three. Suppose we start with a country in
equilibrium and with fixed exchange rates. If the country then embarks on an expansion-
ary monetary policy, or simply loses control of inflation, the money supply is increasing.
In this case, speculators can borrow the country’s currency and exchange it for foreign
money. Such a trade is likely to attract a large number of market participants. In order for
the government to maintain the exchange rate, it can either continue selling its foreign
currency reserves until they run out, at which time the exchange rate peg will collapse, or
impose capital controls.
We have seen many combinations of these policies over time as shown in Table 11.3.
During the time of the gold standard, member countries gave up independent mon-
etary policy in order to get fixed exchange rates and free capital mobility. This was the
single longest period of a stable international currency regime ever recorded. The Bretton
Woods system opted for capital controls, as did China.
Table 11.3 The impossible trinity

Fixed exchange rates International capital Independent
mobility monetary policy
Gold standard ✓ ✓ ✕
Bretton Woods ✓ ✕ ✓
China today ✓ ✕ ✓
EU/US today ✕ ✓ ✓
207
11.5 Perils of overvaluation

Countries rarely set out to have an overvalued currency. Instead, they usually tend to end
up with an overvalued currency as a result of a losing fight with inflation, or the tail-end of
a prolonged crisis when they can finance themselves only by printing money, as Germany
did in the early 1920s.
In some cases, an overvalued currency may be forced on a country from the outside.
The best-known recent example of this is the Plaza Accord in 1985, which was an agree-
ment between France, Germany, Japan, the US and the UK to depreciate the USD relative
to the Japanese yen and the German mark.
In the short run, an overvalued currency is often quite popular with voters because
it makes imported goods artificially cheap. However, in the longer run it hurts exporters
who are no longer competitive, and often ends in a currency crisis.
In order to maintain the exchange rate, the central bank has to stand ready to sell for-
eign currency to all-comers at the artificially cheap rate, which it can do only as long as
it has reserves to sell. Speculators, who are usually well-connected local companies and
political families, observe this and seek to export the domestic currency – in effect, specu-
lating against the currency regime. The end result is a speculative attack.
The government may give in or resort to desperate measures like capital controls or
multiple exchange rates, which in turn are often a recipe for corruption. All of this sug-
gests that it is difficult to the point of impossible for a government to maintain an artifi-
cially strong exchange rate for long without resorting to very costly measures.
11.5.1 Currency crises

We use the same IMF database as in Section 7.3 to identify the frequency of currency cri-
ses, seen in Figure 11.3.
The frequency of currency crises seems to be holding steady over time, with the worst
year being 1994.
The medium- and long-term effects of currency crises have been studied by Bussiere
et al. (2010) who found that three years after a currency crisis the level of GDP is between
2% and 6% lower than in the absence of a crisis.
25
20
15
10
5
0
1975 1980 1985 1990 1995 2000 2005
Figure 11.3 Frequency of currency crises

Source: Laeven and Valencia (2008)
208
11.6 Undervaluation and ‘currency wars’
The interesting part is that these losses tend to materialise mainly before the currency
collapses. The authors conclude that the economic costs of a currency collapse do not
appear to arise from the collapse of the currency itself but reflect other factors.
They find that the collapse seems to actually have positive effects on output. Output growth
tends to slow down prior to a currency crisis, with positive growth rates the norm post-crisis.
11.6 Undervaluation and ‘currency wars’

Deliberate policies of manipulating exchange rates downwards to increase domestic com-
petitiveness have recently been given the name ‘currency wars’, but in years past have
been called competitive devaluations and beggar-thy-neighbour policies.
Countries often deliberately undervalue their currencies in order to increase their com-
petitiveness. Unlike overvaluation, in principle, a country could maintain an undervalued
currency indefinitely, printing money as needed to sell in the currency markets.
The effect of undervaluing a currency in the short run is to make imports more expen-
sive and exports cheaper. It is a subsidy given to exporters paid for by domestic consumers
and foreign companies and their workers because their profitability is eroded.
This does not mean that undervaluing a currency is an easy way to encourage export
industries. On the contrary, the costs can easily become quite significant. First, it will not
make other countries very happy. After all, the country is deliberately manipulating its
exchange rate to favour its industries at the expense of other countries. For very small
countries this may not be a problem, but it usually creates a lot of friction.
In the worst case, it can lead to competitive devaluations, where countries compete to
devalue their currencies. The end result is likely to be high inflation and a huge disruption
to domestic industries. This can also lead to restrictions on trade. Because the country is
trying to gain what is painted as an ‘unfair advantage’, other countries may impose trade
restrictions as a consequence. This can then spiral out of control in the same way as com-
petitive devaluations can. Finally, such a policy can create hidden inflation that makes
later realignment a necessity.
11.6.1 The UK and France

The cases of Britain and France in the interwar years provide an illustrative example of
the effects of a currency misalignment. In 1925, Britain decided to go back on the gold
standard at the same rate of convertibility as before WWI, overvaluing sterling by perhaps
40%. France, on the other hand, having experienced high volatility and depreciation of its
currency in the afterwar years, restored convertibility in 1926 at a devalued rate of around
15–20%. The French policy was analysed by Sicsic (1992).
The results of the different types of misalignments were quite predictable. Gold flowed
from the country where it had less value (the UK), due to an overvaluation of the cur-
rency, to where it had more value (France). Britain, with the overvalued pound and thus
a competitive disadvantage in exports, experienced a persistent balance of payments defi-
cit, draining its gold and FX reserves. In order to deal with the overvaluation, the UK had
to lower domestic costs, by doing what is now called an internal devaluation. This led to
209
significant social disruption and strikes. As a consequence of the overvaluation, Britain

was more or less in continuous recession for six years.
The effects of the undervaluation of the French franc were as expected. The undervalu-
ation meant a highly competitive export industry leading to an export-led boom with
large accompanying inflows of gold reserves. The central bank of France sterilised the
inflows to mitigate the inflationary effects of the gold inflows.
This led to considerable friction between France and the UK, as well as many other
countries. Eventually, other countries devalued their currencies, for example the UK left
the gold standard in 1931, after which Britain became quite competitive. This, however,
meant that France became increasingly uncompetitive because its economy had adjusted
to the weak currency regime and found it difficult to cope when the currency appreci-
ated, leading to significant economic and political instability in the years after.
11.6.2 China and the US

The largest country deliberately maintaining an undervalued currency now is China, and
we can find many parallels between the problems of Britain and France in the interwar
years and the conflict between China and the US today. Just like France, China maintains
a deliberately undervalued currency, accumulating vast amounts of reserves, in the case of
France gold and for China mostly USD, now exceeding $3 trillion. The undervalued currency
has been a key contributor to the economic performance of China and has led to the aura
of invincibility and the feeling that it is on track to become the world’s biggest economy.
It is, however, difficult to gauge exact performance in China. According to official statistics,
inflation is next to zero, but anecdotal evidence indicates it is much higher. That might suggest
that the country’s stellar economic performance might be lower if adjusted for higher inflation.
As a consequence of its FX policy, China has increasingly strained relationships with the
outside world, especially the US.
In the case of France, the end-game proved quite costly. The question remains whether
China will be more fortunate.
11.7 Reserve currency
Definition 11.1 Reserve currency Reserve currency is the currency most

governments hold as reserves and the currency in which most international trade is priced.
Historically, one currency, or asset, tends to assume the role of a reserve currency. In the
nineteenth century it was gold and the pound sterling, but over time, the US dollar has
taken over that function.
There are significant advantages in owning a reserve currency, most importantly that
foreign countries hold it as reserves. That means foreigners have exchanged real goods for
paper, a deal which is quite good for the country with the reserve currency.
It is often argued that the fact that major products are priced in a reserve currency con-
veys some benefits to the owner of the reserve currency, perhaps because it eliminates
210
11.7 Reserve currency
currency risk. That is not true. Prices are set in a competitive international market, and if the
reserve currency weakens, the price of commodities will increase by an offsetting amount,
other things being equal. The risk does not go away; it just moves from FX to prices.
Reserve currencies tend to make other countries unhappy. When reserves are building
up there is a transfer of real goods in return for a nominal asset. Furthermore, if that coun-
try decides to misbehave, creating inflation or doing quantitative easing (QE), it is taxing
reserves held by other countries.
Having a reserve currency can create particular problems when a country is declining
in power and wealth. Even though the USD had long overtaken sterling as a reserve cur-
rency, some countries with particularly close relationships to the UK, like Canada, Hong
Kong and Kuwait, continued keeping reserves in sterling until the early 1970s. This meant
that when the UK was forced to devalue the pound, it adversely affected their closest
allies. To keep these allies happy, the UK was forced to make side payments to them, fur-
ther increasing the British difficulties.
11.7.1 Threats to the US dollar reserve status

When a country enjoys the world’s largest economy, has solid economic growth, low
inflation and a strong military to back it up, the reserve status of its currency is undis-
puted. This was the situation for the US until the 1960s. Today, interest rates in the US
are close to zero, and government debt is close to historically high levels. Its economic
growth is anaemic, and still it is running a trade deficit. If the US did not have a reserve
currency, it would probably have experienced a sharp depreciation, stimulating exporters
and correcting the trade balance.
Because of the reserve status of the dollar, it remains artificially strong, as foreign countries
directly intervene in the currency markets. As a consequence, global imbalances build up.
There is a way out for the US, by creating inflation. That will achieve two goals simulta-
neously, depreciating the currency and reducing the real value of its debt, an increasingly
significant proportion of which is held by foreigners.
This is partly the path the US has taken. Its debt has been increasing sharply, and also it
has been monetising part of the debt by means of QE so that the Federal Reserve System
(Fed) has become a significant purchaser of government debt. We show some of the debt
numbers in Figure 11.4.
China Japan Brazil Taiwan Russia

8% $15
Debt ratio
6% $10
Debt
4%
$5
2%
0% $0
2001 2003 2005 2007 2009 2011
Figure 11.4 Fraction of US government debt held by major creditor countries, and total
debt, USD trillions
Data source: US Treasury
211
By December 2011, US government debt was $15 trillion, having tripled since 2000. The
largest owner of US debt is the Social Security Trust Fund with 19%, followed by the Fed at
11.3%, leaving China in third place at 8%. Overall, 33% of the debt is held by foreign enti-
ties, not necessarily governments. Traditionally, Japan has been the largest holder but it is
now in second place, followed by Brazil, Taiwan and Russia. The US itself does not feel the
need to hold large foreign reserves, which by the end of 2011 amounted to $147 billion.
11.7.2 The Chinese renminbi

Many countries resent the reserve status of the dollar, and if the US continues to rapidly
increase its debt whilst suffering from poor economic growth, some observers have been
trying to identify what the next reserve currency might be.
The most obvious candidate is the Chinese renminbi, as China is the second largest econ-
omy in the world, and if its current growth trend holds, is on track to surpass the US by
the end of this decade. China has been the most vocal complainer about the dollar being
reserve currency, promoting alternatives, such as the SDR. But what about its own currency?
Several factors prevent the renminbi from becoming the reserve currency. So long as
China maintains strict capital controls, that is not possible, because a reserve currency
needs to be freely traded abroad. Foreigners need to be able to hold it without any wor-
ries about converting it into their own currencies at a later date. Related to this is the fact
that the Chinese government maintains a closer control of its economy than any of the
other candidates for reserve status, with widespread fears that the economic statistics it
puts out are not reliable. This means that in the current environment, foreigners would
not trust China with a reserve currency. Finally, at the moment China does not seem to
want to have a reserve currency.
If we look towards the future, and China continues to grow, with its economy becom-
ing more transparent, whilst the US declines, it is possible, if not likely, that China might
regain reserve status over the next few decades.
11.7.3 The euro

Until its recent problems, the euro was touted as the next reserve currency. It is the
currency of the largest economic area in the world, the euro zone, is freely traded and
transparent. Until recently, many countries were increasingly using the euro for foreign
currency reserves and trade was beginning to become priced in euros.
With the recent difficulties facing the euro, this is unlikely to happen soon. Not only is the
future of the euro no longer certain, the European authorities have shown themselves to be poor
stewards of a currency and, therefore, unreliable as the issuers of the future reserve currency.
11.8 Summary
The largest financial market by volume is the market for foreign exchange. Since the level
of exchange rates is vitally important for most countries, it is not surprising that few, if any,
markets see more government intervention than the FX market. Every government needs
212
Appendix: Exchange rate regimes
to have a currency regime in place, ranging from being fully tied to other currencies to
freely floating.
Governments have various mechanisms for intervening in the FX market. If the exchange
rate is overvalued, it can only continue intervening until it runs out of foreign reserves, but
it can intervene indefinitely to maintain an undervalued currency. Governments often
couple such interventions with sterilisation, a simultaneous reduction in the money
supply.
Because of the ‘impossible trinity’, governments do not have full choice in the policy
measures they can implement, and often resort to capital controls. While the traditional
sort of capital controls was aimed at tightly controlling the flow of foreign currency in and
out, ‘capital controls 2.0’ target hot money.
If a country undervalues its currency, this often leads to significant friction with its trad-
ing partners, sometimes called ‘currency wars’.
One country in the world has the reserve currency. Historically, that was sterling, but
over the past century or so the US dollar has been the reserve currency. Its status is under
threat, and other currencies such as the renminbi or the euro may take over.
Appendix: Exchange rate regimes

The following is a presentation of the varieties of exchange rate regimes based on Frankel
(1999) and Gosh et al. (2002).
Currency or monetary union

A group of countries agree to use a common currency issued by a central monetary
authority. The member countries give up an independent monetary and exchange rate
policy to the central monetary authority. This is a multilateral arrangement whilst most
others are unilateral.
The prime example of a currency union is the euro zone of 17 countries that have
adopted a common currency as their sole legal tender. Even a currency union can be
reversed, for example, the Czech Republic and Slovakia gave up the koruna when the
countries split up, but it is the firmest possible commitment to a fixed exchange rate.
Currency board
A country pegs its exchange rate to a foreign currency, setting the exchange rate regime
and the actual exchange rate into law. A key example is Argentina up until late 2001,
which had the peso linked to the dollar.
Single currency peg

A country’s currency is pegged at a fixed rate to the currency of another country. The rate
is generally adjustable and the credibility of the peg is related to the level of FX reserves.
Examples are the East Asian countries linking their currencies to the dollar in the 1990s,
and francophone West African countries that pegged to the French franc. The costs of
readjusting or giving up the peg altogether are lower than in the case of a currency union
or currency board.
213
Adjustable peg
The adjustable peg is similar to a single currency peg but with less commitment and more
open and frequent adjustments. In practice, it is rather difficult because adjustments put
a strain on central bank credibility. The Bretton Woods system was based on an adjust-
able peg setup.
Basket peg
The currency is not pegged to a specific currency but is fixed to a weighted basket of two or
more currencies. This is a useful approach for countries with geographically diversified trade
patterns. The basket can be designed according to country-specific criteria or to a compos-
ite currency (SDR, for example). Theoretically a basket peg can be as rigid as a single cur-
rency peg. In practice, however, countries that use a basket peg frequently keep the weights
secret and can thus adjust the weights or the exchange rate level at their own discretion.
Crawling peg
The exchange rate is set in a rule-based manner and can be regularly reset in a series of
mini-devaluations, which can occur as often as weekly, usually based on a predetermined
rate or a function of inflation-rate differentials. High-inflation countries frequently opt for
a crawling peg, since the exchange rate can be readjusted to allow for inflation differen-
tials without giving up the benefits of a deterministic exchange rate.
Target zones or bands

A target zone establishes fixed exchange rate margins but allows the exchange rate to
fluctuate within those margins. The monetary authorities intervene when the exchange
rate hits the pre-announced margins on either side of the central parity. The European
exchange rate mechanism (ERM) was an example of a target zone.
Managed float or dirty float

The monetary authority allows the exchange rate to move freely in the market, but has a
general view on the broad level and path of the exchange rate, standing ready to intervene.
The government does not defend any particular parity or follow any intervention rule.
Free float
The exchange rate is determined by demand and supply in the FX market without any or
with just minor official intervention. This option requires little or no official reserves but
also mostly prevents the accumulation of reserves. There are no truly freely floating cur-
rencies today.
Dollarisation (or euroisation)

This is a special case of currency arrangement, where a country uses a foreign currency as
legal tender in addition to or instead of a domestic currency. Typically, the foreign cur-
rency is the dollar or the euro.
Dollarisation can be bilateral or unilateral. Examples of the former are Panama and
Kosovo, and of the latter Ecuador and Montenegro. Bilateral dollarisation means that the
214
References
central bank of the country is the central bank of the main country, like the Fed or the
ECB, which makes the arrangement firm and provides for lending at last resort. Unilateral
dollarisation is more unstable, since the country needs to have enough foreign currency
reserves to meet all possible demands for foreign currency, extending to all types of
money including M0, M1, M2 and M3, otherwise the setup is subject to runs.

1 Why is the volume of currency transactions so much higher than the volume of trade?
2 Do you prefer fixed or floating rates?
3 Why do governments sterilise FX interventions, and mechanically how do they go

about doing it?
4 What are traditional capital controls and capital controls 2.0?
5 What is a speculative attack? Do they force the government to abandon a sensible pol-
icy, or make the government see reality?
6 What is the ‘impossible trinity’?
7 What was the Bretton Woods system, and would you want to see it adopted now?
8 What is a currency war, and why is it damaging?
9 Can a country maintain an overvalued exchange rate indefinitely?
10 Can a country maintain an undervalued exchange rate indefinitely?
11 Is China’s foreign exchange policy sensible?
12 What is a reserve currency?
13 Is it likely that the euro will become the next reserve currency?
14 Is it likely that China will have the next reserve currency?
References
Abdelal, R. (2007). Capital Rules: the Construction of Global Finance. Harvard University Press.
Ahamed, L. (2009). Lords of Finance, the Bankers Who Broke the World. The Penguin Group.
Arnason, R. and Danielsson, J. (2011). Capital controls are exactly wrong for Iceland. Vox
EU, 14 November. www.voxeu.org/article/iceland-and-imf-why-capital-controls-are-entirely-
wrong.
Bussiere, M., Saxena, S. C. and Tovar, C. E. (2010). Chronicle of currency collapses re-examining
the effects on output. Technical report, ECB. Working paper series no. 1226, July.
Eichengreen, B. (1996). Golden Fetters: the Gold Standard and the Great Depression, 1919–1939.
Oxford University Press.
Frankel, J. A. (1999). No single currency regime is right for all countries at all times. Working
paper 7338, National Bureau of Economic Research.
215
Gosh, A. R., Gulde, A.-M. and Wolf, H. C. (2002). Exchange Rate Regime: Choices and
Consequences. MIT Press.
Hume, D. (1752). On the balance of trade. In Essays, Moral, Political and Literary. Edinburgh
and London.
Krugman, P. R. and Obstfeld, M. (2006). International Economics: Theory and Policy, 7th edition.
Pearson.
Laeven, L. and Valencia, F. (2008). Systemic banking crises: a new database. Technical report,
IMF. IMF Working Paper.
Lavigne, R. (2008). Sterilized intervention in emerging-market economies: trends, costs and
risks. Discussion paper 2008-4, Bank of Canada.
Marsh, D. (2010). The Euro: the Politics of the New Global Currency. Yale University Press.
McCallum, B. T. (1996). International Monetary Economics. Oxford University Press.
OECD (2002). Forty years’ experience with the OECD code of liberalisation of capital move-
ments. Technical report.
Sarno, L. and Taylor, M. P. (2003). The Economics of Exchange Rates. Cambridge University Press.
Sicsic, P. (1992). Was the Franc Poincaré deliberately undervalued? Explorations in Economic
History, 29: 69–92.
216
12 Currency crisis models
Countries that peg their exchange rates are susceptible to currency crises if the
exchange rate is out of alignment with the underlying economic fundamentals, espe-
cially if the exchange rate is overvalued, the case considered in this chapter. This often
leads to speculative attacks, where speculators aim to force a devaluation by a massive
sale of the currency on the open market. If speculators believe that the exchange rate
is sustainable, and the government is determined to keep the peg, they will act in a
way that supports the currency regime. If, however, they lose that belief, the specula-
tors may force a realignment.
Mechanically, a speculative attack can be implemented by borrowing large
amounts of domestic currency and immediately selling it for a foreign currency, or
alternatively by entering into forward foreign exchange contracts. Once the peg is
abandoned, the speculator can realise her profits by converting foreign currency to
domestic currency at the more advantageous rate and repay the loan, or buy the for-
eign currency at the now cheap rate.
A key factor in whether an exchange rate regime is sustainable is the strength of the
underlying economic fundamentals, referring to variables like inflation, the balance of
payments, unemployment, economic growth and the budget deficit. A country with
strong fundamentals is better able to maintain its exchange rate policies and repel
speculative attacks, which in turn makes it less likely that an attack will be launched
in the first place.
Chapter 12 Currency crisis models

This chapter relates directly to Chapter 11 (currency markets).
Key concepts
■ Speculative attacks
■ First-generation currency crisis models
■ Second-generation currency crisis moments
■ Global games models
■ Argentina
■ ERM crisis

The specific readings in this chapter relate to the 1G, the 2G and the global games
models. Background reading for 1G models is provided in Obstfeld and Rogoff
(1996), pp. 559–566, and the original paper we studied is Flood and Garber (1984).
Our discussion of the 2G models is based on Copeland (2000), with Obstfeld (1996)
providing more detailed mathematical treatments. The global games discussion fol-
lows Morris and Shin (1998, 1999).

The lower-case variables d, m, p and r are expressed as logarithms. Variables with an
asterisk, e.g. p*, are from the foreign country.
c Cost to speculators
d Log of domestic credit
e Spot exchange rate (domestic/foreign)
e Fixed exchange rate
ẽ Shadow exchange rate
en Desired exchange rate
i, i* Domestic and foreign interest
k Threshold level of reserves
/ Proportion of speculators who attack
L Loss function of the government
m Log of domestic nominal money supply
p Log of price level in the domestic country
Q Cost of abandoning the peg
r Log of government foreign exchange reserves
T Timing of attack
v Payoff from attacking
x Noisy signal of fundamentals
Cost(Δe) Indicator function of cost of abandoning the peg
a Elasticity of agents’ demand for money
u Fundamentals
m Growth rate in domestic credit
c, h Parameters
218
12.1 First-generation models

The first modern currency crisis model is the first-generation currency crisis model
(1G model), dealing with speculative attacks. The model is quite simple, assuming the
government pegs the currency whilst running an unsustainable monetary policy. When
the government runs out of foreign currency reserves, the currency devalues. Speculators
know this, and attack the currency before the government runs out of foreign currency.
This was first modelled by Krugman (1979), though the specific model we discuss was
designed by Flood and Garber (1984) and we follow the presentation in Obstfeld and
Rogoff (1996), pp. 559–566.
The model assumes continuous time in a small open economy characterised by pur-
chasing power parity (PPP), uncovered interest rate parity (UIP) and agents that p ossess
perfect foresight. There are no private banks, hence the money supply equals the sum of
the domestic credit provided by the central bank, plus the value in domestic currency of
foreign exchange reserves, which by assumption do not produce any interest.
Setup
The assets available to domestic residents are domestic money, domestic bonds, foreign
money and foreign bonds. The domestic government uses the stock of foreign currency it
holds to fix the exchange rate. The model is built around five equations:
log money market equilibrium: mt - pt = -ait (12.1)

log nominal money supply: mt = dt + kt (12.2)
log domestic credit expansion: dt - dt - 1 = m(12.3)
log PPP: pt = pt* + log et (12.4)
log UIP: it = it* + Et - 1 Δet (12.5)
The money supply, m, and prices, p, are in logs, as well as the growth in domestic money,
m. i is the interest rate and e is the nominal exchange rate, neither in logs. Equation (12.1)
describes the equilibrium condition in the money market. The elasticity a is positive, so the
real demand for money is a decreasing function of the domestic interest rate. Equation (12.2)
shows that the nominal money supply equals the sum of domestic credit and the foreign
exchange reserves. Equation (12.3) is the first derivative of domestic credit in relation to time
and shows that domestic credit is increasing at the positive, constant, rate of m.
The government runs persistent deficits that are financed by money creation at rate m.
Equations (12.4) and (12.5) respectively indicate that the price level follows the PPP rule
and the interest rate obeys the UIP condition.
The model
Under an assumption of perfect foresight, and assuming the exchange rate is fixed and
equal to log e, it follows that:
Et - 1 Δet = 0
it = it*
219
Substituting (12.2) for the nominal money supply in m in (12.1), (12.4) for the domestic
price level pt in (12.1), and (12.5) for the domestic interest rate it in (12.1) leads to:
kt + dt - pt* - log e = -a1it* + Et - 1 Δet2 (12.6)
By assumption, the foreign price level p*t and the foreign interest rate i* are constant, as
is e. According to (12.6), the depletion rate of foreign exchange reserves is -m, so it fol-
lows that:
kt = k0 - mt (12.7)
Equivalently:
dt = d0 + mt(12.8)
k0 and d0 are the initial levels of reserves and credit, respectively, at time t = 0.
Equation (12.7) shows that the foreign exchange reserves decrease proportionally to
the growth in domestic credit in order to keep the nominal money supply and thus the
exchange rate fixed.
Unsustainability
To keep the foreign exchange rate fixed, the central bank uses its foreign currency reserves
to absorb any currency the public does not want to hold at the fixed rate e. Therefore the
foreign reserves decrease at the same speed as the increase in the central bank holdings of
domestic government debt.
This is obviously not sustainable as the central bank will eventually run out of foreign
exchange reserves, supposing it has no means to increase them. At that time, the fixed
exchange rate regime will have to be abandoned. We assume that the government will
support the fixed rate as long as its net reserves remain positive. After the fixed rate regime
collapses, the exchange rate floats freely for ever.
Timing of the attack

The 1G model predicts that the fixed exchange rate regime will be abandoned before the
central bank has completely exhausted its reserves. To solve for the timing of the attack,
we introduce the concept of a shadow exchange rate. The shadow exchange rate, denoted
log ẽ, is the exchange rate that would prevail if the currency was allowed to float. If the
currency was not pegged and the government continued with the domestic credit expan-
sion at rate m every period, the currency would depreciate by m every period. This means
that in the absence of a pegged exchange rate, the rate of depreciation would equal the
rate of money growth.
Under perfect foresight, agents expect that the exchange rate moves in line with the
money growth rate:
Et - 1 Δet = log m
220
Employing the shadow exchange rate, the money market equilibrium in (12.6) becomes:
dt - log ẽ t = -a1Et - 1 Δet2
Solving for the shadow exchange rate log ẽ:
log ẽ t = am + dt
Time of attack, T
First, suppose the attack happens late when log ẽ 7 log e, after which the peg is aban-
doned. In that case we would see a perfectly anticipated discrete rise in the exchange rate
equal to log ẽ - log e and the speculators would reap an instantaneously infinite rate of
profit. This represents an arbitrage opportunity giving speculators the incentive to attack
earlier and pre-empt their competitors, which will lead to an attack happening before
log ẽ 7 log e.
If, on the other hand, speculators attacked early, when log ẽ 6 log e, they would suffer
a loss since the currency would actually appreciate following the attack. Therefore, there
is no incentive to attack and the fixed exchange rate regime will survive.
Therefore, the attack must happen when log ẽ = log e, that is when the shadow
exchange rate equals the fixed exchange rate. Let T denote the time of attack. Substitute
(12.8) for dt and knowing that at T the shadow exchange rate is equal to the peg, we get:
log e = am + d0 + mT
Solving for T:
log e - d0 - am
T =
m
Prior to the crisis, the quantity of money remains constant but its composition var-
ies: domestic credit grows at rate m while reserves decrease at the same rate. After the
crisis, the government has lost all its foreign reserves and the money stock equals domes-
tic credit and thus grows at the same rate m. The solution can be seen graphically in
Figures 12.1 and 12.2.
log e
te
g e ra
Actual exc han
r ate Timing of attack

dow
Sha
T Time
Figure 12.1 Shadow exchange rate
221
k
kt if
no
kT at
ta
c
k
T Time
Figure 12.2 Foreign reserves
Analysis
The 1G model suggests that currency crises originate from an expansionary monetary pol-
icy incompatible with a fixed exchange rate regime. Speculators rationally anticipate that
the fixed exchange rate regime cannot be indefinitely maintained and launch an attack
before the foreign exchange reserves are exhausted. The model relies on strong assump-
tions such as perfect foresight, UIP and PPP.
In this model, all agents are completely rational except the government itself, since its
policy of ongoing monetary expansion eventually destroys the fixed exchange rate regime.
The underlying message of the model, however, is not that speculative attacks can be fore-
seen perfectly, but rather that it illustrates how private agents respond to inconsistent eco-
nomic policies. The 1G model shows that a large asset market event, for instance an attack,
does not necessarily need to be associated with a large shock but can build up gradually.
12.2 The Argentinian crisis

Argentina has long suffered high inflation, at times bordering on hyperinflation. In order
to find stability, the Argentinian government in 1989 embarked on a strong programme of
market-oriented structural reforms.
Part of the agenda was the introduction of a currency board, whereby the exchange
rate was fixed at one-to-one with the US dollar. Initially, this was quite successful: infla-
tion fell to single digits as seen in Figure 12.3, the exchange rate was stable as seen in
Figure 12.4, and economic growth was high in the early years of the programme.
Underneath this success significant vulnerabilities emerged. Not only did countries
competing in the same export markets depreciate their currencies, Argentina was unable
to maintain the fiscal strength necessary for the currency board. Persistent budget deficits
led to a steady increase in the stock of debt, much of which was denominated in dollars,
of increasingly short maturities, and held by foreign investors. The government, facing
an election in 1999, responded by pursuing an expansionary fiscal policy with dire conse-
quences for the already strained fiscal position.
The situation continued to worsen in 2000, leading Argentina to ask the IMF for help
at the end of the year. The IMF did lend $17 billion but stabilisation remained elusive and
Argentina suffered a speculative attack on its currency in 2001. Spreads on Argentinian
bonds rose sharply in the course of 2001, shutting it out of international capital markets
222
12.2 The Argentinian crisis
75%
50%
25%
0%
1991 1993 1995 1997 1999 2001
Figure 12.3 Year-on-year inflation, Argentina 1991–2002

Data source: International Monetary Fund (IMF)
10000
1000
100
10
1
1990 1995 2000 2005
Figure 12.4 Argentinean peso/USD, 1987–2004

Data source: www.globalfinancialdata.com
and making Argentina dependent on IMF financing. The beginning of the end was when
the IMF, dissatisfied with the Argentinian government for not complying with fiscal tar-
gets, held back on more loans.
Argentina partially defaulted on its international obligations in December 2001 and
abandoned the currency board a few weeks later. As a consequence, Argentina went into
a deep crisis, experiencing a GDP decline of 15% in 2002, with unemployment rising to
over 20%. Since then, Argentina has been mostly shunned by the capital markets, and is
still in negotiations with creditors over the sovereign default of 2002.
While Argentina has reached settlement with many of its creditors, it is being pursued
by so-called ‘vulture funds’ who buy distressed sovereign obligations and pursue the debt-
ors in international courts. One example of this was the seizure of an Argentine naval ship
in Ghana in October 2012, at the request of a subsidiary of the American hedge fund Elliot
Capital Management: see BBC (2012a). This is the same fund as in Section 19.3, where we
discuss the enforcement of sovereign debt claims.
Analysis
At the heart of the problem was a fiscal policy inconsistent with the currency board. The
dollar peg eliminated monetary policy as a policy tool and put strong restrictions on fiscal
policy to keep debt sufficiently low to avoid an overvaluation of the peso. Fiscal prudence
was also needed to maintain the credibility of the guarantee to convert pesos to dollars
at parity, which was needed to attract capital inflows, and to support the ability of the
government to act as a lender of last resort (LOLR).
223
This issue was especially delicate in Argentina with its history of irresponsible fiscal
olicy. The fiscal policy was the result of institutional weaknesses, which persistently
p
pushed the government to commit more fiscal resources than it was capable of mobilis-
ing. A key reason was fiscal federalism, whereby regional governments could raise debt
with limited central government oversight. The IMF was an enthusiastic supporter of the
currency board, but after the crisis the chief economist of the IMF, Michael Mussa (2002),
recognised the mistakes:
‘In these kinds of conditions, the choice of the currency board, though effective in
the short-term as a tool to stabilize the economy and the price level, was risky over
the medium- to long-term. The currency board eliminated an expansionary mon-
etary policy and money creation as a mean to raise revenues, while it required long-
term fiscal discipline at the same time. If fiscal discipline was not adhered to, the
medium-term dangers were strongly increased. Adding to these dangers was the fact
that in Argentina only a limited market for long-term debt issued in pesos existed,
forcing the government to finance its deficit by external borrowing in dollars, raising
the costs of a devaluation even further.’
The Argentinian crisis is consistent with the 1G model. The government ran an unsustain-
able monetary and fiscal policy, exhausting all foreign currency resources before aban-
doning the currency board. Towards the end, it was subject to a speculative attack.
While it employed extensive capital controls, they were always quite leaky, for exam-
ple, one could buy stocks in Argentinian companies, exchange them for American deposi-
tary receipts (ADRs) in New York, sell them there and receive dollars. Interestingly, the
ADRs traded at a 40% discount, which was the amount of the eventual depreciation.
12.3 Second-generation models

The 1990s saw a number of currency crises that were inconsistent with the 1G models,
and as a response a new class of models, the 2G models, were developed. Below we
present one of the simpler 2G models, that of Copeland (2000), capturing the essential
intuition whilst leaving out some of the richer features of these types of models.
The underlying premise is that a government pursues a fixed exchange rate, e, but
would prefer a floating regime to achieve its domestic policies. The model focuses on the
government’s continuous comparison of the net benefits from staying in the fixed regime
versus floating the exchange rate.
The government faces three types of concerns:
1 First, it would like to run the economy with a higher level of aggregate demand, requir-
ing a currency devaluation. We can summarise these wishes into a variable denoted en ,
which we call the desired exchange rate, that is, the exchange rate the government
would choose if it had not made a commitment to the fixed rate.
2 Second, there is a cost if the fixed rate peg is to be abandoned. This could be inter-
preted as political pain or loss of credibility of the monetary institutions. We
224
summarise this by the cost of exchange rate change, Cost(Δe). For simplicity we
assume that the function Cost (Δe) can take only two possible values:
0 for Δe = 0
Cost(Δe) = e
Q for Δe 7 0
3 Finally, the peg will be more costly to defend when a devaluation is expected than
when it is not. By the UIP condition, an expected devaluation means that the domes-
tic interest rate will have to rise to defend the peg and this is likely to hurt the
economy.
Decisions
We capture these three concerns in the following loss function that the government
faces:
L = 5c(en - e) + hE(Δe)62 + Cost(Δe) c, h 7 0, en 7 e
The term hE(Δe) reflects the expected economic pain associated with defending the peg
with increasing interest rates. h is a parameter that shows the strength of this loss.
We assume that the desired exchange rate of the government, en , is always higher
(weaker) than the fixed exchange rate, e, which means that the government is bothered
only by an overvaluation of the currency but not by an undervaluation. The term (en - e)
is then the extent to which the currency is overvalued relative to the level which the
government would like. c is a parameter of the magnitude of the loss associated with the
overvaluation.
We assume that if the government chooses to devalue, it will have no further reason to
change the exchange rate further, so from then on, the expected depreciation would be
zero. Hence, the cost of devaluation is simply Q. Figure 12.5 is a graphical description of
this situation, highlighting the multiple equilibria. Two cases can be distinguished, each
with two subcases.
2 1
Q
Defend Abandon
ê
ē A B
Figure 12.5 Multiple equilibria
225
Case 1: Market participants expect the government to resist the pressure to devalue,
therefore, E(Δe) = 0.
The government now has two options. It can either stick to the peg or it can devalue. If it
decides to keep the peg, the cost of abandoning the peg is equal to zero, Cost(Δe) = 0,
and the government faces the following loss function:
L1 = 5c(en - e)62
If, however, the government decides to devalue, it bears the costs of the abandonment
and its total loss becomes:
L = Cost(Δe) = Q
Facing these options, the government will defend if the loss associated with keeping
the peg is lower than the abandonment cost:
L1 6 Q
This condition is satisfied for all values of en between e and B in Figure 12.5. If en sur-
passes point B, the loss function passes the abandonment cost of Q and it becomes
more costly to defend the peg than to devalue.
Case 2: Market participants expect the government to surrender to market pressure and
allow the currency to depreciate, which means that E(Δe) 7 0. The government again
has two options. Either it can decide to defend the peg, leading to the following loss
function:
L2 = 5c(en - e) + h(en - e)6 2

= 5(c + h)(en - e)62
Alternatively, it can decide to give up the peg, suffering the loss of the abandonment
cost function:
Cost(Δe) = Q
In this case the government will defend the peg if
L2 6 Q
The government will now defend its exchange rate for all en between e and A in Figure 12.5.
Analysis
Figure 12.5 shows that for levels of en just slightly higher than e, it is optimal for the govern-
ment to defend the peg, no matter what market participants expect. For very high values
of en the exact opposite is true and it is always optimal for the government to devalue. This
causes trigger effects, whereby a very small change in fundamentals can cause a large
change in exchange rates.
The interesting case occurs when the government finds itself in a situation between
points A and B in Figure 12.5. Here the following condition holds:
L1 6 Q 6 L2
226
In this intermediate situation, the government will find it optimal to validate the
arket’s expectations. If the market expects the government to defend, the government
m
faces L1 and will find it optimal not to devalue. If market participants expect the govern-
ment to abandon the peg, the government is confronted with L2 and will find it optimal
to abandon. In this region there are multiple equilibria and the government will follow
market participants’ expectations, which are, therefore, self-fulfilling. A speculative attack
would then succeed simply because it was expected to succeed.
A country whose desired exchange rate, en , falls between A and B could find it easy or
completely impossible to defend a fixed exchange rate regime, depending on whether the
market expects it to devalue or not.
Role of fundamentals
The state of the fundamentals is reflected in the difference between the desired exchange
rate, en , and the current peg, e. If fundamentals are strong, there is no need to stimulate the
economy by devaluing. On the other hand, a country with weak fundamentals may wish to
depreciate so it can improve its economic situation. Figure 12.5 shows that the smaller the
gap between the desired exchange rate and the fixed parity, the easier it will be to defend.
The state of fundamentals is also expressed through the slope of the loss functions
of the government, which are determined by the level of en . The higher the value of Q
(a country with a sizable political investment in the peg), the easier the government will
find it to defend, ceteris paribus.
The main points of the 2G model are the following:
1 There are self-fulfilling crises in a model with rational investors.

2 Events completely disconnected from the economy may change expectations and trig-
ger a currency crisis.
3 Fixed exchange rate regimes that work well in the absence of speculative attacks may
fail without major fundamental imbalances.
4 Crises are not fully predictable.
These are summarised in Figure 12.6.
Stable peg
Belief of stable peg

Hold
Depreciation
Belief of imminent attack

Sell
Figure 12.6 Multiple equilibria
227
Criticism of multiple equilibria

The relevance of second generation models has been disputed as far as they rely on the
existence of multiple equilibria. Both theoretical and empirical criticisms are levelled
against multiple equilibria.
The main criticism of the 2G models is the notion of multiple equilibria: how can a
model place the economy in two different states at the same time? However, this is not
strictly a multiple equilibria solution, because which equilibrium is chosen depends on
the expectations of the speculators, so conditional on those there is a single solution. This
makes the multiple equilibrium models a useful device to capture the role of expectations
in determining government behaviour and economic outcomes.
More modern versions of speculative attack models are able to dispense with the mul-
tiple equilibria solution by explicitly modelling information available to speculators. We
present one such model later in the chapter.
12.4 The European crisis, 1992–1993

One of the largest currency crises to hit European countries happened in 1992. This
came after the member countries of the European Union (EU) had been experimenting
with ways to reduce foreign exchange (FX) instability in Europe following the collapse of
the Bretton Woods system. From 1979 to the 1990s, EU countries formed a joint system
for coordination of monetary and exchange rate policies consisting of two major com-
ponents: the creation of an artificial unit of account named the European currency unit
(ECU)1 and an exchange rate regime, known as the European exchange rate mechanism
(ERM). The ERM was set up to reduce exchange rate volatility and achieve monetary sta-
bility in Europe and to pave the way for a single currency in the future.
The ERM was a target zone exchange rate regime. A grid of bilateral exchange rates to
the ECU was calculated for each country, where currency fluctuations had to be con-
tained within a margin of 2.25% on either side of the rate. The Spanish peseta, Portuguese
escudo and Italian lira had wider bands of 6%. Member countries had to intervene to
ensure their currencies stayed within the prescribed band. In a crisis, the country with
strong currency – read Germany – was supposed to lend its currency to countries needing
help to strengthen their currency.
In practice, the currency bands were maintained with respect to the most stable cur-
rency of the group, the German mark, which became the effective reserve currency in the
ERM, giving Germany a pivotal role. This arrangement meant that Germany was the only
country able to set its monetary policy independently. All other countries in the ERM
were forced to adopt the German monetary policy if they wanted to stay within the
target zone.
1
This was only a unit of account, not a medium of exchange. It was constructed as a fixed basket of
European currencies. Member countries were Belgium, Denmark, France, Germany, Ireland, Italy,
Luxembourg, The Netherlands, Portugal, Spain and the United Kingdom.
228
12.4 The European crisis, 1992–1993
If, for example, Italy decided to use its own discretion over monetary policy and
expand its money supply, this would result in higher inflation in Italy, weakening the lira.
This gave Italy two choices: either the Italian Central Bank intervened to buy lire on the
open market or it devalued the lira, choosing a new central parity for the lira to fluctu-
ate around. Neither choice was attractive, intervening was costly and may not have been
effective, and devaluing was also costly, causing the Central Bank to lose credibility, set-
ting in inflation and making speculative attacks more likely. Therefore, the best choice for
Italy was to adopt the German monetary policy to ensure the strength of the lira.
The country with the strongest currency did not face these problems, especially if it
chose to contract its money supply. Therefore, the country with the strongest currency
has freedom of action, and the others have to follow.
Start of the crisis – German reunification

The starting point of the ERM crisis was the German reunification. The German govern-
ment embarked on a massive fiscal expansion to transfer resources to the former East
Germany, increasing the government budget deficit from 5% of GDP to 13.2%, whilst East
German marks could be converted to West German marks at a highly favourable rate of
1.8:1, massively increasing the money supply. The German Central Bank became worried
about inflation because the economy was overheating, and opted to raise interest rates
by more than 3%.
Other countries
The German decision to raise interest rates strengthened the mark, putting pressure on
other members of the ERM who felt they had no choice but to also raise their interest
rates, in effect importing the German monetary policy. However, the German decision
was based on the special considerations of the reunification, and other members of the
ERM were in a different situation, with many in a recession. This meant that the optimal
monetary policy for them was an interest rate reduction, not an increase, and the need to
follow Germany meant their economic situations further deteriorated.
The situation could have been alleviated if Germany had lowered its interest rates, but
it refused to do so, recommending realignment instead. Other member countries of the
ERM rejected realignment, increasingly advocating abandoning the ERM altogether.
Speculative attack – Black Wednesday

The tensions between the ERM member states did not escape unnoticed, and specula-
tors recognised that a realignment was increasingly likely. In the context of the 2G mod-
els, some member states had neither the strength of the fundamentals nor the desire
to resist a speculative attack. That made an attack inevitable, and by the beginning of
September 1992 a massive speculative attack against the lira was underway, with the lira
finally devalued by 7% against the mark.
Just a few days later, on 16 September 1992, on what has been termed ‘Black
Wednesday’, speculators turned their attention to the pound. The Bank of England (BoE)
raised the minimum lending rate from 10% to 12%, promising to raise it to 15%. This did
not deter the speculators, and at the end of the day the pound closed below its ERM
229
floor, prompting the Bank to announce a ‘temporary’ withdrawal from the ERM, made
permanent on 19 September. The cost of defending the pound was estimated by the UK
Treasury to be £3.3 billion, around £800 million being due to direct trading losses, with
the rest lost profit from foreign reserves, sold before the pound depreciated.
Italy followed the UK out of the ERM, while Spain, although staying in the ERM, deval-
ued the peseta by 5%. The attacks continued well into 1993, with the franc, peseta,
krone and escudo under pressure. On 30 July 1993, all the ERM currencies except the
Dutch guilder and Irish punt were quoted at the bottom of their bands against the mark.
A thorough revamping of the ERM was announced on 1 August, with the size of the
bands widened from ;2.25% around par to ;15%, which finally managed to ease specu-
lative pressures.
1G analysis
The ERM crisis is not easily explained by the 1G model. The member countries devalu-
ated long before they ran out of reserves, but perhaps most importantly, the curren-
cies of many of the countries recovered over the next 18 months, suggesting that the
initial weakness of the exchange rates was not due to long-run increases in the money
supply and high inflation. Instead, the countries might just have had an overvalued
exchange rate, with a small devaluation necessary for them to become competitive
again. In other words, the devaluation was useful for the affected countries. While
this is contrary to the workings of the 1G model, it does point to the positive benefits
that speculation can bring about in forcing governments to abandon inappropriate
economic policies.
2G analysis
The 2G models were directly motivated by the ERM crisis and consequently describe the
crisis much better than the 1G models. The link between fundamentals and the attack
was weak in the crisis, with the timing random. At this time, the fundamentals were in
an indeterminate state in the language of the 2G models. The only thing that determines
whether an attack takes place is the speculators’ self-fulfilling expectations, and a small
change in expectations can then trigger a crisis.
Choices
Member countries were left with a difficult choice. They could match the German inter-
est rates, but that might not even be sufficient, because if the speculators perceived the
government as being unable to withstand the pressure, the interest rate increase had to
be very high. However, since that would have adversely affected the real economy, the
governments were unwilling to maintain very high interest rates, except perhaps in the
very short run. The speculators realised this, making the attack inevitable. The only way to
prevent the attack was either to maintain a credible monetary policy, matching Germany,
or to have sufficient reserves to repel any attack. With the benefit of hindsight, an attack
on the ERM was inevitable.
The level of fundamentals played an important but indirect role. Countries with weak
fundamentals, like high unemployment, low growth, weak banking systems or high
short-term debt, were the first to be attacked. In the context of the Copeland model in
230
12.5 Global games currency crisis model
Figure 12.5, this means that those countries had a higher slope of the loss functions and,
hence, a smaller defend area e - A and also a smaller indeterminate area A – B.
12.4.1 Implications for the euro

The problems of the ERM are eerily familiar in the current euro crisis. Then as now, the
key problem was that different countries need different monetary policies. Germany, the
Netherlands and other strong economies perform well, needing a strict monetary pol-
icy to keep inflationary expectations in check. Other countries, such as Portugal, Italy,
Greece and Spain, face economic difficulties, their currency is overvalued and they need
an accommodating monetary policy and a devaluation. The same problem applied then
as now, with many of the same countries in the same roles.
There is one crucial difference between the ERM crisis and the ongoing euro crisis. In
1992, the countries with overvalued currencies could revalue or exit. They had an escape
hatch. There is no such possibility now, and their difficulties therefore threaten the mon-
etary arrangement and even a systemic crisis.

One of the main problems with the 2G models is the multiple equilibrium result, which
comes from the assumption of common information among the speculators. Morris and
Shin (1998, 1999) propose a different model for the same problem, called a global games
model, resulting in a unique equilibrium.
The main innovation of their model is that speculators have only imperfect informa-
tion about the state of the fundamentals. This means that the speculators are trying to
guess what other speculators know, and want to attack the currency if and only if all other
speculators also attack.
Setup
The government maintains a peg and gets net benefits from holding the peg:
+ -
f 1u , / 2 = u - /
where u, uniformly distributed, is the underlying strength of the economy – the funda-
mentals. There is a continuum of speculators, where only some number may choose to
attack the currency. Denote / ∈ [0, 1] as the proportion of speculators who attack.
The peg fails regardless of what the speculators do when u … 0, and always survives
when u Ú 1. When 0 6 u 6 1, the peg is ‘ripe for attack’, and is abandoned if and only if
u 6 /, that is, if and only if a sufficiently large speculative attack is launched.
In the context of the Copeland model in Figure 12.5, u 6 1 corresponds to the area
e - A, u 6 0 to the area beyond B to the right, and the intermediate 0 6 u 6 1 cor-
responds to A – B. Unlike the Copeland model, the decision to attack in the immediate
area is explicitly linked to how many speculators attack, which is what provides a unique
equilibrium solution.
231
Decision problem for speculators

The speculators can do one of two things: attack the currency or refrain from doing so.
The payoff from refrain is zero and the cost of attack is c, whilst the profit from the peg
collapsing is 1. The payoff from attacking, v, therefore depends on the state u and the
proportion / who attack:
1 - c if / 7 u
v (u, /) = e
-c if / … u
The speculators face a coordination problem when u ∈ (0,1). If one speculator attacks
and nobody else does, the attack fails and she loses money. If, however, everybody
attacks except the one speculator, she also loses money. The speculators, therefore, want
to attack only when everybody else attacks.
The speculators have imperfect information and receive a noisy signal about the
fundamentals:
xi = u + s i
where si is uniformly distributed over [-e, e]. The distribution over u conditional on xi is
uniform over
[xi - e, xi + e]
The speculator makes a decision based on what she thinks everybody else will do. This
means there is a unique equilibrium in switching strategies, that is switching from refrain-
ing to attacking. Denote the switching point as x*.
Solution
The level of fundamentals where the peg will fail, u*, depends on the switching point x*,
which in turn depends on the failure point u*, solving u = /.
If all follow x* - switching, / is the proportion whose signal is below x* when the true
state is u*:
x* - 1u* - e2
/ =
2e
So, u* = / if and only if
x* - 1u* - e2
u* =
2e
At the switching point x*, the speculator is indifferent between attacking and refraining:
Pr 1peg fails | x*2 11 - c2 + Pr 1peg stays | x*2 1 -c2

= Pr 1peg fails | x*2 - c
= 0
232
The peg fails if and only if u 6 u*. So
Pr1u 6 u* | x*2 = c
u* - (x* - e)
= c
2e
These are two equations in two unknowns – u* and x*. Solving,
u* = 1 - c
x* = 1 - c - e(2c - 1)(12.9)
As e S 0, x* S u*.
Verification of solution
We can verify that when xi 6 x*, the speculator wants to attack, and when xi 7 x*, she
wants to refrain. Suppose xi 6 x*:
u* - (xi - e)
Pr(peg fails ∙ xi) =
2e
u* - (x* - e)
7
2e
= Pr(peg fails ∙ x*)
and conversely for when xi 7 x*. Switching strategy around x* is equilibrium. In fact, it is
the unique equilibrium.
Strategic and fundamental uncertainty

There is a distinction between fundamental uncertainty and strategic uncertainty that arises
from the solution in (12.9). Fundamental uncertainty disappears as e S 0. However, there
is still uniqueness of equilibrium (the difference between e = 0 and the limit as e S 0 ).
This leaves the question of what happens to the strategic uncertainty as e S 0?.
Suppose the signal is exactly x*. What is the probability that proportion / or less of the
speculators attack the currency?
The answer to this question is important, since the fact that a speculator is indifferent
between attacking and not attacking happens due to the uncertainty about the incidence
of attack. The reasoning must take account of the uncertainty over the true state u and the
uncertainty over the incidence of attack.
There are two steps to answer the question. First, if the true state u is higher than some
benchmark level un , then the proportion of speculators receiving a signal lower than x* is
/ or less. This benchmark state un satisfies:
x* - (u - e)
= /
2e
or
un = x* + e - 2e/
233
Second, the answer is given by the probability that the true state is higher than un , condi-
tional on signal x*. This is
(x* + e) - un
2e
(x* + e) - (x* + e - 2e/)
=
2e
= /
Call the proportion of speculators who attack the incidence of attack. The cumulative dis-
tribution function over the incidence of attack is the identity function. Therefore, the den-
sity function over the incidence of attack is uniform over [0, 1].
How is this answer affected by the size of the noise e? Not at all! As e S 0, the uncer-
tainty concerning u dissipates, but the strategic uncertainty is as severe as ever.
Analysis
A key difference between the global games model and the 2G model is that in the former there
is only one equilibrium. Speculators have imperfect information about the state of the funda-
mentals, leading them to guess what other speculators would do, and launching a speculative
attack if and only if they think other speculators will join them – strategic complementarity.
This means that the markets react suddenly, one moment supporting the peg, with an
attack the next moment, causing the peg to fail. This means that a very small change in
the speculators’ beliefs can lead to a very large outcome.
At the same time, the likelihood of an attack is directly correlated with the state of the
fundamentals. This leads to interesting results on the impact of transparency and disclosure.
When fundamentals are weak, greater public disclosure of the state of the fundamen-
tals increases the probability of attack, since the strategic uncertainty dissipates, making
coordinated attack easier. Fundamental uncertainty also dissipates, increasing the incen-
tive for attack. This provides support for a policy of constructive ambiguity.
When fundamentals are strong, greater public disclosure of fundamentals decreases
the probability of attack. Strategic uncertainty dissipates, causing a coordinated pull back
from attack. Fundamental uncertainty also dissipates, further increasing incentives to
refrain from attack.
There is an important difference between ex ante decisions on disclosures and oppor-
tunistic disclosures. In the former case, the authorities are providing regular updates on
the economy outside a crisis state, while in the latter they are disclosing information
opportunistically, with the explicit purpose of affecting the behaviour of speculators.
This, of course, is noted and can backfire on the authorities.
12.6 Summary
The focus in this chapter has been on formal models of currency crises. We started out
with the earliest crisis model, the 1G model, where the authorities run an unsustainable
monetary policy, using reserves to intervene to maintain a peg. Speculators know this, and
234
12.6 Summary
launch an attack on the currency before the authorities run out of reserves. This model is
quite simplistic and is not directly consistent with most crises, except that of Argentina.
In response, the 2G models were developed, where a key ingredient is the expectation
of speculators of the intentions of the authorities. A main result is that we can get self-
fulfilling crises. If speculators believe the authorities will maintain the peg they will not
attack, but if they believe the authorities will not, an attack is forthcoming. Therefore, the
speculators will be seen as correct, regardless of what they do.
These models can be used to analyse the ERM crisis of the early 1990s, where specula-
tors attacked the British pound when they realised that the UK would not stay within the
ERM target zone.
An important problem with the 2G is the presence of multiple equilibria. This is solved
by the global games models, where speculators have imperfect information about the
fundamentals and need to coordinate to launch an attack. Here, the markets can very sud-
denly transit from stability to crisis as a tiny bit of extra information affects expectations.

1 Consider the 1G model presented in the lecture notes. Recall the system of equations
on which the model is built:
log money market equilibrium mt - pt = - ait; (1)
log nominal money supply mt = dt + k t ; (2)
log domestic credit expansion dt - dt - 1 = m; (3)
log PPP pt = pt* + log et; (4)
log UIP it = it* + Et - 1 Δet. (5)
Now suppose the government imposes permanent capital controls, we can capture this
by replacing (12.5) by:
it = (1 - u)(it* + Et - 1 Δet)
where u is the tax rate on profits.

(a) Following the same steps as in the chapter, derive the timing of the speculative attack T
(b) How does your answer differ from the standard case in the chapter? Explain.
2 Was the currency board a mistake for Argentina?
3 Do you think the first-generation models accurately capture speculative attacks?
4 Recall the 2G model by Copeland. At the centre of the model is the loss function of the
government
L = 5c(en - e) + hE(Δe)62 + C(Δe) c, h 7 0, en 7 e(6)
where C(Δe) is an indicator function that can take only one of two values:
0 for Δe = 0
C(Δe) = e
Q for Δe 7 0
For simplicity assume throughout this question that the fixed exchange rate is set
at e = 0, and that if the government devalues, it will directly move to the desired
exchange rate, so that E(Δe) = (en - e).
235
(a) Assuming that the government has a reputation of defending the exchange rate, and
speculators expect that it will stick to that policy in the future, what level of the de-
sired exchange rate, en, would induce the government to give up the peg and devalue?
Derive your results analytically and mathematically and show them in a graph.
(b) Now the government appoints a new president of the central bank who has a
reputation of being ‘soft’ on the exchange rate commitment. Now speculators
expect that a devaluation is probable. What level of the desired exchange rate, en ,
would induce the government now to give up the peg? Again, derive your results
analytically and mathematically and plot them.
(c) Assuming now that speculators have not finally decided whether the appointment
of the new central bank president weakens the exchange rate commitment,
(i) For what levels of en does the government definitely devalue?
(ii) For what levels of en does the government definitely defend?
(iii) What happens in between?
(iv) What role do fundamentals play?
(d) You can now assume that c = 2, h = 3 and Q = 64
(i) For what levels of en does the government definitely devalue?
(ii) For what levels of en does the government definitely defend?
(iii) Show how good fundamentals give the government more leverage in defend-
ing the exchange rate.
5 Was the speculative attack on some European currencies in the early 1990s positive or
negative for the economies of the affected countries?
6 Why does the global games model only have one equilibrium while the 2G model has two?
7 Explain the role of strategic complementarities in how speculators all agree on whether
to attack or refrain.
References
BBC (2012a). Argentina ship in Ghana seized over loans default. www.bbc.co.uk/news/
world-africa-19827562.
Copeland, L. (2000). Exchange Rates and International Finance. Prentice Hall.
Flood, R. P. and Garber, P. M. (1984). Collapsing exchange rate regimes: some linear examples.
J. Int. Econ., 17: 223–34.
Krugman, P. (1979). A model of balance-of-payments crises. J. Money Credit Banking, 11(3): 311–25.
Morris, S. and Shin, H. S. (1998). Unique equilibrium in a model of self-fulfilling currency
attacks. Amer. Econ. Rev., 88: 587–97.
Morris, S. and Shin, H. S. (1999). Risk management with interdependent choice. Oxford Review
of Economic Policy, 15: 52–62, reprinted in Bank of England Financial Stability Review, 7, 141–50.
www.bankofengland.co.uk/fsr/fsr07art5.pdf.
Mussa, M. (2002). Argentina and the Fund: from Triumph to Tragedy. Institute for International
Economics, Washington DC.
Obstfeld, M. (1996). Models of currency crises with self-fulfilling features. Eur. Econ. Rev., 40(3–5):
1037–1047.
Obstfeld, M. and Rogoff, K. (1996). Foundations of International Macroeconomics. MIT Press.
236
13 Financial regulations
Compared to other segments of the economy, financial institutions are special.

Suppose a firm making chocolate goes bust: the shareholders and employees suffer,
but competing chocolate firms will benefit. There is no significant damage to the
economy, and such developments may even be positive, such as argued by Joseph
Schumpeter in his 1942 notion of creative destruction.
Financial institutions are different. The failure of a single bank can have catastrophic
consequences as it may induce domino-style defaults, culminating in a systemic fail-
ure. These negative externalities justify financial regulations aiming at reducing the
incidence of banking crises and the damage caused by crises.
Financial institutions view this differently. They tend to find regulations to be a
costly nuisance. This does not mean they oppose regulation. On the contrary, they
directly benefit from regulations, not least because they act as a barrier to entry,
reducing competition and increasing rent. Regulations also act as a barrier to exit,
because financial institutions subject to regulations are likely to be seen as less risky
than otherwise, and when in difficulty may receive bailouts.
While regulations do mitigate the worst impacts of the externalities arising from
financial sector risk-taking, they can also impose significant costs on society, espe-
cially when poorly conceived. These costs could easily exceed the immediate ben-
efits from the reduction of the externalities. This might happen when banks have to
excessively curtail risk-taking, adversely affecting the sensitive small and medium-
sized enterprise (SME) sector. Because SMEs are often the main creator of new jobs,
it is important they receive funding, but as they are risky, financial institutions need
Chapter 13 Financial regulations
to be able to take risk in order to fund the SMEs. This causes a conflict for supervisors
because they cannot simultaneously curtail risky banking activities and want banks to
participate in economic recovery.
Financial regulations may also have other less visible costs. They may just encourage
financial institutions to continue the same risky activities out of sight – perhaps creat-
ing a shadow banking system – making it even harder for the authorities to understand
and regulate banking activities. Regulations may also encourage financial institutions
to move activities to other jurisdictions, depriving the domestic economy from some
of the benefits banks bring.
Designing and implementing financial regulations is difficult. They are essential
for the protection of society, but all too frequently fail, not providing protection, and
even perversely increasing systemic risk.
Definition 13.1 Regulation and supervision The term regulation encompasses

two different terms in common parlance: the laws passed by the nation’s parliament, and
the specific implementations of those laws, regulations, designed by a regulatory agency.
The term supervision refers to the enforcement of regulations. The term supervisor refers
either to an individual working in banking supervision, or more generally to the supervisory
agency.
The supervisory agency and the regulatory agency can be the same or distinct.

This chapter directly relates to Chapter 3 (endogenous risk) and Chapter 7 (banking
crises). We discuss the failure of regulations prior to the crises from 2007, and the
post-crisis reforms in Chapter 16 (failures in risk management and regulations) and
Chapter 18 (ongoing developments in financial regulation).
Key concepts
■ Motivation for financial regulations
■ Macro-prudential and micro-prudential policies
■ Bank capital and capital ratios
■ Unintended consequences of regulations
■ Basel Committee for Banking Supervision (BCBS)
■ Basel I and Basel II

While a number of books address banking regulation, Dewatripont and Tirole (1994)
stands out. Most banking regulations derive from the activities of the BCBS, whose
history was written by Goodhart (2011). Most information about current financial
regulations is provided on the web pages of the regulatory authorities, for the BCBS
www.bis.org/bcbs, and for the FSA in the UK www.fsa.gov.uk. For an early criticism of
the Basel Accords, see Danielsson et al. (2001).
238
13.1 Banking regulations

A Assets
C Capital
p Probability
T1,T2 Tier 1 and Tier 2
w Risk weight
A Minimum regulatory capital ratio

Banking has traditionally been the most regulated segment of the economy. The history of
the great financial centres in Europe in the thirteenth and fourteenth centuries as told by
Kohn (1999) provides some guidance. At the time, the places to be if one wanted to do
finance were Genoa, Florence, Venice, Bruges and Barcelona. Of those, one of the most
innovative was Barcelona, establishing the principle that banks had to have substantial
capital, and had to pay cash within 24 hours of demand. If banks failed, their owners got
into serious trouble, both with the Almighty and with city authorities. The punishments
could be severe. The banker Francesch Castello was beheaded in front of his bank in 1360,
and a failed banker was permanently prohibited from opening another bank.
Reasons for regulating banks

There are several reasons why it is considered necessary to regulate banks. Left to their
own devices, banks have a tendency to over-extend themselves, take too much risk and
fail. Because efficient and uninterrupted banking services are essential for any economy,
bank failures can impose costs on society far exceeding the private costs of the banks’
owners, employees and counterparties.
This means that banking entails significant externalities where the private incentives
of bankers are not aligned with that of society. For bank employees, working with other
people’s money, there is little downside, at worst dismissal, but they can enjoy a signifi-
cant upside in terms of high salaries and bonuses. This means that banks have an incentive
to take more risks than are desired by either their clients, their shareholders or society at
large. It is this externality that has often been the main motivation for regulating banks.
Banks also have a significant advantage over their clients, especially retail clients. Banks
sell sophisticated financial products to clients who have a very rudimentary knowledge of
finance, perhaps not even understanding basic percentages or present-value calculations.
It is quite easy for a bank to take advantage of clients and it is felt that this merits regula-
tion. The frequent mis-selling scandals in the United Kingdom (UK), most recently with
payment protection, indicate that such concerns are justified.
This advantage of sophistication can also extend to otherwise sophisticated corporate
and public sector clients. History is replete with examples of large entities suffering large
losses from buying complex financial products they did not understand. For example, the
recent bankruptcy of the city of Birmingham in the United States (US) is directly attributed
to its purchase of risky financial products the city’s managers did not seem to understand.
239
Because of the complexity of bank products, and the high costs of monitoring bank
performance, it is beyond anybody, except experts, to understand the underlying risks.
Certainly, for retail clients this is quite infeasible. It is more efficient for a dedicated gov-
ernment agency to monitor financial institutions on the behalf of bank clients.
Laissez-faire
A frequently expressed view on banking regulation is laissez-faire. A bank should be left
alone to prosper and fail like any other private enterprise; the government should exercise
minimum regulations and certainly not protect banks in times of failure.
This has been the prevailing official policy at many times in history, for example, in the
nineteenth and early twentieth centuries. Some forms of this view also affected banking
policy prior to the crisis starting in 2007. However, such a laissez-faire approach to banking
is not credible.
The reason is the externalities arising from bank failures. When the authorities are faced
with large losses being imposed on society as a result of a banking crisis, they have no
choice but to act because the political pressure on the government becomes unbearable.
If it refuses to act, the government faces punishment by the voters. Just three examples
discussed elsewhere in this book are the crises of 1866 in the UK, of 1907 in the US, and of
Argentina in the early 1990s. Deciding not to regulate the financial sector is not a credible
option for the authorities. Being forced to intervene in times of crisis without adequate
preparation is a worst-case outcome. It is better for the authorities to regulate the finan-
cial sector and be prepared for the eventual crisis.
Macro- and micro-prudential regulations

The objectives and scope of financial regulations range from preventing or coping with a
systemic crisis to protecting small retail clients – the widows and orphans, as they often
called. It is now common to refer to those two separately as macro-prudential and micro-
prudential regulations, in the terminology of Andrew Crockett in 2000, then the General
Manager of the Bank for International Settlements (BIS).
Macro-prudential refers to regulations and policies designed to protect the financial
system in its entirety. This may mean something as extreme as coping with a systemic cri-
sis or more commonly the setting of minimum bank capital.
By contrast, micro-prudential regulations are concerned with protecting the individual
small clients of financial institutions. This may include fraud prevention, rules on what
information to provide to savers or the design of mortgage contracts.
Micro- and macro-prudential regulations are quite different from each other, in both
motivation and implementation. The supervisors working on either objective often work
in distinct parts of the same agency or might even belong to different organisations. This
means that the intersection between macro- and micro-prudential supervisors is often
limited.
In some cases the objectives of both coincide, but direct conflict is not uncommon.
For example, a micro-prudential regulator may prefer to prevent an individual financial
institution from failing in order to protect its clients, whilst the macro-prudential regula-
tor would want to see it fail so that moral hazard is minimised.
240
Therefore, it is an open question whether macro-prudential and micro-prudential regu-

lators should work for the same agency. The most recent approach in the UK is to hive the
macro-prudential off to the BoE, leaving the micro-prudential with the Financial Services
Authority (FSA). In the European Union (EU), micro-prudential regulations are by and
large national concerns, while macro-prudential regulations are increasingly the domain
of the EU. In the US, most micro-prudential regulations are in the hands of the individual
states, while macro-prudential regulations are mostly in federal hands.
Implementing macro-prudential regulations

There are various different ways the authorities can implement macro-prudential regu-
lations. They may restrict or prevent certain activities, control day-to-day risk-taking or
resolve a crisis once it is underway.
Activity restrictions is where the authorities limit what financial institutions can do. A
well-known example is the Glass–Steagall Act in the US in 1933 where banks were split
into investment banks and commercial banks. More generally, banks are often restricted
by how much they can lend to any individual borrower, perhaps 10% of the bank’s overall
assets.
Some regulations aim to control day-to-day risk in financial institutions. The Basel
Accords discussed below are a typical example of this. Here the objective is to ensure that
a financial institution does not behave in an excessively risky way.
Finally, macro-prudential activities also relate to the resolution of crises. This may take
the form of LOLR, direct bailouts or even the taking over of financial institutions by the
government. The authorities therefore need to be prepared, have a detailed understand-
ing of the financial system and its risks and even require living wills.
13.1.1 Challenges in banking regulations

The financial system is very complex, and authorities face many challenges in regulating
and supervising finance. Poor regulations can impose unnecessary costs, create perverse
incentives, reduce transparency and even increase risk.
Endogenous risk
Some prudential regulations, especially those addressing risk-taking, can directly increase
endogenous risk. This happens exactly because the regulations aim at preventing excessive
risk-taking by banks, thereby preventing large losses or even bankruptcies. This is some-
times referred to as smoothing the road. Reducing excessive risk is a laudable goal but can
be difficult to implement in practice, and regulations aiming at containing risk taking may
have the perverse consequence of actually increasing risk. There are several reasons why
this may happen.
Risk is very hard to measure. From a statistical point of view, it involves volatility clus-
ters, fat tails and non-linear dependence, which means that the basic statistical problem
of risk forecasting is quite hard. Often, these risk forecasts are then input into complicated
models, where, because of non-linearities, risk can be amplified in a way that is hard to
detect. In addition, banks have an incentive to under-report risk, and even worse, some
employees of banks – the traders – have even stronger incentives to under-report risk.
241
If the road is smooth, and risk is perceived as low, it creates incentives to take more risk.
Because after all, if everything is safe, what is wrong with a little bit more risk? The prob-
lem is that such risk-taking is not immediately visible but is seen only much later. For
example, it was decisions taken in the supposedly low-risk environment between 2003
and 2007 that created the conditions for the subsequent crisis. From a statistical point of
view, it is impossible to detect such hidden build-up of risk.
Smoothing the road is pro-cyclical, encouraging banks to take too much risk when
things are good and too little risk when things are bad.
Incentives of supervisors
A particular problem arises because of the incentives of banking supervisors. When the
banking system is functioning well without headline failures, the supervisors are unlikely
to get much credit whilst bankers and even politicians complain about excessive regula-
tory burdens on this profitable economic activity. If then a big failure occurs, the head of
the supervisory agency may face acrimonious hearings in the country’s parliament and be
pilloried in the press. After all, the supervisor had all the information about the bank but
did not act on it to prevent failure.
There is a danger that the incentives of supervisors are to prevent failure at all costs
and, hence, for the supervisor to become too risk-averse. This means that the incentive
problem of the supervisor is inverse to that of the banker.
It is therefore necessary to have some mechanisms in place to prevent excessive super-
visory risk aversion. One way to do so is by performing cost–benefit analysis on regula-
tions. Unfortunately, that is quite difficult to do in practice.
This problem leads to a particular form of pro-cyclicality. During upswings, regulations
become increasingly lax, amplifying the boom, and after a crisis they become excessively
strict, magnifying the downturn. There are clear signs of this in the current cycle.
‘Tick-the-box’ and legal approaches

After the Icelandic banking system collapsed in October 2008, the head of that coun-
try’s banking supervisor was asked in a TV interview what was the purpose of the super-
visor. He replied ‘to ensure the banks don’t break the law’. This answer is problematic,
because it is important to make a distinction between the spirit and the letter of the
law. The objective of banking regulations is not to ensure banks do not break the law,
rather that they do not behave in a way that harms society, and help in economic
development.
This means that there is a danger of an excessively legalistic or formulaic approach
to banking regulations, often referred to as tick-the-box regulations. In the case of the
Icelandic crisis, as documented by its parliamentary Special Investigation Commission
(SIC) report, when some banks wanted to increase their equity, bank A would sell new
equity to bank B, and then do a contract for difference, that is hedge the exposure of bank B.
Then, bank B would do exactly the same transaction with bank A. From a strict tick-the-box
legal point of view, it looked as if both banks had significantly increased their equity, with
the end result that the banks’ capital ratios looked strong. However, in practice, this bank
capital was purely illusionary, and while legal, did not afford any protection. Even worse,
242
it gave the appearance of protection, encouraging market participants to engage with

these banks as if they were safe. We see one manifestation of this problem in Figure 16.4
in Chapter 16.
This is an example of the conflict between principle-based regulations and tick-the-
box-based regulations. The latter are much easier to implement and often end up being
the default approach. Different agencies often have very differing views on this topic. For
example, in the US, the Securities and Exchange Commission (SEC) takes a legal approach
to regulation, whilst the Federal Reserve System (Fed) prefers a more principle-based
approach. In the ongoing crisis, the SEC has seen responsibilities transferred to the Fed,
because the latter has been seen as a more effective agency.
Transfer of responsibility to government

Any regulation of banks transfers responsibility for the banking system to the authorities.
The supervisors get confidential information about bank activities and powers to prevent
illegitimate activities. This implies that if banks fail, the authorities are partly to blame,
which may mean that the authorities are compelled to use taxpayers’ money to sort out
the crisis.
The banks, of course, fully know this and, therefore, are incentivised to behave in a
way that internalises the possibility of sharing the burden with the government. In other
words, because banks are supervised, they have an incentive to take on more risk than
otherwise.
This means that it is important to consider the secondary impact of regulations. It is
not enough to identify a particular problem and remedy that with regulations. Rather, all
the secondary consequences, such as how the proposed regulations change bank behav-
iour and the impact on the relationship between the government and the banking sys-
tem, need to be considered.
Perverse consequences of regulation

Financial regulations change the behaviour of banks, usually in a positive way; risk is
reduced and the system becomes more stable. In some perverse cases, the outcome can
be the opposite. This may happen because regulations drive risky activities under the
radar. The banks continue as before, but with less oversight. An example of this is shadow
banking.
Often, the banks’ avoidance behaviour takes the form of financial innovation, where
new and more complex financial instruments are created in response to regulation. Such
instruments may be beneficial but just as easily could increase complexity without any
discernible benefit, creating a form of dead weight loss. A clear example of this is capital
structure optimisation. This may in turn lead to additional regulations, causing a ratchet-
ing effect between regulation and bank activity, as seen in Figure 13.1.
Finally, regulations may move profitable and risky activities to other jurisdictions, leav-
ing the country implementing the regulations worse off.
Two early examples of the unintended consequences of financial regulations is regu-
lation Q in the US and the eurodollar market. A more recent case is the emergence of
shadow banking.
243
Avoidance incentive
Financial innovation
New regulation
Figure 13.1 The circle of financial innovation and regulation
Example 13.1 Regulation Q
Regulation Q was a US regulation limiting the interest rates that banks could pay on
deposits. This created incentives to create a parallel banking system, money market
mutual accounts, where market interest rates could be paid.
Example 13.2 Eurodollar market
The Eurodollar market first came into being in the 1950s when the Soviet Union’s oil
revenue – all in US dollars – was being deposited outside the US in fear of being fro-
zen by US regulators. This resulted in a vast offshore pool of dollars outside the con-
trol of US authorities, primarily held in Europe, hence, the term Eurodollar. This was a
major contribution to London becoming a world-leading financial centre.
Regulatory capture
In an ideal world, the supervisors are concerned only with the safety and soundness of
the banking system. In practice, there are often many other reasons why the government
chooses to regulate the banking system. It may have altruistic ulterior motives, such as requir-
ing banks to provide unprofitable banking services to disadvantaged sectors of society, or the
government may want to have national champions. Bank lobbying is also quite strong and
aims at creating banking regulations that favour the incumbents, discouraging entry into the
banking system and providing protection for banks’ profits and even the odd bailout.
In this case, banking supervisors may end up being captured by banks. This means that
the supervisory agency no longer works for society; instead, in effect, it works in the inter-
est of the banks. Such regulatory capture can happen for many reasons, for example,
when banks recruit staff out of supervisory agencies at vastly higher salaries, which might
give the staff an incentive to behave in a way that encourages private-sector recruitment.
In other cases, the banks may bypass the supervisor by going directly to the politicians,
who after all are in charge of the supervisor. Supervisory agencies are often accused of
regulatory capture, but this can be hard to verify. For example, it is often claimed the SEC
has been captured, but that case demonstrates the difficulty in distinguishing between
deliberate laxness in enforcement, incompetence or regulatory capture. A better example
was the supervisor for the S&Ls, as we discussed in Section 7.5.2.
244
13.2 Bank capital
Resource problems
Supervisory agencies have a serious resource problem compared to the banks they are
supervising. The government is likely to pay much less than the banks and have fewer staff
members, meaning the authorities can be seriously outgunned when dealing with the
banks. Any well-performing supervisor is subject to poaching from the private sector, and
the authorities often have real difficulty holding onto staff.
Summary
The challenges in banking supervision are indicative of how hard the problem is. It is not,
and will not be, possible to create anything resembling a perfect regulatory structure;
instead, we have to do our best and have multiple safeguards in place so we can simulta-
neously protect society and avoid excessive risk aversion.
Because each bit of regulation changes behaviour, we may need further regulation
to address the newly created problem, and so forth. While such cycles are inevitable,
they are also a sign of poor regulatory design. Regulations often have unforeseen conse-
quences, and we should consider the regulation of the financial system holistically, tak-
ing into account the entire body of regulations and the financial system and not looking
at individual components in isolation. That is not easy. Regulators are under constant
demand from lobbyists, assorted pundits and politicians, generally demanding action on
the specific point, disregarding the need for an integrated approach to regulations.
13.2 Bank capital

A direct way to improve the resilience of the banking system is bank capital, consisting of
certain financial instruments, most importantly equity but also other instruments consid-
ered to be like equity. Bank capital has two main purposes:
1 To protect a bank against unexpected losses. After all, if the losses are expected, the
bank should just provide for them or write the losses down. The higher the capital, the
more protection the bank has and the less likely it is that the bank fails.
2 To restrict how much risk a bank can take by limiting leverage. Higher capital relative
to assets lowers leverage and risk.
Equity
Consider a simplified balance sheet of a bank.
Assets Liabilities
Assets (loans) Liabilities (deposits)

Equity
The bank has assets (loans), as well as liabilities (deposits). The difference between
these two is net worth or equity:
equity = assets - liabilities (13.1)
245
Equity is the amount of money the owners have invested in the bank, and has no direct
connection with the market value of a bank.
We illustrate the calculation of bank equity by Example 13.3.
Example 13.3 Equity
A bank started five years ago. Assuming the original stock price was 1000, and there
are no dividends or taxes, profits in years 1, 2, 3 and 5 were 100 in each year, while
the loss in year 4 was 250. In this case the shareholders’ equity is
1000 + 100 + 100 + 100 + 100 - 250
There are several different types of equity, for example tangible equity which is cash
raised from shareholders or retained through earnings. Those different types of equity are
a function of the underlying accounting rules, local traditions and national law.
Capital
The concept of capital is broader than that of equity and includes a range of instruments,
so that some of a bank’s liabilities may be a part of its capital:
capital Ú assets - liabilities
It is useful to start with a simple balance sheet of a bank. Suppose we have two cat-
egories of assets, a riskless asset, perhaps a government bond, denoted by A1 and risky
corporate loans denoted by A2. Then the total assets of the bank are A = A1 + A2. Capital
is denoted by C.
The ratio of capital to assets is called the capital adequacy ratio (CAR).
Assets Liabilities
Low-risk assets (A1) Capital (C)

High-risk assets (A2) Non-capital liabilities
Definition 13.2 Capital adequacy ratio
C
CAR = Ú a
A
where a is the minimum CAR specified in banking regulations.
In many cases, the capital ratio is risk-sensitive, so that low-risk assets contribute less to
the denominator than high-risk assets. In this case, the CAR might become
C
CAR = Ú a, w1 6 w2 (13.2)
w1A1 + w2A2
where w1 and w2 are risk weights.
246
13.2 Bank capital
Capital instruments
The most basic form of capital is (common) equity, but certain liabilities, known as capital
instruments, can also be a part of capital. These include preference shares, subordinated debt,
long-term bonds and hybrid instruments which are bonds that can be converted into equity.
Criteria for capital

Not all capital instruments are created equal, and the more equity-like an instrument is,
the better protection it provides. Capital instruments are judged by four different criteria.
First, they need to absorb losses on a going-concern basis, allowing an issuer to avoid
liquidation in times of stress. Second, they need to be relatively permanent so that they
provide a buffer against losses for an extended period of time. Third, they have to allow
for sufficient freedom of action, or discretion over the amount and timing of payments
made on the instrument, and there should be few, if any, covenants restricting the issuer’s
freedom of action. Finally, the instruments need to afford protection to general creditors
in case a bank defaults.
13.2.1 Pro-cyclicality
An important problem with banking and capital is pro-cyclicality.
Definition 13.3 Pro-cyclicality A term that refers to how some economic quan
tity relates positively to economic fluctuations. It is the opposite of countercyclicality. See
Figure 13.2 below for an example.
When the economy is doing well, banks have ample funds which they seek to lend out,
but often find that high-quality borrowers have all the credit they need. Banks therefore
start lending to increasingly low-quality borrowers, often in the real-estate sector, creating
a real-estate bubble. When eventually the bubble bursts and the economy takes a turn
for the worse, banks find themselves faced with large amounts of defaults and sharply
curtail lending, often referred to as a credit crunch. This makes asset prices drop even fur-
ther. This means that bank lending amplifies the business cycle and is hence pro-cyclical.
Pro-cyclicality and capital regulations

The presence of bank capital regulations amplifies this inherent pro-cyclicality in bank-
ing. When times are good, asset values are high, the CAR is high and the bank seems to
employ a low degree of leverage.
Example 13.4 Worked example
Start with the definition of equity from Equation (13.1) and assume equity = $12
and assets = $100. Suppose the authorities impose a minimum CAR of a = 8%. The
CAR is then
$12
= 12% 7 a (13.3)
$100
➨
247
In this case, the bank is comfortably exceeding the regulatory CAR and employing
leverage of
$100
= 8.3
$12
If the economy takes a turn for the worse and the bank suffers losses of 3% of its
assets, that is $3, the CAR becomes
$12 - $3
= 9.3% 7 a (13.4)
$100 - $3
and the bank is closer to violating its regulatory constraint. Now the leverage has
increased to 10.7.
Since the bank prefers to maintain a buffer above the minimum 8%, it will do what it
can to increase its CAR, and the most obvious way to do so is to reduce leverage and sell
off some of its more risky assets.
This means that capital constraints increase the pro-cyclicality of bank capital and
bank leverage. For example, before 2007 banks were excessively free with their lending,
but after the crises started, they sharply curtailed lending, especially to the politically
sensitive SME sector.
Risk sensitivity
The problem of pro-cyclicality is made worse if the capital ratio is risk sensitive because
the same asset would have lower weights at the top of the cycle than at the bottom of the
cycle, amplifying the amplitude. This can be seen in Figure 13.2.
This can be seen clearly from the CAR in (13.2) and Example 13.4. Suppose A2
represents high-risk real-estate lending. At the top of the cycle most signs point to the risk
in real estate being low, and hence w2 will be low, keeping the CAR high. If the bubble
then bursts, real estate becomes more risky and w2 increases sharply, causing the CAR to
fall. The impact can be seen Example 13.5.
Top of credit cycle Bottom of credit cycle

6
5
Level of credit
4
3
2 No risk sensitivity
Risk sensitive capital
1
1 2 3 4 5 6 7 8 9 10
Year
Figure 13.2 Credit cycle
248
13.2 Bank capital
Example 13.5 Continuing from Example 13.4
Suppose before the shock the risk weight is 1, so w = 1. (13.3) therefore becomes
$12 $12
= = 12% 7 a
w * $100 $100
Because of the shock, the risk weight increases to w = 1.5, and (13.4) becomes
$12 - $3 $12 - $3
= = 6.1% 6 a
w($100 - $3) 1.5($100 - $3)
Because of the risk sensitivity, the CAR fell further, and the bank is no longer meet-
ing its regulatory constraint, causing it either to be shut down by the authorities or to
receive a bailout.
13.2.2 Issues with capital

Traditionally, capital levels were quite high, often around 40% in the nineteenth century,
if measured by shareholders’ equity. Over time, this number has fallen dramatically as
shown in Figure 13.3, using data from Alessandri and Haldane (2009), which shows the
capital ratios of banks in the UK and the US from 1880 until recently.
Cliff effects
The presence of a rigid capital ratio may have unexpected adverse consequences. If a bank’s
capital ratio falls below the minimum specified by banking regulators, currently 8%, the
bank most likely would be shut down, and as it gets closer to the 8% it draws increasing
unwanted scrutiny from the authorities. For this reason banks prefer to keep a significant
buffer above the minimum, generally around 12–13% before the crises from 2007.
Suppose a bank finds its capital ratio falling, needing to improve its CAR. The bank can
increase capital by selling equity or other capital instruments on the financial markets.
Because the bank is likely to be in this situation only in times of difficulty, this option is
likely to be quite costly, and if it coincides with a financial crisis, it may be impossible.
20% UK
US
10%
0%
1880 1900 1920 1940 1960 1980 2000
Figure 13.3 Core bank capital ratios in the US and UK

Data source: Chart 2 in Alessandri and Haldane (2009)
249
The bank can also reduce the amount of assets held or the riskiness of its portfolio by
selling risky assets and refusing to provide new loans. This process is termed deleveraging.
If only one bank is in difficulty, with the financial system generally in good shape, such
deleveraging is not all that problematic. However, if it happens during crises the situation
is rather different.
In this case, a general sell-off of risky assets throughout the financial system is likely to
be taking place. This means that the price of risky assets is likely to fall precipitously, sig-
nificantly below their fundamental values, and the bank may get only firesale prices. The
bank may also refuse to provide new loans and to roll over existing loans, perhaps leading
to a credit crunch. Those borrowers most likely to be affected are SMEs. This will lead to
further economic slowdown.
Both actions exacerbate the crisis, and cause an endogenous risk of vicious feedback
between worsening economic conditions and bank difficulties. This shows how fixed capi-
tal ratios can by themselves increase systemic risk.
How capital is used

In designing capital regulations, the authorities are left with the difficult question of
whether a minimum CAR has to be maintained at all times, or can be reduced in times of
difficulty. If capital cannot be allowed to fall below the minimum in any circumstances, it
does not afford protection.
This has been aptly demonstrated by Goodhart’s metaphor (Goodhart, 2009, chapter 8)
of a weary traveller arriving by train to an unknown town late at night. Seeing one taxi
outside the train station, the traveller asks the driver to take her to her hotel. The driver
responds that he cannot do so, and points to a sign on the wall saying ‘local regulations
require that at least one taxi be outside the station at all times’.
The issue of drawing down protective buffers comes up in every crisis. For example,
in their book on the 1907 crisis, Bruner and Carr (2007) relate the following anecdote. A
banker complained to John Pierpont Morgan, of the eponymous bank, about his reserves
falling to 20%. Morgan replied ‘you ought to be ashamed of yourself. Your reserve ought
to be down to 18% or 20%. What is a reserve for if not to be used in times like these?’
As we see below, the existing capital standards do not allow for relaxation of the CAR,
whilst the proposals for the next version allow for more flexible capital buffers.
13.3 International financial regulations: Basel

Historically, banks have operated mostly within a single nation state, with multinational
banks few and far between. This was enshrined in the Bretton Woods era when banks
were heavily regulated and their international activities discouraged. This meant that
banking regulations were mostly domestic, with little need for international coordination
in either regulation or supervision.
Over the past few decades, this has changed significantly and banking is now a truly
international business. When the financial system started opening up, the deficiencies
of a purely domestic approach to financial regulation became apparent, and two bank
250
failures in particular, those of Bankhaus Herstatt in 1974 and Banco Ambrosiano in 1982,
served as a wake-up call to the authorities, making it clear that if we want to have regula-
tions in the first place, they need to be international. This means that the home regulator,
where the bank has its headquarters, needs to have oversight of its banks’ activities every-
where in the world, and the host regulators, countries where the bank operates, need to
cooperate actively with the home regulator and other host regulators.
Supervise everywhere
There are two main reasons why regulation needs to be international. First, there have
been well-documented cases where a financial institution was regulated in one country
but misbehaved in another. This happened because the home supervisor was responsible
only for activities at home, while the host supervisors assumed the home supervisor was
in charge, and at the same time the different national supervisors did not communicate.
This suggests that it is necessary to have international regulations to ensure banks are
regulated wherever they operate, with supervisors cooperating on enforcement.
Regulatory arbitrage
The second problem arises if financial regulations and supervision are not harmonised
across jurisdictions, and financial institutions can shop for the jurisdictions with the most
lax regulations or supervision, a process called regulatory arbitrage. In the best case, this
leads to the most lax environment to be used, but in a worst case, countries may compete
by relaxing financial regulations in order to attract multinational banks.
13.3.1 The Basel Committee

As policymakers came to recognise the challenges arising from international bank-
ing, the governments of the G-10 countries in the late 1960s and early 1970s 1 set up
the Basel Committee on Banking Supervision (BCBS) as a new international organisation
tasked with designing international banking regulations. It is hosted at the Bank for
International Settlement (BIS), whose head office is in Basel, Switzerland, hence the
name. More recently membership has been extended.2 For an in-depth study of the BCBS
see Goodhart (2011).
Countries are represented by their central bank and also by the authority with formal
responsibility for the prudential supervision of banking businesses. Luxembourg has only
one member.
The Committee does not possess any formal supranational supervisory authority.
Rather, it formulates broad supervisory standards and guidelines and recommends state-
ments of best practice in the expectation that individual authorities will take steps to
implement them through detailed arrangements.
1
Belgium, Canada, France, Italy, Japan, Germany, Sweden, the Netherlands, the UK and the US, as well
as Luxembourg. Spain later became a member.
2
To include Argentina, Australia, Brazil, China, Hong Kong SAR, India, Indonesia, South Korea, Mexico,
Russia, Saudi Arabia, Singapore, South Africa, Sweden, Switzerland and Turkey.
251
At times, the Committee uses this common understanding to develop guidelines and
supervisory standards in areas where they are considered desirable. In this regard, the
Committee is best known for the Concordat and the Basel Accords.
One important objective of the Committee’s work has been to close gaps in interna-
tional supervisory coverage in pursuit of two basic principles: that no foreign banking
establishment should escape supervision; and that supervision should be adequate. The
Concordat, originally published in 1975 (in response to the Bankhaus Herstatt crisis), set
down the principles for sharing supervisory responsibility for banks’ foreign branches,
subsidiaries and joint ventures between host and home supervisory authorities. A major
review was made in 1983 following the failure of Banco Ambrosiano.
The most important part of the Committee’s work is the Basel capital accords, a set of
internationally harmonised rules for determining the adequacy of the capital of interna-
tionally active banks. The first Basel Accords, now referred to as Basel I, were proposed in
1988 and implemented in 1992, while their successor, Basel II, was proposed in the late
1990s and at least partially implemented from 2008. The Committee has already pro-
posed the next iteration, Basel III.
13.3.2 Basel I
A key motivation for Basel I was events in the early 1980s when banks competed vigor-
ously in making corporate loans. At the time, bank capital ratios of major banks in Europe
and the US were perhaps 8–10%, while Japanese banks operated with lower bank capital
and saw major gains in market share. This led to competitive pressures to lower bank
capital in Europe and the US. Regulators became alarmed and proposed Basel I with the
primary aim of specifying minimum capital for banks operating in the world’s major finan-
cial centres.
Basel I was broadly successful in achieving its designated purposes, raising capital levels
at a time when they were low and trending down.
Issues with Basel I

Basel I focused primarily on credit risk, employing crude discrete predetermined risk
weights. For example, while corporations may get a 100% risk weight, OECD governments
get zero risk weight. This means that a cash-rich corporation like Microsoft with its AAA rat-
ing might be weighted at 100% and OECD governments like Turkey, South Korea, Mexico
and Greece with their substantially lower credit ratings get a 0% risk weight. This lack of
risk sensitivity distorts economic decision-making by creating an incentive for banks to
undertake transactions whose sole purpose is to reduce capital requirements with no
commensurate reduction in actual risk-taking. The lack of risk sensitivity also impedes
effective supervision, as regulators, rating agencies and market participants tend to focus
on the capital ratios, which limit their information about a bank’s overall risk and capital
adequacy.
A different problem emerged because loans with maturity of less than one year did
not attract a capital charge. This meant that loans with a maturity of 364 days were
cheaper to make than loans with a maturity of 365 days. This is an example of cliff effects
252
in financial regulation that can often lead to surprising outcomes. In this case, 364-day
loans have become quite popular, replacing much longer-maturity loans, often with the
tacit understanding that they will be rolled over after 364 days. They do create, however,
unnecessary liquidity risk and therefore make the financial system more fragile and even
increase systemic risk.
The 1996 Amendment

The BCBS introduced market risk into the Basel I accord with the 1996 Amendment. This
introduced a Value-at-Risk (VaR) approach for the measurement of market risks, requiring
banks to report daily 99% probability 10-day VaR using their internal risk models audited
and permitted by the regulators: see the appendix for more details.
Minimum market risk capital is then the current VaR or the 60-day average VaR, which-
ever is the greater, times a multiplicative factor, plus an add-on. Banks are allowed to
violate their VaR 2.5 times a year, that is, their trading losses can exceed the VaR 2.5 times
a year.
The impact of the market risk amendment was quite small because for most banks
market risk is an order of magnitude smaller than credit risk and, therefore, only margin-
ally affects the banks’ capital.
13.3.3 Basel II
The BCBS embarked on a revision of Basel I in the mid-1990s, which came to be called
Basel II and was introduced to the world around the turn of the century. Since then,
extensive lobbying has taken place, delaying its implementation to 2008 in most relevant
jurisdictions. This means that Basel II reflects concerns and regulatory thinking from the
mid-1990s, so it was already out of date by the time of its implementation.
Basel II improves on the capital regulations in several ways. It substantially improves
the risk sensitivity of the minimum capital requirements. While market risk regulations
remain unchanged, an internal rating based (IRB) approach is introduced for the calcula-
tion of credit risk. Basel II also introduces operation risk.
In the IRB approaches, banks’ internal assessments of key risk drivers serve as primary
inputs to the capital calculation. Because the approach is based on banks’ internal assess-
ments, the potential for more risk-sensitive capital requirements is substantial.
Menu of approaches
While Basel I took a one-size-fits-all approach to capital calculations, Basel II more
explicitly recognises that banks are different. One reason is that Basel I applied mostly to
large international active banks, while the EU intended Basel II to apply to essentially all
banks in its member states, requiring a more nuanced approach for calculating capital,
taking into account that some banks are sophisticated and run their own models whilst
others engage in very basic activities, outsourcing the few needed calculations. Banks
can choose between three different approaches for calculating capital, ranging from
simple ratios to sophisticated internal modelling. The menu of approaches is demon-
strated in Figure 13.4.
253
Credit risk Market risk Operational risk
Advanced internal Advanced
Increased sophistication
ratings based measurement
approach Internal models approach
Foundation
Standardised
internal ratings
approach
based approach
Revised Standardised
approach Basic indicator
standardised
approach
approach
Figure 13.4 Menu of approaches for risk calculations in Basel II
The three pillars of Basel II

Basel II recognises that capital calculations are just one part of the regulatory process and
identifies two other areas of concerns, making up the three pillars of Basel II.
Definition 13.4 The three pillars of Basel II

■ Pillar I. Minimum capital requirement is based on the notion that a bank is able to
communicate its overall risk level via one number to the supervisors.
■ Pillar II. Supervisory review process is designed to ensure that this risk number is
generated in a satisfactory manner.
■ Pillar III. Market discipline depends on the communication of key statistics to the
community at large.
While the main focus of Basel II is still on refining the minimum capital requirements,
the attention has gradually shifted towards the pillars II and III as the current crisis calls for
a more stabilising regulatory regime.
Pillar II forms an integral part of the proposal. It obliges regulators to assess the qual-
ity of the risk models of individual banks, allowing the models to be more flexible with
respect to banks’ particular circumstances whilst encouraging closer cooperation between
the supervisor and the bank.
However, allowing regulators such flexibility raises the need for a mechanism ensur-
ing that pillar II is implemented uniformly across countries and that quality assessments
undertaken under its auspices are consistent across regulators.
Pillar III reflects the conventional wisdom that sufficient transparency in the market is
helpful in increasing stability. This, however, has been somewhat disappointing and has
not proven effective.
254
There are two main factors at work. First, the counterparties who are supposed to do
the monitoring share in the same hubris as the banks that are being monitored. We see
clear examples in a later chapter when we discuss the problem of toxic capital and assets.
Second, it is very hard to make sense of the information being put out. Aggregate sum-
mary numbers can hide a myriad of problems, especially in complex asset classes, while
the understanding of fine-grained disclosures depends on intimate knowledge of a bank’s
operations.
13.3.4 Criticism of Basel II

Basel II has come under considerable criticism, both during the early design stages and
subsequently for its inadequacy exposed during the crises starting in 2007. Danielsson
et al. (2001) submitted a paper to the Basel Committee in response to a call for com-
ments on the initial Basel II proposals. The main parts of their arguments were sum-
marised as:
‘The proposed regulations fail to consider the fact that risk is endogenous. V
alue-at-Risk
can destabilise and induce crashes when they would not otherwise occur.
Heavy reliance on credit rating agencies for the standard approach to credit risk
is misguided as they have been shown to provide conflicting and inconsistent fore-
casts of individual clients’ creditworthiness. They are unregulated and the quality of
their risk estimates is largely unobservable.
Financial regulation is inherent procyclical. Our view is that this set of propos-
als will, overall, exacerbate this tendency significantly. In so far as the purpose of
financial regulation is to reduce the likelihood of systemic crisis, these proposals will
actually tend to negate, not promote this useful purpose.’
Sovereign debt
It is quite common for countries to consider their own domestic currency sovereign debt
as riskless, often writing it into law. Even though very few countries are considered risk-
free, traditionally this may not have been that big a problem because governments in
difficulty with domestic debt payments simply resorted to inflating the debt away rather
than directly defaulting.
The Basel I accord follows in this tradition, considering the debt of OECD member
countries to be riskless, for the purpose of capital calculations. This means that any bank
holding OECD government debt did not have to hold capital against it. This was, of
course, a subsidy to the government, paid for by a tax on regular borrowers.
This practice is not that problematic so long as countries issue debt in their own cur-
rency, but with the establishment of the euro a new problem has emerged. Member
countries of the euro zone issue debt in euros. This means that they cannot resort to
inflating away their debt as they used to before the euro, a common practice in many
countries such as Greece and Italy. While Basel II does not necessarily consider sov-
ereign debt risk-free, the EU does in the case of ‘exposures to Member States central
government . . . denominated and funded in the domestic currency of that central
255
government’.3 This means all member governments of the euro zone issue risk-free
debt, at least for the purpose of bank capital calculations. Note that this does not mean
the governments can borrow at risk-free rates, but rather that loans to governments do
not attract a capital charge.
This has caused particular problems in the euro zone since it sent the signal that gov-
ernment debt was risk-free and in turn meant that banks did not consider the acquisition
of sovereign debt as speculative investments, but rather the byproduct of prudent treas-
ury operations.
13.3.5 Basel capital ratios

Under both the Basel I and the Basel II Accords, the CAR has to exceed 8%. The Accords
define two categories of capital, tier 1 and tier 2.
Tier 1 capital
The most important part of tier 1 capital is common equity. Depending on the jurisdiction,
there are two other types of tier 1 capital, disclosed reserves and preferred stock. These
instruments differ depending on each national regulator, but are always close in nature to
common equity, usually referred to as upper tier 1 capital. In the US, regulators prefer tangi-
ble equity, as it provides the most protection. This is disliked by European banks.
Tier 2 capital
Tier 2 capital contains a much broader range of instruments, and is the lesser of the two
ratios. Note that just like in tier 1, there may be important differences between jurisdic-
tions as to what is allowed as a part of tier 2.
A relatively unimportant part of tier 2 capital is undisclosed reserves, revaluation
reserves and general provisions. The key instruments are subordinated debt and hybrid
instruments. The former has a flaw in that it provides protection only in case of default
and not prior to bankruptcy, when it is most needed. The second instrument had an
important deficiency that manifested itself during the crises from 2007, where the protec-
tion thought to be afforded by the hybrid capital instruments turned out to be illusionary.
Capital adequacy ratio

The Basel capital ratio is the ratio of capital to risk-weighted assets (Definition 13.5).
Definition 13.5 Basel capital ratio
C = T1 + T2
a i = 1w i Ai
CAR = N
Ú 8%
where T1 and T2 denote tier 1 and tier 2 capital, respectively, w1 = 0 is the risk weight on the
safe asset A1 and wi, i 7 1 are the risk weights on the risky assets. N is the number of assets.
3
Article 89(1)(d) of the EU capital requirements directives (CRDs) (amended by Directive 2009/111/EC
or ‘CRD II’), and Annex VI Part 1 paragraph 4.
256
13.4 Summary
For larger banks, the risk weights are calculated by using internal models, whereby
the bank uses historical data and statistical models to forecast the risk of its assets. These
models need to be approved by the supervisor.
Capital calculations
Recall the discussion from Section 13.2 above. We replicate it here in the table below,
having added in tier 1 and tier 2.
Assets Liabilities
Riskless assets (A1) Capital (C = T1 + T2)

Risky loans (A2) Non-capital liabilities
Example 13.6 Capital ratio calculation
Suppose a bank holds $1 billion in sovereign debt and $2 billion in corporate loans,
whilst it has tier 1 capital of $50 million and tier 2 capital of $200 million. Its capital
ratio is then:
50 + 200
= 12%
0 * 1000 + 1 * 2000
13.4 Summary
The focus of this chapter has been on banking regulations. Banking has always been one
of the most regulated parts of the economy because of the externalities induced by bank-
ing, whereby the downside from failure can seriously affect society while the bankers
enjoy the upside.
However, financial regulations carry with them serious side-effects. They can result in
the hiding of risk-taking, prevent socially desirable activities or spur undesirable financial
innovation.
One of the most common ways to regulate banks is by capital requirements, and under
current regulations banks are required to have capital to risk-weighted assets of at least 8%.
Traditionally, financial regulations were the purview of the nation state, but following
the post-Bretton Woods deregulation of the global economy, banking is increasingly global,
and banking regulations have followed with first the Basel I and then the Basel II accords.
Appendix: Value-at-risk
The Basel committee introduced an amendment to Basel I in 1996, stipulating that a
new category of capital requirements was to be created, those arising from market risk.
Fundamental to the amendment is the concept of VaR defined as a loss threshold: for a
given probability, p,
Pr [- losses Ú VaR] = -p
257
For more details see, for example, Danielsson (2011). The probability specified by the
Committee is 99% and the holding period is 10 days. Capital arising from trading activi-
ties is defined as
Capital = 3 VaR99%
10-day + constant
Because it is hard to obtain a proper 10-day VaR, the Committee allows banks to use the
square-root-of-time rule and do the calculation as
99%
Capital = 3210 VaRdaily + constant
This means that the bank capital for trading activities is based on a multiple of the risk of
an event that happens 2.5 times a year on average, the VaR99% daily.

1 What is the difference between regulation and supervision?
2 What are micro-prudential regulations and macro-prudential regulations? When might

the objectives of these two types of regulations coincide and when might they conflict?
3 What are the main challenges in banking regulations?
4 What is the difference between bank capital and bank equity?
5 What are the main two purposes of bank capital?
6 What is pro-cyclicality, and to what extent are regular banking activities pro-cyclical?
7 Outline the specific mechanism by which bank capital and capital regulations are pro-cyclical
8 Can bank capital that cannot be drawn down in times of need still fulfil a useful function?
9 What are the main objectives and achievements of the Basel Committee?
10 What is your view on risk weighing assets for CAR purposes?
11 What you think about the assumption in Basel that sovereigns are risk free?
12 Consider a bank that has $100 in cash, no other assets, and no liabilities. The bank
operates in a country with a 8% minimum CAR requirement
(a) Suppose the bank raises $900 in deposits and invests all of its cash on hand in risk
free government bonds. What will its CAR be?
(b) Suppose instead the bank invests $800 in a AA rated bond that attracts a risk weight
of 1, with the rest in the government bond. Is the bank meeting its CAR requirement?
(c) If the bank instead allocates $500 to a BB rated bond with a risk weight of 2, $300
to the AA bond, with rest to the government bond. Is the bank meeting its CAR
requirement?
(d) The bank invests $800 in a AA rated bond that attracts a risk weight of 1, with the
rest in the government bond. Suppose the markets fear inflation and interest rise
by 4% and as a consequence the market value of all bonds drops by 4%. Is the
bank meeting its CAR requirement?
(e) Comment on the last answer from the point of view of the impact of leverage on
the bank’s capital structure.
258
References
References
Alessandri, P. and Haldane, A. G. (2009). Banking on the state. Mimeo, Bank of England.
Basel Committee on Banking Supervision (1996). Amendment to the Capital Accord to Incorporate
Market Risks. www.bis.org/publ/bcbs24.pdf.
Bruner, R. F. and Carr, S. D. (2007). The Panic of 1907: Lessons Learned from the Market’s Perfect
Storm. John Wiley & Sons.
rr000921b.pdf.
Danielsson, J. (2011). Financial Risk Forecasting. John Wiley & Sons.
Danielsson, J., Embrechts, P., Goodhart, C., Keating, C., Muennich, F., Renault, O. and Shin,
H. S. (2001). An academic response to Basel II. www.bis.org/bcbs/ca/fmg.pdf.
Dewatripont, M. and Tirole, J. (1994). The Prudential Regulation of Banks. MIT Press.
Goodhart, C. (2009). The Regulatory Response to the Financial Crisis. Edward Elgar, Cheltenham, UK.
Goodhart, C. (2011). The Basel Committee on Banking Supervision: a History of the Early Years
1974–1997. Cambridge University Press.
Kohn, M. (1999). Early deposit banking. Mimeo, Dartmouth College, Hanover, NH.
Schumpeter, J. (1942). Capitalism, Socialism and Democracy. Harper, New York.
259
14 Bailouts
We could define a financial crisis as an event where the public is called on to bail out
privately owned financial institutions. The bankers may tell the government that if a
bailout is not forthcoming immediately, the country’s financial system will collapse.
This leaves the political leaders in a difficult situation. Should they take the gamble
that the bankers are exaggerating? Should they believe the bankers and bail them out?
Bailouts create moral hazard, as defined in Section 7.2. This arises when the gov-
ernment is unable to properly price the bailout, something it usually cannot do. If
policymakers appear to be willing to bail out the banks, the banks are encouraged to
take on more risk in the future. If creditors believe that a financial institution will be
bailed out, they will factor that in when pricing credit extended to the institution. In
other words, if a bank can be expected to be bailed out, it can borrow more cheaply
than if bailing out is unlikely. This also reduces the incentives of creditors to monitor
and discipline the bank. This means that bailouts provide perverse incentives and
increase financial fragility.
However, if the authorities do not bail out the banks, the country risks a systemic
crisis. The total breakdown of payments systems and the inability of individuals and
firms to receive funds and make purchases is not pleasant to contemplate. During the
Great Depression, the last time we faced a global financial crisis, it may have been
possible for the economy to function without a banking system. People had access
to basic necessities such as food without banks. In the modern, highly linked world,
financial services are much more interwoven into the economy, and very few peo-
ple or companies could survive long without banking services. This means that the
Chapter 14 Bailouts
authorities have no choice but to do whatever they can to prevent a widespread bank-
ing collapse.
Although it is hard to justify using taxpayers’ funds to support some of the wealthi-
est segments of society, bailouts have been the usual response in crises past, and are
increasingly likely in the future. Of course, policymakers do not like being put in this
position. They may proclaim before a crisis happens that they will never do bailouts,
which if credible would constrain risk-taking. However, this threat is not particularly
credible. Therefore, it is better for the authorities to be prepared for crises, recognise
their likelihood, and have the necessary financial and legal structures in place. If well
prepared, they could aim to impose stringent conditions on the banks, protecting
taxpayers and minimising moral hazard.
Much of the damage caused by the crisis starting in 2007, and many a previous cri-
sis, is due to the authorities being woefully unprepared, because of the complacency
created by years of apparently benign economic conditions.
This suggests that when it comes to regulating the financial industry and provid-
ing bailouts, the interests of the taxpayers are at a serious disadvantage. Not only can
the banks exert targeted lobbying at the government to water down any regulations,
but the technical ability of the government to respond effectively is limited. Senior
government ministers and their advisers are unlikely to understand the underlying
problem in detail. How is the government to know whether a banker is bluffing when
saying ‘if you do not bail me out this afternoon the financial system will collapse’?

This chapter directly relates to Chapter 5 (the central bank) and Chapter 7 (bank-
ing crises). It also connects to the crisis discussion in Chapter 17 (the ongoing crisis:
2007–2009 phase) and Chapter 18 (ongoing developments in financial regulation).
Key concepts
■ Bailouts
■ Liquidity support
■ Lending of last resort
■ Moral hazard
■ Guarantees and bubbles
■ Bailouts in the crisis from 2007
■ The ECB and European banking problems

There are few texts addressing bailouts in a comprehensive way. The first influential
analysis of how the government should deal with financial crisis was made by Bagehot
(1873), and his proposals continue to shape policy response to this day. More recently,
Goodhart (1999) and Capie and Wood (2006) studied lending of last resort (LOLR)
in a more modern context. Little work exists on systematically analysing bailouts, and
this chapter therefore contains more primary analysis than many other chapters in
this book.
262
14.1 Successful and unsuccessful bailouts

K Capital
N Number of institutions
r Returns
A, B Parameters
E Random variable
14.1 Successful and unsuccessful bailouts

Two bailouts in recent history, illustrated in Figure 14.1, stand out as examples of how to
do them right and how to do them wrong.
Sweden, 1992
Perhaps the most highly regarded example of a government bailout of the financial indus-
try is from Sweden in 1992. Sweden employed the approach of good bank – bad bank,
with the government taking over dodgy bank assets in return for significant shareholdings.
Since the government held on to the bad assets, some of which turned out to be more
valuable than envisioned, the eventual cost to the Swedish government was found by
Moe et al. (2004) to be 3.9% of GDP, significantly below the initial cost of the bailout.
Government debt increased sharply following the crisis, as seen in Figure 14.1, reaching
73%, but falling steadily to 38% in 2011.
100%
80%
60%
40% Sweden
20% Ireland
0%
1990 1995 2000 2005 2010
(a) Total (domestic plus external) gross general government debt/GDP
1000% Sweden
800% Ireland
600%
400%
200%
0%
1990 1995 2000 2005 2010
(b) Total (public plus private) gross external debt/GDP
Figure 14.1 Irish and Swedish debt

Data source: Eurostat and www.reinhartandrogoff.com/data/browse-by-topic/topics/9/
263
Chapter 14 Bailouts
Two things from this episode stand out. First, the bailout never threatened the solvency
of the Swedish state; Standard & Poor’s rating on foreign debt dropped from AAA to AA + ,
recovering quickly back to AAA. Second, shareholders lost out since there was no bailout
of bank owners.
Ireland, 2008
The global crisis that started in 2007 was especially damaging to those countries that had
rapidly growing banking systems, like Ireland, paid for by foreign borrowing. The amounts
involved were significant, and growing rapidly as seen in Figure 14.1, where total (public plus
private) gross external debt/GDP has grown by over 12% a year over the past two decades to
over 10 times GDP. This is how Ireland got the funds to grow its banking system to become the
third largest in the European Union (EU) relative to GDP by 2007, as indicated by Figure 1.2.
During the bubble years, this created a virtuous cycle between borrowing abroad, lend-
ing the money out domestically and increasing valuations of domestic assets. Because the
valuations were based on demand fuelled by foreign money, once the money stopped
flowing, prices collapsed, causing a banking and economic crisis.
By 2008, the Irish banks were facing significant difficulties, and unable to roll over
loans. After Lehman collapsed, all funding dried up. The government at the time faced
two choices: it could let the banks default, seriously disrupting the domestic economy but
letting the bulk of the losses fall on foreign creditors; or it could provide guarantees for
bank obligations in the hope that this would reassure foreign creditors so that they would
continue providing funding, thus preventing a banking collapse. The government was pres-
sured to adopt the second option by major creditor countries, such as France, Germany
and the United Kingdom (UK), that wanted to shield their banks from the Irish problems.
The correct decision depended on the nature of the problem. If the problem was a
short-term liquidity crisis, with the Irish economy basically sound and asset valuations real-
istic, the right choice was clearly a bailout, and the government proceeded on this basis,
providing a state guarantee for all bank obligations, initially without taking over sharehold-
ers’ equity. This transferred the private bank debt to the sovereign. The government may
have analysed it in a way similar to the Copeland second generation (2G) currency crisis
model discussed in Section 12.3, considering itself to be in the multiple equilibria region of
the model, and hence being able to determine the outcome by a firm commitment.
Because of the very large scale of banking activities in Ireland, and the inability of the
Irish government to print euros to finance a bailout, substantial losses on the guaranteed
loans called into question government finances. This made the situation much more com-
plex. The debt assumed by the government has proved to be so large that it dominated
all other sources of government debt, creating serious concerns over the government’s
ability to pay. As a result, the guarantees did not restore confidence in the Irish financial
system, because concerns over bank security were replaced by concerns over the state’s
ability to meet its obligations, making it endogenous to the problem. Figure 14.1 shows
that government debt/GDP sharply increased from 25% in 2007 to 109% in 2011, beyond
its debt tolerance, so that Ireland had to be bailed out by the EU. The Irish taxpayers will
have to pay for the bank bailout for a long time to come, suffering a serious contraction in
economic performance and austerity as a consequence.
264
14.2 The historical origins of Lending of last resort (LOLR)
With the benefit of hindsight, the decision to bail out the banks in the way it was done
appears to have been a mistake. If the government was under so much pressure from the
EU that it had no choice but to do the bailout, presumably it could have demanded conces-
sions from the EU, sharing the burden. Although such a confrontational approach would
have been unpopular with other European countries, it appears that the government’s
negotiating position was quite strong. If the government had refused to bail out the banks,
they would have defaulted, but most of the losses would have fallen on foreign creditor
banks. The government could then have implemented a much cheaper bad bank – good
bank model, focusing on continuing to provide uninterrupted banking services but not
providing a blanket guarantee. It may well be that the Irish bailout will be considered as a
textbook illustration of a cardinal rule: do not attempt bailouts you cannot afford.
14.2 The historical origins of Lending

of last resort (LOLR)
A common theme runs through the past two centuries of financial crises. Before a crisis
the authorities proclaim they will not bail out private institutions, but once a crisis is
underway one of two things happens. Either the government gives in immediately, or
it resists, leading to a big economic crisis, with the government subsequently vowing to
provide liquidity in future crises. A transformative event for how the government engages
with financial crises happened in London in 1866.
14.2.1 The Overend and Gurney crisis of 1866

One of the most respected financial institutions in London in the early 1860s was Overend
and Gurney (O&G), called the ‘greatest instrument of credit in the Kingdom’ on 30 Sep
tember 1866, the day after its collapse. It was the largest domestic bank in the UK and the
world in its day. Our discussion of the downfall of O&G follows Elliot (2006).
Its business was based on commercial lending, taking the form of discounting, where a
firm would sell at a discount a note promising to pay a certain amount in the future. The
difference between the two is an interest rate. Modern repos work in a similar way. This
was the major form of lending at the time.
The development of a market for such notes was a major reason why the UK became
the first industrialised country in the world, because budding British capitalists found it
easier to raise money than their counterparts anywhere else in the world. This market
might be called venture capitalism today. At the time, one of the riskiest and most prof-
itable types of ventures was shipping, and the development of new ships and shipping
lines was the high-tech industry of the mid-nineteenth century. For most of its existence,
O&G did not participate in such high-risk lending.
Financial institutions of the nineteenth century were principally partnerships, with the
notable exception of the Bank of England (BoE). This meant that the partners were liable for
all losses – at least in theory. O&G was founded by Quakers who, over time, wanted even
more profits than the immense profits already made and ventured into what now would
be called junk bonds – very high-risk lending. Soon some of the underlying ventures failed,
265
Chapter 14 Bailouts
but O&G did not recognise the losses on its books, instead carrying on as if it had no hidden
losses. When its situation became serious, it sold shares to the general public proclaiming
that the bank was safe and had no bad assets. At the time, this was legal and there was no
obligation for any firm to publish accounts or reveal its true financial situation.
In 1866, O&G collapsed in spectacular fashion. Walter Bagehot, then editor of The
Economist, wrote that the partners of the firm ran their business ‘in a manner so reck-
less and foolish that one would think a child who had lent money in the City of London
would have lent it better’.
In this case, the partners of O&G believed there would be a bailout by the BoE. A del-
egation of three bankers walked from the Bank to O&G, where a brief look at the ledgers
told them all they needed to know. O&G was broke. The Bank faced a delicate decision.
If O&G failed, there would be panic. If it was saved, the many other firms in the ‘finance’
game would also expect to be rescued. The Bank chose the former.
Because the Bank did not bail out O&G, nor provide liquidity support to other banks,
even refusing to grant loans against government securities, panic spread through the
banking system, and liquidity vanished. The market for otherwise safe assets like UK treas-
uries dried up as well.
It is not clear why the Bank took that view. Moral hazard was clearly a significant concern,
but other factors weighed on it. The BoE was a private institution competing with the likes of
O&G, and its future profits were likely to be enhanced by seeing important competitors fail.
The O&G partners were supposed to lose all of the assets in a bankruptcy. However,
this was not the case. One of the senior partners, and a member of the family, was a
Barclay who a few years later started a bank with the eponymous name. The last Overend
became the largest shareholder in that bank.
The partners of O&G eventually faced private prosecution because the government did
not feel they had done anything wrong; financial regulations and rules of conduct were
unknown at the time. The only crime they could be charged with was theft, which was
a narrowly defined concept. In the prosecution the O&G partners hired the UK govern-
ment’s senior lawyer to defend them, so that the same lawyer was representing both the
accused and the government who would judge them. They were acquitted.
14.2.2 Bagehot and the crisis of 1866

At the time, there were no established procedures for dealing with the failure of a big
bank. However, the disruption of this event turned out to be so costly that it was felt that
some steps would be needed to address the externalities associated with the failure of a
big institution.
In reaction to the O&G crisis, the BoE studied how it should respond to future crises, and
the editor of The Economist and future Bank employee, Walter Bagehot, published a White
Paper in 1873 on the topic, establishing the three principles of lending of last resort (LOLR):
1 The central bank should lend freely

2 At a penal rate of interest
3 On good banking securities.
266
14.3 What are bailouts?
Under the gold standard, the monetary arrangement in 1866, liquidity was gold bars.
A country had a limited supply of gold bars, so if a central bank provided liquidity to one
bank, it might not be able to provide it to another. Thus the central bank had to find a
way to ensure that only the banks that really needed liquidity got it. The penalty rate was
just a price to ration scarce liquidity so that it flowed towards the highest-value demand-
ers. The Bank Charter Act of 1844, also known as Peel’s Act, restricted the amount of
paper money that could be created (paper backed by gold), but it could be suspended in
an emergency, which happened in 1847 and 1857.
The BoE successfully dealt with the banking panics of 1878 and 1890 by following
Bagehot’s rules, while the panics of 1847, 1857 and 1866 led to acute financial crises
because the BoE hesitated to engage in LOLR. Bagehot’s analysis of the 1866 crisis, and
his recommendations for how the government should resolve crises and provide LOLR,
was the first modern analysis of financial stability and policy remedies. It has been hugely
influential and has shaped policy throughout the world ever since.
It established the doctrine of LOLR, enabling the UK to successfully deal with the crises
that came after O&G’s. The inability to provide the necessary liquidity support in the 1907
crisis in the United States led that country to finally establish a central bank explicitly to pro-
vide LOLR. The failure to follow Bagehot’s rules was a significant contributor to the Great
Depression.

In a bailout, some entity, be it an individual, a country or a firm, is provided with funds in
order to prevent its bankruptcy, where the money could come from anybody interested in
preventing failure. In this chapter we are concerned with the special case of bailouts where
the recipient of the funds is a privately owned bank and the provider is the government.
Bank bailouts, in principle, can be made at any time when a financial institution is
facing difficulty, but generally happen only when the failure of an institution threatens
the financial system. The reason is that ordinarily, the government has special resolution
mechanisms for dealing with failing institutions that do not involve bailouts.
We classify the main categories of bailouts in Figure 14.2.
14.3.1 Direct bailouts

In a direct bailout, the government provides money to a financial institution facing diffi-
culty. Such funds generally are provided by the treasury, and involve the direct use of tax-
payers’ money to prevent the failure of a private institution. The bailout usually takes the
form of one of five options: the government buying equity or preference shares, making
soft loans, guaranteeing the bank’s obligations, or taking over the bank’s problem assets.
Equity injections
The most obvious way to provide the bailout is by a capital injection by the government,
whereby the bank issues new shares that are bought by the government. This is the best
outcome for taxpayers because it is the only one that allows them to participate in the
upside if the bank performs well after the bailout.
267
Chapter 14 Bailouts
Ways to do bailouts
Liquidity Lending of last

Action Direct bailout
provision resort
Who
Treasury Central bank Central bank
does it
Problem
Solvency Liquidity Liquidity
addressed
Equity injection
Preference shares Let all borrow
Soft loans Lower interest Only lend to
Options
Debt guarantees rates those needing
Buying problem QE funds
bank assets
How hard to hit

Stigma effect
equity holders?
What interest to
Hit debt holders?
Issues Inflation charge?
How to price?
Repo problem
Bad bank – good
assets
bank?
Figure 14.2 Bailout options
The main worry about equity injections is that they represent the partial nationalisation
of the bank, with the government having the same rights as other shareholders. The gov-
ernment might be tempted to use its rights for political purposes, thus distracting the
bank from its normal mission.
Preference shares
An alternative to a direct capital injection is preference shares. These are special types of
shares that do not come with voting rights, but provide a special type of fixed dividends
with priority over regular dividends. They do not have a maturity date, and the company
can skip a payment without being declared bankrupt. We can think of preference shares
as a hybrid between debt and equity. In a bankruptcy, preference shares are prioritised
above common shares.
Loans
The third alternative is for the government to lend funds to the bank directly, replacing
funds not provided by the market. This is related to preference shares, except that the
loan will have a fixed term and will place the government higher in the repayment queue
if the institution fails.
268
Loan guarantees
The government can also guarantee that banks’ creditors get repaid, so that if the bank
fails, the government picks up the tab. If the government has a sufficiently high credit rat-
ing, it lowers the cost of funding of the bank, making it equal to that of the government.
The advantage to the government over directly lending to the bank is that the govern-
ment does not have to come up with funds immediately.
Buying or guaranteeing banks’ problem assets

The final alternative is to take over or guarantee some of the bank’s problem assets. While
this is an efficient way to provide support to a bank, it is troublesome for several reasons. It
is a direct transfer to the banks’ shareholders, and thus carries with it significant moral haz-
ard problems, more than for any other option. Guarantees also cause problems of adverse
selection whereby the government is at the risk of overpaying for the banks’ worst assets.
This might happen because the government is likely to be in a hurry and needs to provide
large amounts of funds to many banks, not having time to do proper due diligence.
A prominent and controversial example of guarantees is the Troubled Asset Relief
Program (TARP) in the US. President Obama called TARP ‘a win–win–win proposal’.
Joseph Stiglitz (2009), in an article for the New York Times, called it a ‘win–win–lose
proposal: the banks win, investors win, and taxpayers lose.’ He argued that the govern-
ment’s plans in effect insured almost all the losses suffered by large financial institutions
and spared private investors’ losses. In effect both had been given a free put option by
the government.
The problem arises because the banks get to choose the loans and securities that they
want to sell to the government – the worst assets and especially the assets that they think
the market overvalues. In principle, the government might participate in the choice of
assets, but in practice it is a complicated task and beyond the technical means of the gov-
ernment, at least if it is to be done quickly enough. The market is likely to recognise this,
driving up the price banks can demand.
Preferences
Of these options, the bank prefers them in reverse order, with the guarantee and buying
problem assets the most desirable and the equity positions the most unpalatable. The
reason is that when the government provides equity the owner’s share is diluted. The tax-
payers have the opposite ordering of preferences. They get most for their money by direct
equity stake and least with the guarantee.
14.3.2 Liquidity provision

Another form of bailout is where the central bank directly increases the money supply
in response to some crisis event, either by lowering its interest rates or more likely by
open-market operations or quantitative easing (QE). Liquidity provisions have been fre-
quently used by the Federal Reserve System (Fed) in recent years, for example, in reaction
to the LTCM crisis and 9/11. Such liquidity injections even got the moniker ‘Greenspan put’
after the then chairman of the Fed, because the Fed was seen as supporting the financial
269
Chapter 14 Bailouts
system when faced with shocks by providing large amounts of liquidity. This could be
seen as a put option, insuring against bad outcomes.
The provision of liquidity is a standard central bank tool, to be used either as a part of
regular monetary policy operations or to increase liquidity in times of stress. It does, how-
ever, have two side effects.
1 It involves an increase in the supply of money and, therefore, is inflationary, especially

if the output gap is not large.
2 As with any support operation, it creates moral hazard. If market participants have
reason to believe that their losses will be covered by the state, they will pay higher
prices for risky assets than would otherwise be the case, so liquidity support can lead
to bubbles.
14.3.3 Lending of last resort

Lending of last resort (LOLR) is where the central bank lends money to a financial institu-
tion facing illiquidity. As pointed out in Goodhart (1999), it is a bilateral loan between
the central bank and an individual institution, for the purpose of assisting that institution,
as opposed to an arms-length liquidity provision by the central bank.
The motivation for doing LOLR is that an illiquid financial institution may face the imme-
diate failure to meet its own funding requirements unless it can monetise its supposedly
safe assets. This is especially relevant when the failure threatens to become a full-blown
financial crisis. In practice, the distinction between LOLR and direct bailouts might not be
very sharp. As a general guide, LOLR relates to the central bank assisting financial insti-
tutions to rapidly monetise their illiquid but otherwise safe assets, while direct bailouts
involve the treasury (and not the central bank) directly funding an institution in difficulty.
Because of their indirect nature, the authorities are often able to do LOLR without any-
body knowing, while a direct bailout is highly visible. For example, both the UK and US
governments provided significant liquidity support to some of their best-known institu-
tions in the second part of 2008, which only became public much later.
LOLR is generally seen as the central bank lending to a financial institution facing liquid-
ity difficulties, not problems of solvency, but it can be difficult to make a distinction. The
central bank may be expected to err on the side of caution and be willing to lend even if
it suspects insolvency. If the objective is a direct bailout of an insolvent bank, the treasury
is a more appropriate institution to do the support.
Terms of lending
This leaves open the question of the terms on which the central bank should lend. On the
one hand, it is preventing the failure of a financial institution and not only should it be
rewarded for doing so, but the financial institution in question should be reprimanded.
This calls for high interest rates.
However, high interest rates can have adverse consequences in that they can induce
financial institutions to take on more risk than they otherwise would do. In addition, the
institution is by definition vulnerable and higher interest rates may increase that fragility,
both by increasing its cost and also because of signalling.
270
Asset choice
Another question facing the central bank is what assets to accept as collateral for the
loans it is making to the bank. Ideally, it provides loans only against very high-quality
assets such as government bonds. However, in a crisis the market is unwilling to accept
assets that normally would be easy to pledge in the markets. This means that if a central
bank insists on accepting only high-quality assets, it would be prevented from performing
its LOLR role, because the banks in need of lending are in this position precisely because
they have insufficient high-quality assets. Indeed, a common pattern in a crisis is that the
central bank starts with stringent requirements on asset quality and then slowly relaxes
them over time, perhaps accepting junk in the end.
An early example is from the 1825 crisis. This is a quote from a former director of the
BoE, Jeremiah Harman, from 1825, commenting on that year’s crisis, from Bagehot (1873):
‘We lent by every possible means and in modes we have never adopted before;
we took in stock on security, we purchased Exchequer bills, we made advances
on Exchequer bills, we not only discounted outright, but we made advances on
the deposit of bills of exchange to an immense amount, in short, by every possible
means consistent with the safety of the Bank. Seeing the dreadful state in which
the public were, we rendered every assistance in our power.’
Jeremiah Harman, 1825
The stigma effect

Financial institutions are usually reluctant to avail themselves of central bank funding
because of reputation risk (see Goodhart, 1999). If a bank requires assistance from the
central bank, it signals that its liquidity needs cannot be satisfied on the wholesale inter-
bank markets. This causes the stigma effect, whereby a financial institution in difficulty is
shunned by the rest of the market, which then is sufficient to cause it to fail.
An example of this was discussed in Chapter 2 on the Great Depression where the US
set up the Reconstruction Finance Corporation (RFC) to help banks, but then insisted that
the loans to banks were publicised. The immediate impact of implementing this require-
ment in January 1933 was that banks receiving RFC loans were hit by bank runs, and other
banks in difficulty therefore were unwilling to seek help from the RFC.
One way to overcome the stigma effect is to keep a LOLR operation secret, but that can
be difficult to do in practice because banks do have a good idea of the liquidity needs of
their competitors. This problem can be bypassed by forcing all banks to borrow from the
central bank, regardless of whether they need it or not. This device was used by the BoE
in the crisis from 2007.
The ECB
The European Central Bank (ECB) has made significant longer-term assistance to banks, as
seen later in Figure 14.7. The scope, and especially the long duration, of this programme
start to blur the distinction between a direct bailout and LOLR. A bank that needs support
for many years is no longer facing liquidity problems, rather it is more likely to be insol-
vent. This signals that the European authorities are unable to recognise the difficulties
271
Chapter 14 Bailouts
in their banking system and instead of properly resolving the crisis are using the ECB to
create ‘zombie banks’. If the authorities want to help the banks, it would be better to
recognise the problems and use taxpayers’ money directly to recapitalise them. Using the
ECB in this manner is not appropriate.
14.4 Alternatives to bailouts

While bailouts are based on using public money to aid banks, there are several other
methods the authorities can use to provide help to banks in difficulty, or to mitigate the
impact of the failure. Of those, perhaps the most important are ‘bail-ins’ and regulatory
forbearance.
14.4.1 Bail-ins
Bail-ins were proposed by Calello and Ervin (2010) at Credit Suisse who argue that the
best way to handle a crisis at a large, systemically important bank is to force creditors,
rather than taxpayers, to assume losses if the bank gets into difficulty. This would happen
at a time when the bank is still operating as a going concern, thus preventing its failure.
The bail-ins are overseen by the regulators rather than bankruptcy courts.
There are no automatic triggers for the bail-in; instead, regulators decide when to
implement them, presumably right before collapse – one minute to midnight. That pre-
vents banks from gaming the triggers, and investors from speculating on their activation.
The idea of bail-ins is controversial, particularly with banks, which naturally expect the
possibility of their adoption to push up their funding costs.
Bail-ins do have several benefits. They prevent bankruptcies and thus circumvent
both the bankruptcy courts and more importantly cross-border issues in resolution.
Furthermore, they shield taxpayers from having to bail out banks.
Bail-ins are designed for the failure of a single bank, where the creditors are sufficiently
robust to take the hit. In a more systemic event, it is unlikely that the creditors would
effectively be able to do so, and if forced, might simply spread the systemic event to other
creditor banks and the wider economy via non-bank creditors. This means that a bail-in is
not a general substitute for a bailout, but can lower the costs associated with bank failures.
14.4.2 Forbearance
Definition 14.1 Regulatory forbearance Regulatory forbearance is where the

regulators refrain from using their full powers to intervene, for example to close an insolvent
bank.
A different way to provide assistance to banks is to use forbearance and allow evergreening,
also known as ‘extend and pretend’ or ‘delay and pray’. Generally, this refers to the prac-
tice of not recognising losses on a loan and instead rolling it over when payments come
272
14.5 Bailouts in the crisis starting in 2007
due. The idea is that a bank can run down bad loans slowly and earn its way out of dif-
ficulty. While this can stave off the closure of the bank, it does not come without cost.
Evergreening makes banks’ balance sheets look arbitrarily healthy, thereby deceiving
shareholders and counterparties. Since the bank is using its limited funds to refinance bad
loans instead of using those funds to make good new ones, it reduces the overall credit in
the economy, holding back economic growth.
Evergreening is of course not a bailout, but contravenes accounting rules and prevailing
laws, and normally would be discouraged by banking supervisors. If the authorities instead
actively encourage evergreening, they are imposing a cost on the rest of the economy to
support the banks. Furthermore, evergreening can easily lead to zombie banking, as in
Japan from 1990. Of course, with near-zero interest rates the practice becomes much easier.

The largest international financial crisis since the Great Depression started in 2007. One
of the first manifestations of the brewing crisis was the freeze-up of the interbank market
in early August 2007, and central banks soon had no choice but to inject funds into the
interbank market. The early interventions were not sufficient to get the interbank market
functioning again, and after the failure of Lehman in September 2008, difficult financing
conditions degenerated into a total freeze as all banks attempted to hoard cash and the
authorities felt they had no choice but to provide much more widespread support to the
financial sector.
14.5.1 The United States

While bailouts in various forms happened in many countries, perhaps the most visible
and best documented type of bailouts happened in the US, primarily with TARP and the
Fed credit and liquidity facilities.
Troubled Asset Relief Program (TARP)

The US government signed the TARP into law in early October 2008, primarily to address con-
sequences of the subprime mortgage crisis. TARP originally authorised expenditures of $700
billion, but the eventual amount disbursed was $432 billion. Estimates in mid 2012 indicate
that the cost to taxpayers may be $19 billion, a significant proportion of TARP loans have
already been repaid with interest, and the government expects most of the rest to be repaid
as well. The main outstanding amount is from AIG, as shown in Figure 14.3. The fact that the
eventual loss was so small suggests that the crisis in the US was about liquidity, not solvency.
Credit and liquidity facilities

TARP was highly visible, with financial institutions not surprisingly reluctant to make use
of it. This was not the case for the secret, and much larger, credit and liquidity facility pro-
gramme of the Fed. The programme was initially meant to be secret, and the Fed main-
tained that the recipients should not be publicly disclosed, because of stigma effects,
273
Chapter 14 Bailouts
Citigroup
JP Morgan
Wells Fargo
GMAC
Morgan Stanley Repaid

Goldman Sachs Received
0 10 20 30 40
USD billions
Figure 14.3 Largest TARP bailout recipients in the private financial sector, USD billions
Data source: propublica.org, as of 19 March 2012
700
600
Maximum
Bear Stearns
500
merger
USD billions
$687
400
300
200
Lehman
failure
100
0
2008 2009 2010
Figure 14.4 Fed emergency lending, USD billions

Data source: The Board of Governors of the Federal Reserve System
no doubt considering RFC. However, the Fed was forced to do so under the Freedom of
Information Act, after a lawsuit brought by Bloomberg.
This programme, in its various forms, was much more comprehensive than TARP.1
Below, we present information from three of the largest facilities,2 with Figure 14.4 show-
ing the aggregated amounts.
The programme started at the end of 2007, and peaked at the end of 2008, the worst
time of the global crisis. At that time, the Fed had around $700 billion outstanding, but
starting from early 2009 the numbers tapered off, and by the second quarter of 2010 all
loans had been paid off.
The largest amounts outstanding were on 4 December 2008. Figure 14.5 shows the
largest borrowers at that time. Citigroup was not surprisingly the largest, with $61 billion,
1
See www.federalreserve.gov/newsevents/reform_transaction.htm for more information.
2
Term Auction Facility, Primary Dealer Credit Facility, and Term Securities Lending Facility.
274
Citigroup
RBS
Bank of America
Wachovia
Wells Fargo
Merrill Lynch
Credit Suisse
Barclays
Goldman Sachs
Deutsche Bank
0 10 20 30 40 50 60
USD billions
Figure 14.5 Largest borrowers from Fed emergency lending, outstanding loans on
4 December 2008, USD billions
Data source: The Board of Governors of the Federal Reserve System
Citigroup
Bank of America
RBS
Wells Fargo
Barclays
JP Morgan
Dexia
HBOS
0.0 0.5 1.0 1.5

USD billions
Figure 14.6 Profit of largest recipients from Fed emergency lending, USD billions
Data source: Bloomberg (2011)
but more surprisingly was followed by a foreign bank, the Royal Bank of Scotland (RBS).
Below that, well-known American and European institutions were significant borrowers,
for example Goldman Sachs at $28 billion and Deutsche Bank at $24 billion.
The interest rate on the loans was generally below market rates, in some cases only
0.01%, so the banks were able to make significant profits. Bloomberg (2011) has calcu-
lated those profits, and we report their numbers in Figure 14.6. Citigroup comes in largest
at $1.8 billion, but many others made large profits. This constituted a direct subsidy of
financial institutions at the expense of taxpayers.
14.5.2 The United Kingdom

The UK has a relatively large financial system, as can be seen in Figure 1.2, and is a sig-
nificant exporter of financial services. It is therefore not surprising that it faced one of the
earliest financial institution failures with Northern Rock in 2007. It was, however, with the
275
Chapter 14 Bailouts
failure of Lehman Brothers in September 2008 that the British banks faced their largest
difficulties. The government took stakes in RBS, Lloyds TSB and HBOS in October 2008,
injecting £37 billion of capital into the three banks, leaving it with more than a 50% stake
in RBS and 43% of the combined HBOS and Lloyds TSB, after facilitating their merger.
The government subsequently allocated an additional £19 billion of capital into RBS,
increasing its stake to 84%. In addition, the BoE provided covert liquidity to some banks:
£36.6 billion for RBS and £25.4 billion for HBOS.3
14.5.3 The European Central Bank

Within the euro zone, direct capital injections were left to the nation states, but the ECB
also played a significant role. It started providing liquidity support to banks in 2007, and
since then has been a significant provider of liquidity in the euro zone. In some cases it
has been just about the only wholesale provider of liquidity to entire countries.
This can be seen from Figure 14.7 which shows total assets held by the ECB, as well as
lending to euro area credit institutions. From before the crisis in January 2007 until mid-
2012, both variables have seen a growth of about 160%.
What is interesting is that while the US was able to stop lending to financial institutions
by mid-year 2010 (see Figure 14.4), and the largest amount of lending took place at the
height of the crisis in the fall of 2008, the ECB has been both accelerating its lending and
increasing the maturities of funds lent, most recently to three years.
Such long maturities go along way beyond what is commonly understood as liquidity
support. This in part reflects the trajectory of the balance sheets of commercial banks,
which in Europe continued to expand long after the US peak. It also reflects the huge
geographical imbalance between savers (notably in Germany) and borrowers, in the crisis
countries, which are intermediated via the ECB. This suggests that banks need not only
current liquidity but also a commitment to future liquidity over the coming years.
1200 3000
1000 2500
Total assets
800 2000
Lending
600 1500
400 1000
200 Lending 500
Total assets
0 0
2007 2008 2009 2010 2011 2012
Figure 14.7 ECB lending to banks and total assets, billion euros
Data source: ECB
3
www.publications.parliament.uk/pa/cm200910/cmhansrd/cm091125/debtext/91125-0004.
htm#09112522000002.
276
14.5.4 Analysis
When the crisis started in 2007, the major central banks and financial institutions were
not prepared for a liquidity crisis. After all, such a global event had not happened in
developed economies for almost a century. The financial markets reacted in a way typical
of crises past, and liquidity disappeared. The central banks eventually reacted by provid-
ing liquidity, at first quite reluctantly but eventually supplying seemingly infinite amounts
of funds to financial institutions.
The defining event was the bailout of AIG and the failure of Lehman Brothers in
September 2008. The latter was transformational and signalled the start of the most
severe phase of the crisis, when central bank lending increased sharply. At the height of
the crisis, bank lending by both the Fed and the ECB exceeded 5% of GDP, and direct
bailouts in various forms were quite substantial.
Asian crisis analogies

The previous international liquidity crisis was in Asia in 1997. In that crisis, the diagnosis
came too late, and initial policy responses focused on moral hazard and weak funda-
mentals. If more liquidity had been forthcoming, the severity of the Asian crisis would
have been limited. However, the countries affected were unable to provide the necessary
liquidity support because it had to take the form of US dollars, not domestic currency.
Since there was no effective international lender of last resort (ILOLR), and the inter-
national community, led by the International Monetary Fund (IMF), reacted slowly and
provided insufficient liquidity, the optimal crisis response was not forthcoming.
The US and European policymakers did not repeat that mistake. Their task was simpler
because the liquidity support was in domestic currency, directly within their control.
Downsides
The authorities sent inconsistent signals, first bailing out Bear Stearns, creating the impres-
sion that any similar bank failure would be treated in the same way. When AIG failed,
they had no choice but to bail it out, but the failure of Lehman demonstrates the main
problem. While the question of whether Lehman should have been bailed out is debat-
able, and something we discuss later, it did establish the principle that the failure of a
sufficiently large bank, one that is too big to fail (TBTF), will result in a bailout.
The fact that the liquidity support was highly profitable to the receiving banks is quite
worrying, because it amplifies the moral hazard problems created by the bailout.
The efforts to draw a line under support and curtail moral hazard were unsuccessful,
instead underlining the fact that the authorities could not in practice deny support. This
weakened their bargaining position, and the generous terms of post-Lehman bailouts are
a direct result of this miscalculation.
14.5.5 Europe and the ongoing ECB liquidity support

The amount of liquidity support provided by the ECB dwarfs that of the Fed, and is still
increasing in size whilst the US was able to fully wind down its liquidity support by early
2010. The US is, of course, still providing liquidity support by maintaining near-zero interest
277
Chapter 14 Bailouts
rates. While the US problem was more liquidity based, the amount of support in Europe
implies that the problem is deep-seated solvency difficulties amongst euro zone banks.
However, the European authorities are still addressing the problem as if it were a liquid-
ity crisis and seem very reluctant to face up to the real difficulties. The historical experi-
ence of banking crises suggests that decisive action is necessary to minimise the costs, and
by using the ECB to delay the inevitable, Europe is sharply increasing the eventual costs
of this banking crisis. Apparently, large parts of the European banking system would not
survive without ECB funding, meaning that these banks are, in effect, zombie banks.
There are several reasons why this might be the case. A single nation state, with its own
central bank, can act much more decisively and quickly than a group of countries with
divergent interests. This is why the US, the UK and Switzerland have addressed the prob-
lems of their banking systems much more robustly than the EU. While the EU squabbles
over which measures to take, the banking crisis continually deepens, creating a solvency
crisis out of a liquidity crisis.
If this crisis becomes another Great Depression, the inability of the EU to address its
banking and sovereign debt problems will be the main reason. The EU has the resources
to solve the problem, but what is lacking is the political will to implement the necessary
programmes.
14.6 Bailouts, moral hazard and politics

An important ingredient in maintaining unsustainable credit expansion, or a bubble, is
the unwillingness of everybody involved to stop what seems to be a mutually beneficial
situation. Such sentiments are succinctly summarised by the former Citigroup chief execu-
tive officer (CEO) Charles Prince’s infamous comment before the US housing crisis: ‘As
long as the music is playing, you’ve got to get up and dance.’
The general public sees the benefits brought by a bubble and does not generally have
the expertise or the inclination to note it is unsustainable. Politicians follow. Most of the
bankers enjoy the short-term benefits, and even if they should know better, are more
likely to follow Charles Prince.
Those who understand the dangers are more likely to be technocrats within the banks,
government agencies or academia, and not those in power. By sounding a warning, they
risk being denounced as spoilsports, possibly jeopardising their livelihoods, and may even
be threatened prosecution as has happened in both South Korea and Greece.
This may even apply to top policymakers. Alan Greenspan made his famous ‘irrational
exuberance’ comment in 1996, warning about a build-up of risk, which in retrospect
seems very well judged. However, he did not act on it in a way that reflected a belief in
his analysis. Perhaps, if he had done so, he would now be remembered as inflicting totally
unnecessary damage on the economy rather than having presided over a golden era.
14.6.1 Morality of bailing out bankers

When a financial institution is bailed out, it involves a transfer from the relatively poor
to the very rich. Bloomberg (2011) reports that those working for the six largest banks in
the US earn twice as much as the average worker in the US, with a salary increase of 20%
278
14.7 Model of asset bubbles
between 2005 and 2010, while the salary of the average worker has increased by only
15%. Average bank pay was the same in 2010 as it was before the crisis in 2007.
While it might be surprising that the bankers earn the same before and after the crisis,
it is not that unexpected when one considers how aggressive they can be in getting paid.
For example, after AIG failed, the very employees in the credit default swap (CDS) divi-
sion that brought down AIG and almost the world’s financial system demanded, and got,
bonuses in the millions of dollars if they were to continue solving the mess they created.
Some got the retention bonuses and promptly left. The division’s chief executive told the
Wall Street Journal (2009a) that the ensuing public outcry ‘stunned people [working for
him] such that our wind-down has slowed down’.
Similarly, at the height of the European sovereign debt crisis, the BBC4 interviewed a
trader who said that:
‘he had been looking forward to a recession in order to profit from it. “I dream
of another moment like this”, adding: “Anybody can actually make money. It’s an
opportunity.” ’
Politicians, therefore, need to consider the moral dimension in providing bailouts in addi-
tion to what they consider rational economic arguments.
14.6.2 Challenges
The government’s problem in understanding the complexity of the financial system is
made worse by the fact that its task is more difficult than that of the banks. A bank has
only to worry about its own risk, while the government has to worry about the risk of each
and every bank, individually and in aggregate. The reason is that the endogenous risk cre-
ated by the banking system can be identified only if the financial system in its entirety is
studied. The only body with the power to do that is the supervisor.
The government’s task of effectively providing bailouts is complicated by the close con-
nections between government and industry. The financial system in many countries is
in effect an oligopoly of very large, powerful and well-connected financial institutions.
Indeed, one could say the more banks a country has, the less political power they have.
The opposite is now happening: the number of banks is falling while their size is increas-
ing, making this problem worse. The revolving door between the industry and govern-
ment in some countries can exacerbate this problem. Eventually, these problems may
lead to regulatory capture.

Government guarantees may cause bubbles as in the simple model of Krugman (1998),
aiming to explain the underlying workings of the Asian crisis. We adapt it here to the more
general context of asset bubbles. The main intuition of the model helps to explain many
of the crises discussed in this book.
4
www.bbc.co.uk/news/magazine-15095191.
279
Chapter 14 Bailouts
Model
Suppose banks are perceived to operate under an implicit government guarantee but are
otherwise unregulated entities and therefore subject to severe moral hazard.
The government guarantees lead to excessive lending, creating an asset price bubble,
in a self-reinforcing process of more lending and further-rising asset prices. This will make
the financial conditions of the banks seem sounder than they actually are. At some point
the bubble bursts and asset prices plunge, making the problems facing the banks visible
and forcing them to stop operating, causing a further fall in asset prices.
Model setup
The production function – a function that provides the output of a firm, an industry, or an
entire economy for inputs under consideration – has the following quadratic form:
f1K2 = 1a + P2K - bK2 (14.1)
where P is a random variable and a and b are parameters. The economy is able to borrow
at a fixed world interest rate, which we can, without loss of generality, set to 0. Capital
will earn its marginal product, so the first derivative of (14.1) with respect to capital, the
rental per unit of capital, will be
df1K2
= a + P - 2bK
dK
We assume investors are risk neutral. In the absence of any distortions, capital will be
invested up to the point where the expected return, r, equals the cost of funds, the world
interest rate we set to 0. Therefore:
1a + E3P42 - 2bK = r = 0
Solving for K:
1a + E3P42
K =
2b
Moral hazard
At this point a large number of guaranteed financial intermediaries enter the picture.
Their liabilities are guaranteed by the government and their owners do not need to put up
any capital of their own and can simply walk away if their institutions fail.
The earnings of the financial intermediaries depend on e, with a profit for any realisa-
tion of P leading to r 7 0. The intermediaries need not consider any negative outcome
because the guarantee ensures their returns cannot be negative. Since any economic profit
will necessarily be competed away among the intermediaries, two conclusions follow:
1 All available capital K will be purchased by guaranteed intermediaries in the end and
all other agents will be driven out. This captures the extreme level of leverage of the
guaranteed intermediaries.
280
2 Investment is pushed up to the point where r = 0 for the highest possible value of P.
This captures the over-optimistic investment behaviour of the intermediaries.
Moral hazard decisively influences the behaviour of financial intermediaries in such a set-
ting. In normal economic conditions, investors are responding to the expected value of a cer-
tain investment. Guaranteed intermediaries, however, will focus not on the expected value of
an investment but on what Krugman terms the Pangloss value – referring to a character from
Voltaire’s Candide – which is the return that would be achieved in the most favourable case.
Example 14.1 Pangloss value
Suppose that a = 1, b = 0.25 and P is either 0 or 1 with both values occurring with
equal probability so that the expected value of P is E3P4 = 0.5. The undistorted
level of investment is
1 + 0.5
K = = 3
210.252
But guaranteed intermediaries will base their investment decision not on the expected
value of P = 0.5 but on the Pangloss value and therefore assume that P = 1. The
level of investment will be pushed up to
1 + 1
K = = 4
2(0.25)
The result is over-investment which lowers expected welfare because the increased loss
in the bad state will not be offset by increased gain in the good state.
Impact of perfectly inelastic assets like land

Up until now, we have assumed that the supply of capital goods is perfectly elastic, mean-
ing that any increase in the demand for capital goods will be satisfied.
If we now assume that capital assets are perfectly inelastic, which would be the case for
land, the increase in demand that accompanies the presence of guaranteed intermediar-
ies will have an impact not on the quantity of capital but on its price.
Consider a three-period economy. In the first period, t = 0, investors bid for land; in
the second and third periods they receive returns. In each period the return on one unit
of land can be either 100 with a probability of 1>3 or 25 with a probability of 2>3. Table 14.1
shows the returns and probabilities for each period.
Table 14.1 Investment decisions
t = 1 t = 2
p Outcome p Outcome
Good state 1>

3 100 1>
3 100
Bad state 2>
3 25 2>
3 25
281
Chapter 14 Bailouts
In an undistorted economy we can solve backwards for the price that risk-neutral inves-
tors are willing to pay. The expected return in the last period is 50, which is consequently
the price of land purchased at the end of t = 1. At t = 0, therefore, the expected return
on land purchased is the expected return at t = 1 of 50 plus the expected selling price
at the end of t = 1, which is also 50, for an expected value at t = 0 of 100, the total
expected return over the periods.
Guaranteed intermediaries will base their investment decision not on the expected value
but on the Pangloss value – the best-outcome value independent of probabilities. Again,
working backwards, at the end of t = 1, the intermediaries are willing to pay the Pangloss
value of the third period return, which is 100. At t = 0 they will therefore be willing to
pay the most they could hope to realise in both subsequent periods, which is the Pangloss
return at t = 1 plus the Pangloss resale value at the end of t = 1, which is 100. Guaranteed
intermediaries will therefore pay 200 for one unit of land, so that in the end they will have
driven out all other investors and raised the price to double what it would have been in an
undistorted economy. Figure 14.8 shows the Pangloss values for the three periods.
Implications for banking crises

In this example the probability of the guarantee being exercised is 8>9, (1 minus the prob-
ability of the two good states, 1>3 * 1>3 ) and if additional periods are added the likelihood
rapidly approaches certainty because the guarantee is required on the first failure.
This remains the case even if we modify the odds of the two outcomes so that the poor
outcome is unlikely. The asset will be priced ignoring the poor outcome, so if the number
of periods is a sufficient multiple of the probability of a bad outcome, the guarantee is
almost certain to be required.
This underlines the almost unavoidable nature of banking crises. We cannot exclude
the possibility of a damaging mistake. This is clear when we consider that a large fraction
of the US banking industry was until 2007 prepared to use models that assumed that US
house prices could not fall, yet a small possibility is all that is required to guarantee that a
crisis will eventually occur. Viewed from this perspective, the recurrence of crises at roughly
79-year intervals (2007 - 1929) could be seen as implying a roughly 1.3% chance on aver-
age that investors are unaware of some adverse possible outcome to their decisions.
Analysis
The central message of this simple model is the adverse impact caused by government
guarantees of banks. Because creditors assume they will be bailed out, they are willing
200
Expected value
150
Pangloss value
100
50
0
Period 0 Period 1 Period 2
Figure 14.8 Pangloss values
282
14.8 Summary
to lend to banks at low interest rates even though they think the banks’ investments
are bad.
In addition, these guarantees give banks an incentive to price risky investments with
over-optimistic assumptions. Any kind of activity involving uncertain but diminishing
returns to capital – almost all activities – will as a consequence see higher levels of invest-
ment than would be made by equity investors focusing on expected returns. Nor do the
banks’ creditors have any incentive to withdraw funding.
This has two adverse impacts for society. First, it encourages excessive risk-taking at the
expense of the taxpayer, a classic form of moral hazard. Second, it leads to asset price
bubbles that are destined to burst, destabilising financial markets on the way up and
causing a crisis on the way down.
14.8 Summary
Governments often find it necessary to bail out private financial institutions in order to
prevent an individual failure becoming a full-blown financial crisis. There is a right way
and a wrong way to implement bailouts, with the Swedish government’s actions in 1992 a
model of a successful bailout, and the Irish bailout of 2008 less successful.
The main issue in bailouts is moral hazard where the expectation of government sup-
port induces financial institutions to take more risk.
The first formal government bailout policy was LOLR, formulated after the 1866 crisis.
Bailouts have been used frequently, both before and after that date, for example, during
the crisis from 2007 where central banks in effect replaced the interbank markets and
large financial institutions received bailouts, most significantly AIG.
Bailouts can take many forms, in particular direct bailouts, LOLR and liquidity pro-
vision. The damage from bank failures can be mitigated by bail-ins and banks can be
helped by regulatory forbearance. In implementing bailouts, it is important to distinguish
between illiquidity and insolvency, even though conceptually their boundary is a broad
grey area and identifiable only in extreme cases.
We demonstrated a simple model of how government guarantees can lead to asset
price bubbles, which contributed to the 2007 crisis and is of continuing concern given the
apparent determination of governments to prevent private sector bank losses.

1 Comment on the relationship between the size and credibility of a bailout package to
its effectiveness.
2 Why did the Irish bailout of the banks end up causing the bailout of the Irish government
itself?
3 What are the main ways a government can do a direct bailout? Which way is preferred
by taxpayers and which by the banks?
4 Why should the treasury and not the central bank do a direct bailout?
5 Why does the central bank and not the treasury do liquidity injections and LOLR?
283
Chapter 14 Bailouts
6 Should the central bank charge a penalty rate when doing LOLR?
7 Does the existence of LOLR encourage excessive risk taking?
8 How could the problem of moral hazard associated with bailouts be reduced?
9 Should the bailout be based on banks’ size or interconnectedness? Offer one example
for each form of bailout.
10 How can we justify using taxpayers’ money to bail out those who own and run banks,
especially when the bank owners and managers represent one of the richest segments
of the population, who then receive a transfer from their poorer countrymen?
11 What are the main arguments for keeping bailouts secret and the main arguments for
making them public?
12 If a central bank provides liquidity assistance to banks, what should the time span for
the assistance be?
13 What is your view on the ECB liquidity programmes?
14 What is a Pangloss value?
15 How can the concept of Pangloss value be used to explain the behaviour of investors?
Is it rational?
(a) Consider a two-period model with an initial investment at t = 0 and returns in
p
eriods t = 1 and t = 2. With probability p the returns will be y, otherwise the
returns will be x. The following holds: x 7 y and 0.5 7 p 7 1. What is the value
of the return that investors can expect? What is the Pangloss value?
(b) The government explicitly guarantees all investments, so that investors at t = 0
bid up the price to the Pangloss value. Now the government announces that it will
bail out investors only once and will then withdraw all guarantees. Show that this
leads to a situation of multiple equilibria.
(c) Prove that the government guarantee at t = 0 makes losses worse if things turn out to
be bad in t = 1 than they would have been without any explicit or implicit guarantees.
References
Bagehot, W. (1873). Lombard Street: a Description of the Money Market. H.S. King, London.
Bloomberg (2011). Secret Fed loans undisclosed to Congress gave banks $13 billion in income.
www.bloomberg.com/news/2011-11-28/secret-fed-loans-undisclosed-to-congress-gave-
banks-13-billion-in-income.html.
Calello, P. and Ervin, W. (2010). From bail-out to bail-in. The Economist, 28 January.
Capie, F. and Wood, G. E. (2006). The Lender of Last Resort. Routledge.
Elliot, G. (2006). Overend & Gurney, a Financial Scandal in Victorian London. Methuen.
Goodhart, C. (1999). Myths about the lender of last resort. Technical report FMG SP 120,
London School of Economics. http://fmg.lse.ac.uk/pdfs/sp120.pdf.
Krugman, P. (1998). What happened to Asia? Mimeo, MIT.
Moe, T. G., Solheim, J. A. and Vale, B. (2004). The Norwegian banking crisis. Technical report,
Norges Bank. Occasional Paper no. 33.
Stiglitz, J. (2009). Obama’s ersatz capitalism. www.nytimes.com/2009/04/01/opinion/01stiglitz.
html?_r=1&scp=2&sq=stiglitz&st=cse.
Wall Street Journal (2009a). Action on AIG unit may cost taxpayers. Wall Street Journal, 13 April.
online.wsj.com/article/ SB123957925343711995.html.
284
15 Dangerous financial instruments
Financial institutions are continually creating ever more complex financial instru-
ments, motivated by potential profits from new types of trading strategies, the ability
to market new instruments to their clients, to move risk off their balance sheets, or
to hide the risk-taking activity from other parts of the institution, supervisors and cli-
ents. While the new instruments may be highly profitable during boom times, they
also act as enablers for excessive risk-taking and as accelerators for small events to
spill out of control. They can contribute directly to financial crises, just one exam-
ple being their role in the crisis from 2007 which included new instruments like
CDOs, CDSs, SIVs and conduits. This is why we might term them dangerous financial
instruments.
These dangerous instruments undermined the extensive monitoring process we
had implemented to prevent the excessive build-up of systemic risk-taking. They did
that by enabling banks to spin off a part of their balance sheet, helping banks to
do more business and take on more risk with the same amount of capital, and even
replacing bank supervisors with more emollient rating agencies.
The main problem arises because of the complexity of the financial instruments.
Not only is it hard for non-experts to understand the inherent riskiness, but the finan-
cial instruments often depend on a number of assumptions about the stochastic
behaviour of market variables and the estimation of complicated statistical models.
In this process there are many things that can go wrong. The designer of the product
may not understand all the issues, the underlying assumptions are likely to be inad-
equate, and the statistical models will be estimated with significant degrees of uncer-
tainty. In short, these instruments are priced and their risk is assessed with complex
Chapter 15 Dangerous financial instruments
statistical models, and since every statistical model is wrong by definition, prices and
risk will similarly be incorrectly imputed.
The underlying modelling weaknesses are masked by high profits during boom
times. A large adverse shock may be required to lay bare the mistaken model assump-
tions. The main problem is that this is most likely to happen when the financial sys-
tem is already heading into a crisis, so that a reassessment of the riskiness of these
instruments becomes part of a vicious endogenous risk feedback loop between reas-
sessment, institutional difficulty and prices.
Fundamentally, the main problem of the dangerous financial instruments is that
they ignore the likelihood of systemic losses.

This chapter relates directly to Chapter 10 (credit markets), Chapter 9 (trading and
speculation) and Chapter 17 (the ongoing crisis: 2007–2009 phase).
Key concepts
■ Complexity
■ Derivatives
■ Credit default swaps
■ Collateralised debt obligations
■ Special investment vehicles
■ Conduits
■ Rating agencies

The underlying mathematics of complex credit instruments are discussed in detail
in several good texts. We made use of Crouhy et al. (2000), Duffie and Singleton
(2003) and Hull (2011, 2012). Murphy (2009) provides a good discussion of the
role played by credit instruments in the crises, whilst the Board of Governors of the
Federal Reserve System (2009) has a good overview of the various credit facilities
and their role in the crisis.

F, f Factors
K, k Number of defaults
N Number of assets
X, x Random outcomes
pd Probability of default
P Idiosyncratic shock
g
R Default correlation
Covariance matrix
𝚽1 # 2 Standard normal distribution
f1 # 2 Standard normal density
A 1R
286
15.1 Complexity kills
15.1 Complexity kills

Before the ongoing crisis that began in 2007, sophisticated financial models and intricate
assets structures enabled many banks to reap significant profits. Complexity is profitable
because complex products offer more scope for differentiation and profit than commod-
itised ones. Financial institutions have an incentive to maximise the complexity of their
products and operations, and even in special cases may have as one of their objectives the
minimisation of the quality of assets.
There is a clear advantage for financial institutions to move towards complexity.
It makes it harder for others to see what they are up to, be they competitors, clients
or the authorities, while at the same time complex structures help in exploiting arbi-
trage opportunities. Even better, as the more complex and less transparent an instru-
ment becomes, the bigger the potential for high fees. The greatest profit opportunities
often lie at the edge of chaos. Unfortunately, at that point it takes little to end up over
the edge.
While lack of knowledge makes it hard to know how safe you are, it takes some-
thing else (perhaps greed and/or competition) to force you to the edge, which hap-
pens because of the over-reliance on a mathematical description of history. Complexity
became a virtue, but not without costs. The banks often did not have a clear view of
their risk exposures.
The impact of complexity is even worse, as it creates incentives to maximise low-quality
investments and exploit model uncertainty. When financial institutions set up structured
credit products, certain tranches were more profitable than others. The highest profits
could be found where banks used assets with high loss given default and higher correla-
tions than assumed in the models, assets often colloquially referred to as ‘garbage’. It
was the demand for garbage assets that was behind the demand for subprime assets. The
enabling factor was complexity.
In such complex financial models, mathematics often assumes far greater importance
than an accurate depiction of reality. The models generally ignored important risk factors,
like liquidity and non-linear dependence, even disregarding the possibility of economic
downturns.
Consequently, the valuations and risk assessments of complex products became increas-
ingly out of sync with the economic fundamentals and underlying assets. Unfortunately,
few mechanisms exist for identifying the looming problems. If the models indicate that
everything is fine, backed up by mark-to-market accounting practices, it is not surprising
that the market for structured products becomes over-inflated.
When the credit markets collapsed and liquidity disappeared in 2007, complexity
became a vice. In a crisis, banks gain access to liquidity by being able to demonstrate
solvency. If assets are so complicated that nobody – not the regulators, not the clients
and not even the banks – is able to get any realistic assessment of valuations and risk,
of course, investors will refuse to supply liquidity. Seen with a newly sceptical eye and a
radical change in assumptions, it became clear that banks had simply become too sophis-
ticated for their own good.
287
15.2 Derivatives
‘Derivatives are financial weapons of mass destruction, carrying dangers that, while
now latent, are potentially lethal.’
Warren Buffett (2003)
Derivatives are securities that derive their value from something else, called the underlying
asset. The underlying asset is commonly another security, perhaps shares in IBM, a physical
asset like bushels of wheat or a non-traded financial asset such as a bundle of mortgage loans.
The underlying asset can also be something abstract, like the occurrence of an event or the
prevalence of a special state of the world. For example, there are weather derivatives whose
value depends on the temperature or rainfall in some place over a particular period, and credit
derivatives whose value hinges on whether a firm files for bankruptcy in a set timeframe. Fixed-
income instruments are one type of derivatives, their underlying asset being the interest rate.
Different sorts of derivatives are designed for different purposes. Many derivatives con-
tracts were originally designed to bring together parties exposed to offsetting risks so that
they could net them out, a practice called hedging. Others provide optionality, build in
leverage, or create exposures to risks otherwise not tradable in financial markets.
Derivatives carry with them significant hidden dangers. Not only do they allow highly
leveraged bets on financial markets, but also the risk is often deeply hidden.
Volume
The volume of trading in derivatives markets has grown significantly over the past few
years. The Bank for International Settlements (BIS) collects data on the volume of deriva-
tives trading, shown in Figure 15.1. Figure 15.1(a) shows the total notional amounts, or
the face amounts, used to calculate payments. This amount does not change hands.
60
400 FX
FX, CDS
Interest
Interest
300 40
CDS
200
20
100
1998 2000 2002 2004 2006 2008 2010

(a) Notional amounts
20 5
FX
4
FX, CDS
15 Interest
Interest
3
10 CDS
2
5 1
1998 2000 2002 2004 2006 2008 2010

(b) Gross market values
Figure 15.1 Amounts outstanding of over-the-counter (OTC) derivatives, USD trillions

Data source: BIS
288
15.3 Credit default swaps
By contrast, Figure 15.1(b) shows the gross amount, which is the sum of the value of contracts
owned by a financial institution, without taking netting into account. Therefore, it shows the
worst-case situation of what happens when counterparties default and loss is 100%.
The amounts are shown in trillions of dollars, and we see that the notional amounts
significantly exceed the global GDP of about $60 trillion. The more important gross
amounts are slightly less than half the global GDP. The rapid growth over just a few years
is a significant concern, and the lack of concern by the world’s policymakers prior to the
crisis starting in 2007 is itself a cause for concern.

The derivative that saw the most rapid increase in volume in Figure 15.1 is the credit
default swap (CDS). It is one of the two derivatives most closely associated with the crisis
that began in 2007, not least because it was directly responsible for the largest ever single
financial institution failure, that of AIG. It also more indirectly contributed to the failure of
Bear Stearns and Lehman Brothers.
As a consequence, it is not surprising that CDSs have come under considerable criti-
cism. Eric Dinallo, the insurance superintendent of New York state, who supervised AIG,
called the CDS a
‘catastrophic enabler of the dark forces that have swept through financial markets’.
The Economist observed on 6 November 2008:
‘They are, says a former securities regulator, a “Ponzi Scheme” that no self-respecting
firm should touch. [. . .] Alan Greenspan, who used to be a cheerleader, has dis-
owned them in “shocked disbelief”. They have even been ridiculed on “Saturday
Night Live”, an American TV show.’
Much of the criticism relates to the fact that the CDS market had blossomed without
much oversight, allowing the sellers of CDSs to amass excessive amounts of risk. At the
same time, the fact that CDSs are OTC instruments implies that the market for them is
fairly non-transparent, so it is difficult to identify net exposure.
What is a CDS?
A CDS is an instrument providing insurance that protects the buyer of the CDS against the
default of some third party called the reference entity. Suppose a bank owns a high-risk
bond and wishes to hedge that risk. The bank, known as the protection buyer, would enter
into a contract with another institution, the protection seller. The protection buyer pays a
regular fee to the protection seller until the CDS expires, unless a credit event happens,
which could be a default but might also include other occurrences. In this case, the protec-
tion seller would buy the risky bond at par value from the protection seller and no more
payments take place. This is described in Figure 15.2. For detailed information on how
CDSs are created, see Appendix A. In some cases, when the reference entity is high risk, all
the fees may be up front. This, for example, has been the case for some Greek CDSs.
289
Annual
payments
Protection Protection
buyer seller
(a) No default
Par value
of bond
Protection Protection
buyer seller
(b) Default
Figure 15.2 Simple CDS payment flow
Benefits of CDSs
In principle, CDSs are quite useful instruments, allowing users to trade and manage credit
risk separately from other risks and cash flows. They provide an efficient way to hedge
credit risk, spreading risk to those best able to assume that risk, and allowing users to effi-
ciently diversify and tailor risk-taking. CDSs also enable speculation, in principle, increas-
ing the efficiency of financial markets.
15.3.1 The problem with CDSs

Individual risks created by CDSs
One of the two major risk factors for those buying CDSs is counterparty risk, arising
because they do not settle on a clearinghouse, or a central counterparty (CCP), but are
instead traded and settled OTC. Since CDSs are often sold by major financial institutions,
such counterparty risk directly relates to the institutions’ failure probabilities. For exam-
ple, if one had purchased a CDS protection from Lehman Brothers in January 2008, that
protection would have lost effectiveness in September 2008 with Lehman’s bankruptcy.
CDSs and network vulnerabilities

CDSs create their own types of risk. We illustrate this in Figures 15.3 and 15.4 and outline
the key steps below. The arrows indicate exposure.
Step 1 Bank A makes a loan to a creditor and consequently is exposed to risk of

default.
Step 2 To hedge the probability of default, bank A buys a CDS from bank B. The
initial credit risk is now transferred to B, whilst A and B are exposed to each
other; however, the degrees of exposure of A and B are not symmetrical, as it
is much easier to make a stream of small payments in good times than a large
payment in bad times.
Step 3 The creditor pays back the loan early to bank A, which is left with paying fees
to bank B. To minimise costs, it sells a CDS to bank C. Now A and C are also
exposed to each other.
290
Step 1: Initial exposure
Creditor Bank A
Step 2: Credit default swap (CDS) Bank A
Creditor Bank B
Step 3: New CDS

Bank A Bank C
Creditor Bank B
Figure 15.3 CDS network risk creation
Bank C
Information
asymmetry
The Each bank

system knows only
Bank A nets out its own
obligations
Bank B
Figure 15.4 Netting out
Step 4 To hedge its exposure, bank B buys a CDS from bank C. Now A, B and C are all
exposed to each other. This is shown in Figure 15.4.
Throughout this process, the gross amount increases with each step, but the net
amount remains the original exposure to the creditor, who after all has paid back the loan
so the net exposure in the system is zero. This problem is shown in Figure 15.4, showing
the gross exposures and the zero net. This creates a chain of vulnerabilities, contributing
to the problem of interconnectedness in the financial system, creating systemic risk.
Consequently, the failure of a major dealer would have major repercussions for the rest
of the market participants, exactly the problem created by AIG and Lehman. Furthermore,
because CDS contracts are not cleared on a clearinghouse, they are not government regu-
lated, nor standardised, making it very difficult to calculate net obligations. There are no
public records showing whether sellers have the assets to pay out if a bond defaults.
291
It is difficult to develop a clear picture of which institutions are the ultimate holders of
some of the credit risk transferred. It can even be difficult to quantify the amount of risk
that has been transferred. This is the main systemic risk posed by CDSs.
Related to this is the problem that the protection seller might need to come up with
significant funds during a crisis episode, when it and the financial system are under stress.
Perverse incentives
A CDS affects normal incentives in financial markets. It can create moral hazard since
the initial lender has less incentive to ensure the quality of the underlying loans and
monitor the borrower if the risk is sold off to a third party, who probably has less infor-
mation about the borrower than the initial lender. This can lead to a lower quality of
loans.
CDSs also provide perverse incentives for creditors. Pragmatic lenders who hedge their
economic exposure through CDSs can often make higher returns from CDS payouts than
from out-of-court restructuring plans. Bankruptcy codes assume that creditors always
attempt to keep solvent firms out of bankruptcy; however, if a creditor holds a CDS it may
be more profitable to trigger a bankruptcy even if that significantly reduces the amount of
money received by bondholders.
This may have been behind the bankruptcy of Six Flags, a failed American theme-park
operator, as noted by The Economist (2009) where the Fidelity mutual fund turned down
an offer that would have granted unsecured creditors an 85% equity stake, and as a con-
sequence uninsured bondholders will receive less than 10% of the equity now that Six
Flags has filed for protection.
15.3.2 Naked CDSs

Trading CDSs for purely speculative reasons as easily without owning the underlying
assets is called trading naked CDSs. The buyer could just make a deal with the protection
seller, whereby the buyer pays fees to the seller until a credit event occurs, at which time
the seller pays a certain amount to the buyer. Insurable interest is missing for naked CDSs,
so they are different from typical insurance contracts. A comparison is sometimes made
with what might happen if a person could buy fire insurance on their neighbour’s house.
There is no insurable interest and the person would have a motive to burn the neighbour’s
house down to trigger an insurance payout.
Naked CDSs have become quite controversial, especially in Europe, as some European
leaders have blamed speculators using them for undermining certain European economies.
‘I think that derivative products [such as] the CDS on sovereign debt have to be at
least very, very regulated, rigorously regulated, limited or banned [. . .]’
Christine Lagarde, former French minister and now managing
director of the International Monetary Fund (IMF),
quoted in the Financial Times (2010a)
The EU has decided to impose a permanent ban on naked CDSs, strongly supported by
Germany and opposed by the UK.
292
15.4 Collateralised debt obligations
‘These balanced measures will ensure that sovereign CDS are used for the purpose
for which they were designed, hedging against the risk of sovereign default, without
putting at risk the proper functioning of sovereign debt markets [. . .] Short selling
did not cause the crisis, but can aggravate price declines in distressed markets.’
Michel Barnier (2011), commissioner for the single market
Naked CDSs, however, are not the evil instrument the politicians sometimes make
them out to be. A European Commission report concluded that CDSs had no adverse
impact on bond markets, where the prices of CDSs and sovereign debt largely moved in
tandem. These issues are discussed by Duffie (2010).
Banning naked CDSs could also deepen a crisis. Market participants might have an
indirect exposure to a counterparty. Furthermore, by banning naked CDSs, market par-
ticipants might attempt to reconstruct the hedge by other means, perhaps by shortening
government bonds or bank equities or otherwise reducing their exposure. Because such
activities would be less visible, it would be typical of regulations driving risk-taking under-
ground, reducing transparency and financial stability.

A collateralised debt obligation (CDO) is a debt structure holding, whereby a portfolio
of fixed-income assets is held in a special purpose vehicle (SPV). Investors buy rights to
the payment flow from the underlying assets in the SPV. The sole purpose of the SPV is
to collect collateral cash flows and pass them to CDO investors. CDOs typically allocate
interest income and principal repayments from the underlying assets into what is known
as tranches.
Definition 15.1 Tranche A tranche, French for slice, is a particular slice of the risk,
and hence payments flow from a CDO. Tranches often get the same credit rating as regular
bonds. The main tranches are:
■ Super senior (AAA)
■ Senior notes (AA)
■ Mezzanine (A to BB)
■ Equity (unrated), a.k.a. toxic waste.
Higher-rated (senior) tranches get paid before lower-ranked mezzanine tranches, with any
residual cash flow paid to an equity tranche.
Tranches are usually sold separately. The CDO structure allows the senior tranches to
obtain credit ratings in excess of the average rating on the collateral pool as a whole,
because of an assumption of diversification within the pool of assets.
Maximising the size of desirable tranches is a key objective of the designers of CDOs.
Often, the mezzanine tranches are the easiest to sell, with the sponsoring bank left hold-
ing the super senior and equity tranches. This depends on buyer preference, however,
293
and in many cases the AAA and AA tranches are easiest to place because buyers simply
count the As (see safety) while mezzanine investors do more modelling. By retaining AAA
tranches, the sponsoring bank might obscure its profits on the structure, by selling it later
and booking it as separate profit.
15.4.1 Example CDO

A bank creates a SPV which buys 10 bonds. Each bond has a maturity of 1 year and a total
face value of $10 million, so the total face value of the assets held by the SPV is $100 mil-
lion. The bonds have a 12% annual interest rate, while the probability of default over the
next year is 25%. In default there is no recovery.
Suppose an AAA corporate bond has a probability of default pd = 0.3% and an interest
rate of 6% whilst a BBB bond has pd = 7% and an interest rate of 8%. It is then straight-
forward to create three tranches (see Appendix B for details), assuming defaults are uncor-
related. We show the size of the various tranches in Table 15.1, with the CDO structure
summarised in Figure 15.5.
In this sample CDO, the sponsor has been able to create $30 million worth of AAA-
rated securities out of $100 million worth of high-risk instruments. This is achieved by
Table 15.1 Tranching
Rating Interest rate Number of bonds Value of bonds Interest payment
AAA 6% 3 $30 million $1.8 million

BBB 8% 2 $20 million $1.6 million
Equity 17.2% 5 $50 million $8.6 million
Total 12% 10 $100 million $12 million
Assets Tranches
AAA $1.8 million
Special
$100 million
purpose Mezzanine
of subprime
vehicle BBB $1.6 million
mortgages
(SPV)
Equity
(toxic waste)
$8.6 million
Figure 15.5 Payments from the CDO structure
294
concentrating the probability of default in the equity tranche, composed of $50 mil-
lion. Provided the underlying mathematical assumptions are correct, in particular the
probability of default is actually 25% and the bonds are uncorrelated, the tranching is
correct.
Sensitivity to probability of default

CDO tranching is quite sensitive to the underlying probabilities of default. We show the
tranche sizes for the example CDO, as we vary the probability of default, in Figure 15.6.
As the probability goes from 0 to 50%, the AAA tranche is reduced from all 10 bonds to
none, and by pd = 80% the BBB tranche has essentially disappeared as well. Because of
the binary nature of the CDO, even a small change in the probability of default can trigger
large changes in the tranche sizes.
15.4.2 Correlated defaults

If defaults are independent events, that is, one bond defaulting has no bearing on the prob-
ability of other bonds defaulting, we can calculate each tranche’s probability of default
using the binomial distribution as in the simple example in Section B.1 of Appendix B.
But it is usually the case that defaults across bonds are correlated. Firms default more
when there is an economic downturn or there is a shock affecting a whole sector. It is
quite straightforward to incorporate correlated defaults into the calculation, perhaps by
using the Gaussian copula approach, made (in)famous by Li (2000) whose contribution
enabled the pricing of structured credit products (like subprime CDOs) that subsequently
got blamed in the crisis that began in 2007. We present one approach to incorporating the
Gaussian copula in Section B.3 of Appendix B.
Note the crucial impact that different assumptions about correlations have on the
probability of default of each tranche. As correlations increase, the probability that all
bonds will default increases, and the same applies to other tranches too. Increasing cor-
relations increase the risk of all tranches.
Figure 15.7(a) shows the impact of changing default correlations for the example CDO.
As the correlations increase from 0% to 100%, the probability that all 10 bonds will default
rises from zero to 25%. Figure 15.7(b) zooms in on the bottom part from Figure 15.7(a),
enabling us to see in detail the impact on tranches.
10 AAA BBB Equity

Tranche size
8
6
4
2
0
0% 1% 10% 20% 30% 40% 50% 60% 70% 80%
Probability of default
Figure 15.6 Effect of probability of default on tranche size
295
60% rho = 0 rho = 0.4 rho = 0.8
Probability
rho = 0.2 rho = 0.6 rho = 1
40%
20%
0%
0 1 2 3 4 5 6 7 8 9 10
Defaults
(a) All outcomes
4%
Probability
2%
0%
0 1 2 3 4 5 6 7 8 9 10
Defaults
(b) Focus on bottom 5%
Figure 15.7 Effect of correlations on the probability of default
Table 15.2 Correlations and default probabilities
r Probability of defaults
8 - 10 6 -7
0% 0.04% 1.93%
20% 1.46% 6.84%
40% 4.77% 8.83%
60% 9.15% 8.59%
80% 14.47% 6.70%
100% 25.00% 0.00%
Table 15.2 presents the same information in a different way, focusing on what happens
to the sizes for the AAA and mezzanine tranches when the correlations increase, com-
pared to the benchmark case where the correlations equal 0.
15.4.3 Issues with CDOs

Correlations in the crisis from 2007
As we saw above, single-tranche CDOs are quite sensitive to small changes in correla-
tions. Unfortunately, the data samples used to rate the CDOs containing subprime mort-
gages were not long enough to include a recession; the subprime market started taking
off in the early 1990s right after the recession ending in 1991. The United States (US) was
then recession-free until 2007, except for a few months in 2001, according to the NBER
296
which keeps track of US recessions. The biggest boom years in the subprime market were
entirely recession-free. Since mortgage defaults are highly dependent on the state of the
economy, the estimation of default correlations would have been skewed towards zero.
Even if the sample contained a downturn, it would be difficult to estimate the correla-
tions because, as noted by Duffie (2007), there is a serious shortage of good models for
estimating default correlations.
The problem of correlated defaults is worse than this, because even if accurate correl-
ation estimates had been available, so long as the industry used Gaussian copula meth-
ods the dependence in mortgage defaults would still have been underestimated. The
reason is that the Gaussian copula assumes constant correlations regardless of the state of
the economy, while in actuality the correlations are much higher in downturns. The tech-
nical name for this is non-linear dependence (discussed in Section 3.3.1) and an asym-
metric copula, properly taking this into account, is much more accurate. Unfortunately,
it is very challenging to use such asymmetric copulae, not least because it is difficult to
identify the best copula to use unless very large amounts of data are used, much more
than are available.
CDOs as catastrophe bonds

A related problem is noted by Coval et al. (2008a, 2008b) who point out that some CDOs,
such as those containing subprimes, can be considered a form of catastrophe bonds.
A traditional catastrophe bond is a security whose default probability is constant and
independent of the economic state. Catastrophe bonds are typically used by insurers,
and deliver their promised payoff unless there is a natural disaster, such as a hurricane
or earthquake, in which case bond payments are no longer made. Investors are willing to
pay a relatively high price for catastrophe bonds because their risks are uncorrelated with
other economic indicators and, therefore, can be reduced through diversification.
Similar factors were at work with structured credit products containing subprime mort-
gages. In boom times, mortgage defaults tend to be idiosyncratic events, caused by a per-
sonal tragedy leading to a default. This means that if an individual defaults, it says little
about the probability of anybody else defaulting.
In a recession it is different. Then an individual default may be caused by factories
closing, causing widespread distress in a community, where an individual default is the
harbinger of multiple defaults. The practice of geographically concentrating mortgages
structured into CDOs amplified this effect.
For this reason, one can consider structured credit products containing subprime
mortgages as a type of catastrophe bonds. In other words, the securitisation process sub-
stituted risks that are largely diversifiable for risks that are highly systematic. Structured
finance products have far less chance of surviving a severe economic downturn than tradi-
tional corporate securities of equal rating.
Optimisation
Certain tranches of CDOs are more valuable than others. This means that banks might
want to maximise the size of these tranches at the expense of the rest. This is known as
optimising the CDO structure. In order to facilitate this process, the rating agencies supplied
297
the banks with computer software enabling them to optimise the CDOs in-house before
they were submitted to the rating agency.
This undermines the integrity of the whole process. The assignment of probabilities to
the various tranches using the mathematical approach discussed above is correct condi-
tional on the assets put into it. If the assets are preselected to maximise certain tranches,
the probability assignments are biased. This is a variant of data mining as studied in
econometrics.
A related and perhaps even more important problem is that the optimisation of the
structure also minimises the quality of the asset pool. In other words, if the sponsor wants
to optimise ratings, the cheapest way is to use assets with high loss given default and
assets that have higher default correlations than are assumed in the modelling. This in
turn affects the true value of all the tranches.
The end result is that before 2007, sponsors demanded assets that were high risk, with
low expected recovery rates and high default correlations and probability of default, in
other words, garbage assets. If one of the criteria for high profits is maximising the use of
garbage assets, it should not come as a surprise that this all ends in tears.
Synthetic CDOs and CDOs squared

Before the crisis, the demand for CDO tranches significantly outstripped the supply,
which led to frantic efforts to provide assets into those structures. This was a key reason
for the boom in subprime lending in the US. Another manifestation of this demand was
the creation of instruments that provide CDO-type risk by using financial engineering to
create a synthetic CDO. When the supply of assets into CDOs was not sufficient, some
financial institutions resorted to a concept called CDO squared (and cubed, etc.), CDO2,
CDO3 or CDO4. The input into a CDO2 is not subprime mortgages or other bonds; instead
it is tranches from other CDOs.
The logic behind this is to enable financial institutions to create yet more instruments
and more finely tune the structures. However, such instruments amplify the errors in cal-
culations. After all, if the initial CDO is subject to significant model risk, any structure
using the equity tranches of the CDO will just magnify the model risk.
Example 15.1 CDOs and fractional reserve banking
A CDO has many parallels with fractional reserve banking, and one could easily view
a CDO as a type of bank. For example, both are subject to the same fundamental fra-
gility in an economic downturn. A bank is vulnerable to a run and a CDO is sensitive
to downgrades on the ratings of the tranches and funding problems as the probability
of default or the correlations increase. This applies especially to conduits and SIVs
(see below).
Both fractional reserve banks and CDOs substitute idiosyncratic risk with systemic
risk. The reason is that both are vulnerable to idiosyncratic risk factors, perhaps
298
problems facing an individual bank or bad CDO modelling. This may enable a rela-
tively small idiosyncratic problem to undermine confidence in the banking system,
spreading and amplifying idiosyncratic risk to systemic proportions.
Rating agencies
The rating agencies are private firms whose job it is to provide investors with information
about the credit quality of financial instruments being offered.
For corporate bonds, especially in the US, investors can rely on over 80 years of rat-
ings history to assess the quality of ratings, and thereafter make an independent deter-
mination of ratings quality. For bonds in other countries, both corporate and sovereign,
the ratings history is much shorter, usually much less than two decades, making it much
harder to ascertain ratings quality.
It is, however, with the various credit instruments where the ratings have been most
severely criticised. The ratings on the many CDO-type instruments were significantly
lacking in quality, ignoring the impact of correlations and how such instruments con-
centrate catastrophe risk. The market for dangerous credit instruments was fuelled by
over-optimistic ratings, and the ratings agencies failed in their mission of providing inde-
pendent quality advice to investors.
There is evidence that rating agencies made significant mistakes. For example, in May
2008, Moody’s acknowledged that it had given AAA ratings to billions of dollars of struc-
tured finance products due to a bug in one of its rating models. In March 2007, First Pacific
Advisors discovered that Fitch used a model assuming home prices could only increase.
15.4.4 Conduits and SIVs

While the misuse of CDOs was an important contributor to the crisis dating from 2007,
two CDO-type structures were especially damaging, conduits and SIVs.
Conduits
A conduit is a simple structure involving a bank that sells some assets to a SPV, for example
credit card receivables, loans or mortgages, in other words, various asset backed securi-
ties (ABSs). The conduit issues an asset-backed commercial paper (ABCP) to fund perhaps
90% or more of the purchases, the rest coming from its equity. The payments from the
assets first go to pay interest on the ABCP; any excess belongs to the equity holder. The
ABCP pays interest based on short-term rates, whereas the return from the assets is based
on longer-term rates. The conduit’s net interest income is comprised of a term structure
component (long rates – short rates) and a credit spread component. The ABCP is usually
AAA-rated due to over-collateralisation, so the cost of funding is low.
The sponsor guarantees funding for the SPV, so if it is unable to borrow sufficient
amounts of money at reasonable terms, the sponsor steps in and funds the SPV. These
liquidity guarantees provide a steady profit to the sponsor when times are good, but sub-
ject it to significant liquidity risk, realised at the worst time when the markets are already
in turmoil.
299
Example 15.2 A conduit as a bank
We can compare traditional banking with conduit banking, by considering the main
balance sheet items:
Old-style bank Conduit
Equity capital Provided by shareholders Provided by sponsor

Debt Short-term deposits Short-term ABCP
Assets Long-term loans Various long-term loans
The conduit bank has direct parallels with a traditional bank, and is subject to
the same main risk factors: default risk, bank runs (liquidity risk) and maturity mis-
matches. The problem is that the vulnerabilities of the conduit bank are less visible
and understood than the vulnerabilities of the traditional bank, and the regulatory
ways to address them less developed.
In good times, structures like conduits allow banks to enhance their earnings by ben-
efiting from the return of the conduit’s assets without having to hold the asset themselves.
This means that banks can avoid regulatory capital requirements on those assets.
If the bank retains the equity portion, it retains most of the risk, both from losses on the
assets and also from the liquidity guarantees. However, the exposure to the conduit was
not covered by bank regulations before 2007, Basel I. This enabled the bank to take more
leverage and more risk than regulations would otherwise allow.
SIVs
A structured investment vehicle (SIV) is similar to a conduit, but rather than just having
equity and ABCP in its capital structure, it also has a mezzanine level, so that senior debt
can be highly rated despite there being only a small amount of equity. This allows greater
leverage. The mezzanine notes are typically of longer duration. They provide higher-
yieldingand longer-term instruments for investors seeking something riskier than ABCP.
Liquidity risks
Perhaps the main danger in these structures from a financial stability point of view is the hid-
den liquidity risk. The sponsoring banks generally provide backup liquidity lines, guaranteeing
funds to the structure if it is unable to raise money elsewhere at reasonable rates. This has two
purposes: first, it increases the security and attractiveness of the structure, and secondly it
provides fees to the sponsor for providing the backup line. Even though the backup lines were
quite risky to provide, they did not attract regulatory capital. This meant that these structures
had a direct advantage over other forms of risk-taking, which generally require capital.
SIVs and conduits face serious funding liquidity risks. Their funding is short term, being
rolled over frequently, often every three months. They are also highly sensitive to lenders’
confidence in their structure, as the ability to borrow depends on the credit quality, since
ABCP investors want to be convinced that the vehicle’s assets are worth significantly more
than the ABCP. If asset values dip even a little, it may suddenly be impossible to roll funding
300
Appendix A: Mechanics of CDSs
over. An ABCP buyer’s strike is therefore the equivalent of a bank run. In this case, the vehicles
may be forced to draw upon the backup liquidity line provided by the sponsoring bank.
In the end, it was the need to fund the structures at a time when the crisis had already
started and funding was scarce, that caused significant problems in 2007, for example
triggering the failure of IKB.
Example 15.3 IKB
While the crisis started in the US, it quickly crossed the Atlantic, and seemed to affect
medium-sized European banks especially badly. IKB, a small, partly state-owned
German bank, was the first European victim of the subprime crisis. It had created a
conduit some years earlier, and at the start of the crisis the assets held by the conduit
suffered losses. Consequently, the conduit was unable to roll over its loans and had
to draw on its credit line from IKB. The amount of money required was beyond the
ability of IKB to provide, and it eventually received a bailout from private entities and
the government amounting to €5 billion.
15.5 Summary
Complexity in financial instruments can be quite profitable when things are good, but
make financial institutions much more vulnerable both to an individual shock and espe-
cially to a systemwide shock.
The growth in volume of the various complex instruments has been exponential in
recent years, especially that of CDSs. This is an instrument that significantly contributed
to the crisis beginning in 2007, not least because of its OTC nature and general lack of
transparency.
Another complex instrument that has proven to be quite dangerous is the CDO. The rea-
son is that CDOs often contained a significant number of subprime mortgages, and mistakes
in modelling created the illusion that high-risk assets could be repackaged into safe assets.
Two financial instruments caused significant problems in the early days of the crisis,
SIVs and conduits, not least because of the embedded but hidden liquidity guarantees.
The rating agencies rated the tranches of the structured credit products with the same
rating categories as corporate bonds, but it turned out that the risk characteristics of these
two asset types were fundamentally different.
Appendix A: Mechanics of CDSs

A plain vanilla CDS is an OTC bilateral contract where in exchange for the payment of pre-
mium, the protection seller (the one writing the CDS) agrees that if any one of a number of
credit events occurs on a reference instrument, they will compensate the protection buyer
(the one buying the CDS) for the difference between the value of the reference instru-
ment after the credit event (recovery value) and par (face value).
301
Terms and definitions

Reference entity The legal entity which borrows money; this may be a sovereign,
financial institution, corporation or another specified entity.
Reference obligation Any holding, obligation, debt, or other credit instrument that
is ‘referenced’ in the transaction.
Notional principal The hypothetical underlying quantity upon which interest or
other payment obligations are computed.
Trade date The date on which the parties agree to the terms of a contract.
Effective date The date on which the parties begin calculating accrued obligations.
Termination date The date on which the CDS expires.
Credit event Any event that happens in respect of the reference entity that triggers
payment under the CDS.
Settlement method The way in which the protection buyer is compensated after the
occurrence of a credit event.
Typically, the reference instrument will be a debt security or a loan, where the allowed
credit events include default, so the protection seller will compensate the protection
buyer if the reference entity defaults. After a credit event, the holder of the reference
instrument will get the recovery value, so the protection seller will suffer a loss of par
minus recovery. Unlike a typical option, the premium on a CDS is usually paid periodi-
cally and this payment terminates in the event of a default.
The range of credit events on a CDS is usually chosen from bankruptcy, obligation
default, violation of covenant, repudiation/moratorium, restructuring, reduction in inter-
est or principal, or the lowering of seniority.
Terminology
‘Buy’ means buying protection:
■ Pay premium
■ Receive default payment if credit event occurs
■ Sells/hedges credit risk
■ Equivalent to selling a bond.
‘Sell’ means selling protection:
■ Receive premium
■ Pay default payment if credit event occurs
■ Buys/takes on credit risk
■ Equivalent to buying a bond.
An example
Suppose that two parties enter into a five-year credit default swap on 1 September 2012.
Assume that the notional principal is $100 million and the buyer agrees to pay 90 basis
points 190100 of 1%2 annually for protection against default by the reference entity. The
CDS is shown in Figure 15.8.
302
Appendix B: CDO calculations
90 basis points per year
Protection buyer Protection seller
Payment if default by
reference entity
Figure 15.8 Credit default swap
If there is no credit event, the buyer receives zero payoff and pays $900,000
1$100,000,000 * 0.0092 on 1 September in each of the years 2013, 2014, 2015, 2016 and
2017. If there is a credit event, a substantial payoff is likely. Suppose the buyer notifies the
seller of a credit event on 1 April 2014. If the contract specifies physical settlement, and
the reference entity is a bond, the buyer has the right to sell bonds to the writer issued by
the reference entity at the full face value of $100 million.
If the contract specifies cash settlement, an independent calculation agent will poll
dealers to determine the mid-market value of the cheapest deliverable bond a predesig-
nated number of days after the credit event. Suppose this bond is worth $35 per $100 of
face value, then the cash payoff would be $65 million 1$100 million - $35 million2. The
total amount paid per year, as a percentage of the notional principal, to buy protection is
known as the CDS spread. We can calculate the implied probability of default from the
CDS spreads.

B.1 Simple example
Consider three bonds, A, B and C, each with a default probability pd = 0.1, where defaults
are independent. What is the chance that exactly two bonds will default? The possible
outcomes with two defaults are A, B or A, C or B, C. We can calculate the number of com-
binations by using the binomial coefficient:
n
a b =
n! , a3b = 3
k k!1n - k2! 2
where n is the total number of elements and k the desired number of elements we want
in a combination, K is the random variable, and k the outcome.
But how to get the probabilities? The outcomes are binomially distributed. The prob-
ability of exactly k outcomes in a sample of size n with independent probabilities pd is
n
g1k n, pd2 = Pr1K = k2 = a b pdk 11 - pd2n - k (15.1)
k
303
Table 15.3 Probability of defaults
Number of defaults Probability of defaults Cumulative probability
0 0.729 0.729
1 0.243 0.972
2 0.027 0.999
3 0.001 1.000
where g() denotes the binomial density. This gives 2.7% in our example. In our case, we
are more interested in at least k defaults in a sample of n, that is, the binomial distribu-
tion, G(), with the outcomes shown in Table 15.3:
G1k n, pd2 = Pr1K … k2 = a a b pd 11 - pd2

k n i n-i
i
i=0
B.2 Worked example from Section 15.4.1

We show the probabilities and cumulative probabilities in Table 15.4 and use the default
probabilities of benchmark corporate bonds’:
■ AAA has pd = 0.3%, interest = 6%

■ BBB has pd = 7%, interest = 8%.
For simplicity, assume that if the benchmark corporate bond defaults, there is no recov-
ery, so for a $1 million AAA bond:
$60,000 + principal 1$1 million2 with probability 99.7%

payment = e
0 with probability 0.3%
Table 15.4 Defaults and probabilities
Number of Probability Cumulative Cumulative probability

defaults of defaults probability from largest to smallest
0 0.05631 0.05631 1.00000

1 0.18771 0.24403 0.94369
2 0.28157 0.52559 0.75597
3 0.25028 0.77588 0.47441
4 0.14600 0.92187 0.22412
5 0.05840 0.98027 0.07813
6 0.01622 0.99649 0.01973
7 0.00309 0.99958 0.00351
8 0.00039 0.99997 0.00042
9 0.00002861 0.99999905 0.00002956
10 0.00000095 1.00000000 0.00000095
304
The SPV has a net inflow of $12 million 110 * $10 million * 0.122 if all goes well, but
there is considerable risk that some of the loans will default so that it is very unlikely that
all the money will come in.
From Table 15.4, the probability of all 10 bonds defaulting is 9.54 * 10-7. The prob-
ability of 9 or 10 bonds defaulting is 9.54 * 10-7 + 2.86 * 10-5. This number is recorded
in the last column of Table 15.4. Going up the table, the probability of 8 to 10 bonds
defaulting is 4.158 * 10-4 6 0.003, while the probability of 7 to 10 bonds defaulting is
0.00351 7 0.003. To summarise:
Pr38 to 10 defaults4 6 Pr3AAA defaulting4 6 Pr37 to 10 defaults4
This means that the probability of getting $3.6 million 13 * $10 million * 0.122
in interest payments from the SPV, 99.958%, is higher than the probability of get-
ting paid from the AAA bond, 99.3%, while the probability of getting $4.8 million
14 * $10 million * 0.122 in interest payments from the SPV, 99.65%, is lower than the
AAA default probability. We conclude that the payment flow from the first three bonds
gets an AAA rating. Similarly for the BBB:
Pr36 to 10 defaults4 6 Pr3BBB defaulting4 6 Pr35 to 10 defaults4
Of the flow from the first five bonds that is less risky than the corresponding BBB, three
are AAA, so the flow from the next two bonds gets BBB.
Tranching
We can tranche the payments in the following way. The SPV has a net inflow of $12 mil-
lion from interest payments 110 * $10 million * 0.122, plus principal, if all goes well.
Flow from the first three bonds, $3.6 million 13 * $10 million * 0.122 + principal, is
safer than AAA, so that it gets AAA rating. An investor will get the AAA interest rate, or 6%
plus the principal back ($30 million).
Flow from the next two bonds, $2.4 million 12 * $10 million * 0.122 + principal,
gets a BBB rating. An investor will get the BBB interest rate, or 8% plus the principal back
($20 million).
The rest of the flow, $12 - $1.8 - $1.6 = $8.6 million + principal, goes to the equity
investor. As she has rights to the last $50 million, her interest rate is 17.2%.
B.3 Correlated defaults

The probability of default can be captured by a random variable whereby if the outcome
of that random variable is below a certain threshold, a default occurs. For example, sup-
pose the random variable X is standard normally distributed:
X ∼ n10, 12
Suppose the probability of default is pd. We can then say that default happens if the out-
come of the random variable is below the inverse normal distribution at the probability pd;
if X … Φ-1(pd2 default
if X 7 Φ-11pd2 no default
305
where Φ is the normal distribution (CDF). Therefore, we have a simple way to deal with
the probability of defaults mathematically, most importantly since this framework ena-
bles us to incorporate correlated defaults, via the Gaussian copula, quite easily. In this
case, we need to make the random variable X correlated across multiple assets. This
means that if we observe a low value of X for one asset, and hence observe a default, we
are more likely to see the same for different assets (if the correlations are positive).
Generally speaking, suppose we have N assets, then we would say that the random
variable X is an N * 1 vector, with each element corresponding to an asset. Therefore, X
X ∼ n10, g2
has the distribution
where g is the covariance matrix. However, we want to consider a simpler case, where the
correlations are all constant, denoted by r. We then have in the three-asset case:
g = c rs2s1
s21 rs1s2 rs1s3
s22 rs2s3 s
rs3s1 rs3s2 s23
This means it is relatively straightforward to set up the problem using common factors,
where the outcome for each asset has two components: the first is the common factor
and the second is idiosyncratic. In our setup, we have constant correlations, so we get:
a = 1r
xi = af + 21 - a2Pi outcome for bond
(15.2)
F ∼ n10, 12 common factor
Zi ∼ n10, 12 idiosyncratic shock
where the factor is F, with common outcome f, whilst the idiosyncratic shock is ei. We
then have that the probability of default conditional on the factor f is
Pr1xi 6 Φ-11pd2 f2
From (15.2) we have that the probability of default, conditional on the factor f, is
Φ-11pd2 - af
Pr1xi 6 Φ-11pd2 f2 = Φ a b
21 - a2
= pd f
We are still left with the factor. Conditional on f we can get the expected cashflow from
the portfolio (and indeed the entire distribution, perhaps for risk analysis). Suppose we
want the probability of k defaults. From (15.1) we get:
n
g1k n, pd, f2 = a b 1pd f2k11 - pd f2n - k
k
306
This is the distribution of outcomes conditional on the factor. Of course, we do

not see the factor and we eliminate it from the equation by integrating it out over its
density:
∞
L-∞
g1k n, pd, f2f1f 2df

1 Who are the main beneficiaries from the complexity of financial products, and who are
the losers?
2 Is the rapid growth of derivatives over the past few decades a cause for concern?
3 There are some proposals for using structured credit products to solve the EU sovereign
debt problem. What sort of product could they be, why are they proposed, and would
they work?
4 Do you think CDSs should be banned or heavily restricted?
5 The EU has effectively prohibited CDSs on European sovereign debt. It argues that the
CDS market contributes to the crisis. Do you agree?
6 What are naked CDSs?
7 How can we use CDOs to turn high-risk assets into safe assets?
8 What is the role of correlations in the creation of CDOs?
9 Some commentators have said that CDOs, especially those containing subprime assets,
are the equivalent of catastrophe insurance, in particular, the highest rated tranches.
Identify the arguments in favour of the statement.
10 Spreads on AAA bonds have traditionally narrowed in times of financial uncertainty,

and when the economy is in a deep recession yield on AAA tend to be very close to
yields on government bonds. Has this been the case for AAA rated tranches of struc-
tured credit products?
11 What is your view on CDO2 and CDO3?
12 When the crisis started in 2007 it was often blamed on US structured credit products
that crossed the Atlantic and adversely affected some European banks. What were
those products, and why were some banks so badly affected?
13 We have three subprime mortgage backed securities, S1, S2, S3, each with a par value
of 1 million dollars. Each of these securities has a 20% default probability, where the
defaults are independent across securities.
Suppose an AAA bond has a 1% probability of default, an A bond 5% and a BB bond 8%.
(a) Calculate the probability and cumulative probability of 0, 1, 2 and 3 defaults.
(b) Suppose you package the securities into a collateralised debt obligation (CDO).
How big would you make the AAA tranche?
307
References
Barnier, M. (2011) quoted in Barker, A., EU ban on ‘naked’ CDS to become permanent. Financial
Times, 19 October, www.ft.com/cms/s/0/cc9c5050-f96f-lle0-bf8f-00144feab49a.html.
Board of Governors of the Federal Reserve System (2009). Revenge of the steamroller: ABCP as
a window on risk choices. Unpublished paper, Division of International Finance.
Buffett, W. (2003). Avoiding a mega-catastrophe. Fortune, 24 March.
Coval, J., Jurek, J. and Stafford, E. (2008a). The economics of structured finance. Harvard
Business School Working Paper 09-060.
Coval, J. D., Jurek, J. W. and Stafford, E. (2008b). Economic catastrophe bonds. Mimeo, Harvard
University.
Crouhy, M., Galai, D. and Mark, R. (2000). Risk Management. McGraw-Hill.
Duffie, D. (2007). Innovations in credit risk transfer: implications for financial stability. Mimeo,
Stanford University.
Duffie, D. (2010). Is there a case for banning short speculation in sovereign bond markets?
Financial Stability Review, 14, Banque de France.
Duffie, D. and Singleton, K. (2003). Credit Risk. Princeton University Press.
Financial Times (2010a). The benefits of naked CDS. Financial Times, 2 March.
Hull, J. C. (2011). Options, Futures, and Other Derivatives, 8th edition. Prentice Hall.
Hull, J. C. (2012). Risk Management and Financial Institutions, 3rd edition. Prentice Hall.
Li, D. X. (2000). On default correlation: a copula function approach. J. Fixed Income, 9: 43–54.
The Economist (2009). No empty threat. Credit-default swaps are pitting firms against their own
creditors. The Economist (US), 18 June.
308
16 Failures in risk management and
regulations before the crisis
The most damaging global financial crisis since the Great Depression started in 2007.
While the underlying causes are varied, failures in financial regulations and financial
risk management played a key role.
By 2007, the prevailing view by market participants and banking supervisors was
that the problem of financial risk had been reduced to an engineering-type exercise.
So long as risk was measured correctly and the appropriate correction mechanisms
were employed, either imposed by banking supervisors or built into internal risk man-
agement, the financial system would remain in an almost permanent state of safety.
This was incorrect. The nature of financial risk was misunderstood because the risk
controls not only failed to protect but also created conditions for excessive risk-taking,
because economic agents have strong incentives to find areas of weakness and exploit
those to make profit.
This happens clandestinely, outside the scrutiny of banking supervisors or even sen-
ior bank management and internal risk control. We learn about excessive risk-taking
only when it is too late and things blow up. To the outside world, the system looks
stable and the general perception of stability creates incentives to take on more risk,
because, after all, ‘what could possibly go wrong?’ In the words of Hyman Minsky,
stability breeds instability.

This chapter relates directly to Chapter 10 (credit markets), Chapter 9 (trading and
speculation) and Chapter 17 (the ongoing crisis: 2007–2009 phase).
Chapter 16 Failures in risk management and regulations before the crisis
It also draws heavily on previous chapters in this book, such as Chapter 1 (sys-
temic risk), Chapter 3 (endogenous risk), Chapter 13 (financial regulations) and
Chapter 15 (dangerous financial instruments). It also connects to Chapter 18 (ongo-
ing developments in financial regulation).
Key concepts
■ Why the safeguards failed
■ Complexity and financial engineering
■ Prudential versus system regulations
■ Toxic assets and capital fragility

A large number of studies have been made of regulation and risk management fail-
ures leading to the crisis, with just about every relevant government body and pri-
vate institution publishing a contribution. Of those, two are particularly important,
the Larosiere (2009) EU report and the the Turner Review by the Financial Services
Authority (2009b). Goodhart (2009) has a good analysis of the regulatory failures.
Since the crisis is not over, no authoritative study has yet been published, and this
chapter is mostly based on the author’s own analysis.
A Assets
C Capital
T1, T2 Tier 1 and Tier 2
w Risk weight
16.1 Regulatory failures

In 2007, most financial institutions were highly regulated and practised sophisticated
risk management. It was felt by both the authorities and the industry that this was
sufficient to prevent bank failures and protect the real economy. After all, we had not
seen a global financial crisis since the 1930s, and subsequent crisis episodes had a well-
understood cause and effect. It might not be an exaggeration to say that both bank-
ing supervisors and risk managers thought they had regulations and risk management
down to science, some tinkering needed on the margins, but generally the system was
thought in good shape. We had seen some crises, for example in Asia in 1997, and LTCM
in 1998, but those were either in developing countries, caused by imbalances thought
not to affect more developed countries, or arising from specific deficiencies quickly
identified and fixed.
Events since have demonstrated the folly of such views; the banks and supervisors were
complacent, excessively trusting in quantitative approaches to regulation and risk man-
agement, and blind to the build-up of risk.
310
16.1.1 Was there excessive deregulation?

One view that is often expressed is that a key cause of the crisis was excessive deregula-
tion. That is not strictly true, and in many aspects the financial system had never been
more regulated than in 2007. During the Bretton Woods era, many governments exercised
strong controls over their banking systems, but this was often more in the area of certain
activities, especially capital flows. The most visible part of the post-Bretton Woods dereg-
ulation was global capital flows, liberalised from the 1950s, and essentially completely
free since the 1980s. Over time, regulations have taken new forms, and sweeping restric-
tions on general activities have been replaced by a much more intrusive micro-approach,
where even minute bank activities are liable to be closely supervised and controlled.
Until the 1980s, national regulations were a patchwork of conflicting and often inef-
fective rules, allowing regulatory arbitrage and casting a blind eye on dodgy practices
taking place in other jurisdictions. Since then, regulations have been generally tightening,
most directly due to the various activities of the Basel Committee for Banking Supervision
(BCBS). While the capital accords are the best known of its activities, the Committee has
also been effective in establishing common international standards for banking supervi-
sion, the Concordat. These have been implemented by member countries, as well as the
European Union (EU) and a number of other countries.
The crisis from 2007 did not happen because of a lack of regulations or deregulation; it
is rather that the regulations were not effective.
16.1.2 The focus of risk management and regulations

If we have to pick a single newspaper headline cause for the ongoing crisis it would be
excessive risk-taking. If the banks had not taken on too much risk, the crisis would not
have happened. While this is of course true, the relevant question is different: why wasn’t
this excessive risk-taking detected and prevented?
Based on the prevailing approaches of risk managers and banking supervisors at the
time, risk-taking was not excessive. How did the excessive pre-crisis risk-taking pass unde-
tected? There are two fundamental reasons. First, both regulations and risk management
systems targeted day-to-day risk and not extreme risk. Perhaps the main reason for this is
that it is much harder to model and manage extreme risk than day-to-day risk. The second
reason is endogenous risk. Most risk models are based on forecasting risk based on histori-
cal samples, which explicitly excludes risk that is not visible in observed data. Both rea-
sons suggest that risk controls focused on what was easy rather than what was relevant.
This applies equally to the industry and to supervisors.
Risk target levels

Financial regulations, and especially the Basel Accords, have a close connection with how
financial institutions internally manage risk. The 1996 amendment to Basel I and espe-
cially Basel II generally require banks to accurately forecast and manage risk with internal
models, and create dual-use risk systems, whereby models are to be used both for inter-
nal risk control and for regulatory purposes such as reporting and capital calculations.
The target risk levels were not very extreme, and ignored tail risk.
311
This can be seen in the calculation of market risk, specified in the 1996 amendment to
Basel I as a multiple of value-at-risk (VaR). Bank capital for trading activities is based on
a multiple of the risk of an event that happens 2.5 times a year on average, VaR99% daily. Such
events do not seem very systemic or a threat to financial stability.
Day-to-day risk and extreme risk
Definition 16.1 Tail risk or extreme risk We refer to the risk of very large and
very frequent adverse outcomes in financial markets as extreme risk. Another name for it is
‘tail risk’ because extreme outcomes happen in the far left tail of return distributions. While
there is no consensus as to what constitutes extreme risk, a useful definition is an event that
happens no more than once every five years on average. We might observe two five-year
events close to each other and none for 10 years. There is no specific time period over which
the adverse event happens. It could play out in 10 minutes as in a flash crash type of event,
in one day as in the 1987 crash, or even over several months as in 2008.
This regulatory and internal risk management focus on particular aspects of risk explains
why the financial industry was able to take on excessive levels of risk undetected. Financial
risk models have become quite successful at forecasting day-to-day risk and banks in turn
have become very good at managing such risk. This success was noted by the supervisors.
Unfortunately, it is much harder to model and understand extreme risk.
This focus on day-to-day risk levels impacted on the distribution of financial returns in
a particular way. Borrowing language from the statistical literature on risk modelling, we
can say that active risk management may cause the volatility to decrease and the tails to
thicken. We see a depiction of this in Figure 16.1 (repeated from Figure 3.9). The blue line
shows what the distribution of market outcomes could be if financial institutions were not
actively trying to manage day-to-day risk, whilst the red line shows what happens when
they do. The impact of active risk management is to reduce the probability of uncommon
With active Without active

risk management risk management
0.5
Risk of very large
Risk level targeted

and uncommon
risk management
0.4
Distribution of risk
outcomes
by active
0.3
0.2
0.1
0.0
−3 −2 −1 0 1 2
Market outcomes
Figure 16.1 Impact of active risk management
312
events, shown by the red line being below the blue line in the sides of the distribution,
but at the expense of the red line being higher in the tails, signalling a high probability of
extreme outcomes.
This approach to risk management led to an attitude well described by an old joke. A
policeman sees a drunk man crawling on the ground at night and asks him what he is up
to. The drunk responds that he is looking for his keys. The policeman says why are you
looking there? The drunk says that is where the light is. The lesson from this story is that
financial institutions and banking supervisors became blinded by their success in finding
risk where it is visible and neglected to search for risk elsewhere.
Unfortunately, those who really want to maximise risk, like bank traders, were actively
searching for the dark areas, so almost by definition, if the supervisors or the bank risk
managers were looking in one area, excessive risk-taking took place elsewhere.
This meant that prior to 2007, the financial system had all the outward signs of low
risk, and the era became known as ‘the great moderation’. Alas, that very perception cre-
ated conditions for excessive risk-taking and the subsequent crisis.
16.1.3 Normality and non-linear dependence

Financial returns are often assumed to be normally distributed. Empirically, that is not
true; instead they adhere to three stylised facts, as discussed in Danielsson (2011). They
have fat tails, volatility clusters and non-linear dependence. The first refers to the obser-
vation that outcomes in financial markets are more extreme than is consistent with the
normal distribution, and the second that volatilities cluster and go through time periods
with high volatility followed by periods of low volatility. Finally, non-linear dependence
(see Definition 3.1) refers to the fact that correlations are inadequate in describing how
financial returns behave, especially in times of stress.
Fat tails
Since many methods in portfolio theory and derivative pricing assume that returns are nor-
mally distributed, they break down in the presence of fat tails. It is, however, in the field of
financial regulations and risk management where the lack of normality is crucially impor-
tant. An inappropriate assumption of normality leads to the gross underestimation of risk:
‘[. . .] as you well know, the biggest problems we now have with the whole evolu-
tion of risk is the fat-tail problem, which is really creating very large conceptual dif-
ficulties. Because as we all know, the assumption of normality enables us to drop
off the huge amount of complexity in our equations [. . .] Because once you start
putting in non-normality assumptions, which is unfortunately what characterizes
the real world, then these issues become extremely difficult.’
Alan Greenspan (1997)
The summer of 2007

The apparent sudden emergence of fat tails and non-linear dependence caused diffi-
culties for many financial institutions in the crisis from 2007. One of the first manifesta-
tions of the crisis was the quant event of 2007, causing some banks to lose large sums of
313
money on quantitative trading. The markets began to move in a way that was inconsist-
ent with the computer models used by banks. This triggered selling, resulting in price
drops, further creating difficulties typical of the feedback loops associated with problems
of endogenous risk. Goldman Sachs’s flagship Global Alpha fund, which used quantita-
tive strategies across a range of asset classes, had lost 27%:
‘“We were seeing things that were 25-standard deviation moves, several days in a
row”, said David Viniar, Goldman’s chief financial officer. “There have been issues
in some of the other quantitative spaces. But nothing like what we saw last week.”’
Financial Times (2007)
Lehman also had problems:
‘“Wednesday is the type of day people will remember in quant-land for a very long
time”, said Mr Rothman, a University of Chicago Ph. D. who ran a quantitative fund
before joining Lehman Brothers. “Events that models only predicted would happen
once in 10,000 years happened every day for three days.”’
Wall Street Journal (2007)
Volatility and fat tails

A common measure of financial risk is volatility or the standard deviation of financial
returns. This is correct if and only if returns are normally distributed. If they are fat, volatil-
ity significantly under-represents risk. If we consider the 25-standard deviation moves that
Goldman suffered a few days in a row, the only appropriate distributional assumption,
in the absence of other information, is normality. A 25-standard deviation loss under the
normal distribution has a probability of 3 * 10-138. By contrast, the age of the universe is
estimated to be 5 * 1012 days whilst the earth is only 1.6 * 1012 days old. This indicates
that Goldman only expected to suffer a one-day loss of this magnitude fewer than one
every 1.5 * 10125 universes.
The explanation is probably that Goldman’s models never considered such extreme
losses and, therefore, were unable to properly quantify the probability. This does point to
the challenges presented by fat tails in risk modelling.
Value-at-risk
Perhaps the most common measure of financial risk is VaR. While it has been widely criti-
cised, most of the commentators do not seem to properly understand the nature of the
measure, in particular those claiming it is dependent on normality. That is not true, even
if actual implementations frequently do so. The fault there lies with the modellers, not
the concept.
VaR does, however, have several flaws, making it inappropriate in most cases for appli-
cations in financial stability.
■ First, it is not sub-additive, meaning it is not relevant for assets that most of the time
deliver steady returns but very rarely suffer large losses. An example of such assets is
high-risk bonds. A misuse of VaR in this context has been at the heart of many losses,
for instance, those suffered by UBS in 2007, as noted by UBS (2008).
314
■ Second, VaR typically applies only to a single-day loss, independent of other days. This
means that a series of small losses individually might not seem too worrying if judged
by VaR, even if the accumulated losses might be so large as to cause a bank to fail.
■ Third, VaR is only as good as the underlying statistical model, most of which are rather
poor. In particular, attempting to model extreme outcomes or multi-day losses with
VaR is generally impossible if an accurate risk assessment is desired.
■ Finally, VaR can capture only perceived risk, not actual risk, because it depends on using
historical data for the risk forecast.
We saw one example of this in the discussion of the JP Morgan Chase ‘whale’ in
Example 9.1. The risk does not seem to have been picked up by the bank’s internal models:
the average VaR for the first three months of 2012 for chief investment office (CIO) activi-
ties was $129 million, compared to $60 million the year before. In other words, the per-
ceived risk was only a small fraction of the actual risk, a clear example of how risk builds
up out of sight. This also demonstrates the problems of relying on risk measures like VaR,
because it is conceivable that JP Morgan Chase had many days with large losses but never
violating its VaR limit on a single day.
16.1.4 Prudential versus systemic regulations

While the problem of non-extreme risk levels and endogenous risk comes from how risk
modelling is practised, a bigger failure arises from the prevailing approach to financial
regulations. Neither the financial industry nor the authorities seemed very bothered by
systemic risk prior to the crisis. Financial regulations were almost exclusively prudential
in nature. That meant banking supervisors were primarily concerned with the risk of each
bank individually, rather than how the risk of all banks could aggregate up to systemic
risk. They disregarded the potential for the prudential approach to banking supervision
leading to the creation of hidden endogenous risk.
Systemic risk was ignored

If the interconnectedness between financial institutions is limited, the safety of the system
can be ensured by each bank being prudentially run. However, in a world where financial
institutions are highly interconnected and interdependent, such an institutional focus can
lead to a situation where each and every institution is perceived as safe, and perversely
because of that prudence, the system becomes unstable.
This can be explained by a simple stylised scenario. Perhaps some asset values fall for
exogenous reasons. The self-preservation principle fundamental to any prudentially run
institution will make it want to sell high-risk assets. But exactly because of that behav-
iour, prices fall further, causing further distress. In other words, because of the exogenous
shock, risk is no longer exogenous and becomes highly endogenous. This is manifested by
the firesale effect, seen in Figure 16.2 (repeated from Figure 1.6).
It is the self-preservation behaviour that is destabilising the system. The only way to
properly prevent excessively large crisis events is to understand the interconnectedness
between financial institutions and design mechanisms to mitigate the implied danger.
Such views, however, ran contrary to the prevailing views of both financial institutions
315
External
shock
Financial difficulties
Prices fall Risk increases

Firesale
Forced
sales
Figure 16.2 Firesales and acceleration effects
and banking supervisors. The banks wanted to focus on their own activities, and for the
supervisors it was hard enough to ensure that each institution was prudently run without
having to consider the more complicated issues of interconnectedness and its contribu-
tion to systemic risk.
Liquidity and asymmetric information

The first sign of the crisis starting in 2007 was the drying up of liquidity. When investors
went on strike and the interbank market effectively ceased, financial institutions through-
out the world, which had been accustomed to effectively infinite amounts of liquidity,
found themselves in serious difficulties. Eventually, the central banks had to step in as
liquidity providers. This liquidity crisis caught the supervisors and the banks unawares.
The focus on prudential regulations also meant that the authorities had ignored liquidity.
The world had not seen a global liquidity crisis since the 1930s and the last regional liquidity
crisis was in Asia in 1997, with liquidity also playing a big role in the LTCM crisis in 1998. One
example of how liquidity was ignored was that some banks were able to create products like
conduits and charge fees for liquidity guarantees for those products, disregarding the fact
that any drying up of liquidity would probably cause the failure of the sponsoring bank.
A related problem arose because of interlocking exposures in over-the-counter (OTC)
derivative contracts like credit default swaps (CDSs) and collateralised debt obligations
(CDOs). The perception at the time was that banks were able to use high-quality internal
models to both price and risk-assess these derivative instruments, and were always able to
hedge or monetise them, that is, liquidity was thought infinite. As it turned out, liquidity
was far from infinite and a key reason was asymmetric information: market participants
stopped trusting valuations and risk assessments made by other institutions.
Liquidity is a system-wide phenomenon, and is therefore easy to miss when regula-
tions focus on individual banks rather than on the system. Liquidity has been absent in
the Basel regulations and while some countries did impose liquidity constraints, such as
the United Kingdom (UK), other main financial centres took a more ambivalent attitude.
316
16.2 Capital and the crisis
16.1.5 Complexity, incentives and resources

The complexity of the banking system directly contributed to the failure of financial regu-
lations before 2007. The reason is that complexity makes it hard for the authorities to have
an overview of risks being created, both within banks and system-wide. While this com-
plexity arises naturally in finance, the problem is made worse because banks have a direct
incentive to maximise complexity.
Complexity is attractive to banks for several reasons. It reduces transparency, mak-
ing it harder to scrutinise their activities and replicate successful trading strategies,
whilst increasing fees. It can create an aura of sophistication, helping in marketing.
Furthermore, complexity directly contributes to financial institutions becoming too
big to fail (TBTF), and hence increasing the probability of bailouts and lowering fund-
ing costs.
The complexity of the financial industry leaves banking supervisors at a distinct dis-
advantage. Banks throw a significant amount of resources at creating complex finan-
cial products and trading activities, effectively maximising complexity. The supervisors
charged with monitoring bank activities do not have access to the same resources even
though their problem is much harder than that of the banks, because the supervisors
have to understand the aggregated risk across the financial system, which includes the
complexity of one institution interacting with another.
The supervisory approach to complexity was inadequate. Instead of addressing
individual practices and their inherent complexity, the supervisors opted to focus
on the end result, either prices or risk forecasts. This misses the big picture because
the problems are in the calculation of the risk forecasts and prices, so the only way
to understand the danger being built up is to study the individual products and
model assumptions, and then aggregate the resulting information across the firm and
industry.

The main body of financial regulations in place when the crisis started was Basel I, but
most jurisdictions and banks had already made a significant step towards Basel II. It
quickly became apparent that many of the criticisms of Basel II, which had been dis-
missed by the BCBS prior to the crisis, were indeed valid. Perhaps the largest failure of the
Basel Accords was demonstrated by bank capital.
Banks around the world found themselves confronted with an unexpected prob-
lem when the crisis started. At the time, they seemed generally to be highly capital-
ised, comfortably exceeding the Basel II capital adequacy ratio (CAR), many at 12–13%.
However, when the crisis was underway, this turned out to be insufficient, not because
a CAR of 13% is not respectable, but rather because the markets did not believe that
13% represented the banks’ true CAR. What happened was that banks were faced with
a twin problem, toxic assets and capital fragility, aided by the practice of capital structure
optimisation.
317
16.2.1 The undermining of capital

Recall the Basel CAR from Definition 13.5:
C = T 1 + T2
CAR = Ú 8% (16.1)
w1 * A 1 + w2 * A 2
It is in the interest of any bank to make the CAR look as large as possible, for a given set
of assets and liabilities. To this end, a bank might employ capital structure optimisation to
maximise the numerator of (16.1) whilst minimising the denominator.
Toxic assets in the denominator

There are two categories of variables in the denominator in (16.1), the assets and the
risk weights. It can be difficult to identify the value of assets because only a fraction of
them are actively traded on the open market and consequently have a market price. The
rest are illiquid, without a market price, so the only way to obtain valuation is by using a
financial model to find the most likely price given the nature of assets and general market
conditions.
It is even harder to calculate the risk weights because, unlike prices, risk cannot be
measured. Instead, it has to be inferred from historical information about the assets, their
inherent properties and general market conditions. In the language of statistics, risk is a
latent variable. The only way to identify risk is by using a statistical model to forecast the
risk. Since every model is incorrect by definition, every risk forecast is incorrect and highly
sensitive to the underlying assumptions of the model.
An example of an asset that has both problems is collateralised debt obligation (CDO)
tranches, since both the valuation and risk assessment come from a pricing model that is
highly sensitive to the underlying assumptions. Before 2007, both the banks and the rat-
ing agencies were too optimistic in their assumptions, assigning too high values and too
low risk to the CDO tranches.
As a consequence, many of the assets held by banks were referred to colloquially as
toxic, meaning that their market value was highly uncertain and their risk weights under-
stated. This meant the markets did not trust the quality of the denominator, and often
ignored the risk weights during the crisis from 2007, assuming the worst.
Capital fragility and the numerator

The numerator in (16.1) was also mistrusted. It is composed of the two main categories
of capital, tier 1 and tier 2. Of those, tier 1 is closer to the ideal qualities of bank capital,
especially the equity part. By contrast, tier 2 capital is imperfect, especially the hybrid
capital instruments.
Hybrids are financial instruments that, at least in theory, provide almost the same pro-
tection as equity but at a lower cost. They broadly define a range of securities which are
positioned between equity and senior debt, being long-term bonds with a special feature
enabling the issuing bank to convert them into equity or defer payments in times of dif-
ficulty. This means that such instruments are able to absorb losses without triggering the
financial institution into liquidation.
318
In principle, hybrid instruments sound like a good idea. They provide stable long-term
funding to a bank, whilst providing protection when things get rough. They are, of course,
inferior to equity. Not only is equity the banks’ own money while hybrids are just bor-
rowed money; equity also is money owned by the banks’ owners, which presumably
incentivises them to behave prudently.
The main problem with hybrids is how they were set up; in particular, it was generally
at the discretion of the bank when a payment could be deferred or the bond converted.
In other words, there was no explicit trigger for conversion, and the experience during the
crisis showed that banks were quite reluctant to trigger conversion.
An important reason seems to have been that banks sold these instruments to pre-
ferred clients with the promise of a high return, and were reluctant to see them suffer
losses. In other words, reputation risk held them back. The banks converted only when
forced to do so by the supervisors during the crisis.
Another key reason was that a conversion into equity is dilutive and thus runs against
shareholder and perhaps management interests.
Finally, doing a conversion signals that the financial institution is in difficulty and, there-
fore, creates the potential for a run, not only on the bank converting but also on other banks.
This meant that the protection afforded by the hybrid capital instruments turned out
to be illusory for the most part; the markets recognised this and ignored hybrids when
calculating bank capital and evaluating the strength of financial institutions.
Capital structure optimisation

The banks are directly to blame for this predicament. Before the crisis, they wanted to
demonstrate they were highly capitalised, and to achieve that they used capital structure
optimisation to maximise the visible part of their capital whilst at the same time keeping
the cost of capital as low as possible. In effect, the number representing capital was max-
imised and the protection minimised.
Both the regulators and the banks seem to have assumed that such optimisation was
benign, or at least not damaging given other sources of protection, such as management
expertise, lending of last resort (LOLR), modern risk management, diversification, the
secular decline in risk, and the like. In other words, capital structure optimisation was
condoned, not because the potentially adverse consequences were not understood, but
rather because other protection factors were thought to be in place.
16.2.2 Case study: European banks in the crisis

When the crisis started in 2007, the financial markets became increasingly concerned
about banks failing, and started to look for signs of weakness. The markets dismissed the
optimistic reported assessments of bank capital and instead focused on the most con-
servative way to calculate bank capital. The banks with the biggest disparity between the
loose and strict ways of calculating capital were deemed most at risk. In turn, this caused
difficulties for institutions that were highly capitalised by the Basel ratio but not by the
conservative measures. These problems were the subject of many studies by financial
institutions, and we draw on numbers from Lehman Brothers (2008).
319
Four capital adequacy ratios

We focus on four different ways of calculating a CAR, where going down this list we find
increasingly restrictive or conservative measures of the capital ratio:
■ Tier 1/RWA This is the more strict Basel ratio, tier 1 capital divided by risk-weighted
assets (RWA).
■ Core tier 1/RWA This focuses only on the equity part of tier 1, since core tier 1 is com-
posed of shareholders’ equity and retained earnings.
■ Tier 1/TA This replaces RWA with total assets (TA).
■ Tangible equity/TA This replaces tier 1 with tangible common equity, the subset of
shareholders’ equity that is not preferred equity and not intangible assets.
Aggregate results
Figure 16.3 shows the four ratios. Of those, tier 1/RWA is trending upwards, therefore, the
Basel capital ratio was sending the signal that the banks were becoming more capitalised
and, hence, safer over time. However, we see that from 2002 equity has been reduced and
other forms of tier 1 have become an increasing part of tier 1 capital.
8.5%
8.0%
7.5%
Tier 1/RWA
7.0%
Core tier 1/RWA
6.5%
2000 2001 2002 2003 2004 2005 2006 2007
(a) Risk-weighted assets (RWA)
3.6%
3.4%
3.2%
3.0% Tier 1/TA
2.8% Tangible equity/TA
2000 2001 2002 2003 2004 2005 2006 2007

(b) Total assets (TA)
3.4% 2000
2002 2003
Tier 1/RWA
3.2% 2001
3.0% 2004
2007 2006
2.8%
2005
7.5% 7.7% 7.9% 8.1% 8.3%

(c) Trend Tangible equity/TA
Figure 16.3 Capital ratios over time

Data source: Lehman Brothers (2008)
320
This is in line with Figure 16.3(b); tier 1/TA first increases sharply but then starts drop-
ping from 2004, whilst the most conservative capital measure, tangible equity over TA, is
steadily dropping throughout the sample period. These results support the analysis of the
impact of financial engineering on the capital structure; the visible reported component
is increasing, but a look at the components that are harder to manipulate shows that the
banks’ capitalisation was getting worse over time.
Finally, Figure 16.3(c) shows a cross-plot of tangible equity/TA to tier 1/RWA, along
with a regression line. We see here the relationship is trending downwards, albeit with a
high degree of uncertainty. This demonstrates how Basel capital and conservatively calcu-
lated capital were evolving at a different rate.
Individual results
A similar picture emerges by looking at individual European institutions in Table 16.1 dur-
ing the first quarter of 2008, when the crisis was becoming serious and considerable focus
was on bank capital. What is interesting is the change in rankings of different institutions
when we switch from using the Basel ratio all the way to tangible equity/TA. Figure 16.4
shows the ranking of capital, identifying four categories of banks.
CS went from being the best capitalised bank to below average, whilst UBS went from
being the tenth lowest capitalised bank in the sample to the lowest. Both of these banks,
especially the latter, experienced significant difficulties in the crisis. DB, similarly, went
from 35 to 2. Banks like HSBC maintained a fairly consistent ranking across all the measures.
The sample contains some institutions that ran into difficulties. Interestingly, AIB,
which was now in serious trouble with the sovereign debt crisis in Ireland, was looking
relatively healthy, whilst RBS was showing clear signs of financial engineering.
Figure 16.4 shows this same information in a different way, as a cross-correlation
between tier1/RWA and tangible equity/TA, where we identify banks that have received
some government assistance. There seems to be little or no connection between needing
help and having a weak capital base; if anything, the most highly capitalised banks with
the best capital were more likely to get into trouble.
Table 16.1 Capital ratios and rank amongst peers in Q1 2008
Tier 1 Tangible equity Distress

RWA TA
Allied Irish Bank (AIB) 7.6% (23) 5.8% (43) Fatal

Credit Suisse (CS) 11.5% (50) 3.6% (22) Considerable
Deutsche Bank (DB) 8.7% (35) 1.3% (2) Considerable
HSBC 9.1% (37) 4.2% (33) Moderate
Royal Bank of Scotland (RBS) 7.3% (19) 2.2% (7) Fatal
Santander 7.77% (28) 5.3% (39) Moderate
SocGen 6.7% (8) 2.3% (10) Moderate
UBS 6.9% (10) 0.4% (1) Considerable
Note. Ranks: 1 worst capitalised, 50 best capitalised

Data source: Capital Advisory Group: Current topics, Technical Report (Lehman Brothers 2008), Epiq Systems, Inc.
321
Little capital Highly capitalised

structure optimisation Alpha Bank
Eurobank
BES
40
BPM
AIB
MPS
Kaupthing
Santander
Landsbanki
High conservative
Sabadell
DnBN NOR
Stan Char
capital
Glitnir
30 HSBC
Rank by tangible equity/TA
Banesto
Anglo Irish
BPI
BBVA Popolare
Nordea
ISP CS
Low capital
20
High capital
Erste Bank structure optimisation
BCP
SEB
HBOS
Lloyds
BNPP
A&L
BankInter
Commerz
10
Low conservative
DB SocGen
Dexia
capital
RBS B&B
Barclays
CASA
DB
UBS
0
Low Basel capital High Basel capital
0 10 20 30 40
Rank by tier 1/RWA
Figure 16.4 Capital rankings of European banks in Q1 2008. Those that failed or got
significant government help between 2007 and July 2012 in red
Data source: Lehman Brothers (2008). Note: 1 worst capitalised, 42 best capitalised
Analysis
These results do not provide much indication that bank capital is doing what is expected
of it; at least, it is hard to reconcile the fact that many of the most highly capitalised insti-
tutions failed, while many of the low capitalised banks survived the crisis, and amongst
those that seemed most active in capital structure optimisation, most survived.
However, using TA, as in the leverage ratio is also problematic. After all, if two banks
have the same amount of capital and one buys government bonds and the other junk
bonds, both would have the same leverage ratio, but the second bank is a lot more risky.
For this reason, it is better to have banks meet multiple capital ratios at the same time,
which is indeed what Basel III aims to do.
These results are consistent with those of Blundell-Wignall and Atkinson (2010) who con-
tend that banks’ ability to arbitrage the capital weights to reduce capital and expand leverage
is very extensive. They give a simple example showing that the Basel risk-weighting approach
has allowed banks to expand their leverage almost without limit for all practical purposes.
322
References
Ultimately, this provides a good argument for transparency. Instead of trusting the
bank to put assets into the right risk bucket and apply the correct formula for diversifica-
tion and unlikely events, the market could make that judgement independently. What the
crisis demonstrated is that if the market is not given information, it is likely to make crude
worst-case assumptions, which does not benefit anybody.
16.3 Summary
While there are many reasons for the crisis that started in 2007, failures in financial regula-
tions and risk management were a direct contributor.
If regulations and risk management function as intended, risk-taking by banks is con-
tained and bank failures, and certainly near-systemic crises, are highly unlikely. A key rea-
son why the crisis happened in 2007 is that financial regulations and risk management
at the time focused on frequent non-extreme events, neglecting tail risk, whilst ignoring
endogenous risk and the systemic aspect of bank activities.
Perhaps the single most identifiable failure of regulations was in the Basel Accords and
especially bank capital regulations. Prior to the crisis, banks appeared to be highly capi-
talised, but this was mostly illusory, because banks used capital structure optimisation to
manipulate their CAR.

1 Was excessive deregulation and liberalisation of financial markets a key contributor to
the crisis from 2007?
2 It is sometimes said that the outcome of active risk management is to lower volatility and
fatten the tails. What does this mean, and do you agree?
3 Does the focus of prudential regulations potentially increase systemic risk?
4 How did financial engineering undermine the integrity of bank capital?
5 What is the problem of toxic assets in bank capital?
6 Did the hybrid capital instruments work effectively in the early stages of the crisis from
2007?
7 What you think about the assumption in the Basel accords and EU laws that sovereigns
are risk free?
References
Blundell-Wignall, A. and Atkinson, P. (2010). Thinking beyond Basel III: necessary solutions for
capital and liquidity. OECD Journal: Financial Market Trends, Volume 2010, Issue 1.
Danielsson, J. (2011). Financial Risk Forecasting. John Wiley & Sons.
Financial Services Authority (2009b). The Turner review: a regulatory response to the global
banking crisis. Technical report, European Central Bank. www.fsa.gov.uk/pubs/other/turner_
review.pdf.
323
Financial Times (2007). Goldman pays the price of being big. Financial Times, 13 August.
Goodhart, C. (2009). The Regulatory Response to the Financial Crisis. Edward Elgar, Cheltenham, UK.
Greenspan, A. (1997). Discussion at symposium: Maintaining Financial Stability in a Global
Economy, p. 54. Federal Reserve Bank of Kansas City.
Larosiere, J. (2009). The high-level group report on financial supervision in the EU. Technical report.
http://ec.europa.eu/commission_barroso/president/pdf/statement_20090225_en.pdf.
Lehman Brothers (2008). Capital advisory group: current topics. Technical report, Lehman
Brothers.
UBS (2008). Shareholder report on UBS’s write-downs. Technical report.
Wall Street Journal (2007). One ‘quant’ sees shakeout for the ages – ‘10,000 years’. Wall Street
Journal, 11 August.
324
17 The ongoing crisis:
2007–2009 phase
The worst global financial crisis since the Great Depression started in August 2007,
reached a peak in the second part of 2008 and seemed to be over by 2009. However,
that turned out to be wrong — a second phase started in 2010 and, at the time of
writing, it is getting worse by the day. In this chapter the focus is on the first phase,
whilst Chapter 19 addresses the European sovereign debt crisis.
The underlying causes of the crisis are familiar. Banks took on too much debt whilst
not recognising the increasing underlying risks, in the short run making everybody
feel better off. The complexity and the hidden risk and the shadow banking system
enabled those desiring to ignore problems to continue investing.
It all ended in tears. Crises require people to sell assets below the price at which
they bought them. Market participants came to realise that their assumptions were
wrong, that valuations had been too high, so out of sync with reality that a correction
became inevitable. This happened in 2007.
This chain of events is well known to policymakers who had set in place safeguards
both to prevent such excesses building up and to protect us from the eventual failure.
The safeguards failed. Developments in the financial system had outpaced the regula-
tory and supervisory structure. Financial institutions were increasingly taking advantage
of globalised markets, the structure of banks was changing and they were increasingly
dependent on models and complexity for profits whilst relying more and more on short-
term wholesale funding. All of this meant that financial institutions were becoming more
interconnected and dangerous. Regulations did not keep up; the Basel II Accord, imple-
mented in 2007, represents state-of-the-art regulatory thinking from the late 1990s.
As financial institutions became ever more international, the potential for cross-
border banking failures was mostly ignored, with no effective internationally
Chapter 17 The ongoing crisis: 2007–2009 phase
coordinated supervision or resolution regimes. Even in Europe, with its common mar-
ket in financial services, supervision was a strictly national affair, and the supervisors
in different countries did not speak much with each other. The regulatory and super-
visory structure in place in 2007 was inadequate for the banking system at the time.

Many aspects of the crisis have been discussed in detail in other specialised chapters.
There are parallels with Chapter 2 (the Great Depression, 1929–1933), Chapter 14
(bailouts), Chapter 15 (dangerous financial instruments) and Chapter 16 (failures in
risk management and regulation). Chronologically, the discussion here builds directly
on those chapters and does not repeat their content.
Key concepts
■ First and second globalisms
■ Hidden risk
■ How banking has changed
■ Investors’ strike
■ Liquidity and bailouts
■ Bear Stearns, AIG and Lehman
■ The fall of 2008
■ Subprime
■ Response to the crises

Many books and papers have been written about the crisis. We made especial use of
the following. Murphy (2009) provides a good discussion of the role played by credit
instruments in the crisis, Bitner (2008) discusses the role of subprime mortgages,
Lewis (2011) discusses the overvaluation of structured credit products, while Huertas
(2010) gives a perspective of a senior official of the FSA. Brunnermeier (2008) has a
blow-by-blow description of early events.
17.1 Build-up to a crisis

It is difficult to identify the causes of the crisis of 2007–2009 because of the number of
different factors that came together. Commentators will invariably disagree about the
causes. We are still debating the causes of the Great Depression 70 years after it ended,
and are likely to debate the causes of this crisis for decades to come. This is a measure of
the complexity of the financial system and the heterogeneity of commentators.
17.1.1 Globalism
The crisis was a product of the relatively liberalised economic environment that fol-
lowed the collapse of the Bretton Woods system in the early 1970s. The global economy
has been gradually opening, with increasingly free trade and capital flows, along with
326
deregulation and privatisation. If there is one defining term for the general economic envi-
ronment, it is globalism. This is not, however, the first but the second time globalism has
been the prevailing philosophy.
Throughout this period of ever increasing globalism, we did not suffer a global finan-
cial crisis, and when this crisis happened, everybody was ill-prepared.
The first globalism

In understanding what went wrong, it is useful to draw lessons from the first globalism,
from the second part of the nineteenth century up until the first World War (WWI).
Globalism was even more prevalent in many aspects in that epoch. The exchange rates
of the major currencies were fixed because of the gold standard, capital flows between
countries were mostly unhindered, and passports had yet to be invented.
Then, as now, in a highly interconnected financial system, a relatively small liquid-
ity crisis could trigger significant and widespread losses, simply because of the two key
vulnerabilities in a financial system: confidence and interconnectedness. This was at the
heart of three crises discussed earlier, in 1866, 1906 and 1914, and many others. The prob-
lem of liquidity was well understood by bankers and policymakers, and policy responses
reflected that understanding.
The second globalism

Throughout the Bretton Woods era, the global economy was heavily regulated, with trade
and capital flows restricted. While this did not prevent economic crises, it was more effec-
tive in preventing financial crisis, as noted by Reinhart and Rogoff (2009), with crises mostly
national affairs. If links between countries are limited, crises stay within national borders.
The break-up of the Bretton Woods system in the early 1970s signalled the start of the
second era of globalism, both trade and capital flows were liberalised, countries became
ever more interconnected and banking truly international. This created new challenges
for policymakers, challenges that were not adequately recognised.
Money was being made in new places, in the Middle East, whose petrodollars were
‘recycled’ to countries seeking capital inflows, typically Latin America. From 1975 to 1982
Latin American borrowings from abroad increased from $75 billion to $313 billion. This
meant that finance became much more important than before, and the financial system
had become a major source of systemic risk for the first time since before the second
World War (WWII).
History was forgotten

The lessons of the first globalism, in particular the potential for international liquidity
crises, seem to have been forgotten by many of the architects of financial regulation and
financial stability policies in the era of the second globalism. In particular, the design of
financial regulations and our approach to financial stability did not adequately recognise
the dangers inherent in the global economy. Financial regulations remained focused on
the prudent behaviour of individual institutions, and central banks were more concerned
with monetary than financial stability. It is almost as if international liquidity crises were
thought impossible. History was forgotten.
327
The Washington consensus and anti-globalism

Two conflicting philosophies defined the attitude towards the second globalism, the
Washington consensus and anti-globalism. The International Monetary Fund (IMF) was
in charge of global currency arrangements during the Bretton Woods era, and after it
collapsed, the Fund became the champion of the Washington consensus. As capital
flows surged in the 1970s, sovereign debt problems became increasingly common but
individual banks felt ill-equipped to deal with them. During the first globalism, this
problem was solved by gunboat diplomacy, but during the second globalism, the IMF
assumed the role of debt enforcer. It provided emergency loans for countries in diffi-
culty, demanding in return what it called structural adjustment programmes. Generally,
these programmes followed the Washington consensus. While the structural adjust-
ment programmes might have been sensible, the way they were implemented became
quite controversial and often counterproductive, as we saw in the discussion of the
Asian crisis. That set the stage for a backlash against the Washington consensus –
anti-globalism.
Whilst the prevailing economic orthodoxy until 2007 was quite consistent with the
Washington consensus, the crisis has changed that, and we are increasingly seeing the
various elements of anti-globalism impacting on policy. Even the IMF has referred to
itself as ‘new and cuddly’, eschewing old policies, for example, by advocating capital
controls.
17.1.2 Monetary policy and bubbles

The financial system was heavily regulated in the 1970s; the big enemy was inflation. In
order to conquer it, central bankers, encouraged by academic research, developed the
doctrine of central bank independence and the primacy of monetary policy. The problem
is that the success in fighting inflation, and the importance of maintaining stable monet-
ary policy, let central banks drop their guard and neglect financial stability.
We now see, with the benefit of hindsight, that the central bankers became a little like
generals who dwell on the winning tactics of the last battle and thus get blindsided by a
new enemy. They became complacent because they were able to maintain low interest
rates and enjoy low inflation.
The factors enabling this success were not properly recognised, not least the impact of
China. Its production costs were steadily falling, not only making China a strong competi-
tor but also keeping prices low. In effect, China was exporting deflation. This worked to
counteract the inflationary impacts of the low interest rates.
The Greenspan put

After inflation was conquered, some central banks, especially the Federal Reserve System
(Fed), proceeded to use monetary policy as a means to contract financial shocks by pro-
viding massive amounts of liquidity. Just two examples are the reaction to the LTCM crisis
and 11 September 2001.
Those liquidity injections became known as the Greenspan put – after Alan
Greenspan, then chairman of the Fed – as the Fed stood by to bail out the financial
328
markets when a shock hit, providing a put option. The impact was to increase moral
hazard and stimulate asset bubbles, both in equities but perhaps more importantly in
American real estate.
Risk-averse short-term investors

There was another impact from the low-interest policy. Significant amounts of money are
held by traditional risk-averse short-term investors, such as municipalities and other gov-
ernment authorities facing asynchronous tax revenues and expenditures, money market
mutual funds, and the like.
While this type of investor traditionally might have placed its funds in government
paper, the low interest rate and low perceived risk environment encouraged such
investors to seek more yield in still ‘very low risk’ assets, which in turn stimulated the
market for the various highly rated structured credit products. When the crisis hit in
the summer of 2007, the first manifestation was the withdrawal of those investors from
the asset backed securities markets. It was said that they ‘went on strike’, but more
accurately they simply rebalanced back to their traditional investments – short-term
government securities.
17.1.3 Savings glut

It was quite easy prior to 2007 to raise very large amounts of money on the capital mar-
kets, and some commentators have claimed that this is due to the fact that there were
more savings in the world than could be invested wisely. Two ingredients are the expan-
sion of pension funds and the massive build-up of currency reserves, for example, in the
East Asian countries reacting to the crisis of 1997. Governor Ben Bernanke in 2005 coined
the term global savings glut, which indicates that worldwide there is too much saving rela-
tive to available investment opportunities. This means that the global savings glut directly
contributed to the problem of global imbalances.
This assertion was challenged by Borio and Disyatat (2011) who argue against the pre-
sumption that net capital flows from current account surplus countries to deficit nations
helped finance credit booms there and that a rise in saving relative to investment in sur-
plus countries depressed world interest rates. Their argument is that available data give
only a partial picture of capital flows, revealing little about financing and excluding finan-
cial asset-based transactions which are the main part of cross-border activities. They fur-
ther argue that the impact of interest rates is based on the stock of debt, reflecting the
interaction between central bank rates, expectations about future policy rates and risk
premiums.
17.1.4 Efficient markets

Some observers have argued that the efficient market hypothesis (EMH) is one of the
causes of the crisis. It is a theory developed by financial economists, saying that market
prices are efficient in the sense that one cannot on average achieve systematic excess prof-
its by forecasting prices with available data and technology.
329
A narrow notion of the EMH captures the inability to systematically make money
by forecasting. This is not problematic and is supported by empirical evidence. After
all, if markets were not efficient in that sense, one could easily make almost infinite
profits. This is the view that is most prevalent in academic studies referring to efficient
markets.
There is, however, another interpretation of the EMH which is the view that a belief
in efficient markets makes people blind to the faults of the market. For example, former
Fed chairman, Paul Volcker, said in 2011 that it was ‘clear that among the causes of the
recent financial crisis was an unjustified faith in rational expectations [and] market effi-
ciencies’. This view is more common amongst non-economists, who often attribute the
wider meaning to those using it formally in the narrow sense.
17.2 Hidden and ignored risk

Prior to the crisis, it is almost as if all of the relevant stakeholders had what we could call
the 2003–2006 mindset. The general perception was that we had solved the problem of
financial risk, by applying ever more sophisticated statistical models, modelling the finan-
cial system on an almost atomistic level, where risk was understood and controlled. This
period is sometimes called the great moderation.
Underneath, the situation was different: the very fact that risk was perceived as low
motivated market participants to assume ever more risk. The authorities charged with
overseeing risk did not understand what was going on. This is a clear example of Hyman
Minsky’s observation that stability is destabilising. The very fact that we perceive the
situation as being stable motivates us to take on more risk, which then destabilises the
system.
Why risk was undetected

Recall a quote from the first chapter of this book by Andrew Crockett in 2000:
These words were very prescient, written seven years before the crisis erupted. But
what are the mechanisms that allow risk to build up whilst hidden away?
Pro-cyclicality and endogenous risk

Perhaps the strongest factor is pro-cyclicality, capturing the tendency of the components
of some system to amplify cyclical behaviour, be it the business cycle or the valuation
cycle. It makes booms bigger and busts deeper. Pro-cyclicality is fundamentally a mani-
festation of endogenous risk.
Pro-cyclicality is an inevitable part of finance. Banks have a tendency to over-extend
themselves when things are going well, lending too much to borrowers engaged in spec-
ulation and contributing to an asset price bubble. A virtuous feedback loop is formed
330
17.3 The changing nature of banking
between credit expansion, speculative investments and rising prices. Eventually, the bub-
ble bursts.
The complexity of finance and the use of models add to the pro-cyclicality. The com-
plexity enables financial institutions to create ever more categories of instruments, sup-
posedly increasing efficiency, but often this complexity just serves to hide risk-taking. A
clear example of this is shown by the postmortem analysis done by UBS into its losses at
the beginning of the crisis (UBS, 2008). Amongst the many findings is inappropriate use of
value-at-risk (VaR) models for credit exposures.
Financial models contribute to pro-cyclicality. They are generally based on short price
histories, and so long as prices are going up and risk is perceived as low, the models give
a signal to buy, creating a bubble. Risk estimates are always influenced by recent experi-
ence, which can be especially dangerous when the pace of innovation is high. At that
time, new products are introduced that did not exist in more stable times; innovation is
aimed at avoiding perceived risks and gives a pretext for the idea that this time is different
and risks are really lower.
Leverage and capital

The 8% regulatory minimum capital adequacy ratio (CAR) allows banks significant lev-
erage. During normal times this is not a problem, but it does make them vulnerable to
shocks affecting their capital and assets. One reason is that one would usually expect indi-
vidual losses to be relatively uncorrelated, so that a full loss of the 8% capital would seem
quite unlikely. However, an assumption of diversification is crucial for this assumption to
work. In a crisis, correlations between assets tend to increase sharply, and diversification
does not work as well – an example of non-linear dependence. That makes it much more
likely that a bank will suffer losses large enough to lose most or all of its capital, as did
happen to some in the crisis.
17.3 The changing nature of banking

After most financial crises, the postmortem analysis identifies the failures allowing the
crisis to happen, often resulting in safeguards preventing a repeat of events. Considering
the long history of banking crises, we might have expected the underlying risks in bank-
ing to be sufficiently well understood so that large-scale failures are prevented. One rea-
son why this turned out to be incorrect in 2007 is that banking had been changing in
nature in a way that was poorly understood, creating hidden risks, whilst regulations did
not keep up.
Funding
The most expensive way for a bank to fund itself is via capital, followed by long-term bor-
rowing. The shorter the maturity of funds, the cheaper they are. For that reason, prior to
the crisis, financial institutions, driven by competitive pressures, were increasingly resort-
ing to the shortest possible funding – overnight (Brunnermeier, 2008, reports up to 40%
for large American banks). Of course, any bank’s assets have much longer maturities, so
331
that such a funding strategy gives rise to a significant maturity mismatch. We see the fund-
ing options in the following table:
Capital (very expensive)

Long-term financing (expensive)
Short-term financing (cheap)
Three-month commercial paper
Repo (one day) cheapest
While overnight funding is the cheapest for banks in normal times, it maximises liquidity risk.
One of the first things to happen when the crisis started in 2007 was a drying up of the inter-
bank market, causing banks relying on overnight funding to face a significant liquidity squeeze.
In principle, borrowing overnight from the interbank market is not much different from
depending on retail deposits, which are payable on demand. In practice, the differences
are significant: overnight repos are much more dangerous.
While both sorts of funding are subject to bank runs, sophisticated investors in the
interbank market are much more likely to do a run than retail depositors. The reason is
that they monitor their counterparties very carefully, and spot problems quickly. By con-
trast, retail depositors tend only to react after the problems hit the news. A good example
of this was the failure of Northern Rock, where the institutional run took place over a
month before the retail run.
Shadow banking
A traditional, even caricature, view of a bank has it collecting demand deposits and
making long-term loans. While such a picture has always been over-simplistic, in recent
years it is increasingly outdated, not least as banks have embraced what is called shadow
banking. This enables banks to move risk into entities directly under their control but
not a part of the balance sheet. This can help in tax and risk management, and aid in
capital structure optimisation. One example of this is conduits as banks as discussed in
Example 15.2.
Shadow banking is not a term with a single precise meaning, but generally refers to
institutions and banking practices that exist outside the traditional regulated banking
sector.
Shadow banking enabled regulated financial institutions to conduct business outside
the glaring eye of bank supervisors, shareholders, accountants and other stakeholders, in
a way that was legal and compliant with extant regulations.
The move towards shadow banking is a clear example of Goodhart’s Law. By putting
heavy regulations on bank capital and assets, regulators created the incentives that made
bank capital and assets poor measures of financial solidity. Shadow banks were one of the
mechanisms used.
It is worth noting that many structures such as SIVs and conduits that were off-balance
sheet and, hence, a part of the shadow banking system, under regulations prior to 2007,
no longer enjoy that status under the Basel II Accord, which took effect in 2007 in many
countries.
332
17.4 Crisis, 2007–2008
Given the long time it took to move Basel II from proposal to implementation – seven
years – if that process had been faster, some of the problems causing the crisis might have
been prevented, since banks would not have been able to take on so much risk without
it being noticed.
17.4 Crisis, 2007–2008

There were increasing signs that something was underfoot early in 2007. HSBC revealed
losses of $10.5 billion on mortgage-backed securities in February 2007, soon to be fol-
lowed by other firms. The first financial institution to fail – the canary in the coalmine –
was IKB in July 2007.
The impact was predictable and major deleveraging was soon underway. Investors con-
verted risky assets into cash, inevitably creating a downward pressure on prices. As always
happens at the onset of crisis, historical correlations began to increase sharply, exactly
what endogenous risk predicts.
The VIX, often called the ‘markets fear gauge’, started increasing in early 2007, as seen
in Figure 17.1, but the volatility levels remained below those of earlier in the decade.
The quant event of 2007

The IKB failure was quickly followed by a quant crisis, where many prop desks and
hedge funds using statistical arbitrage methods saw very large losses. One reason is
that they underestimated the non-linear dependence in return distributions, whereby
correlations sharply increase in a crisis. The high dependence across trading strategies
and funds meant that trades got crowded, and the heterogeneity in behaviour wasn’t
all that high.
In July 2007, shares began to move in ways that were the opposite of those predicted by
computer models. Trigger selling by the funds as they attempted to cover their losses and
meet margin calls exacerbated the share price movements. Just one example is Goldman’s
flagship Global Alpha fund which lost 27% of its value. It was eventually closed in the
autumn of 2011 after sustaining more losses.
40
30
20
10
0
2000 2001 2002 2003 2004 2005 2006 2007
Figure 17.1 Intraday high VIX in 2000–2007 and long-run mean

333
The eye of the storm

After the main crisis event in August 2007, it appeared that the worst was over. However, fun-
damental problems had been exposed. Investors went on strike. In the autumn of 2007 and
the winter of 2008, more and more financial institutions started to face liquidity problems.
Bear Stearns
The weakest of the American investment banks, Bear Stearns, started facing serious dif-
ficulties in early 2008. As losses mounted, in March 2008 the New York Federal Reserve
Bank (NYFed) gave a $29 billion loan to JP Morgan to facilitate its takeover of Bear Stearns,
buying it at $10 a share. Bear traded at $172 in January 2007, and at $93 in February 2008.
The Chairman of the Fed, Ben Bernanke, defended the bailout in the Senate Banking
Committee, saying ‘Given the exceptional pressures on the global economy and finan-
cial system, the damage caused by a default by Bear Stearns could have been severe and
extremely difficult to contain.’
Temporary calm
While the bailout temporarily prevented widespread disruption in financial markets, it
was highly controversial and created the expectation that the authorities should similarly
bail out other banks that were considered too big to fail (TBTF). This temporarily calmed
the markets, but by September 2008 two important financial institutions were facing dif-
ficulties, Lehman Brothers and AIG.
17.4.1 Lehman Brothers

Perhaps the most disruptive event of the crisis was the failure of Lehman Brothers on
15 September 2008. It was one of the largest financial institutions in the world, an invest-
ment bank under US rules, and suffered large losses on real estate. Its bankruptcy was a
turning point in the crisis. It triggered a collapse in asset prices and an almost complete
drying up of liquidity. The US came under heavy pressure to bail Lehman out, but it did
not do so, maintaining that there was no legal authority for a bailout. It remains contro-
versial whether that was a mistake.
DILEMMA 17.1 The controversial decision not to bail out Lehman Brothers
Those arguing that this was the right decision maintain that if Lehman had been bailed
out, it would have cemented expectations by the markets that any bank would be
bailed out, encouraging risk-taking and simply creating ingredients for a larger failure
down the road – a pure example of moral hazard. Lehman’s failure crystallised a prob-
lem already present, making both market participants and the authorities realise the
gravity of the situation and creating the conditions for the eventual crisis response.
Those arguing in favour of a bailout say that because of the failure, global liquid-
ity almost completely dried up, with interbank lending, trade financing and the like
falling to zero, setting the world economy on the road to collapse. The failure created
334
17.4 Crisis, 2007–2008
unnecessary uncertainty, not only causing the severe crisis of the last part of 2008, but
also leading to the European sovereign debt crisis. By bailing out Lehman, the authori-
ties would have been given more time to implement measured, negotiated and less
damaging crisis-fighting methods.
17.4.2 AIG
The day after Lehman Brothers failed, one of the world’s largest insurance companies,
AIG, received a bailout from the NYFed. It was considered more systematically important
than Lehman Brothers because it was the largest writer of CDSs in the world, and it was
feared that its default would trigger a systemic crisis, especially coming right after the
failure of Lehman. For an overview of how AIG got into such problems with CDSs, see, for
example, Lewis (2009).
AIG and CDSs
‘It is hard for us, without being flippant, to even see a scenario within any kind of
realm of reason that would see us losing one dollar in any of those transactions.’
Joseph J. Cassano, the AIG executive in charge of the
credit default swap (CDS) unit, August 2007
AIG was one of the world’s largest insurance companies with a trillion-dollar balance
sheet, 116,000 employees and operations in 130 countries. It set up a London-based bank
that quickly became the world’s largest seller of CDSs. The reason why AIG was able to
become such a large writer of CDSs was that it had an AAA rating, making it the safest pos-
sible counterparty, requiring the smallest haircuts. It wrote about $450 billion worth of
corporate CDSs, which suffered small losses, and about $75 billion of subprime-mortgage
CDSs which suffered more losses after the crisis started. This eventually caused AIG to be
downgraded, increasing its funding costs and haircuts, in a typical vicious feedback loop.
Given the danger posed by the failure of AIG, the authorities felt they had no choice
but to provide a bailout, eventually amounting to $130 billion. Note, however, that the
losses to taxpayers will be much smaller, even though the final amount will not be known
for some time. The US taxpayer also gave tens of billions of dollars to financial institutions
that were counterparties to AIG, most to Goldman Sachs and Deutsche Bank.
While the collapse of AIG could be handled more surgically by avoiding bailing out
its counterparties, the existing resolution regime did not allow for that option. Given the
alternatives, the bailout was the lesser of two evils, as the failure of AIG might have trig-
gered a systemic crisis.
17.4.3 The fall of 2008

The failure of Lehman and the bailout of AIG triggered the worst phase of the crisis in the
autumn of 2008. Global liquidity dried up and financial institutions depending on the
interbank market found themselves without funding. This was the high point of the global
335
credit crunch. The extreme risk levels are clearly visible in the VIX, seen in Figure 17.2,
which shot up in September.
The crisis threatened to cause a repeat of the Great Depression, however, as we dis-
cussed in Section 2.4, the authorities took on board the lessons from the Depression and
did the necessary to stop the liquidity crisis in its tracks.
A clear manifestation of the determination of the authorities was the scale of the vari-
ous bailouts provided to the financial sector, as discussed in Section 14.5. In the absence
of these bailouts, it seems likely that large parts of the European and American banking
systems would have failed, with catastrophic consequences for the real economy.
Figure 2.6 showed the collapse in world trade during the Great Depression. We add to
that figure current trade data in Figure 17.3, and show both time periods scaled to begin
at 100. World trade fell significantly in the second half of 2008, but quickly recovered.
17.4.4 The crisis appeared over by 2009

By the middle of 2009, it seemed that the worst was over, the VIX returned to its long
mean by year end, as seen in Figure 17.2, and the stock markets were improving, as indi-
cated in Figure 17.4.
Banks were no longer in difficulty, the economy started to recover and the stock markets
were rising. This, however, was just the eye of the storm; the second round was coming.
80
60
40
20
0
2007 2008 2009
Figure 17.2 Intraday high VIX in 2007–2009 and long-run mean

140
120
100
80
60 Great Depression
40 Current crisis
1929 1930 1931 1932 1933

2007 2008 2009 2010 2011
Figure 17.3 World trade (USD), scaled to start at 100

Data source: World Trade Organization (WTO) and League of Nations
336
17.5 Was it a subprime crisis?
100
90
80
70
SP 500
60 FT 100
50 DAX
2008 2009 2010
Figure 17.4 Major stock indices, 1 August 2007 = 100

17.5 Was it a subprime crisis?

In the early stages of the crisis, it was often called a ‘subprime crisis’ because of the preva-
lence of US subprime mortgages in the most infamous structured credit products. This
begs the question of whether the subprime mortgages were to blame or whether their
role in the crisis was coincidental.
In 2007, total credit in the US was $50 trillion; $14 trillion was owed by the household
sector, of which $11 trillion were mortgages. However, subprime mortgages constituted
only $1.3 trillion and their delinquency rate hit 13% at the height of the crisis in 2008,
or $170 billion. Therefore, the subprime mortgage sector accounted for only 2.6% of US
debt and subprime delinquencies for only 0.3% of overall debt. By comparison, a 3%
change in the US stock market amounts to perhaps $500 billion.
The amount of damage apparently caused by such a small sector was quite impressive,
and to understand how that happened one has to look at how the subprime mortgages
were used and the acceleration effects created by the resulting instruments. One mecha-
nism was the $4.5 trillion of derivative contracts linked to subprime mortgages.
The role of subprime was only ancillary; the banks demanded high-risk assets during
the bull market prior to the crisis. This did not seem all that worrying because the falls
in risk premiums made historical returns look excellent. Perceived risk was low. One of
the most obvious ways to exploit this was through structured credit products containing
subprime mortgages. If it had not been subprimes, the market would have used some-
thing else. The real destruction was caused by the accelerator effects inherent in these
products.
Conduits, structured investment vehicles (SIVs) and the assumption of diversification
gave banks the opportunity to expand their exposures to extreme levels, whilst simulta-
neously hiding the underlying risks. The sudden vast amounts of capital invested in risky
assets forced up prices, whilst lowering perceived risk, creating a classic bubble and laying
the groundwork for the crisis.
The crisis itself was inevitable because the structured products gave the appearance
of low risk — low enough to fool both the banks and their supervisors. Because of the
337
tendency of the industry to extrapolate such low-risk environments to infinity, eventually

the underlying risk would have been so high and the valuations so unrealistic that the
crisis was bound to occur. The fact that it happened to involve subprime mortgages or
structured credit products was ancillary.
17.6 Policy response

After the crisis started, it took the authorities some time to realise what was happening.
The first authority to respond was the Fed, but the European authorities, both central
banks and supervisors, were much slower to act. While we can only guess as to the rea-
son, an important reason seems to be that the European authorities did not understand
the various financial instruments that played a central role in the crisis.
In the UK, the Bank of England (BoE) had been downplaying the importance of finan-
cial stability in the preceding years, whilst the Financial Services Authority (FSA) was more
concerned with micro-prudential regulations. In Europe, the European Central Bank (ECB)
did not have any supervisory mandate, with no European supervisor or agency concerned
with financial stability.
Example 17.1 The United Kingdom
One example is from Alistair Darling (2011), the then UK Chancellor (minister of
finance) who described the governor of the BoE, Mervyn King, as ‘impish’ and ‘incred-
ibly stubborn’ and excessively focused on moral hazard, maintaining that ‘a penalty
interest rate must apply to any help given by the Bank’, reflecting Bagehot’s rules.
Darling further quotes King as saying ‘He regretted not having confronted these issues
before’, referring to the need for liquidity in the second part of 2007.
Lessons from history

Most financial crises are fundamentally the same, only the details differ. This crisis is
no different. There are many historical parallels, where events have often been eerily
similar – the wilful ignorance of risk build-up, the failure of protection mechanisms,
and early fumbling policy responses. Financial crises always have liquidity as a central
theme.
The historical event that resonates most with this crisis is the Great Depression. Even
though it had faded from memory, the institutional safeguards put in place after the
Depression are, for the most part, still in place and policymakers were quick to brush up
on their Depression lessons. In this, we are fortunate that the governor of the Fed, Ben
Bernanke, has written extensively about the Depression.
When the authorities finally woke up to the seriousness of the crisis, it was as if they
started with a list of mistakes made in the 1930s and said ‘we are going to do the correct
thing this time around’.
338
References
17.7 Summary
The focus in this chapter has been on the first part of the ongoing global crisis, 2007
to 2009. The background to the crisis is the era of the second globalism following the
collapse of the Bretton Woods system, with ever increasing international capital flows,
especially to the US, helping to maintain very low interest rates. This was not a cause for
concern because large countries exported deflation, keeping inflation low.
The central theme is that it was a classical financial crisis, but with many unique ingredients
such as inadequacies in financial regulations and excessive amounts of financial engineering.
The visible crisis was the manifestation of inappropriate investment decisions, poor
risk estimates and mistaken assumptions on liquidity, all made over the preceding years.
Subprime was the trigger, but its role was only ancillary.
Fortunately, the authorities had studied the policy mistakes during the Great
Depression, and been determined not to repeat them.

1 Identify the main economic factors leading to the crisis.
2 Describe a typical crisis/bubble and analyse the 2007–2009 crisis in the context of that.
3 What is the role of ‘accelerators’ in crisis?
4 What was the role of complex assets and marking to market/model/magic in the crisis?
5 How has banking changed over the past decades and how did that make banks more
vulnerable?
6 What were the initial signs of the crisis in the summer of 2007?
7 What were the main symptoms of the crisis when it reached its peak?
8 Was it correct to bail AIG out?
9 Many commentators maintain the failure to bail Lehman out was the main factor in the
crisis becoming so severe. Today, European authorities use the example of Lehman as a
justification for supporting European banks. Do you agree?
10 Was the crisis response of the authorities appropriate?
References
Bernanke, B. S. (2005). The global saving glut and the US current account deficit. Speech deliv-
ered at the Sandridge Lecture, Virginia Association of Economists, Richmond, VA.
Bitner, R. (2008). Confessions of a Subprime Lender: an Insider’s Tale of Greed, Fraud, and
Ignorance. John Wiley & Sons.
Borio, C. and Disyatat, P. (2011). Global imbalances and the financial crisis: link or no link?
Technical report, BIS. Working Paper 346.
Brunnermeier, M. (2008). Deciphering the 2007–08 liquidity and credit crunch. J. Econ.
Perspect., 23(1): 77–100.
339
rr000921b.pdf.
Darling, A. (2011). Back from the Brink: 1,000 Days at Number 11. Atlantic Books, London.
Huertas, T. F. (2010). Crisis: Cause, Containment and Cure. Palgrave Macmillan.
Lewis, M. (2009). The man who crashed the world. Vanity Fair, August.
Lewis, M. (2011). The Big Short: Inside the Doomsday Machine. Penguin.
UBS (2008). Shareholder report on UBS’s write-downs. Technical report.
Volcker, P. (2011). Financial reform: Unfinished business. www.nybooks.com/articles/archives/
2011/nov/24/financial-reform-unfinished-business.
340
18 Ongoing developments in
financial regulation
Reforms of financial regulations tend to come after crises. Then, after some decades
have passed and those who stood watch during the crises have retired, standards
relax, until it all starts over again.
We have already seen some examples of this in this book. The Overend and Gurney
(O&G) crisis of 1866 led to lending of last resort (LOLR) policies, and the Great
Depression motivated a number of new regulations like deposit insurance and the
Glass–Steagall Act, while regulatory failures in the United States in the early 1980s led
that country to develop prompt corrective action. Such crisis response is not always
positive, as there is great political pressure on policymakers to ‘do something now’
and quick policymaking is often bad policymaking.
One of the main causes of the crisis from 2007 was the failure of financial regu-
lations. In response, G20 member countries, represented by the Financial Stability
Forum (FSF), issued an influential report in 2008, identifying five priority areas of
what it saw as the post-crisis regulatory reform process: ‘Strengthened prudential
oversight of capital, liquidity and risk management. Enhancing transparency and val-
uation. Changes in the role and uses of credit ratings. Strengthening the authorities’
responsiveness to risks. Robust arrangements for dealing with stress in the financial
system.’
Not surprisingly, a large number of initiatives have been launched since 2007
by various nation states, government organisations, financial institutions, lobbying
groups, academics and other assorted pundits. The mother of all reform battles has
been launched, and we are seeing the most substantial restructuring of financial regu-
lations since at least the 1930s.
Chapter 18 Ongoing developments in financial regulation
At a risk of overgeneralisation, there are three different approaches being dis-

cussed. Some claim that the entire financial system is unsound, requiring a root and
branch reform, perhaps even to the extent of being moved to state hands. Opposed to
that are those who argue the system is fine as it is, even identifying state interference
as the cause of the crisis. Such views, whilst not common in the early stages of the
crisis, have gained increased traction more recently. The mainstream view is between
those extremes, maintaining that financial regulations need significant reform, and
the structure of the financial industry needs to be altered, but without fundamentally
altering the landscape.
It is often very difficult to disentangle the motives of those participating in the
debate, and frequently debate participants are more motivated by personal gain or
deep-seated ideologies than a desire to improve the system. For some, the crisis is an
opportunity to press some agenda, succinctly described by Rahm Emanuel, former
White House Chief of Staff to President Barack Obama, in 2007: ‘You don’t ever want
a crisis to go to waste; it’s an opportunity to do important things that you would oth-
erwise avoid.’
The two main parties to the regulatory debate are the authorities and the indus-
try. Regulators and supervisors have come under intense pressure from their political
masters to ‘do something about the financial system’ and in response have launched
a number of initiatives. The G20 has taken a lead on the international stage in guid-
ing the regulatory developments, with the Basel Committee for Banking Supervision
(BCBS) playing a key role.
The governmental responses have been far from uniform, exposing national differ-
ences. Countries with large financial centres – large exporters of financial services –
want to protect their financial sectors, whilst others may prefer to cut the financial
sector down in size. The way the various governments approach the reform process
shows a lot of opportunistic behaviour, since a particular regulatory initiative may
adversely affect a competing financial centre and favour the sponsoring country.
Within Europe, the crisis has provided opportunities for state aid to national champi-
ons that might under more normal circumstances not be allowed.
The financial industry has responded with a furious lobbying effort, arguing that
the various initiatives increase complexity, are not effective, are unfair, are discrimi-
natory, or prevent financial institutions from fulfilling their role of supporting the
economy, for example, because of the adverse impact of new capital regulations on
the small and medium-sized enterprise (SME) sector.
Underneath it are fears about new regulations adversely affecting profitability and
the competitive landscape. Regulations do not affect banks uniformly. A bank find-
ing that its competitors will be more strongly and adversely affected by new regula-
tions than it will, is likely to support those regulations. This has exposed the different
preferences of various parts of the system, with banks and non-banks, large and
small institutions, and institutions in different countries all having different agen-
das. For example, the main technical advocacy group of large internationally active
banks, the Institute of International Finance (IIF), has a very different agenda than
smaller banks.
342
Regulatory reform is technical, and the language often impenetrable. Add to that
the multiple agendas being pursued, the many special interests being protected,
and the fact that the discussion tends to take place in closed meetings, and it is
clear that the debate is complex and difficult to follow.
However, at the time of writing, half a decade has passed since the 2007 crisis
started and a picture is starting to emerge on the post-crisis regulatory landscape. For
a roadmap, see Baudino (2011).
New government agencies have been tasked with the study of systemic risk, even-
tually to develop regulations and strategies for addressing systemic risk. This falls
under the general heading of macro-prudential regulations. The Basel Accord has been
identified as needing significant reform, and it will see changes in the calculation of
capital and the introduction of liquidity constraints. Enhanced monitoring and crisis
response is a central plank of the new regime. Individual countries have also launched
important initiatives, and we have seen a reshuffle of supervisory agencies.
Unfortunately, it appears that some areas identified as key contributors to the crisis
from 2007 are unlikely to be effectively addressed. The most important of these is
the too big to fail (TBTF) problem. TBTF institutions played a central role in attenu-
ating market discipline and breeding excess confidence before the crisis. A signifi-
cant effort is going into this area by the various authorities, but the core problem is
likely to remain. In fact, it has been made worse, by encouraging mergers and further
entrenching the assumption that national champions will receive national support.
Less clarity exists on risk-taking. We hear politicians complain in one sentence
about excessive risk-taking and the need to do something about risk, and in the next
sentence demanding more lending to the politically sensitive SME sector. The view
of the policymakers, reasonably enough, is that the financial system should be taking
risk primarily in areas that benefit the economy rather than just speculating. In prac-
tice, no such clear delineation exists.
At the moment, chastened by recent and painful experience, financial institutions
are content to run with modest amounts of risk. However, past experience indicates
that this reticence is unlikely to survive competitive pressures for long, nor that regu-
lators will be very effective in enforcing it against the will of market participants.
Different regulatory initiatives may sometimes be in conflict with each other. Since
insurance companies need to invest large amounts of reserves, they are a large tra-
ditional buyer of bank bonds. One issue has arisen with Solvency II, the European
insurance regulations to be implemented in 2014. It will induce insurance companies
to hold shorter-dated instruments, while banks will need to issue longer instruments.
This interferes with the insurers’ ability to fulfil their social role, and raises questions
where bank funding is going to come from. However, this remains controversial and
several studies, such as that by Gorter and Bijlsma (2012), maintain that the impact
will be relatively small.
Such conflicting views highlight the difficulty in quantifying the impact of new reg-
ulatory initiatives, as debate participants often have compelling arguments, backed
up by detailed studies, but with widely different conclusions and even methodologi-
cal approaches.
343
We have already discussed several regulatory initiatives in Chapter 9, in particular,

central counterparties (CCPs), financial transaction tax (FTT), restrictions on bonuses
and narrow banking, and will not repeat that discussion here.
The online chapter at www.GlobalFinancialSystems.org contains the latest devel-
opments in the regulations debate.

This chapter directly builds on the previous two chapters on financial regulations,
Chapter 13 (financial regulations) and Chapter 16 (failures in risk management and
regulations), as well as Chapter 9 (trading and speculation). The motivation for these
reforms is the ongoing crisis, discussed in Chapter 17 (the ongoing crisis: 2007–2009
phase) and Chapter 19 (sovereign debt crises).
Key concepts
■ Post-crisis regulatory reform
■ New institutions and the changing role of old institutions
■ Basel III
■ Flexible capital buffers
■ SIFIs, G-SIBs and too big to fail

A large number of proposals and studies for regulatory reform have been made over
the past few years and we have drawn on these reports, many of which can be found
on the websites of the various agencies, for the BIS www.bis.org, the BCBS www.bis.
org/bcbs/, the FSA www.fsa.gov.uk, the Fed www.federalreserve.gov/, the FSB www.
financialstabilityboard.org and the IIF www.iif.com. A good collection of articles on
the ongoing developments is Quagliariello and Cannata (2011).
18.1 New and changed institutions

The crisis that started in 2007 demonstrated that the institutional setup of the various
government agencies charged with overseeing the financial system was inadequate. Many
agencies seemed to live in their own silos, focusing on narrow mandates and ignoring the
overall objectives of financial regulations, perhaps inevitably for bureaucratic institutions
established with specific legal mandates.
18.1.1 National institutions

Central banks and supervisors
The role of central banks has changed significantly. Before the crisis, the central banks
tended to focus on monetary policy, with supervision moved to separate institutions.
Financial stability was a low priority and the concept of central bank independence sacro-
sanct. After the crisis started, some central banks, most prominently the Bank of England
(BoE), were initially quite reluctant to engage with a broader mandate of financial stabil-
ity, but in the end did not have much choice.
344
The crisis demonstrated that the central banks have to play a key role in financial sta-
bility because they have a monopoly on printing money, one of the vital instruments in
support and resolution operations. However, they can play this role effectively only if they
also have appropriate supervisory information and powers, and the willingness to use
them. The result is that the central banks have been gaining new responsibilities.
This is the case in the United Kingdom, where many of the functions of the Financial
Services Authority (FSA) are to be merged with the BoE, leaving only micro-prudential regu-
lations. Similarly, in the US, the Securities and Exchange Commission (SEC) used to supervise
investment banks, but the remaining investment banks have become bank holding compa-
nies and, hence, are subject to prudential regulation by the Federal Reserve System (Fed).
The SEC remains a significant regulator in issues relating to securities markets and trading.
However, the central banks have also lost some power to the treasury/ministries of
finance, as any bailouts need government approval, and the new systemic risk institutions
enshrine the role of the treasury, especially in the US.
The treasury
The crisis demonstrated the key role played by the treasury because it manages the public
purse. Vast amounts of public money were allocated to fighting the crisis and, since the
treasury is ultimately responsible for public money, it necessarily took a pivotal role. This
does not mean the treasury is better equipped to do this than the central banks or the
supervisors; far from it, the technical expertise lies with the latter. One could say that the
treasury has regained some of its historical role lost the when the central banks became
independent.
While it is inevitable that the treasuries are intimately involved in the regulatory reform
process, and need to take the lead in any bailout operation, this does not come without
cost. The role of the treasury has led to an unfortunate politicisation of financial regula-
tion. The reason is that government agencies like the supervisor or central bank are one
step removed from the political process. The treasury is directly under political control.
This inevitably means that the treasury is more populist, more susceptible to lobbying.
The political leanings of the minister, and the political advisers he or she makes use of,
can directly affect the regulatory reform agenda. Not only can this lead to sharp differ-
ences between various countries, it may also contribute to regulation risk since we may
expect rapid policy shifts when governments change.
18.1.2 International institutions

Financial regulations have been designed and coordinated internationally at least since
the early 1970s. This is mostly done under the auspices of the BCBS, most importantly
with the Basel Accords.
G20 and FSB

Since the crisis, the lead in coordinating the international efforts in financial regulations
and crisis response has been taken by the G20. In a summit in 2009, the G20 called for
the establishment of the Financial Stability Board (FSB) based on an older body called
the FSF. The mandate of the FSB is to monitor the global financial system, coordinate with
345
other relevant bodies like the International Monetary Fund (IMF) and make recommenda-
tions to member states.
IMF
One international organisation has taken on a new life with the crisis, the IMF. By early
2007, the Fund was in serious financial difficulties, laying off staff, and there was specula-
tion it might be shut down altogether. Its reputation had been damaged by the previous
crises it engaged with, in Asia in 1997, and in Turkey and Argentina in the early 2000s.
No large economy had requested Fund assistance after Argentina, and because interest
on loans to countries in crisis is the major source of revenue for the IMF, it was suffering
financially from the lack of crises. By 2006, it had found itself without a mission or money.
All of this changed in 2007. As individual countries got into serious difficulties, it was felt
that the only body with sufficient expertise, authority and funds to help those countries was
the IMF. It has taken a pivotal role in the European crisis, being involved with Cyprus, Iceland,
Hungary, Greece and other countries, and has become one of the main players in formulating
the crisis response. It has a very limited role, however, in formulating regulatory policy.
EU
The crisis exposed significant weaknesses in the approach of the European Union (EU)
to financial regulation and supervision. Financial regulations are to a significant extent
designed and implemented on a Union level, but implemented only on a state level.
Before the crisis, different national supervisors had limited interaction with each other
and often applied widely differing standards to regulations. The European Central Bank
(ECB) was focused only on monetary stability, and there was almost no macro-prudential
policymaking within the EU. The EU has been addressing this problem since 2007, and
has taken major steps towards more centralised European supervision via new institutions
like the European Systemic Risk Board (ESRB), the European Banking Authority (EBA) and
the European Securities and Markets Authority (ESMA). The EU is currently planning the
Single Supervisory Mechanism (SSM) as an embyonic pan-EU supervisor.
18.1.3 New systemic risk institutions

The crisis has led to the creation of new types of institutions explicitly responsible for sys-
temic risk. The US established the Financial Stability Oversight Council (FSOC) in 2010,
whose mandate is to identify and monitor risk to the US financial system and respond to
threats to financial stability. The chairperson is from the Treasury. The counterparty in the
UK is the Financial Policy Committee (FPC), with a different institutional setup because
the Treasury has only a non-voting role.
The situation is somewhat different in the EU because it is unable to create a systemic
risk body with the same powers as its counterparts in the UK and the US. Therefore, the EU
version, the ESRB set up in 2010, has a much more limited role. It is supposed to prevent or
mitigate systemic risk, taking into account macroeconomic developments and contribut-
ing to the smooth functioning of the internal market. The ESRB is hosted and supported
by the ECB.
346
This does leave a vacuum, and at the time of writing it is unclear how it will be filled.
The EU would like to fill it with a new powerful institution tasked with Union-wide super-
vision and macro-prudential regulations, but this has run into significant opposition from
various member states. If the EU is unsuccessful, the power vacuum will probably be filled
by the ECB at Union level and various state bodies like the FPC.
18.1.4 European banking union

To overcome the problems caused by the lack of a single European supervisor, the EU
established the EBA in 2011, to increase transparency in the European banking system
by identifying capital structure weaknesses. It is best known for the stress tests it runs
on European banks. Those have been somewhat hampered by the problem of sovereign
debt, which as we noted in Chapter 13 is considered risk-free by the EU for the purpose
of bank capital calculations, when the debt is funded and issued in the currency of the
issuer. This means that sovereign debt is not included in the stress tests, undermining
their credibility.
The EU announced in June 2012 its intention to establish a banking union in Europe,
creating a pan-European supervisor, and explicitly link discipline and control from the
EU to assistance to individual member states. The ECB is involved with the mechanism,
implying that it will be to some extent separate from it. The motivation for the establish-
ment of the banking union is to maintain financial stability, preserve the single market in
financial services and avoid competitive distortions in the single market.
The EU would want the banking union to have at least four key elements. The first is
what is known in European Union jargon as the single rulebook, meaning that the same
rules are to be applied uniformly throughout the Union. Second, banking supervision will
be pan-European. The third element is a pan-European resolution regime, and finally a
harmonised deposit insurance scheme. The SSM is a step towards this goal.
It remains unclear at the time of writing what the banking union will look like, when it
will be established, or even if it will be established. At the moment, each member state
maintains its own supervisor, and establishing a new supervisory agency is not a trivial
undertaking.
The notion of a banking union has run into significant opposition, led by Germany. A
banking union might lead to stronger states having to pay for deposit insurance payoffs or
provide bailouts to banks in weaker member states. This means that an effective banking
union requires a transfer union, a political step many countries are reluctant to take.
18.1.5 International cooperation and colleges of supervisors

Many regulations are harmonised internationally to some extent, for example via the activi-
ties of the International Organisation of Securities Commissions (IOSCO) and the BCBS, and
various other similar arrangements. Within the EU, a significant number of regulations are
pan-European, even if there are important national differences in specific implementations.
The situation is different for the enforcement of banking regulations, and the crisis
from 2007 demonstrated the fragmented nature of banking supervision. Until that point,
347
supervision was essentially a national affair. Supervisors in different countries had a lim-
ited overview of the overseas activities of financial institutions operating in their juris-
diction. This even extended to the EU, where it became especially problematic because
banks were able to operate throughout the Union because of the common market, but
without any centralised overview of their activities.
In response to this, various initiatives are underway to coordinate the activities of super-
visors. An important example of this is the activities of the Senior Supervisors Group, rep-
resenting 10 major banking nations. Within this work, colleges of supervisors have regained
a renewed importance, referring to multilateral working groups of relevant supervisors
working together to ensure consolidated supervision of internationally active banking
groups. Colleges of supervisors have become one of the key components in the ongoing
reform of financial regulation. However, enforcement is the last frontier of coordination
and cooperation, and there is some way to go.
While cooperation among supervisors has improved considerably since the crisis, there
have been some hiccups, most famously in the case of UK bank Standard Chartered,
accused by the New York Department of Financial Services of money laundering. It
informed the FSA only 90 minutes before announcing the allegations, and appears not
to have cooperated much with other US supervisory agencies, even though some of them
had supervisory powers over the bank.
18.2 Basel III

The most important changes in financial regulations are in the Basel Accords. After the
start of the crisis, the BCBS moved with surprising speed and in short order introduced
the next revision of the capital accords, called Basel III, at the end of 2009 in a series of
consultative documents, which have been steadily updated since then. The Committee
identifies the system as a main priority:
‘The objective of the [regulatory] reforms is to improve the banking sector’s ability
to absorb shocks arising from financial and economic stress, whatever the source,
thus reducing the risk of spillover from the financial sector to the real economy.’
Basel Committee on Banking Supervision (2011b)
These priorities are consistent with the definitions of systemic risk in Section 12.1.
The main impetus of the reforms is on capital, prioritising equity and introducing sev-
eral categories of capital buffers. In addition, the proposals include liquidity regulations
and the management of counterparty risk. At the time of writing, the latest version of the
Basel III was published by the BCBS in June 2011, and we follow that here.
18.2.1 Capital
The Basel III proposals make several important changes to how capital is calculated.
Under Basel II, the main parts of capital are tier 1 and tier 2, and since most of the direct
problems caused by capital relate to tier 2, it is not surprising that it is to be significantly
reduced in importance and most emphasis put on tier 1, especially equity.
348
18.2 Basel III
Types of capital
Regular bank capital will continue to be split into tier 1 and tier 2, and the minimum capital
adequacy ratio (CAR) will remain at 8%. However, the relative amount of tier 1 will increase,
with especially more emphasis put on equity, that is the common equity tier 1 (CET1). For
technical details on the terminology of the capital regulations, see BCBS (2011c).
Several new categories of equity are to be introduced. A capital conservation buffer can
be used to absorb losses during stress, while a countercyclical buffer fluctuates with the
general state of the economy, and global systemically important banks (G-SIBs) are to get
a separate capital charge. Therefore:
total regulatory capital = tier 1 + tier 2

+ conservation buffer
+ countercyclical buffer
+ G-SIB surcharge
By 2019, if all goes according to schedule, the capital ratios will be as indicated in
Table 18.1, with the transition in steps over the coming years as seen in Table 18.2.
Table 18.1 Eventual components of capital

Category Ratio
Tier 1 6%(CET1 = 4.5%)

Total capital (does not include those below) 8%
Add-ons, all CET1:

Conservation buffer 2.5%
Countercyclical buffer 0%–2.5%
Systemically important banks 1%–2.5% ( + 1%)
Data source: Basel III: a global regulatory framework for more resilient banks and
banking systems, Technical Report, 2011b, Annex 4 (Basel Committee on Banking
Supervision 2011), Bank for International Settlements
Table 18.2 Timing of the transition
Year of Common Conservation Minimum Minimum total +

introduction equity buffer tier 1 buffers
2013 3.5% 4.5% 8%

2014 4.0% 5.5% 8%
2015 4.5% 6.0% 8%
2016 – 0.625% – 8.625%
2017 – 1.25% – 9.25%
2018 – 1.875% – 8.875%
2019 – 2.5% – 10.5%
Data source: Basel III: a global regulatory framework for more resilient banks and banking
systems, Technical Report, 2011b, Annex 4 (Basel Committee on Banking Supervision 2011),
Bank for International Settlements
349
Conservation buffer
The idea behind the conservation buffer is to solve one of the worst problems with capital
buffers. They do not provide any protection if they cannot be used in times of crisis. It
therefore makes sense for financial institutions to be able to reduce their capital levels in
times of difficulty, and that is the purpose of the conservation buffer, allowing for up to a
2.5% reduction in capital when needed.
Drawing down the buffers is not without consequences. Banks are restricted in pro-
viding distributions to shareholders, other capital providers and employees, because
these stakeholders are supposed to bear the risk that recovery will not be forthcoming.
Furthermore, the banks are likely to be given a deadline for rebuilding their buffers.
Countercyclical capital buffer

The risk sensitivity of capital amplifies the problem of pro-cyclicality in banking. This has
become especially pertinent after the crisis in the debate of risk-taking by banks and lend-
ing to SMEs. This problem is partly addressed in Basel III by the introduction of a counter-
cyclical buffer that should be maintained when the economy is doing well but is reduced
during recessions.
While such a buffer is sensible in theory, in practice, it can be difficult to decide on
what measure should be used to signal where we are in the business cycle. Many propos-
als have been made, such as credit/GDP, various asset prices, funding spreads and credit
default swap (CDS) spreads, credit conditions surveys, real GDP growth and data on the
ability of non-financial entities to meet their debt obligations on a timely basis. There is
a difficult trade-off in implementing those. The most accurate, and the hardest to put
numbers to, are based on national accounts, such as GDP, but that is measured only with
a significant lag: it takes many months for the state of the economy to be known and such
numbers may even be revised for a number of years. As a consequence, a GDP-based
buffer might not be very countercyclical; it might even become procyclical, allowing for a
reduction in capital when the economy started to grow, or requiring an increase in capital
as the economy is heading for a recession.
While other measures are more up-to-date, they are still lagging behind the state of
the economy and, hence, may have unforeseen consequences for cyclicality. In addition,
such measures are likely to be more noisy, not least because they achieve shorter lags by
relying on data from markets, and these may not always reflect the broader economy in
the desired manner.
It is not surprising, therefore, that the formulation of the countercyclical buffer has
been controversial – see Borio et al. (2010) for more discussion, and Repullo and Saurina
(2011) for critical analysis. The BCBS converged to a credit-to-GDP ratio with respect to its
trend. This has problems of its own, and we can look forward to banks arguing that trend
rates have increased during future upswings, reducing their need for capital.
Systemically important financial institutions (SIFIs)

Financial institutions that are so large that their failure threatens a systemic crisis are
denoted in the Basel jargon as systemically important financial institutions (SIFIs). When
applied only to banks, we use the term global systemically important banks (G-SIBs).
350
18.2 Basel III
The reason why AIG was bailed out, and why the failure of Lehman was so damaging, is
exactly because they were systemically important.
Not surprisingly, G-SIBs are singled out in the Basel process:
‘The assessment methodology for global systemically important banks [incorpo-

rates] an indicator-based approach and comprises five broad categories: size, inter-
connectedness, lack of substitutability, global activity and complexity.’
BCBS (2011c)
G-SIBs are to have an extra buffer of 1–3.5% of CET1, though the last 1% of that is cur-
rently an empty bucket designed to be a disincentive for banks to get bigger.
The G-SIB surcharge is intended, at least in part, to reflect the extra risk created by
G-SIBs and to neutralise the advantage the TBTFs are felt to enjoy. This is part of several
policy initiatives aimed at this category of banks, discussed in Section 18.6 below.
Table 18.3 shows the latest list of institutions deemed to be G-SIBs, along with their
capital charges.
CoCos
The problems associated with hybrid bank capital, as discussed in the last chapter, have
led to the introduction of capital instruments with a more explicit trigger called contin-
gent convertibles (CoCos). While there are several different versions of CoCos, one of the
most common is bonds that turn into equity if a bank’s capital ratio becomes too low,
while another type involves write-downs. A key distinction between CoCos and other
capital instruments is that the former focuses on explicit early triggers and hence early
interventions, while the latter is more related to resolution and disincentives.
Several banks have already issued CoCos, but the only nation requiring their use is Swit-
zerland, where UBS has issued write-down CoCos and Credit Suisse (CS) convertible CoCos.
While CoCos are not a direct part of the Basel III proposals, there are some indications
that they may become included in some form, as hinted by the statement that ‘all classes
of capital instruments fully absorb losses at the point of non-viability before taxpayers are
exposed to loss’ (Basel Committee on Banking Supervision, 2011a).
If CoCos are to become a key part of bank capital, the interesting question is who
holds them. If it is only other financial institutions, systemic risk is not reduced, and may
Table 18.3 G-SIBs as of November 2012
Bucket CAR Banks
5 3.5% (Empty)
4 2.5% Citigroup, Deutsche Bank, HSBC, JP Morgan Chase
3 2.0% Barclays, BNP Paribas
2 1.5% Bank of America, Bank of New York Mellon, Credit Suisse, Goldman Sachs,
Mitsubishi UFJ FG, Morgan Stanley, Royal Bank of Scotland, UBS
1 1.0% Bank of China, BBVA, Groupe BPCE, Group Crédit Agricole, ING Bank,
Mizuho FG, Nordea, Santander, Société Générale, Standard Chartered,
State Street, Sumitomo Mitsui FG, Unicredit Group, Wells Fargo
Data source: Update of group of global systemically important banks (G-SIBs) Annex 1 (Financial Stability Board 2012),
Bank for International Settlements
351
even increase because of cross-exposures. It is important, therefore, to have restrictions on

CoCo ownership, and to ensure that the owners are willing and able to hold the equity if
the conversion takes place.
A major issue is whether there would be a market interest in CoCos, given their possible
conversion, legal-investment issues, and the like. For example, natural buyers of bank debt,
such as insurance companies and pension funds, might find the potential for conversion
unpalatable at a time when their other assets are already under stress. However, banks have
found little difficulty placing CoCos in the market, suggesting that these fears are misplaced.
18.2.2 Leverage ratio

One of the reasons why many banks got into so much difficulty in 2008 was a general
mistrust of the risk weights on assets in the calculation of the Basel CAR, as discussed in
Section 16.2. One way this can be addressed is by doing away with the risk weights in the
calculation of the CAR. The resulting ratio is known as the leverage ratio (LR):
equity tier 1 capital
LR = = Ú 3%
assets total assets
The 3% is a rule introduced by Basel III with a phased-in implementation.
Before the crisis, the LR was the preferred way of specifying capital requirements in the
US, whilst Basel-regulated banks elsewhere used the risk-weighted ratio.
The LR does have some advantages over the Basel CAR, in particular, that by disregard-
ing the risk weights, there is less model risk in the calculation of the ratio, and less scope
for manipulation and capital structure optimisation aiming at maximising the visible ratio
whilst minimising the cost and protection afforded by the ratio.
There are also considerable disadvantages to the LR. Because it does not distinguish
between the most safe government bonds (like those of the UK, Germany and the US)
and junk bonds, it creates a disincentive to hold safe assets. This also puts LR in conflict
with the objectives of the liquidity coverage ratio (LCR) aiming at encouraging banks to
hold liquid government bonds. The LR is especially problematic for banks in countries
holding very large structural positions in government bonds, like Japanese banks.
If the choice was between either the LR or the Basel ratio, it would be hard to decide
between the two. Both ratios are flawed, but do meet a useful objective. Therefore, the
decision by the Committee to require both ratios simultaneously is sensible, providing
a useful compromise between the various issues, and reducing the scope for gaming.
18.3 Liquidity
Some national jurisdictions, like the UK, have long required banks to hold liquidity buff-
ers, but such requirements have not extended to international regulations. In particular,
Basel II is essentially silent on the question of liquidity.
Events from 2007 have demonstrated the fallacy of such complacency, because the
sudden disappearance of liquidity was at the heart of the crisis, and remains a major
problem in the ongoing sovereign debt crisis.
352
18.3 Liquidity
First, it was assumed that banks could always obtain funding liquidity, which meant
that they naturally migrated to the cheapest forms of liquidity, overnight repos. When the
interbank market dried up in 2007, banks that relied heavily on short-term funding found
themselves in significant difficulty. Because funding problems happen very quickly –
literally overnight – the banks, the authorities and everybody else were caught unpre-
pared, and governments had to rapidly provide bailouts. Longer and more stable funding
would have made it easier to deal with these problems.
Second, it was assumed that some types of funding were more stable than they turned
out to be. For example, bank deposits coupled with deposit insurance have historically
represented funding that was considered unlikely to evaporate quickly. Deposit insurance
schemes were thought to provide sufficient assurance to depositors, and hence prevent
bank runs. However, the limitations of deposit insurance schemes have made depositors
more skittish, especially those holding high-interest Internet savings accounts. Events in
the crisis demonstrated that depositors were sensible enough to withdraw large amounts
of money at the slightest whiff of trouble.
For these reasons, liquidity regulations have became a priority in the Basel III proposals,
and the Committee has proposed two liquidity ratios, the 30-day LCR and the one-year
net stable funding ratio (NSFR).
18.3.1 Liquidity coverage ratio

The LCR is based on the idea that banks should have sufficient high quality liquid assets to
survive an acute stress scenario lasting one month:
Stock of high quality liquid assets
LCR = Ú 100%
Net cash outflows over a 30-day time period
Numerator
The assets need to be unencumbered, meaning that they have not been pledged or allo-
cated to other purposes, and hence are ready to be monetised.
A key challenge lies in the definition of high quality liquid assets. Two categories of
assets are defined, level 1 and level 2. Level 1 assets are traditionally highly safe and liquid
assets, such as cash, demand deposits with central banks and marketable government
and international agency securities that meet certain conditions. Level 2 assets contain
marketable government and international agency securities that are more risky than their
level 1 counterparts as well as low-risk liquid corporate bonds.
The included assets should be monetisable, at a low cost, providing a relatively certain
amount to a bank facing funding problems.
Denominator
It is harder to define what to include in the denominator. The basic idea is that banks
have certain amounts of outflows and expect to receive some amounts of money. The
difference between the two, the net cash outflow, needs to be sufficiently low. There are
several challenges in the determination of net flows. A bank may have a reasonably clear
353
idea of outflows, but less of potential inflows, especially in times of stress. For example,
while it may have positions that are fully netted out, actually receiving funds depends on
the solvency of the counterparty.
Other problems include how to account for net flows from derivative positions and
other complex assets, where valuations may be model driven and assets only traded over-
the-counter (OTC).
At the time of writing, it is unclear how these issues will be resolved.
National discretion
National supervisors have significant discretion in how they implement the ratio, giving
rise to fears of inconsistent implementation, regulatory arbitrage and a race to the bottom.
Costs
Maintaining a large buffer of liquid assets is costly for the banks. These assets cannot be
used for other, more lucrative, purposes and the LCR, therefore, will directly affect the
banks’ profitability. They maintain that this will reduce their lending capacity, especially
in the context of the LR, with the sectors most affected presumably the SMEs.
This leads to one of the typical dilemmas in the regulatory reform process. More liquid
assets make the banks more resilient, but at the expense of less lending.
Sovereign bonds as liquid assets

The liquidity ratios are subject to the same problem as bank capital in that government
bonds are considered safe assets. Because some of the distressed European sovereign
bonds are considered riskless for the purpose of the ratios, a bank complying could actu-
ally end up being more risky than a bank holding safer liquid assets that do not count as
safe liquid assets in LCR calculations.
Dilemma
The dilemma for the policymakers is that true stability is strongly associated with high
costs. When banks are the most overconfident – right before a crisis happens – they are
least willing to pay the premium associated with stable sources of funding, and prefer to
take more risk by funding short term. Whatever definition of ‘stable funding’ is adopted,
in the next crisis banks will be meeting this requirement from the cheapest available
sources. This funding will be cheapest because it is the least stable.
While seeking stable sources of funding is a sensible idea, like liquidity it seems almost
certain to disappoint when tested in the next crisis, because banks’ incentive to reduce
their funding cost is too strong.
Gaming
Perhaps the main concern about the LCR is the same as worries about the integrity of
bank capital. Banks were able to use capital structure optimisation to significantly game
their capital levels. There is no reason to believe that there is any less scope for gaming
when it comes to liquidity. If anything, the problem might be worse because with capi-
tal we have three decades of observations on how it can be gamed and, hence, have an
354
18.3 Liquidity
understanding of how to minimise the problem. The worry is that the banks would be
more inventive and more nimble in the gaming process than the supervisors.
There are many ways to undermine the LCR. A bank might enter into contracts that
effectively encumber a large part of its liquid resources, but not in a way considered
encumbering by the supervisors.
Another way might be contracts that give rise to large cash requirements when a
bank comes under stress. So long as these do not give rise to any net cash flows (how-
ever defined) under normal circumstances, these might leave the LCR intact, but also
mean that at the onset of a crisis liquid resources would collapse much faster than
anticipated.
While these loopholes are obvious, there are many ways to achieve the same purpose
in a less visible way.
18.3.2 Net stable funding ratio

The NSFR is designed for longer time horizons, creating incentives for banks to better
maturity-match liabilities to investments.
Available amount of stable funding
NSFR = Ú 100%
Required amount of stable funding
Stable funding is defined as the fraction of equity and liability financing expected to
be reliable sources of funds over a one-year horizon, during extended stress. At the time
of writing, the definition is still quite vague, but more details are expected when the final
version of Basel III is published.
Issues
The NSFR demonstrates the common dilemma between safety and credit creation. One
of the main functions of banks is maturity transformations, simultaneously providing sav-
ers with demand deposit accounts and making long-term loans. Therefore, maturity mis-
matches are inherent in any bank, and not something that should be eliminated, or even
significantly reduced.
Therefore, a strict interpretation of the NSFR might seriously undermine the maturity
transformation function of banks, to the detriment of society.
Another concern relates to cliff effects. One result from the Basel capital ratios was the
creation of 364-day loans, one day short of one year, triggering capital requirements. A
one-year NSFR might do the same, with banks preferring 364-day loans.
This has two important disadvantages. First, it increases liquidity risk because borrow-
ers will have to worry about rolling over loans every 364 days. This is the type of risk that
is most likely to manifest itself during turmoil, at exactly the worst time. The second prob-
lem is that it disrupts business planning, because companies cannot count on long-term
funding for long-term projects. It might lead to funding migrating away from banks onto
the capital markets, where larger companies can issue longer-term bonds. This in turn
disadvantages the SMEs, because their access to the capital markets is more limited than
that of their larger competitors.
355
18.3.3 Danger
Perhaps the main worry about the liquidity ratios is that they may create an impression of
protection that is not veridical. This would endogenously encourage more risk-taking and
therefore undermine financial stability. It would be disastrous if the authorities assumed
the liquidity ratios were successful, and hence that liquidity crises could not happen.
In the next crisis the authorities should be prepared for reported liquidity to prove just as
illusory as reported capital. The ratios came under heavy criticism from the industry and
in response an amended, less strict, version was announced at the start of 2013.
18.4 How much capital?

Individual countries have drawn different conclusions from the crisis. Some have said
that banking was excessively risky and that it is important to increase capital to reduce
risk-taking. Others have taken a different view, arguing that banks are currently weak and
by increasing capital their banks might be pushed into more difficulties or even default.
Dilemma
There are two main problems for the authorities. First, there is the classical one whereby
a safe and resilient financial system may be over-conservative when it comes to lending.
Second, for a country with weak banks to make the rules stricter may perversely cause
banks to collapse, whilst maintaining weak regulations and low capital levels may give
the banks sufficient breathing room to improve their financial positions over time. The
various national positions on capital levels can be understood in the context of those two
concerns.
National positions
Perhaps surprisingly, the countries that most advocate more restrictive Basel III rules,
including more capital, are the US, the UK and Switzerland. The last two are countries
with particularly large financial sectors and all three share in a relatively laissez-faire tradi-
tion on financial regulation. They are joined by Sweden, Denmark and the Netherlands
within the EU, and by Canada and Australia among other BCBS members.
Even more surprisingly, the countries that most strongly advocate weak regulations
and low capital are Germany and France, especially the latter. If we consider the public
statements of the leaders of those countries, one would get the impression that they are
against risk-taking and speculation, advocating strict regulations, but in the committees in
which the rules are being drafted, those countries advocate the opposite.
The difference in national attitudes can partly be explained by the fact that countries
with relatively stronger banks, and countries whose banks have been more forthcoming
in recognising problem assets and recapitalising, find it easier to require more capital. If
banks are weak and have been allowed to hide losses, they will find it more difficult to
increase capital. The attitude of the Germans and the French depends in large part on the
role of banks in their economies, and for the French the amount of dirigiste influence the
state has over banks’ lending and general economic activity.
356
18.4 How much capital?
The Basel III proposals are the outcome of an agreement amongst all member countries
of the BCBS, and in the end, the capital levels were much lower than desired by certain
members. For some countries, this is not necessarily that problematic, because they can
simply unilaterally insist on more capital, a position taken by several countries, including
the US and Canada. The situation is different in Europe, because of the maximum harmo-
nisation principle, and as usual is more complex.
18.4.1 Maximum harmonisation principle

Amongst the European members of the BCBS, some want more capital than eventually
agreed to. Of those, the most vocal are the UK, Sweden, the Netherlands and others.
Initially, they aimed to follow the non-European members and unilaterally increase their
capital requirements. That ran into significant opposition, especially from the European
Commission, and to a lesser extent France and Germany, who claim it runs counter to the
so-called maximum harmonisation principle.
Besides a desire for more capital, the opposition to maximum harmonisation is gener-
ally rooted in a preference for national flexibility for macro-prudential policies, and many
member states, and even the ECB, have expressed a desire for such flexibility.
Single rulebook
This principle is well stated by Andrea Enria (2011), head of the EBA, who highlights the
great importance, when implementing European regulations, of consistent implementa-
tion and of the single rulebook. This obviously makes sense because inconsistent rules and
implementation can be a cloak for protectionism, encourage regulatory arbitrage and
increase frictional costs.
This has implications for the enthusiasm for more capital, because the single rulebook
implies that all member states need to have ‘exactly the same rules on the definition of
capital and maintain them up to date in a coordinated fashion in light of financial innova-
tion’. This means that member countries should not be allowed to require more capital
for their banks than agreed to by the EU, and cannot seek to establish an independent
brand based on higher regulatory standards.
Main dilemma
Maximum harmonisation is quite important for the EU, not least because of the impor-
tance of the common rulebook. It is also a defence against further tiering of the market,
which happens when the stronger countries rush ahead, undermining the market image
of the weaker ones.
The eventual outcome is problematic, however, especially if the eventual capital
requirements in Europe will be lower than outside the Union. This will send a sig-
nal that North American and Asian BCBS member countries will have more resilient
banking systems. It is possible that this will cause higher-quality clients to migrate to
these banks, leaving poorer-quality clients and riskier activities to the EU. Lower capital
requirements will at least suggest that Europe has a more fragile banking system than
the rest of the world.
357
Perhaps the real problem is that it signals both a deep current malaise and an unwill-
ingness to tackle it, guaranteeing the crisis will fester.
Eventually, enough wiggle room was found to allow member countries to adopt any
minimum capital ratios they desire.
18.4.2 Economic costs

Basel III significantly increases the direct cost of doing banking but hopefully will make the
banking system more resilient, and perhaps reduce the total cost to society from banking.
The question of which of these factors dominates remains highly controversial, and whilst
many studies have been made, the conclusions are not consistent.
The main dilemma for policymakers is the elusive cost and benefit problem that is so
hard to analyse, especially the growth versus stability trade-offs.
Studies
Admati et al. (2010) argue that increased bank equity brings with it significant economic
benefits, whilst lowering leverage and the risk of bank failures. Other studies find that
Basel III will be costly, for example studies by the New York Federal Reserve Bank (NYFed)
(Angelini et al., 2011) and the BIS (2010). The latter finds that a 1% increase in equity will
lead to a 0.19% reduction in GDP over four years. An FSA study (de Ramon et al., 2012)
finds that quantifying the pros and cons is difficult. The Institute of International Finance
(2010) finds a much stronger impact, where a 2% increase in capital, along with other
elements of Basel III, will cost 3.1% of GDP over five years.
GDP impact
It is very hard to identify the costs and benefits of Basel III. By a first approximation,
increased capital increases the cost of doing business, reducing lending and increasing
the cost of lending. However, if the banking systems become more resilient, requiring
fewer and smaller bailouts and, hence, are less prone to boom and bust cycles, these
costs are partially, and even fully, offset.
In addition, standard Modigliani–Miller analysis would predict that reduced funding
costs because of higher capital would provide greater security for well-capitalised banks.
This implies that the GDP impact of higher capital would be very small. After all, every
time a bank issues new capital, the economic impact is non-detectable.
It is hard to see how a relatively modest increase in bank capital, tiny compared to
GDP, could have anything but a very moderate economic impact. This suggests that the
studies with the low impact numbers are more credible.
GDP and stability trade-off

It is important how the main issues are compared. Are we comparing the question of
short-term growth against more short-term resiliency, or multi-decade growth compared
to the potential for systemic crisis?
The empirical studies tend to focus on the former, not least because it is easier to ana-
lyse. The main question for society is the second. Systemic crises are so costly that we
358
18.5 Recovery and resolution
would do almost anything to prevent them. More capital would be a cheap way to pre-
vent systemic crises, so long as it was effective.
And herein lies the relevant question. What is the best way to prevent systemic risk,
taking into account the effectiveness and cost? If the answer is capital, it is obvious that
capital requirements should be increased. If the answer is not necessarily capital, then the
capital increases are harder to justify.
To this day, this question has not been answered in a satisfactory way.
18.5 Recovery and resolution

In response to the crisis, the authorities have put significant emphasis on crisis resolu-
tion, as discussed in Bassani and Trapanese (2011). For an industry view, see Institute of
International Finance (2012). Resolving the failure of an individual financial institution is
difficult in the best of times.
Resolving bank failures within existing bankruptcy laws

Bankruptcy law moves very slowly and intends to protect the interests of the creditors. In
any bankruptcy, the clients of the failed company will suffer, but when a bank fails the
effects may be particularly severe.
This provides the motivation for special bank resolution regimes, for example the
FSB Key Attributes, the Federal Deposit Insurance Corporation (FDIC) resolution
regime and the Dodd–Frank orderly liquidation authority. The objective is to take bank
insolvencies out of the usual bankruptcy procedures and make quick administrative
action possible.
18.5.1 Resolution regimes

Every financial centre has some resolution regime in place. These are most routinely
applied to domestic failures, and work reasonably well, provided the institution is small
and operates only in the home country.
Definition 18.1 Resolution regimes These refer to the mechanisms implemented

when a financial institution fails, like bankruptcy proceedings, dealing with creditors and
debtors, provision of continuous banking services and the like.
When financial institutions become large, the problems of resolution become much
harder, and even more so when they involve multiple jurisdictions.
Before the crisis, the problem of cross-border resolution had been to a large extent
ignored because of the many inherent complexities. It is hard to see how an international
agreement could effectively override markedly different national regimes. The failure
of Lehman Brothers demonstrated the complacency inherent in the extant resolution
regimes.
359
Example 18.1 Lehman Brothers
When Lehman brothers failed, there was a scramble to file for bankruptcy. The bank-
ruptcy filing in New York occurred just after midnight on Monday 15 September,
beating the London filing by a few minutes. Even if London had been first, it would
not have made much difference, because liquidity in Lehman Brothers was han-
dled centrally in New York. It swept to New York each night, only to be released
to the regional operations the next day. The Friday sweep took $8 billion out of
London.
At the point of failure, a single integrated operation with a common treasury func-
tion, the situation on Friday, was replaced by around 100 discrete entities, all insol-
vent and with their bankruptcies being handled by their respective national regimes.
Each administrator has a duty to maximise recovery for the creditors to its entity,
which places them in a competitive situation, and four years after the bankruptcy
there are still fresh, multi-billion-dollar claims being made by one administrator
against another. As most of these will need to be addressed in court, it is clear that for
most creditors substantial repayment remains a distant prospect, and also that the
frictional costs of the bankruptcy will be very high.
While the scramble for filing was not all that important, the bigger problem was
that proceedings in the US, UK and other jurisdictions all went on independently,
without adequate legal grounds to establish one, consolidated proceeding.
Source: based on information from Financial Times (2008) and Sorkin (2010)
Lehman Brothers was international in life but national in death, as noted by Thomas
Huertas.
The administrators of a single bankrupt entity are not allowed to place cooperation
with other administrators above the interests of the creditors whose money they are
spending, so it is sometimes difficult for them to justify cooperative information shar-
ing unless it is clear that their creditors will benefit. An external entity could assist in
promoting efficient information sharing and cooperation and reduce duplicated effort.
In 2009, the FSB announced principles for cross-border and crisis management. Based
on the Lehman experience, this should be of some help. Building on supervisory colleges,
a new entity called a crisis management group (CMG) should be in place, ready to handle
crisis management. The CMGs are meant to improve information sharing, build-up of
trust and cooperation between various authorities.
However, beyond that it is not clear how much a CMG can achieve, in practice, as
the fundamental problem is one of conflicting fiduciary duty. By law, administrators are
not allowed to give preference to the interests of other group companies over those of
their creditors, so while better information sharing would perhaps save time it would not
prevent the attempts by administrators to maximise their outgoing claims against other
group companies, and to minimise the incoming claims they accept. Intragroup arrange-
ments will be complex, so this will remain a recipe for lengthy wrangling even if a CMG
can provide improved access to company data.
360
18.6 What about too big to fail?
18.5.2 Living wills

One way to reduce the danger created by the failure of banks is by the concept of living wills.
The objective of living wills is to require banks to demonstrate how they could be wound
up during financial turmoil without requiring bailouts. In effect, living wills serve to provide
a manual to deal with difficulties and, if necessary, unwind failed banks painlessly without
recourse to public funds.
Living wills were originally proposed by the FSA in October 2009 and have two
phases, recovery and resolution. The recovery plans are managed by the firm and the
resolution plans by the authorities. The recovery plans are put into effect before insol-
vency or resolution, involving perhaps the sales of businesses, the winding down of
exposures, and the like. The resolution plans address how to handle a bank in resolu-
tion. A lot of that is just collecting information so that all concerned can understand the
legal entity structure, IT issues, business exposures, where assets are in the group, etc.
The idea of living wills is sensible, but it remains to be seen how it can be executed
effectively. Furthermore, they may create interesting legal problems because bankruptcy
law tends to take precedence over pre-existing contracts and can even unwind recent
management decisions if these are deemed improper. This will be especially relevant
across jurisdictions, and one could imagine a court in one country refusing to recognise
some of the steps and demanding that their own rules take precedence when their sub-
sidiaries are involved.
It also seems likely, given management incentives, that supervisors will be slow to rec-
ognise that resolution is required. It would be unwise to assume that, in practice, resolu-
tion will commence swiftly enough to prevent the requirement of public funds.

An important failure of financial regulations before 2007, and one that has been getting
worse since then, is the problem of TBTF financial institutions, especially the G-SIBs.
A bank that is TBTF is a bank whose failure would be so catastrophic that it would
probably trigger a systemic crisis. The economic and political consequences of the failure
of a TBTF institution are so big that a government has no choice but to try to bail it out,
even if it may be beyond the government’s means to do so.
Bank incentives
Banks have direct incentives to become TBTF. Any bank that is perceived as TBTF is
more likely to be bailed out than smaller banks. Creditors notice this, and provide
funds to the TBTFs at lower costs than to their smaller competitors, because those
loans are implicitly underwritten by the government. This funding cost advantage gives
the TBTFs a competitive advantage over their competitors. After all, if your failure trig-
gers a systemic crisis, you will be bailed out. Your funding costs will reflect that bailout
and so your competitive situation is improved. Finally, in order to become TBTF, it
helps to be seen as big, interconnected and dangerous. If a bank is badly run, it is more
361
dangerous and more likely to become seen as TBTF. This creates perverse incentives for
bank management.
18.6.1 The view until 2007

Having financial institutions of this type does not seem very sensible. Regardless, prior to
the crisis, banks were allowed and often encouraged to grow and merge, eventually creat-
ing many TBTF institutions. This happened without much public scrutiny, and with the
acquiescence, or even direct encouragement, of the authorities.
Supervisors
In the past, banking supervisors have not been very concerned about the TBTF institu-
tions. They were not subject to more stringent prudential requirements, nor was their
range of activities restricted. On the contrary, if anything, the tendency of policymakers
was to treat TBTF banks more lightly than other banks, reflecting the perceived benefits of
diversification and management sophistication. In other words, because of their size, they
were considered safer than smaller institutions.
Government
Politicians are also to blame for the TBTF problem. Countries often like to have national
champions, big and powerful banks, visible on the global stage, demonstrating the prow-
ess of the country. Therefore, governments actively encourage banks to become large, to
become TBTF.
A view frequently expressed by governments before 2007, and even to this day, is that
what is good for the national banking champion is good for finance and hence the country.
National champions are, almost by definition, TBTF, and frequently identified as a major
contributory factor to the crisis from 2007. They are one of the main reasons why we have
so far been unable to resolve the banking crisis in Europe.
An interesting observation from Table 18.3 is how many European countries are repre-
sented. The US has its fair share of G-SIBs, eight, but it is within the means of such a large
economy to bail out one or even more of its G-SIBs. Banking assets in the US are much
smaller than in the EU. Furthermore, of the top banks in Table 18.3, JP Morgan’s assets
equal 15% of GDP and Citigroup’s 13%. Overall, the assets of US banks aggregate to less
than the country’s GDP.
The same cannot be said for the European countries; for example, the UK and France
each have four G-SIBs. Two European countries facing a sovereign debt crisis have some
banks in the G-SIB group, Spain with two and Italy with one. The European banks are
also much larger relative to their national GDP than in the US: Deutsche Bank has 84%
of Germany’s GDP, Santander 92% of Spain’s GDP, and RBS, Barclays and HSBC together
have 337% of the British GDP.
Summary
The crisis demonstrated the problem of TBTF institutions by identifying the importance
of systemic risk compared to idiosyncratic risk. In principle, a large, globally diversified
firm may be less likely to fail than smaller and more concentrated firms. Problems in one
362
business line are offset by profits elsewhere, and diversification may in principle protect
a large financial institution against idiosyncratic risk, so long as its idiosyncratic risk was
different from that of other financial institutions.
Instead, the crisis showed that financial institutions were exposed to the same
risk to a much greater extent than envisioned and, therefore, exposed to similar self-
reinforcing collapses in confidence as their counterparts. If the markets have doubts
about the institution’s solvency, diversification will not protect it. Lehman was but one
example.
18.6.2 What is being done?

Emergency measures in crisis
The problem of TBTF has become worse since 2007 because several institutions that
were already very large were allowed to merge, because that was thought to be the
best emergency response at the time. Some examples from the US include Bank of
America with Merrill Lynch, JP Morgan Chase with Bear Stearns, and Wells Fargo
with Wachovia, and in the UK, HBOS with Lloyds TSB. In all of these cases, the TBTF
problem was made worse and competition was reduced. For example, post-merger,
Lloyds has 21% in SME banking and provides one in five mortgages to new borrowers
in the UK.
New initiatives
The authorities have recognised that crisis mergers were unfortunate, and have proposed
a range of new measures intended to avoid them in the future.
A number of initiatives are aimed at the TBTF problem, most within the FSB. This
includes special resolution regimes, additional regulations and increased intensity of
supervision, and the G-SIB capital surcharges.
18.6.3 The problem is not really getting better

Whilst many of the initiatives aimed at the TBTFs make significant contributions towards
mitigating some of the worst aspects of the problem, even to the extent of preventing
more TBTFs from being created in the future, the signs do not point to the problem get-
ting better.
First, the main reason why the problem exists in the first place is that governments like
to have large, internationally known banks that demonstrate the prowess of the country.
That means they implicitly agree to have TBTF banks.
Second, large national banks are ideally placed as lobbyists, being few, wealthy,
and with similar interests. Because state support is so valuable to banks, they devote
a great deal of effort to ‘maintaining close links’ with governments, via various forms
of influence and lobbying. We have seen senior managers of large banks migrate
to equally senior jobs in government, and some governments employing bankers
as senior advisors. Similarly, senior government officials often migrate to high-level
bank jobs.
363
Finally, we can expect competition for G-SIB status to be intense because of the lower
funding costs it will bring. The G-SIB designation officially identifies the banks as such.
It sends a signal that the banks are under a special watch, and most likely would not be
allowed to fail. In turn, it makes them a preferred counterparty, lowering their cost of
funding and giving them a competitive advantage. In other words, the G-SIB surcharge
institutionalises the category of institutions that were perhaps the main contributor to the
crisis, and are most likely to be bailed out in the future.
18.7 Summary
Following the crisis that started in 2007, a significant revamp of financial regulations has
been underway. One large set of changes relates to the Basel process and the work of the
FSB. The next iteration of the capital accords, Basel III, is expected to be implemented
from 2013, tightening the calculation of bank capital, introducing drawdown buffers,
countercyclical buffers and special charges for G-SIBs. The accord also specifies new regu-
lations on liquidity.
Several government regulatory institutions have been created as a consequence of the
crisis; some have acquired more responsibility and others less. The relationship between
the central banks, the supervisors and the treasuries has been altered, with the central
banks taking on more responsibility but also losing some of their independence to the
treasury.
Many other ideas have been discussed but have not made it into the reform process.
Some problems, most importantly what to do about TBTFs, are not being dealt with in a
satisfactory manner.
Chapter 9 discussed several regulatory initiatives, CCPs, FTT, restrictions on bonuses
and narrow banking, and we have not repeated that discussion here.
Until a new crisis emerges, the initiatives discussed here, and in Chapter 9, are likely
to be the final word on the financial regulatory process, as there seems to be insufficient
political will to tackle more fundamental issues.

1 How has the role of the central banks changed in the crisis?
2 How has the role of the treasury changed in the crisis?
3 How has the role of the supervisor changed in the crisis?
4 Compare contrast the systematic risk institutions in the US, UK and EU.
5 What is the problem the European banking union is meant to solve and what is the
main reason why it may not happen?
6 What are the main changes to capital in Basel III?
7 Which do you prefer, the risk-weighted Basel CAR or the leverage ratio?
364
References
8 Do you expect the LCR to be effective?
9 Does the NSFR positively or negatively affect the problem of maturity mismatches?
10 What is the maximisation harmonisation principle, and do you think it should be used
to set European capital levels?
11 What are the main legal problems in resolution?
12 Do you think that the problem of too big to fail is being adequately addressed?
References
Admati, A. R., DeMarzo, P. M., Hellwig, M. F. and Pfleiderer, P. (2010). Fallacies, irrelevant facts,
and myths in the discussion of capital regulation: why bank equity is not expensive. Working
Paper, Graduate School of Business, Stanford University.
Angelini, P. L., Clerc, L., Curdia, V., Gambacorta, L., Gerali, A., Locarno, A., Motto, R., Roeger, W.,
den Heuvel, S. V. and Vlcek, J. (2011). Basel III: long-term impact on economic performance
and fluctuations. Technical report, Federal Reserve Bank of New York.
Basel Committee on Banking Supervision (2011a). Basel Committee issues final elements of the
reforms to raise the quality of regulatory capital. www.bis.org/press/p110113.pdf.
Basel Committee on Banking Supervision (2011b). Basel III: a global regulatory framework for
more resilient banks and banking systems. Technical report.
Basel Committee on Banking Supervision (2011c). Global systemically important banks:
assessment methodology and the additional loss absorbency requirement. www.bis.org/
publ/bcbs207.pdf.
Bassani, G. and Trapanese, M. (2011). Crisis management and resolution. In Quagliariello, M.
and Cannata, F., editors, Basel III and Beyond. Riskbooks, London.
Baudino, P. (2011). The policy response: from the G20 requests to the FSB roadmap; working
towards the proposals of the Basel Committee. In Quagliariello, M. and Cannata, F., editors,
Basel III and Beyond. Riskbooks, London.
BIS (2010). Assessing the macroeconomic impact of the transition to stronger capital and
liquidity requirements. www.bis.org/publ/othp10.pdf.
Borio, C., Drehmann, M., Gambacorta, L., Jimenez, G. and Trucharte, C. (2010). Countercyclical
capital buffers: exploring options. Technical report, BIS. Working Paper 317.
de Ramon, S., Iscenko, Z., Osborne, M., Straughan, M. and Andrews, P. (2012). Measuring
the impact of prudential policy on the macroeconomy. A practical application to Basel III
and other responses to the financial crisis. Technical report, FSA. Occasional Paper
Series 42.
Enria, A. (2011). The future of EU regulation. www.eba.europa.eu/cebs/media/aboutus/
Speeches/The-Future-of-EU-Regulation–British-Bankers-Association–29-June-2011.pdf.
Financial Stability Forum (2008). Report of the financial stability forum on enhancing market
and institutional resilience. Technical report.
Financial Times (2008). Winding up Lehman Brothers. Financial Times, 7 November.
Gorter, J. and Bijlsma, M. (2012). Strategic moves. www.risk.net/life-and-pension-risk.
365
Institute of International Finance (2010). Interim report on the cumulative impact of proposed
changes in the banking regulatory framework. Technical report.
Institute of International Finance (2012). Making resolution robust. Technical report.
Quagliariello, M. and Cannata, F. (2011). Basel III and Beyond. Riskbooks, London.
Repullo, R. and Saurina, J. (2011). The countercyclical capital buffer of Basel III: a critical assess-
ment. CEMFI Working Paper 1102, June 2011.
Sorkin, A. R. (2010). Too Big to Fail: Inside the Battle to Save Wall Street. Penguin.
366
19 Sovereign debt crises
Sovereign debt crises happen when a national government – the sovereign – is unable
to service its debt. This form of crisis was common in the developed world until the
second World War (WWII), but after that has mostly been consigned to emerging
markets countries.
The reason is that developed countries are able to borrow in their own currency,
and so can usually find ways to avoid defaulting, perhaps by inflating the debt away.
Recently, however, sovereign debt has become a serious policy issue for some mem-
bers of the euro zone.
This chapter focuses on giving a general background to sovereign debt crises in gen-
eral, as well as the causes of the European sovereign debt crisis. The crisis in Europe is
still ongoing, and because of the rapid developments, we maintain a separate chapter
on that crisis online at www.GlobalFinancialSystems.org.

This chapter mostly stands on its own, but the most direct connections are to
Chapter 17 (the ongoing crisis: 2007–2009 phase).
Key concepts
■ Newfoundland
■ External and domestic debt
■ Sovereign defaults and corporate defaults
■ Enforcement of sovereign debt
■ Currency unions and sovereign debt crises
■ European sovereign debt crisis
Chapter 19 Sovereign debt crises

For a good overview of debt crises, see Sturzenegger and Zettelmeyer (2007). Reinhart and
Rogoff (2009) present a comprehensive collection of facts and figures. For legal analysis,
see, for instance, Gulati and Triantis (2007), and for a legal view on how to do restructuring,
see Buchheit (2011). For statistical analysis of the cost of sovereign default, see Borensztein
and Panizza (2008). For an overview of the euro, Marsh (2010) is comprehensive.
19.1 Newfoundland
In the long history of sovereign debt crises, creditors have resorted to a variety of extreme
means to enforce their claims. However, they have never used liquidation, in the same
way as a corporation is liquidated, except in the case of Newfoundland. We follow this
story as told by Hale (2003).
Newfoundland is the extreme easternmost part of Canada. The British established a
settlement there in 1497 and the economy was primarily based on fishing. It became the
first self-governing part of the British Empire in 1855 and by the end of the nineteenth
century it enjoyed most of the trappings of sovereignty.
The Newfoundland government borrowed heavily to finance its military expenditures
in the first World War (WWI), and continued increasing its debt throughout the 1920s.
By 1933, its public debt was about three times GDP. By that time the economy had col-
lapsed, not least because the Catholic countries of Latin America stopped buying fish
because of the Great Depression.
The government of Newfoundland asked the British government for help, which
obliged by sending a Royal Commission to investigate the situation. While Newfoundland
preferred to default, this was not considered palatable to the Commission, which
declared: ‘No part of the British Empire has ever yet defaulted on its loan obligations [. . .]
bankruptcy is at best an ugly word and carries a stigma which a nation even more than an
individual would do well to avoid.’
The Deputy Leader of the (opposition) Labour Party, Clement Attlee, suggested that default
was preferable to giving up democracy. Referring to Britain’s own default on its wartime loans
from the United States (US), he said: ‘All the best countries default nowadays’ (Hansard
1933). But in the early 1930s it was impossible to imagine a British dominion defaulting.
The Commission’s proposal was that Newfoundland temporarily give up its independ-
ence to a UK-appointed administration – a dictatorship in all but name – and this was
agreed to by Newfoundland’s population in 1933.
Following WWII, the UK government felt compelled to resolve the issue. The alterna-
tives were for Newfoundland to become a part of the UK, to become a part of the United
States (US), to become independent or to merge with Canada. The first choice was unpal-
atable to the UK, and the second disliked by Canada, so only the last two were considered.
The UK-run administration held a referendum, and a slim majority rejected confedera-
tion with Canada. The administration repeated the referendum several weeks later, this
time getting a majority vote for confederation. The British North America Act required
that the local parliament solicit the federation with Canada, but as Newfoundland had
368
19.2 Sovereign debt
no parliament, the British government chose to overlook this requirement. It appointed

a negotiations team in short order and by 1949 Newfoundland had became a part of
Canada. The confederation treaty was an act of Canada and Britain, not of Newfoundland.
In dealing with the crisis, the British government opted to liquidate Newfoundland, the
only case in modern history where a sovereign nation has suffered that fate.
19.2 Sovereign debt

Sovereign debt is owed by a sovereign authority, often including government entities such
as the central bank. It does not include the debt of regional and local governments, nor
the debt of government-owned agencies or enterprises, unless explicitly guaranteed by
the central government. A sovereign default occurs when the government fails to meet
payments on its sovereign debt.
Historically, governments were the biggest borrower from banks, and the success and fail-
ure of financial institutions was directly dependent on the ability of the sovereign to meet
its obligations, often in turn dependent on its success in war. As noted by Homer and Sylla
(1996), during the Renaissance, important financial institutions, including the Medici’s,
failed because of the inability or unwillingness of the kings to repay their loans. In turn, those
sovereigns who were more likely to wage wars were considered more risky, paying a spread
over those more peaceful. After all, if the king loses, the creditor will not be repaid.
Sovereign default has been a part of virtually every country’s history. France defaulted
eight times from 1500, but eventually managed to emerge from its status as a serial
defaulter, and has not defaulted since 1788. Spain defaulted seven times in the nine-
teenth century after having defaulted six times in the preceding three centuries.
The UK defaulted frequently a few centuries back. It defaulted indirectly in the late
1800s by unilaterally adjusting the coupons on its debt, and did not fully repay its WWI war
debt to government and private investors in the US, owing $4.4 billion in 1934. Updating
to current prices,1 this is $74 billion if adjusted by the consumer price index (CPI) and
$1,010 billion if measured as a share of GDP.
As the developed nations stopped defaulting on their debt in the twentieth century,
countries that had recently gained independence started to have trouble with their debt
burden. For example, India has defaulted four times and practically every newly inde-
pendent African country has defaulted at least once. The most defaults in any continent
is in Latin America where every country has defaulted at least twice. China also defaulted
twice in the twentieth century.
19.2.1 External and domestic debt

Governments can issue debt in either domestic or foreign currency. Generally, domestic
debt is issued in local currency, under local law and held by local citizens, whilst external
debt is issued in foreign currency, under foreign law and held by foreigners. Many excep-
tions exist, and this distinction does not apply to debt issued by euro zone governments.
1
See www.measuringworth.com.
369
External defaults get much more attention than domestic defaults. One reason is that
they are more common, but it is also because external defaults are highly visible in inter-
national capital markets and, hence, the international press; whereas domestic defaults
are less visible, and often more indirect.
External debt
Most countries prefer to issue domestic currency debt, but this is often not a viable option,
perhaps because the domestic debt market is underdeveloped and not very deep. In this
case, a country might resort to issuing foreign currency debt to overseas investors. We
refer to this as external debt. Of course, anything will sell if priced appropriately, so the
implication is that the yield at which domestic debt can be sold appears unattractively
high compared to that of foreign debt.
Countries default regularly on external debt, often because of adverse external factors,
such as falling commodity prices, increases in interest rates or the sudden stop or reversal
of capital flows. External borrowing, not surprisingly, is strongly pro-cyclical. High com-
modity prices are often correlated with increased borrowing, setting the stage for a debt
crisis when commodity prices eventually drop.
Many emerging markets countries prefer short-term borrowing because it is cheaper
than longer maturities. Investors also prefer shorter maturities, both because that mini-
mises their risk and also because it gives them more ability to discipline borrowers.
However, short-term borrowing introduces roll-over risk and makes a country more vul-
nerable to a liquidity crisis, as we have seen in the East Asian case.
Domestic debt
Domestic currency debt, mostly held by local citizens, has traditionally been the major
part of the overall stock of public debt, perhaps averaging at around two-thirds of total
debt. This proportion is much higher for developed economies. Unlike external debt, out-
right defaults of domestic debt are much less common than defaults on external debt.
Governments, however, have numerous indirect options for reducing the domestic debt
burden, for example inflation, the forced conversion of debt to lower coupon rates, and
a unilateral reduction in principal. Both the UK and the US have resorted to such meas-
ures. Another possibility is financial repression in the form of an interest rate ceiling with
a simultaneous rise in inflation. India, for example, introduced interest rate ceilings in the
wake of its external debt restructuring in 1972–1976.
The many alternatives for reducing the debt burden mean that it is often difficult to deter-
mine exactly what constitutes a default and what is just an event that happens to reduce the
debt burden. The extensive ‘wiggle-room’ this provides borrowers is one reason why foreign
lenders tend to prefer foreign-currency debt, issued in developed countries’ jurisdictions.
Debt reduction through inflation

Perhaps the most common way for a government to reduce its debt burden is by reduc-
ing the value of its currency by inflation. Before the advent of paper money, govern-
ments would regularly debase their coinage, which was usually based on silver or gold,
by mixing in cheaper metals, shaving down coins or reissuing smaller coins in the same
370
19.2 Sovereign debt
denomination. For example, between 1542 and 1547 the pound lost 83% of its silver con-
tent. The tools of inflating away debt have changed over the years and the introduction of
fiat money made it much easier for governments to create inflation.
High inflation by itself does not have to imply default, so long as interest rates com-
pensate or the debt is inflation indexed, but a rapid increase in inflation when debt is not
indexed strongly signifies that a government is deliberately deflating debt away. Some
commentators have claimed that the inflation in the US in 1947–1948 was deliberately
created to reduce WWII debt, and once that debt had reached manageable levels, the
government slowed down the inflation. One could argue that the various quantitative
easings (QEs) are a form of inflating debt away.
What about European sovereign debt?

While the distinction between internal and external debt has never been absolute, the
boundaries were further blurred with the euro zone. Is Greek debt, issued in euros, exter-
nal or internal debt? Since a significant portion of the debt is held by foreigners, it is exter-
nal, but it is issued in the Greek domestic currency and, therefore, domestic. Furthermore,
86% of Greek debt is subject to Greek law, with the rest mostly issued in euros in London
and, hence, subject to UK law.
This means that in the case of the euro zone, the distinction between domestic and
external debt depends very much on context.
19.2.2 Difference from corporate defaults

Sovereign defaults are inherently different from corporate defaults. A company goes broke
when it breaches a covenant with its creditors, perhaps by failing to pay interest, so that
creditors petition a court for a formal bankruptcy process. In that process, a company may
be restructured or in a worst case liquidated. The creditors can draw on a well-defined
legal framework providing them with clearly drawn-up rights, and rely on bankruptcy
courts to enforce their claims.
Countries do not generally go out of business and get liquidated, the single exception
being Newfoundland. The creditors cannot use a formal legal process to enforce their
claims against a country in the same way as they can use the legal system in a corporate
bankruptcy, because there is no proper supranational legal framework for enforcing sover-
eign debt contracts across borders.
However, the trigger is the same: default means failure to honour a contract, perhaps
through failure to repay principal or to make an interest payment or through some unilat-
eral adjustment in contract terms.
Unlike a corporation, a country defaults because the government has made a strategic
decision to default, presumably after cost–benefit analysis, weighing the cost of debt ser-
vice against the losses incurred because of default. It could perhaps better be compared
to the US ‘Chapter 7’ reorganisation because the existing management may well remain
in place. Consequently, countries generally default long before they run out of the abil-
ity to service debt. This means that in sovereign defaults the lenders depend on both the
ability and the willingness of the government to honour its debt.
371
19.2.3 Why repay sovereign debt?

Defaulting on foreign currency debt held by foreigners imposes costs on the foreigners
and saves the nation-state money. So why pay back at all? The reason is that it is costly to
default. Not only does it adversely affect the reputation of a country, but creditors often
have a variety of means to enforce their claims, relying on carrots and sticks.
Reputation risk
A country that defaults suffers a loss to its reputation, which may hamper its access to
international capital markets or increase future costs of borrowing. Since the ability to tap
the international capital markets at a low cost is quite valuable, it provides direct incen-
tives not to default.
One example of this was provided by South Africa repaying apartheid-era debt. The
Economist (1999) had strong opinions about the consequences: ‘[South Africa’s] credit
rating would be wrecked as it came to be lumped in with other deadbeats. Foreign inves-
tors would be deterred and South Africa would have to pay more for future borrowing.’
Analysis of default costs

A country that defaults on its obligations is also likely to suffer disruption in international
trade. Access to trade financing might seize, and creditors may attempt to enforce their
claims in friendly jurisdictions. Similarly, loss of reputation may discourage foreign direct
investment, since a country that defaults may appear more likely to seize the assets of
foreigners or cause difficulties for their operations.
Borensztein and Panizza (2008) analyse the costs of sovereign defaults, considering rep-
utational costs, international trade exclusion costs, costs to the domestic economy through
the financial system, and political costs to the authorities. They find that reputational costs
are significant but short lived. There is little impact on trade, growth in the domestic econ-
omy suffers, default episodes seem to cause banking crises and not vice versa, and finally
governments suffer a significant political cost from the consequences of debt crises.
Contagion
In Europe, banks are exposed to high-risk European sovereigns, and this may trigger their
default and transmit a sovereign debt crisis to their home country. This works as a channel
for contagion.
Default immunity and economic development

The impact of sovereign defaults varies greatly between countries: some seem quite immune
whilst others are strongly affected. Less developed countries generally are not as sensitive
to sovereign defaults than their more developed counterparts. After all, if a country is rela-
tively self-sufficient and most consumers buy from domestic companies, where the export
sector is small, a sovereign default on external debt might not have a huge impact.
For example, one reason why the Latin American countries have defaulted so frequently is
that such an event affects the population only indirectly, perhaps because they have adapted
their economies to become default immune, a classic case of a self-fulfilling prophecy.
372
19.2 Sovereign debt
However, the more exposed a country is to the global economy, the better its previous
reputation, and the less its citizens are prepared for the event, the bigger the impact of a
sovereign default. This is why it is so difficult for European countries to contemplate a default.
The costs can be expected to be much larger than, say, for Argentina which defaulted in 2002.
19.2.4 Debt (in)tolerance
Definition 19.1 Debt intolerance Even though a lot of commentary focuses

on debt/GDP, that is not as relevant as some other variables indicating the ability to
service debt, for example debt/foreign income.
Some countries manage easily with very high debt levels, perhaps exceeding 240% of
GDP as in Japan, whilst others default with debt levels of 40%. The inability to maintain
high debt levels was called debt intolerance by Reinhart et al. (2003). This happens when a
weak institutional structure, and a problematic political system, make external borrowing
a tempting way to avoid tough decisions regarding fiscal spending and taxing.
What matters is the point at which lenders cease to lend, which depends on their
belief on whether the loan will be repaid. This will be influenced mainly by how much
damage they perceive the borrower is exposed to if they default. If this is high for what-
ever reason, for example because the lender has a big army, the borrower does a lot of
international trade, or the borrower has a big reputation to protect for some other reason,
then the debt/GDP ratio can get high.
Debt intolerance manifests itself in the extreme duress many emerging markets experi-
ence at external debt levels that would seem quite manageable by the standards of more
developed countries. Default can occur in emerging markets at levels of debt well below
the 60% ratio of debt to GDP that is enshrined in the Maastricht Treaty, though the com-
parison is not entirely fair. Maastricht applies to highly developed countries and some
less developed countries that receive credit enhancement via the euro and associated
political machinery; as a result its signatories can be expected to be highly debt tolerant.
Mexico defaulted in 1982 with a ratio of debt to GNP of 47% and Argentina in 2001
at around 50%. Reinhart and Rogoff (2009) find that external debt exceeded 100% of
GNP in only 16% of defaults of emerging markets during 1970–2008. More than half the
defaults occurred at levels below 60%, and 20% occurred below 40%. The threshold level
is determined by the history of the country. A serial defaulter with persistent inflation and
weak institutions and in which citizens are inured to crises will have a lower threshold
than a country that appears to have lower credit risk.
Example 19.1 Japan
The country that successfully manages the world’s highest debt level is Japan, where
the government debt reached 230% in 2011. The reason such a high debt level is sus-
tainable is that all of the debt is in domestic currency, interest rates are almost zero and
➨
373
most of the debt is held by domestic agents. This high debt is not without problems.
The Japanese economy has been stagnating for two decades, often suffering from bor-
derline deflation. If the government attempts to stimulate the economy, it may cause
interest rates to increase, sharply increasing the debt burden. Any attempt to inflate
the debt away would be opposed by the electorate, an unusually high proportion of
whom are beyond retirement age and consequently living primarily on savings.
Vulnerabilities
As a country’s debt gets closer to the intolerance level, it becomes increasingly vulner-
able to shocks. It may not have the funds for a fiscal stimulus, and any negative shock will
reduce tax revenues and, hence, increase the relative debt burden. Even though policy-
makers may find that their debt levels appear manageable under the current – favourable –
economic situation, high debt may mean the country is unable to respond to a crisis. We
see some of this in the European sovereign debt crisis.
This means that it can take very little to trigger a sovereign debt crisis, and it can hap-
pen without any significant changes in wealth or income. All that is needed is a change in
confidence. This means that many of the same forces are at work in a sovereign debt crisis
as in a currency crisis, and we can easily analyse them with one of the second-generation
(2G) currency crisis models discussed earlier.
If the market begins to suspect that a government may not be able to pay back its debt,
it will demand a higher compensation for taking on the higher risk of default. A spike in
bond yields, driven by fear that a government will fail to honour its debt, makes debt refi-
nancing for a country difficult to impossible. This is a sovereign debt crisis. We illustrate
the mechanisms in Figure 19.1.
Funding ample
Below debt threshold

Economy grows
It can take very little to

go from one to the
other
Sudden stop
Debt threshold exceeded
Crisis; GDP falling
Figure 19.1 Towards a sovereign debt crisis
374
19.3 Enforcement
19.3 Enforcement
Even though it is just not possible to legally enforce sovereign debt as efficiently as corpo-
rate debt, creditors have a wide variety of means at their disposal to compel governments
to repay.
Institutions
There are several supranational mechanisms for addressing sovereign debt. Perhaps the
most prominent of those is the Paris Club, which is a group of finance officials from 19 of
the world’s largest economies, meeting in the French Ministry of Economy, Finance and
Industry. They provide specialised financial services such as war funding, debt restructur-
ing, debt relief and debt cancellation. Resorting to the Paris Club is often the last course
of action for highly indebted poor countries. Recent examples include the cancelling of
some of the debt of Nigeria, Liberia and the Democratic Republic of Congo.
The International Monetary Fund (IMF) (see Krueger, 2002) proposed a formal legal
mechanism for dealing with sovereign debt restructuring. This would have established
similar legal structures for sovereign debt as exist for corporate debt. However, this initia-
tive was rejected by almost everybody.
Military enforcement
In centuries past, some states resorted to rather extreme measures to guarantee their
claims. Egypt reneged on its obligations to France and Britain in 1882 and as a response
both countries invaded Egypt, making it a British ‘protectorate’. Just a few years before,
the British invaded Istanbul in 1876 in the wake of one of Turkey’s defaults.
Similarly, the US interventions in Venezuela in the 1890s and the Dominican Republic
in 1916 and the occupation of Haiti in 1915 were at least partly motivated by debt repay-
ment concerns. Haiti was effectively run as a military dictatorship by the US Marines, who
used their control of the country’s institutions, such as the customs house, to enforce US
interests.
The French and Belgian occupation of the Ruhr region in Germany between 1923 and
1925 was motivated by German difficulties in making war reparations. The formal pretext
was that Germany was 10 days late in delivering 100,000 telegraph poles!
Vulture funds and extreme legal steps

Claims on governments can live on for a long time, even centuries, as the examples of
Haiti and Russia show. If the prospects of repayment appear low, such debt may trade on
the secondary market for pennies to the dollar, attracting so-called vulture funds which
specialise in buying sovereign debt at very large discounts and pursuing impoverished
countries in international courts.
One example was reported in The Guardian newspaper in 2007 (Kaseki, 2007): ‘billion-
aire Paul Singer, who in 1996 paid $11m for discounted Peruvian debt and then threatened
to bankrupt the country unless they paid him $58m [. . .] In order to keep a good stand-
ing in international financial markets Peru paid. Singer and his New Y ork-based invest-
ment fund, Elliot Associates, have since sued the Republic of Congo (Congo Brazzaville)
375
for $400m for a debt bought at $10m.’ Peru was forced to pay because it sent funds to
Euroclear to pay bond holders participating in its restructuring, and the Court of Appeals
in Brussels enjoined Euroclear from making the payments because they violated pari passu
clauses. A similar case was reported by the BBC (2012a) in which Ghana impounded
an Argentinian naval vessel at the request of Elliot Capital Management, to enforce the
fund’s claim against Argentina arising from that country’s sovereign default in 2002.
Induce restructuring – carrots and sticks

Buchheit (2011) identifies two ways for a debtor to restructure sovereign debt, labelled
carrots and sticks.
■ Carrots. The creditor can offer to stretch out maturities or raise interest rates to coun-
teract the costs of a haircut. The sovereign can also offer credit enhancements, such as
collateral securities or guarantees. This is quite costly for the sovereign, perhaps even
removing the advantages of defaulting in the first place. Also these sweeteners may run
foul of existing covenants and may violate pari passu clauses.
■ Sticks. If the sovereign is unable or unwilling to fund a restructuring plan, it can
employ a variety of encouragements to get maturity extensions, interest rate reductions
or haircuts. The sovereign can either threaten default or actually default. In every sov-
ereign restructuring over the past 30 years until the Greek restructuring, the sovereign
either suspended payments or threatened default on debt that was not a part of the
restructuring.
Encouraging lenders to agree

One problem in debt restructuring is to get all of the creditors to agree to a settlement, and
several features have been introduced into restructuring processes to facilitate agreement.
Exit consents have become a common feature of debt restructuring, after Ecuador in
2000. Bondholders that agree to restructuring give the sovereign a proxy vote that may
strip away valuable features of old bonds, making them less attractive to holdouts. This
often requires only a simple majority of bondholders.
Bonds may have collective action clauses (CAC) that permit the majority or superma-
jority of creditors to modify key features of the terms of bonds, including principal and
interest payments.
If sovereign debt is governed by local law, the government may just change the law to
facilitate restructuring. While emerging market borrowers often are not able to issue bonds
governed by their own law, European sovereign debt generally is governed by local law.
19.3.1 History lessons

Buchheit (2011) draws the following six lessons from history.
First, don’t let a sovereign debt problem become a banking crisis. Sovereign debt is fre-
quently held by domestic or foreign banks, and there is always the potential that a large
sovereign default may trigger a banking crisis domestically and even abroad. The restruc-
turing process should aim to prevent this. One simple approach would be to discourage
banks from buying government debt, something most governments find hard to resist.
376
19.4 Background to the European sovereign debt crisis
Second, don’t delay recognising that debt levels are unsustainable. The longer govern-
ments try to prevent an inevitable restructuring, the higher the costs. Mexico in 1982 and
Argentina in 2001 are good examples of such mistakes. Of course, with governments usu-
ally out of power following restructuring, it can be tempting for them to hold on as long
as possible.
Third, keep track of government obligations. In most cases, the government has a rela-
tively clear idea of its own direct liabilities, but the problem is caused by unmonitored bor-
rowing by other government entities, like local governments or state-owned enterprises,
enjoying a sovereign guarantee. Such borrowing can often explode under the radar. This
was a major problem for Argentina, and is now emerging as a problem for Spain and China.
Fourth, ask for enough relief. If a country comes out of a restructuring process with debt
levels that are too high, the problem remains, and the country may have to go through
many rounds of debt restructuring. A cautionary tale is the repeated debt restructuring of
Latin American countries in the 1980s, and the ongoing problems with Greece.
Fifth, be ruthlessly efficient. Sovereign debt crises don’t come alone, but are usually
the final part of significant underlying problems and are often accompanied by other
crises. It is in everybody’s interests to resolve the crisis as quickly as possible. This is likely
to require dealing with some uncomfortable underlying causes (such as reduced competi-
tiveness or unsustainable expenditure) as well as the crises that result from them.
Finally, be even-handed and treat all creditors the same way. Creditors will always jockey
for advantage. Commercial banks may say they have always provided funds and govern-
ment creditors may use geopolitics. Debtors should not discriminate amongst its creditors,
unless absolutely necessary. One exception might be trade and supplier debt. However,
for every creditor accorded preferential treatment, the remainder must be treated more
harshly, and overall resentment will be increased by any perceived lack of fairness.
19.4 Background to the European sovereign

debt crisis
The global crisis that started in 2007 seemed to some commentators to be mostly over
by 2009, but towards the end of that year the second phase of the crisis started, the
European sovereign debt crisis. It started with Greece, but soon Portugal and Ireland were
also in crisis, to be followed by Spain and Cyprus. As these problems remain unresolved,
it is worth examining the historical precedents.
19.4.1 The Faroese crisis of 1992

The first sovereign debt crisis in the European Union (EU) is not the current one; several
member states have been in serious difficulties and have even received bailouts, such as
Britain from the IMF in 1976. There is, however, one sovereign debt crisis from 20 years
ago that has many parallels with the current problems of some countries in the Union,
the Faroese crisis of 1992. We follow the story as told by Danielsson and Oskarsson (2012)
and show key statistics in Figure 19.2.
377
Real GDP index Debt/GDP

140%
110
120%
Real GDP index
Debt/GDP
100 100%
80%
90
60%
80 40%
1986 1988 1990 1992 1994 1996 1998
Figure 19.2 Real Faroese GDP and government debt/GDP
The Faroe Islands are a Danish possession situated between Iceland and Scotland. Even
though they are part of Denmark, and use the Danish currency, they are not a member
of the EU. The economy of the Faroes is primarily based on fishing. As the neighbour-
ing countries expanded their economic zones to 200 miles, the Faroese fishing fleet lost
access to its traditional fishing grounds, and profits plummeted. In response, the govern-
ment of the Faroes resorted to subsidising its only export sector – fishing – to the tune of
34% of total exports throughout the 1980s.
Prior to the crisis in 1990, government net debt never exceeded 50% of GDP. Private
debt, however, had been increasing rapidly, reaching 125% of GDP in 1990. The private
borrowing was facilitated by government guarantees, capital subsidies and production
guarantees. Some of the loan guarantees were deliberately kept secret from the Faroese
parliament, the Danish government and the general public, while bad book-keeping
shielded other support from scrutiny.
The guarantees meant the foreign creditors only worried about the ability of the Faroese
government to repay the loans. Creditors implicitly assumed the Faroese sovereign debt
was underwritten by the Danish government, with its AAA rating, even though no explicit
guarantees were issued. The Danish government did not seem all that concerned and did
nothing to disabuse the creditors.
The money was not put to good use, export subsidies directly added to the debt bur-
den and many, if not most, investments turned out to be sour. In addition to financial
intermediation, the local banks actively engaged in interest rate and tax arbitrage between
the Faroe Islands and Denmark, reaping significant profits.
Eventually, it all got too much, and as the economy was hit by exogenous shocks due to a
slowdown in the global economy and reduced fishing catch, a sovereign debt crisis ensued.
The first affected were the banks, and the Danish financial supervisor became concerned
and demanded that the banks shut down their arbitrage schemes and sharply increase
loan provisions. This was the final straw for the banks. They defaulted and the Danish
government forced the Faroese government to take the banks over at its own expense. This
caused debt to reach 140% of GDP, mostly borrowed from the Danish government.
To the chagrin of the Faroese, the Danish deposit insurance fund and foreign banks,
like Danske Bank, got off relatively scot-free. As the domestic banks defaulted, the GDP
378
contracted by 40%, and 15% of the population emigrated. The crisis soured relations
between the Faroese and their colonial masters. The Faroese maintained that Denmark
had acted with malice in forcing them to assume too much debt in the crisis and imposing
too much austerity, a view vindicated by the Danish state commission report on the crisis
in January 1998 and a Danish parliamentary decision to ask the government to renegoti-
ate the settlement terms in favour of the Faroese. The Danes worried about moral hazard
and claimed the Faroese had been irresponsible and should not have an open-ended
claim on the Danish treasury.
The conflict simmered until March 1998, when the Social Democrats won a majority in
the Danish parliament, provided that the Faroese Social Democratic MP supported them,
which he did, securing a settlement of the dispute worth DKK 1.5 billion to the Faorese
authorities, or 20% of the islands’ GDP.
After the crisis, the Faroese economy recovered swiftly, with an annual growth of 4%
between 1995 and 2010. The Faroese ran a capital account surplus amounting to 13%
of GDP per year for a decade, eliminating the sovereign debt. This was helped by the
re-exports of the excess investment goods acquired prior to the crisis, and a sharp fall in
imports, not least because of widespread emigration.
There are many parallels between the Faroese crisis and the ongoing European sover-
eign debt crisis, especially for Greece. Both countries got into difficulty because of excess
borrowing, facilitated by belonging to a currency union with an AAA-rated partner. The
creditors implicitly assumed that the debt was somehow underwritten, and that fiscal
misconduct was prevented by the rules of the greater community. In neither case did the
senior partner seem all that concerned, even as the sovereign debt spiralled upwards.
In the Faroese case, the crisis in state relations was eventually resolved when political
necessities outweighed the cost of the bailout. It did help to have representation in the
parliament of the main protagonist.
19.4.2 Currency union

The origins of the European sovereign debt crisis can be found two decades earlier in the
formulation of the Economic and Monetary Union (EMU) and the euro.
History of monetary unions in Europe

There is a long history of monetary unions in Europe. For example, in the nineteenth cen-
tury, we had the Latin Monetary Union, established in 1866 with Belgium, France, Italy
and Switzerland as members, soon to be joined by Spain and Greece, and others later.
This was not a true monetary union because each country issued its own currency, but
they maintained pegged exchange rates with each other. The Latin union survived until
1914. The UK discussed joining, but two stumbling blocks prevented it. First, it had to
devalue slightly, to make one pound sterling equivalent to 25 French francs. The second
was that it had to decimalise, giving up shillings and pence. Another nineteenth-century
attempt was the Scandinavian monetary union between Denmark, Sweden and Norway,
established in 1873 and also lasting until 1914. These two attempts were just the last in a
long chain of monetary unions, mostly doomed to failure.
379
The German monetary union in 1857 was a success. It replaced the currencies of the
many German states with a dual system that lasted until the German unification of 1870,
with the two currencies themselves replaced by the mark in 1875.
The other main success was the Belgium–Luxembourg monetary union of 1922. This
came under stress on occasion, especially in the 1980s when Belgium unilaterally deval-
ued the franc, infuriating the Luxembourgers. In response, Luxembourg set up a standby
central bank, ready to switch over to Luxembourg francs within 24 hours if Belgium mis-
behaved again.
19.4.3 Why do monetary unions succeed and why do they fail?

Transfer unions
Definition 19.1 Transfer union A transfer union is an arrangement whereby a

number of countries come together in some organisation, where the stronger members of
the group directly and significantly subsidise the weaker members on a long-term continuous
basis, and where the member countries have strong political connections between them.
Intergovernmental transfers are very common, as, for example, in development aid. Most
of these do not imply that the countries involved are members of a transfer union. For
example, development aid fails both of the tests in the definition above.
Is the EU a transfer union or not? If we look at the EU budget for 2010 (European Union,
2011, Annex 2c, p. 75), we see that the net amount received by member countries is
around 0.3% of the GDP of the Union, even though for some member countries it might
be as high as 4%. It is harder to identify net transfers within individual countries, but one
report attempts to do that for Spain for 2005 (Spanish government, 2008, table 1, p. 13),
finding that intra-region transfers amount to 2.2% of the GDP of Spain, in some regions
amounting to 16% of that region’s GDP.
This suggests that a transfer union is defined by the magnitude of the transfers.
Individual nation states transfer an order of magnitude more funds between regions than
the EU. Consequently, individual nation states are generally transfer unions, but the EU
is not.
Success and failure

Whilst there is no set formula for why monetary unions succeed, they tend to bring
together people who belong together culturally, in a single nation state, where peo-
ple move freely between different parts of the country, the same rules and regulations
apply generally across the union, economic development is not deeply out of sync and
there is general acceptance of the notion of a transfer union. The main exception to this
relates to situations in which small countries use the currency of a much bigger one, like
Luxembourg and Belgium or Panama and the US. This directly relates to analysis of opti-
mum currency areas, first discussed by Mundell (1961).
Monetary unions fail when these things are not in place. In times past, different mem-
bers might be prone to war between themselves, activity not conducive to running joint
380
monetary policy. Also, over time, it is likely that some countries prosper and others are
left behind. If so, they will come to require different monetary policies. This suggests
that unless a transfer union is in place, currency unions among near equal parties are
unstable.
19.4.4 European currency arrangements since Bretton Woods

The EU has gone through several attempts at currency arrangements. The first was the
‘snake in the tunnel’ in the 1970s. This was followed by the European Monetary System
(EMS), whereby a new virtual currency was created, the European Currency Unit (ECU).
Exchange rates were supposed to fluctuate within a 2.25% target zone, except 6% for Italy.
The German mark was the effective reserve currency of the arrangement. This caused the
EMU crisis in the early 1990s as discussed in Section 12.4.
The European authorities decided to adopt a common currency in 1995 – the euro –
to be implemented as an accounting currency on 1 January 1999, with coins and bank-
notes entering circulation on 1 January 2002. For an overview of the euro, see Marsh
(2010). The area where the euro is the legal currency, and whose members are a part of
the European Central Bank (ECB), is referred to as the euro zone, currently consisting of
17 members.2 The euro is also used in six other countries,3 but they are not represented
at the ECB.
The motivation for setting up a monetary union need not be primarily economic. The
reason the EU was set up in the first place is because it was felt that if Europe became
integrated, war would be prevented. Based on that metric, the Union has been a success,
the past 60 years having been about the most peaceful in European history. The establish-
ment of the EMU was the next logical step in the integration process. This means that the
motivation for the euro was more political than economic.
The Maastricht criteria

The designers of the euro did recognise the danger facing monetary unions and, in
response, established the Maastricht criteria, which have four requirements. Two are
related to restrictions on exchange rates and interest rates prior to joining.
Definition 19.2 Maastricht debt criteria The deficit must not exceed 3% annually,
with the possibility of temporary suspension in exceptional cases. Government debt must
not exceed 60% of GDP, and if it is higher, must approach the 60% at a satisfactory pace.
Table 19.1 shows the debt and deficits of member countries in the year they joined.
More than half violated Maastricht criteria that year, and of those that qualified, two are
now in a state of crisis.
2
Austria, Belgium, Cyprus, Estonia, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg,
Malta, the Netherlands, Portugal, Slovakia, Slovenia and Spain.
3
Montenegro, Andorra, Monaco, San Marino, Vatican City and Kosovo.
381
Table 19.1 Euro zone deficit and debt in first year of euro membership and Maastricht compliance
Primary Compliance with

Country Year of joining Deficit Debt balance Maastricht
Austria 1999 -2.3, 66.8% 1.1% :

Belgium 1999 - 0.6% 113.6% 6.2% :
Cyprus 2008 0.9% 48.9% 3.8% ✓
Estonia 2011 1.0% 6.0% 1.1% ✓
Finland 1999 1.7% 45.7% 4.7% ✓
France 1999 - 1.8% 58.9% 1.2% ✓
Germany 1999 - 1.6% 61.3% 1.6% :
Greece 2001 - 4.5% 103.7% 2.0% :
Ireland 1999 2.7% 46.6% 5.0% ✓
Italy 1999 - 1.9% 113.0% 4.6% :
Luxembourg 1999 3.4% 6.4% 3.7% ✓
Malta 2008 - 4.6% 62.3% -1.4% :
Netherlands 1999 0.4% 61.1% 4.7% :
Portugal 1999 - 3.1% 51.4% -0.2% :
Slovakia 2009 - 8.0% 35.6% -6.6% :
Slovenia 2007 0.0% 23.1% 1.2% ✓
Spain 1999 - 1.2% 62.4% 2.3% :
Data source: Eurostat, http://epp.eurostat.ec.europa.eu/portal/page/portal/eurostat/home/ © European Union
19.4.5 Evolution of European debt

Table 19.2 shows that debt levels vary quite significantly across the euro zone. Greece
and Italy have the highest debt, with Estonia the least. Only three countries satisfied the
Maastricht criteria in 2011.
Figure 19.3 shows the change in debt over the past decade and a half. Two countries in
crisis, Greece and Portugal, have had the biggest increase in debt, while Belgium has been
the most successful in reducing its debt.
The figures indicate that the level of debt tolerance varies widely across Europe.
Table 19.2 shows that there is not a particularly strong correspondence with debt levels
and credit rating, otherwise Estonia would be the most highly rated country in Europe
and Belgium and Italy would be rated much lower, with even the AAA ratings of France
and Germany under threat. This is a clear example of how debt tolerance varies across the
Union.
Figure 19.4 shows the evolution of the ratings of the main crisis countries. Deficits seem
to do a good job in predicting ratings, suggesting that a debt tolerance concept based
solely on debt/GDP ratio is incomplete.
Total (public plus private) gross external debt/GDP is reported in www.reinhartandro-
goff.com for a number of countries, and we report the results for the euro zone members
in Figure 19.5, where available. Ireland has the highest level of external debt, exceeding
382
Table 19.2 Euro zone debt to GDP, Maastricht compliance, and ratings in August 2012
2011 Rating,
primary August Compliance with
Country 1995 debt 2007 debt 2011 debt 2011 deficit balance 2012 Maastricht
Austria 68% 60% 72% - 2.6% 0.0% AAA :

Belgium 130% 84% 98% - 3.7% - 0.4% AA :
Cyprus 52% 59% 72% - 6.3% - 3.8% BB + :
Estonia 8% 4% 6% 1.0% 1.1% A+ ✓
Finland 57% 35% 49% - 0.5% 0.6% AAA ✓
France 56% 64% 86% - 5.2% - 2.6% AAA :
Germany 56% 65% 81% - 1.0% 1.6% AAA :
Greece 97% 107% 165% - 9.1% - 2.2% CCC :
Ireland 80% 25% 108% - 13.1% - 9.7% BBB + :
Italy 121% 103% 120% - 3.9% 1.0% A- :
Luxembourg 7% 7% 18% - 0.6% - 0.1% AAA ✓
Malta 35% 62% 72% - 2.7% 0.4% A+ :
Netherlands 76% 45% 65% - 4.7% - 2.6% AAA :
Portugal 59% 68% 108% - 4 .2% - 0 .4% BB + :
Slovakia 22% 30% 43% - 4 .8% - 3 .2% A+ :
Slovenia 19% 23% 48% - 6 .4% - 4 .5% A :
Spain 63% 36% 68% - 8 .5% - 6 .1% BBB :
Data source: Eurostat and Fitch
Belgium
Netherlands
Finland
Estonia
Italy
Austria
Spain
Luxembourg
Cyprus
Slovakia
Germany
Ireland
Slovenia
France
Malta
Portugal
Greece
−30% −15% 0% 15% 30% 45% 60%
Figure 19.3 Change in debt to GDP in the euro zone, 1995–2011

Data source: Eurostat
383
B− B−
B+ Greece Ireland Italy B+
BB+ Portugal Spain Cyprus BB+
BBB BBB
A− A−
A+ A+
AA AA
AAA AAA
1995 2000 2005 2010
(a) 1995 to July 2012
Greece Spain
B− B−
Portugal Italy
B+ Ireland Cyprus B+
BB+ BB+
BBB BBB
A− A−
A+ A+
AA AA
AAA AAA
2010 2011 2012
(b) August 2009 to July 2012
Figure 19.4 Ratings

Data source: Fitch
Austria Italy
Belgium Netherlands
1000% France Portugal 1000%
500% Germany Spain 500%
Ireland
100% 100%
50% 50%
10% 10%
1980 1990 2000 2010
Figure 19.5 Total (public plus private) gross external debt/GDP

Data source: Eurostat and www.reinhartandrogoff.com
10 times the GDP, and Italy the smallest at 1.2. In round terms, external debt has increased
10 times since the collapse of Bretton Woods.
19.4.6 Individual countries

Of the member states of the euro zone, six countries are in particular difficulties. Three
have already received bailouts – Ireland, Greece and Portugal. Cyprus may be next, while
Spain got a bailout for its banks and a government bailout may follow. The biggest worry
is Italy, which has received only indirect assistance via the ECB liquidity facility but is
384
Estonia
Slovakia
Slovenia
Ireland
Luxembourg
Greece
Finland
Cyprus
Spain
Austria
Netherlands
Belgium
Germany
France
Portugal 1999−2007
Italy 2008−2011
−2% 0% 2% 4% 6%
Figure 19.6 Real average annual economic growth

Data source: Eurostat. No data available for Malta
Estonia
Slovakia
Slovenia
Greece
Finland
Ireland
Austria
Cyprus
Netherlands
Luxembourg
Portugal
France
Belgium
Italy 1999−2007
Spain 2008−2011
−1% 0% 1% 2% 3% 4% 5% 6%
Figure 19.7 Average annual productivity growth

Data source: Eurostat. No data available for Malta
in a precarious position. We can see economic and productivity growth in Figures 19.6
and 19.7, respectively.
It is important to note that the benefits of the euro zone do not all flow in one direc-
tion. The presence of uncompetitive countries in the euro zone has over many years weak-
ened the euro and so improved external (to EU) terms of trade for the most competitive;
385
these have also benefited from growth at the expense of their less competitive neighbours.
Germany’s full employment and relatively healthy domestic finances are due in part to
the currency union. Below we present background analysis on main crisis countries. More
information is in the online chapter at www.GlobalFinancialSystems.org.
Greece
The crisis had its origins in Greece, which was one of the fastest growing economies in the
euro zone in the 1990s and early 2000s, averaging an annual growth rate of close to 4%.
This was fuelled by a rapid increase in debt, 68% between 1995 and 2011, the highest in
the euro zone. At the time it joined the euro, it already violated the Maastricht criteria, but
unlike the other two high-debt countries, Italy and Belgium, its primary balance was low.
Furthermore, the debt numbers were manipulated downwards, and are reported as provi-
sional. To comply with the monetary requirements for joining the euro, the Greek govern-
ment consistently and deliberately misreported the country’s official economic statistics.
For example, Greece paid several investment banks to execute deals that moved a high level
of debt off-balance sheet in order to conceal the actual level of debt. Greece never enjoyed
a high credit rating; still, until 2009 its rating was A, and as recently as 2004 it was A +.
While there are many reasons for the crisis in Greece, prior to joining the euro its ability
to borrow too much was limited by its high inflation and perceived structural weaknesses.
The governments would use devaluations to stimulate the economy when needed. After
it joined, it was able to borrow at almost the same interest rate as Germany, as investors
assumed in the early years of the currency union that bonds of various euro zone coun-
tries carried equal risks.
After a general election in 2009 swept a new government into office, the incoming
prime minister, George Papandreou, revised the actual deficit figures of the country from
an originally estimated 6% to 12.7%, later increased to 13.6%. Accumulated sovereign
debt reached 165% of GDP in 2011.
As the true situation of Greece’s public finances became apparent, Greek public debt
was downgraded several times, leading to rapid increases in bond yields as a default
started to appear like a real possibility. The new government embarked on a course of
harsh austerity measures but did not manage to convince the markets.
Ireland
The second country to run into difficulties was Ireland. We have already discussed its prob-
lems in Section 14.1. The difficulties in Ireland were quite different in nature from those in
Greece. The government seemed quite prudent in its public financing; the problems were
in the banking sector, mostly financed with borrowing from abroad. After the government
failed in its attempt to use sovereign liquidity guarantees to the banks as means to restore
confidence, the Irish debt levels increased from 25% in 2007 to 108% in 2011, and are still
rising. The Irish government obtained an €85 billion bailout from the EU in 2010.
Portugal
Soon after Greece got into difficulty, Portugal followed, for many of the same reasons but
not on as extreme a scale as Greece. Eventually, it got a €78 billion bailout from the EU and
the IMF in 2011. In return, it has to implement austerity measures, such as privatisation,
386
increasing sales taxes, freezing or cutting benefits, cutting school spending, freezing pen-
sions and reducing the number of civil servants. Unemployment reached 14.8% in the
beginning of 2012. To date, Portugal has been able to comply with the requirements, and
the current consensus suggests that its problems are less profound than those of Greece.
Cyprus
Cyprus is the latest country to get into difficulty, caused primarily by the exposure of its banks
to Greece. It requested a bailout from the EU in the summer of 2012, but delays in coming to
an agreement led to a banking crisis in March 2013, which ended with a bailout agreement
with the EU and the IMF, with large haircuts imposed on bank creditors and depositors.
Spain
The situation in Spain is in many ways similar to that of Ireland. In both cases the govern-
ment had low debt, and the problems were created in the banking sector. In both coun-
tries, real-estate speculation played an important part, especially in Spain.
The Spanish government was slow to wake up to its problems; as problems in the
banking sector mounted, and as unemployment sharply increased, it was perceived as sit-
ting on its hands. This undermined investor confidence in the government, in spite of its
low debt levels at the time, and the cost of borrowing increased sharply, along with debt
downgrades. In 2012, finally recognising the problems in the banking sector, the govern-
ment got a banking bailout from the EU. In exchange, Spain gave up some supervisory
control over its banking sector, the first step towards a banking union. After all, ‘he who
pays the piper calls the tune’!
Italy
If we had been studying the statistics presented here in 2007, none of the crisis countries
would have shown particular weaknesses compared to other member states. The only
exception is Italy. It was not compliant with the Maastricht criteria when it joined, with
debt of 113%, and its debt levels are essentially the same now as they were in 1995. It had
the lowest economic growth in the euro zone in the first eight years of the euro, the sec-
ond lowest productivity growth after Spain (almost zero) and is ageing rapidly. The main
positive factor is that Italian debt is mostly held domestically.
Italy so far has avoided a sovereign debt crisis, but there are considerable fears that it
might get to that state if it is not able to effectively address its structural problems.
19.4.7 The causes of the sovereign debt crisis

If there is one single cause it is the euro. Before the European countries had a common
currency, each country borrowed in its own currency, where the borrowing rates reflected
the fundamentals of the country. If the country got into difficulty, it always had the option
of devaluing, which simultaneously made its economy more competitive and reduced the
real value of its sovereign debt, because of the resulting inflation. In this situation, the sov-
ereign debt of the country was its own affair, not spilling over to other European countries.
This all changed with the euro. When a country has a common currency, weaker coun-
tries no longer have the ability to devalue their currency or inflate away their debt, while
387
at the same time their borrowing rates are likely to improve. Initially, such countries find
it easier to borrow, while at the same time they have less ability for dealing with a crisis. It
seems optimistic of the designers of the euro in the 1990s to discount such an eventuality.
The problem has become so extreme because no member country (supposedly) can
leave the euro zone, so that the fate of the weakest member automatically becomes a
direct concern of the strongest number. Lenders’ belief in this proposition was the reason
for the sharp reduction in interest rates enjoyed by the weaker countries, contributing
greatly to the economic booms preceding the crisis.
19.4.8 Why is it so difficult to solve the crisis?

Unfortunately, the resulting integration of the economies of member states now poses
substantial problems. If a country leaves, it will cause huge economic disruption for other
member countries. Politicians’ credibility will suffer, and the benefits of currency union to
the remaining members will be reduced.
This gives the potential leavers a strong negotiating position. If they are unable to sur-
vive in the currency union, they can point out that a transfer union is necessary to prevent
them from leaving.
The problem is that the wealthier countries find the transfer union unpalatable. Even
though the political leaders may be contemplating it, it is less clear whether they have the
support of their voters.
The differing agendas and incentives of the European authorities make it impossible
for them to speak with one voice on the crisis and, hence, directly contribute to the insta-
bility. There is also an element of brinkmanship at work: if the crisis gets bad enough,
perhaps a transfer union will be seen as the only solution, taking Europe directly towards
unification – a desirable outcome for many. To the extent that wealthy countries are com-
mitted to no country leaving the union, they cede negotiating power to poor countries
that can threaten to do so.
Whilst this is opposed by voters in wealthy countries (and whilst the loss of sovereignty
is opposed by voters in poor countries), this opposition is not entirely clear-cut. If the
question is posed slightly differently, there is also strong support for ‘solidarity’ with other
European countries, suggesting that the current state of indecision and muddled compro-
mise could persist for a long time.
19.4.9 Global impact

In principle, there is no reason why a sovereign debt crisis in a handful of countries should
threaten a global crisis. Even in a European context, the GDP of the crisis countries is still
relatively small, and it would cost the euro only zone about 2% of its GDP to do a com-
plete bailout of Greece.
This suggests that in order to understand the nature of the problem we should not
focus on the detail of what is happening in Greece but rather how it affects the rest of the
global economy. As with other crises discussed in this book, the trigger event for a crisis
is less important than the mechanism allowing it to spiral out of control. In Europe, the
388
19.5 Summary
enabling factor is the euro, the cross-border economic activity it (by design) increased,
and the result that a major crisis in any member country can cause a crisis of confidence
elsewhere.
This is why the crisis response of the European authorities is so important. To date, the
response has not been sufficient to stop the crisis, and the situation is at best stable but
in a very unsatisfactory state. More probably it will continue to deteriorate, as the funda-
mental problems remain unaddressed. It seems likely that some trigger may be required
to bring about fundamental change.
19.5 Summary
We started by discussing one of the extreme cases of dealing with the sovereign debt cri-
sis, when the country of Newfoundland was liquidated.
The topic of this chapter is sovereign debt crises, perhaps the most common form of
crisis throughout history. We discussed the nature of sovereign debt crises, what makes
sovereign debt special and different from corporate debt, how the interests of creditors can
be protected, and how a country should respond when faced with a sovereign debt crisis.
Even though there is no international court that enforces sovereign debt, creditors have
a variety of means at their disposal, historically military intervention and more recently a
range of sanctions.
We followed this by a discussion of the European sovereign debt crisis, starting with the
first sovereign debt crisis in the EU, and analysis of monetary unions, why they succeed and
why they fail. This is key to understanding the European sovereign debt crisis, and we ana-
lysed to what extent the necessary conditions for a successful monetary union have been
present in the euro zone. Finally, we discussed some of the European countries that are expe-
riencing economic difficulty, why it is so difficult to solve the crisis, and the global impact.
We leave a detailed analysis of current events to a chapter that is published online at
www.GlobalFinancialSystems.org.

1 Is it too strong to say Britain liquidated Newfoundland?
2 How can a government default on its debt without actually declaring default?
3 Yields on government bonds of the US are currently trading at record lows. How can
this be reconciled with the fact that the US is approaching its greatest level of debt to
GDP in its history, apart from in wartime?
4 Do you think a return to gunboat diplomacy would be useful in dealing with sovereign
debt crises?
5 The role of entities buying sovereign bonds on the secondary market at a big discount is
controversial, especially the so-called ‘vulture funds’. Do you agree with their methods
and, more generally, what is your view on the enforcement mechanisms that should
exist for sovereign debt?
389
6 What is the main reason why the euro zone was established?
7 What are main conditions for a successful monetary union, and did the euro zone fulfil
those criteria?
8 What is the cause of the European sovereign debt crisis?
9 Would Greece be better off if it left the euro zone?
10 Would the euro zone be better off if Greece left?
11 Does a country suffer more when it experiences a currency crisis or when it experiences
a banking crisis or a sovereign debt crisis?
12 Many commentators talk of contagion when it comes to sovereign debt crises. Identify
how this could happen.
References
BBC (2012a). Argentina ship in Ghana seized over loans default. www.bbc.co.uk/news/
world-africa-19827562.
Borensztein, E. and Panizza, U. (2008). The costs of sovereign default. Technical Report
WP/08/238, IMF.
Buchheit, L. C. (2011). Sovereign debt restructuring – the legal context. Cleary Gottlieb Steen &
Hamilton LLP, New York.
Daníelsson, J. and Oskarsson, H. (2012). The first sovereign debt crisis in the EU. Vox EU, 11
September. www.voxeu.org/article/europe-s-pre-eurozone-debt-crisis-faroe-islands-1990s.
European Union (2011). EU budget 2010 financial report. Technical report.
Gulati, M. and Triantis, G. (2007). Contracts without law: sovereign versus corporate debt.
University of Cincinnati Law Review, 75: 977–1004.
Hale, D. (2003). The Newfoundland lesson. The International Economy, 17(3): 52–61.
Hansard (1993) http://hansard.millbanksystems.com/commons/1933/dec/12/newfoundland-
bill#S5CV0284PO_19331212_HOC_267.
Homer, S. and Sylla, R. (1996). A History of Interest Rates. Rutgers University Press.
Kaseki, J.-R. (2007). Preying on the poor. The Guardian, 27 September.
Krueger, A. O. (2002). A new approach to sovereign debt restructuring. Technical report, IMF.
Marsh, D. (2010). The Euro: the Politics of the New Global Currency. Yale University Press.
Mundell, R. A. (1961). A theory of optimum currency areas. Amer. Econ. Rev., 51: 657–65.
Reinhart, C. M., Rogoff, K. and Savastano, M. A. (2003). Debt intolerance. Technical report,
NBER.
Spanish government (2008). Las balanzas fiscales de las CC.AA. españolas con las aa. Públicas
Centrales 2005. Technical report, Ministerio de Economía y Hacienda.
Sturzenegger, F. and Zettelmeyer, J. (2007). Debt Defaults and Lessons from a Decade of Crises.
MIT Press.
The Economist (1999). South Africa’s debt: unforgivable. The Economist (US), 24 April.
390
Glossary
1G first-generation currency crisis model. CEO chief executive officer.

2G second-generation currency crisis model. CET1 common equity tier 1, a part of bank capital.
ABCP asset-backed commercial paper is a short-term CFTC Commodity Futures Trading Commission is an
debt instrument, typically with a maturity between independent agency of the United States government
90 and 180 days, backed by physical assets such as that regulates futures and option markets.
trade receivables. CIO Chief Investment Office.
ABS asset-backed securities derive their value and CMG Crisis Management Group, a cooperative multi-
income payments from a pool of underlying assets. national and multi-organisational set up to deal with
ADR American depositary receipts. the failures of large cross-border financial institutions.
AIB Allied Irish Bank is an Irish bank that got bailed CoCo contingent convertibles.
out by its government.
CPI consumer price index.
AIG American International Group, a US insurance
CRA credit rating agencies are international firms
company that failed in 2008 because of its CDS writ-
that provide evaluations of the creditworthiness of
ing activities.
various types of borrowers.
Basel I the first international capital accords, in effect
CS Credit Suisse is a large Swiss bank.
from 1992 until 2007.
DAX the main German stock market index.
Basel II the second international capital accords, in
effect from 2007. DB Deutsche Bank is a large German bank.
Basel III the next version of the international capital DJIA the Dow Jones industrial average index is a popu-
accords, intended to be implemented from 2013 but lar United States stock market index.
being delayed. EBA European Banking Authority.
BCBS Basel Committee for Banking Supervision is a ECB European Central Bank.
group of senior officials that design international finan- ECU European Currency Unit.
cial regulations, best known for the Basel Accords.
EEA European Economic Area.
BCCI Bank of Credit and Commerce International, a
EMH the efficient market hypothesis maintains that
Luxembourg-registered bank, with head offices in
one cannot systematically earn excessive profits from
Karachi and London. It was established in 1972 and
exploiting public information.
collapsed in 1991 after massive fraud.
BIS Bank for International Settlements. EMS European Monetary System.
BoE Bank of England. EMU synonymous with the euro zone.
BoJ Bank of Japan. ERM European Exchange Rate Mechanism.
CAC collective action clauses permit the majority or ESMA European Securities and Markets Authority.
supermajority of creditors to modify key features of ESRB European Systemic Risk Board.
the terms of bonds, including principal and interest EU European Union.
payments.
EURIBOR Euro Interbank Offered Rate, a reference
CAPM capital asset pricing model. rate calculated from the averaged interest rates at
CAR capital adequacy ratio. which euro zone banks offer to lend unsecured funds
CBT computer based trading. to other banks.
CCP central counterparty. FDIC Federal Deposit Insurance Corporation.
CDO collateralised debt obligation. Fed the Federal Reserve System is the central bank of
CDS credit default swap. the United States.
391
Glossary
FPC Financial Policy Committee. NYFed the New York Federal Reserve bank is the most
FSA Financial Services Authority. important branch of the Fed. It is the Fed’s main
interface with financial markets.
FSB Financial Stability Board.
O&G Overend and Gurney.
FSF Financial Stability Forum, a group of financial autho
rities of the G7 countries. It is superseded by the FSB. OECD Organisation for Economic Cooperation and
Development, a Paris-based international organisation.
FSOC Financial Stability Oversight Council.
OTC over-the-counter.
FT100 the main British stock market index.
PPP purchasing power parity.
FTT financial transaction tax.
QE quantitative easing.
FX foreign exchange.
RBS Royal Bank of Scotland is a large British bank
GATT General Agreement on Tariffs and Trade.
that got bailed out by the government in the crisis
G-SIB global systemically important banks that have
from 2007.
been identified by the FSB as being especially large
RFC Reconstruction Finance Corporation.
and whose failure poses a special danger to the
world economy. RWA risk-weighted assets of a bank, used in capital
calculations.
G20 a group of large countries that include Argentina,
Australia, Brazil, Canada, China, France, Germany, India, S&L savings and loans, a part of the United States
Indonesia, Italy, Japan, Mexico, Russia, Saudi Arabia, financial system that was in a crisis in the 1980s.
South Africa, South Korea, Turkey, the United Kingdom, S&P Standard & Poor.
the United States and collectively the European Union. S&P-500 the Standard & Poor’s 500 index is the most
GDP gross domestic product. representative United States stock market index.
GNP gross national product. SDR special drawing rights, a virtual currency created
HBOS a large British bank that failed in the crisis from by the IMF.
2007. SEC Securities and Exchange Commission.
HFT high-frequency trading. SIC Special Investigation Commission.
HSBC a large British bank. SIFI systemically important financial institution.
IIF the Institute of International Finance is the main advo- SIV structured investment vehicle.
cacy group (lobbyists) of internationally active banks. SME small and medium-sized enterprise.
IKB a German bank that was the first financial institu- SNB Swiss National Bank.
tion to fail in the crisis in 2007.
SPV special purpose vehicle.
ILOLR international lender of last resort.
SSM Single Supervisory Mechanism, the embryonic
IMF International Monetary Fund. pan-EU supervisor.
IOSCO International Organization of Securities TA total assets of a bank, used in capital calculations.
Commissions is an association of organisations that
TARP Troubled Asset Relief Program.
regulate the world’s securities and futures markets.
TBTF too big to fail.
IRB internal rating based refers to banks developing
UBS a large Swiss bank.
their own methodologies for assessing risk.
UIP uncovered interest rate parity.
LCR liquidity coverage ratio.
UK United Kingdom.
LIBOR London Interbank Offered Rate, average inter-
est rate from banks calculated by the British Bankers’ US United States.
Association. USD US dollar.
LOLR lending of last resort. VaR Value-at-Risk is a common statistical technique
LR leverage ratio. for forecasting market risk.
LTCM Long Term Capital Management, a celebrated VIX implied volatility of the S&P-500 index.
hedge fund that failed in 1998, triggering a global crisis. WTO World Trade Organization.
MPT modern portfolio theory. WWI First World War.
NSFR net stable funding ratio. WWII Second World War.
392
Bibliography
Abdelal, R. (2007). Capital Rules: The Construction of Barnier, M. (2011) quoted in Barker, A., EU ban on ‘na-
Global Finance. Harvard University Press. ked’ CDS to become permanent. Financial Times,
Acharya, V. and Pedersen, L. H. (2005). Asset pricing 19 October, www.ft.com/cms/s/0/cc9c5050-f96f-
with liquidity risk. J. Finan. Econ., 77: 375–410. 11e0-bf8f-00144feab49a.html.
Admati, A. R., DeMarzo, P. M., Hellwig, M. F., and Pflei- Basel Committee on Banking Supervision (1996).
derer, P. (2010). Fallacies, irrelevant facts, and myths Amendment to the Capital Accord to Incorporate
in the discussion of capital regulation: why bank Market Risks. www.bis.org/publ/bcbs24.pdf.
equity is not expensive. Working Paper, Graduate Basel Committee on Banking Supervision (2004). Bank
School of Business, Stanford University. failures in mature economies. Working paper no. 13,
Adrian, T. and Shin, H. S. (2010). Liquidity and leverage. Bank of International Settlements.
J. Finan. Intermed., 19: 418–37. Basel Committee on Banking Supervision (2006). The
Ahamed, L. (2009). Lords of Finance, The Bankers Who management of liquidity risk in financial groups.
Broke the World. The Penguin Group. Technical report, pp. 1–25.
Akerlof, G. (1970). The market for ‘lemons’: quality un- Basel Committee on Banking Supervision (2011a). Ba-
certainty and the market mechanism. Quart. J. Econ., sel Committee issues final elements of the reforms to
84(3): 488–500. raise the quality of regulatory capital. www.bis.org/
Alba, P., Bhattacharya, A., Claessens, S., Ghosh, S. and press/p110113.pdf.
Hernandez, L. (1998). Volatility and contagion in a Basel Committee on Banking Supervision (2011b). Ba-
financially-integrated world: lessons from East Asia’s sel III: a global regulatory framework for more resil-
recent experience. Paper presented at the PAFTAD 24 ient banks and banking systems. Technical report.
Conference, ‘Asia Pacific Financial Liberalization and Basel Committee on Banking Supervision (2011c).
Reform’, 1998. Global systemically important banks: assessment
Alessandri, P. and Haldane, A. G. (2009). Banking on methodology and the additional loss absorbency re-
the state. Mimeo, Bank of England. quirement. www.bis.org/publ/bcbs207.pdf.
Allen, F. and Gale, D. (1994). Limited market participa- Bassani, G. and Trapanese, M. (2011). Crisis manage-
tion and volatility of asset prices. Amer. Econ. Rev., 84: ment and resolution. In Quagliariello, M. and Canna-
933–55. ta, F., editors, Basel III and Beyond. Riskbooks, London.
Allen, F. and Gale, D. (2005). From cash-in-the-market Baudino, P. (2011). The policy response: from the G20
pricing to financial fragility. J. Eur. Econ. Assoc., 3: requests to the FSB roadmap; working towards the
535–46. proposals of the Basel Committee. In Quagliariello,
Anderson, G. M., Shughart, W. F., II and Tollison, R. D. M. and Cannata, F., editors, Basel III and Beyond. Risk-
(1988). A public choice theory of the great contrac- books, London.
tion. Public Choice, 59: 3–23. BBC (2012a). Argentina ship in Ghana seized over loans
Angelini, P. L., Clerc, L., Curdia, V., Gambacorta, L., Gerali, default. www.bbc.co.uk/news/world-africa-19827562.
A., Locarno, A., Motto, R., Roeger, W., den Heuvel, S. V. BBC (2012b). Barclays’ widening Libor-fixing scandal.
and Vlcek, J. (2011). Basel III: long-term impact on eco- www.bbc.co.uk/news/ business-18671255.
nomic performance and fluctuations. Technical report, Benediktsdottir, S., Danielsson, J. and Zoega, G. (2011).
Federal Reserve Bank of New York. Lessons from a collapse of a financial system. Eco-
Arnason, R. and Danielsson, J. (2011). Capital controls nomic Policy, 26: 183–231.
are exactly wrong for Iceland. Vox EU, 14 November. Bernanke, B. (1995). The macroeconomics of the Great
www.voxeu.org/article/iceland-and-imf-why-capital Depression: a comparative approach. J. Money Credit
controls-are-entirely-wrong. and Banking, 27(1): 1–28.
Bagehot, W. (1873). Lombard Street: a Description of the Bernanke, B. S. (2005). The global saving glut and the
Money Market. H.S. King, London. US current account deficit. Speech delivered at the
Bandt, D. and Hartmann, P. (2000). Systemic risk: a sur- Sandridge Lecture, Virginia Association of Econo-
vey. Working paper no. 35, European Central Bank. mists, Richmond, VA.
393
Bibliography
BIS (2010). Assessing the macroeconomic impact of the Brunnermeier, M. and Pedersen, L. H. (2009). Market li-
transition to stronger capital and liquidity require- quidity and funding liquidity. Rev. Finan. Stud., 22(6):
ments. www.bis.org/publ/othp10.pdf. 2201–38.
Bitner, R. (2008). Confessions of a Subprime Lender: an Buchheit, L. C. (2011). Sovereign debt restructuring –
Insider’s Tale of Greed, Fraud, and Ignorance. John the legal context. Cleary Gottlieb Steen & Hamilton
Wiley & Sons. LLP, New York.
Black, F. (1971). Toward a fully automated exchange, Buffett, W. (2003). Avoiding a mega-catastrophe. Fortune,
part I. Financial Analysts J., 27: 28–35. 24 March.
Black, F. (1995). Hedging, speculation, and systemic Buffett, W. (2012). Why stocks beat gold and bonds.
risk. J. Derivatives, 2: 6–8. Fortune, 27 February.
Black, F. and Scholes, M. (1973a). The pricing of op- Bussiere, M., Saxena, S. C. and Tovar, C. E. (2010).
tions and corporate liabilities. J. Polit. Econ., 81(3): Chronicle of currency collapses re-examining the ef-
637–54. fects on output. Technical report, ECB. Working pa-
Black, F. and Scholes, M. (1973b). The valuation of per series no. 1226, July.
option contracts and a test of market efficiency. J. Fi- Cagan, P. (1956). The monetary dynamics of hyperinfla-
nance, 27: 399–418. tion. In Friedman, M., editor, Studies in the Quantity
Bloomberg (2011). Secret Fed loans undisclosed to Theory of Money. University of Chicago Press.
Congress gave banks $13 billion in income. www. Calello, P. and Ervin, W. (2010). From bail-out to bail-
bloomberg.com/news/2011-11-28/secret-fed-loan- in. The Economist, 28 January.
sundisclosed-to-congress-gave-banks-13-billion-in- Capie, F. and Wood, G. E. (2006). The Lender of Last Re-
income.html. sort. Routledge.
Blundell-Wignall, A. and Atkinson, P. (2010). Thinking Caprio, G. and Honohan, P. (2009). Banking crises. In
beyond Basel III: necessary solutions for capital and Berger, A., Molyneux, P., and Wilson, J., editors, The
liquidity. OECD Journal: Financial Market Trends, Vol- Oxford Handbook of Banking. Oxford University Press.
ume 2010, Issue 1. Committee on the Global Financial System (1999). A
Board of Governors of the Federal Reserve System review of financial market events in autumn 1998.
(2009). Revenge of the steamroller: ABCP as a win- Technical report, Bank for International Settlements.
dow on risk choices. Unpublished paper, Division of Copeland, L. (2000). Exchange Rates and International
International Finance. Finance. Prentice Hall.
Borensztein, E. and Panizza, U. (2008). The costs of sov- Coval, J., Jurek, J. and Stafford, E. (2008a). The econom-
ereign default. Technical Report WP/08/238, IMF. ics of structured finance. Harvard Business School
Borio, C. (2009). Ten propositions about liquidity cri- Working Paper 09-060.
sis. Working paper no. 293, Bank of International Coval, J. D., Jurek, J. W. and Stafford, E. (2008b). Eco-
Settlements. nomic catastrophe bonds. Mimeo, Harvard University.
Borio, C. and Disyatat, P. (2011). Global imbalances Crockett, A. (2000). Marrying the micro- and macro-
and the financial crisis: link or no link? Technical re- prudential dimensions of financial stability. The
port, BIS. Working Paper 346. General Manager of the Bank for International Settle-
Borio, C., Drehmann, M., Gambacorta, L., Jimenez, G. ments, www.bis.org/review/rr000921b.pdf.
and Trucharte, C. (2010). Countercyclical capital Crouhy, M., Galai, D. and Mark, R. (2000). Risk Manage-
buffers: exploring options. Technical report, BIS. ment. McGraw-Hill.
Working Paper 317. Danielsson, J. (2011). Financial Risk Forecasting. John
Brady Commission (1988). Report of the presidential Wiley & Sons.
task force on market mechanisms. Technical report, Danielsson, J. and Keating, C. (2009). Bonus incensed.
Government Printing Office, Washington, DC. Vox EU, 25 May. http://voxeu.org/index.php?q=node/
Brown, W. A. (1940). The international gold standard 3602.
reinterpreted, 1914–1934. Technical report, National Danielsson, J. and Oskarsson, H. (2012). The first sov-
Bureau of Economic Research. ereign debt crisis in the EU. Vox EU, 11 September.
Bruner, R. F. and Carr, S. D. (2007). The Panic of 1907: www.voxeu.org/article/europe-s-pre-eurozone-debt-
Lessons Learned from the Market’s Perfect Storm. John crisis-faroe-islands-1990s.
Wiley & Sons. Danielsson, J. and Shin, H. S. (2003). Endogenous risk.
Brunnermeier, M. (2008). Deciphering the 2007–08 liquid- In Modern Risk Management – a History. Risk Books,
ity and credit crunch. J. Econ. Perspect., 23(1): 77–100. London. www.riskresearch.org.
394
Bibliography
Danielsson, J., Embrechts, P., Goodhart, C., Keating, C., Enria, A. (2011). The future of EU regulation. www.
Muennich, F., Renault, O. and Shin, H. S. (2001). An eba.europa.eu/cebs/media/aboutus/Speeches/The-
academic response to Basel II. www.bis.org/bcbs/ca/ Future-of-EURegulation–British-Bankers-Association–
fmg.pdf. 29-June-2011.pdf.
Danielsson, J., Shin, H. and Zigrand, J.-P. (2012a). En- Epstein, G. and Ferguson, T. (1984). Monetary policy,
dogenous and systemic risk. In Haubrich, J. G. and Lo, loan liquidation, and industrial conflict: the Federal
A. W., editors, Quantifying Systemic Risk. University of Reserve and open market operations of 1932. J. Econ.
Chicago Press for NBER; Download from www.risk His., 44: 957–83.
research.org. European Union (2011). EU budget 2010 financial re-
Danielsson, J., Shin, H. and Zigrand, J.-P. (2012b). En- port. Technical report.
dogenous extreme events and the dual role of prices. Fabozzi, F. J. (2009). Bond Markets, Analysis, and Strate-
Ann. Rev. Econ., 4(1): 111–129; download from www. gies, 7th edition. Pearson.
riskresearch.org. Ferguson, N. (2008). The Ascent of Money. The Penguin
Danielsson, J., Shin, H. and Zigrand, J.-P. (2012c). Pro- Group.
cyclical leverage and endogenous risk. Mimeo, LSE, Financial Services Authority (2009a). Reforming re
www.riskresearch.org. muneration practices in financial services. Technical
Darling, A. (2011). Back from the Brink: 1,000 Days at report.
Number 11. Atlantic Books, London. Financial Services Authority (2009b). The Turner re-
Davis, E. P. (1999). A reappraisal of market liquidity view: a regulatory response to the global banking cri-
risk in the light of the Russian/LTCM global securities sis. Technical report, European Central Bank. www.
markets crisis. Bank of England. fsa.gov.uk/pubs/other/turner review.pdf.
de Ramon, S., Iscenko, Z., Osborne, M., Straughan, M. Financial Services Authority (2012). Final notice. Techni-
and Andrews, P. (2012). Measuring the impact of cal report. www.fsa.gov.uk/static/pubs/final/barclays-
prudential policy on the macroeconomy. A practical jun12.pdf.
application to Basel III and other responses to the fi- Financial Stability Board (2012) Update of global
nancial crisis. Technical report, FSA. Occasional Paper systemically important banks (G-SIBS) Annex 1, Bank
Series 42. for International Settlements.
de Soto, J. H. (2009). Money, Bank Credit, and Economic Financial Stability Forum (2008). Report of the financial
Cycles. Ludwig von Mises Institute, Auburn, AL. stability forum on enhancing market and institution-
Dewatripont, M. and Tirole, J. (1994). The Prudential al resilience. Technical report.
Regulation of Banks. MIT Press. Financial Times (2007). Goldman pays the price of be-
Diamond, D. and Dybvig, P. (1983). Bank runs, deposit ing big. Financial Times, 13 August.
insurance, and liquidity. J. Polit. Econ., 91: 401–19. Financial Times (2008). Winding up Lehman Brothers.
Duffie, D. (2007). Innovations in credit risk transfer: Financial Times, 7 November.
implications for financial stability. Mimeo, Stanford Financial Times (2010a). The benefits of naked CDS. Fi-
University. nancial Times, 2 March.
Duffie, D. (2010). Is there a case for banning short spec- Financial Times (2010b). HSBC in clearest warning over
ulation in sovereign bond markets? Financial Stability relocation. Financial Times, 3 September.
Review, 14, Banque de France. Financial Times (2012a). DoJ probes $2bn JPMorgan
Duffie, D. and Singleton, K. (2003). Credit Risk. Prince- trading loss. Financial Times, 15 May.
ton University Press. Financial Times (2012b). JPMorgan whale harpooned.
Duffie, D. and Zhu, H. (2011). Does a central clearing Financial Times, 11 May.
counterparty reduce counterparty risk? Stanford Uni- Flood, R. P. and Garber, P. M. (1984). Collapsing ex-
versity Working Paper. change rate regimes: some linear examples. J. Int.
Eichengreen, B. (1996). Golden Fetters: the Gold Stand- Econ., 17: 223–34.
ard and the Great Depression, 1919–1939. Oxford Uni- Frankel, J. A. (1999). No single currency regime is right
versity Press. for all countries at all times. Working paper 7338, Na-
Eichengreen, B. and Irwin, D. (2009). The slide to pro- tional Bureau of Economic Research.
tectionism in the Great Depression: who succumbed Friedman, M. (1969). The optimum quantity of money.
and why? NBER Working Paper 15142. In Friedman, M., editor, The Optimum Quantity of
Elliot, G. (2006). Overend & Gurney, a Financial Scandal Money and Other Essays, chapter 1, pp. 1–50. Adline
in Victorian London. Methuen. Publishing Company, Chicago.
395
Bibliography
Friedman, M. and Schwartz, A. (1963). A Monetary His- Hull, J. C. (2011). Options, Futures, and Other Derivatives,
tory of the United States: 1867–1960. Princeton Univer- 8th edition. Prentice Hall.
sity Press. Hull, J. C. (2012). Risk Management and Financial Insti-
Goldstein, M. (1998). The Asian Financial Crisis: Causes, tutions, 3rd edition. Prentice Hall.
Cures, and Systemic Implications. Institute for Interna- Hume, D. (1752). On the balance of trade. In Essays,
tional Economics, Washington, DC. Moral, Political and Literary. Edinburgh and London.
Goodhart, C. (1999). Myths about the lender of last re- Institute of International Finance (2010). Interim report
sort. Technical report FMG SP 120, London School of on the cumulative impact of proposed changes in the
Economics. http://fmg.lse.ac.uk/pdfs/sp120.pdf. banking regulatory framework. Technical report.
Goodhart, C. (2002). The organizational structure of Institute of International Finance (2012). Making reso-
banking supervision. Economic Notes, 31(1): 1–32. lution robust. Technical report.
Goodhart, C. (2009). The Regulatory Response to the Fi- International Monetary Fund (1997). Camdessus an-
nancial Crisis. Edward Elgar, Cheltenham, UK. nounces IMF support for Indonesia’s economic pro-
Goodhart, C. (2011). The Basel Committee on Banking gram. News brief no. 97/19, 8 October 1997, www.
Supervision: a History of the Early Years 1974–1997. imf.org/external/np/sec/nb/1997/nb9719.htm.
Cambridge University Press. International Monetary Fund (1998). World economic
Goodhart, C. and Schoenmaker, D. (1995). Should the outlook and international capital markets: interim as-
functions of monetary policy and banking supervi- sessment. Financial turbulence and the world econo-
sion be separated? Oxford Economic Papers, 47(4): my. Technical report. www.imf.org/external/pubs/ft/
539–60. weo/weo1298/index.htm.
Gorter, J. and Bijlsma, M. (2012). Strategic moves. International Monetary Fund (2008). Global financial
www.risk.net/life-and-pension- risk. stability report: containing systemic risks and restor-
Gosh, A. R., Gulde, A.-M. and Wolf, H. C. (2002). Exchange ing financial soundness. April 2008, pp. 1–211.
Rate Regime: Choices and Consequences. MIT Press. International Monetary Fund, Bank for International Set-
Graeber, D. (2011). Debt: The First 5,000 Years. Melville tlements, Financial Stability Board (2009). Report to
House Publishing, New York. G20 finance ministers and governors. Guidance to as-
Graham, B. and Dodd, D. (1934). Security Analysis. sess the systemic importance of financial institutions,
Whittlesey House (McGraw-Hill). markets and instruments: Initial considerations. Tech-
Greenspan, A. (1997). Discussion at symposium: Main- nical report.
taining Financial Stability in a Global Economy, p. 54. Ito, T. (2007). Asian currency crisis and the Internation-
Federal Reserve Bank of Kansas City. al Monetary Fund, 10 years later: Overview. Asian
Gulati, M. and Triantis, G. (2007). Contracts without Econ. Policy Rev., 2: 16–49.
law: Sovereign versus corporate debt. University of Kaminsky, G. L. and Reinhart, C. M. (1999). The twin
Cincinnati Law Review, 75: 977–1004. crises: the causes of banking and balance-of-payments
Hale, D. (2003). The Newfoundland lesson. The Interna- problems. Amer. Econ. Rev., 89: 473–500.
tional Economy, 17(3): 52–61. Kane, E. J. (1989). The S&L insurance mess: how did it
Hanke, S. H. and Kwok, A. K. F. (2009). On the meas- happen? Technical report, Urban Institute Press.
urement of Zimbabwe’s hyperinflation. Cato J., 29(2): Kaseki, J.-R. (2007). Preying on the poor. The Guardian,
353–64. 27 September.
Hansard (1933). http://hansard.millbanksystems.com/ Kashyap, A. K., Rajan, R. G. and Stein, J. C. (2008). Re-
commons/1933/dec/12/newfoundland-bill#S5CV thinking capital regulation. http://online.wsj.com/
0284PO_19331212_HOC_267. public/resources/documents/Fed-JacksonHole.pdf.
Herring, R. J. and Litan, R. E. (1995). Financial Regula- Kaufman, G. G. (1996). Bank failures, systemic risk, and
tion in the Global Economy. The Brookings Institution, bank regulation. Cato J., 16(1): 17–46.
Washington, DC. Keynes, J. M. (1920). The Economic Consequences of the
Homer, S. and Sylla, R. (1996). A History of Interest Rates. Peace. Harcourt Brace, New York.
Rutgers University Press. Keynes, J. M. (1924). A Tract on Monetary Reform.
Honohan, P. and Klingebiel, D. (2003). The fiscal cost Macmillan.
implications of an accommodating approach to Keynes, J. M. (1936). The General Theory of Interest,
banking crises. J. Banking Finance, 27(8): 1539–60. Employment and Money. Macmillan.
Huertas, T. F. (2010). Crisis: Cause, Containment and Kindleberger, C. (1986). The World in Depression, 1929–
Cure. Palgrave Macmillan. 1939, 2nd edition. University of California Press.
396
Bibliography
Kindleberger, C. P. (1996). Manias, Panics, and Crashes: Macrae, R. and Watkins, C. (1998). A disaster waiting to
a History of Financial Crises, 3rd edition. John Wiley & happen. Mimeo, www.cs.rhul.ac.uk/home/chrisw/
Sons. Disaster.pdf.
Kohn, M. (1999). Early deposit banking. Mimeo, Dart- Madura, J. (2010). Financial Institutions and Markets,
mouth College, Hanover, NH. 9th edition. South Western College, Cincinnati, OH.
Kohn, M. (2003). Financial Institutions and Markets, 2nd Marsh, D. (2010). The Euro: the Politics of the New Global
edition. Oxford University Press. Currency. Yale University Press.
Kovacevich, R. M. (1996). Deposit insurance: It’s time McCallum, B. T. (1996). International Monetary Eco-
to cage the monster. www.minneapolisfed.org/ nomics. Oxford University Press.
publications papers/pub display.cfm?id=2682. Merton, R. (1995). A functional perspective of financial
Krueger, A. O. (2002). A new approach to sovereign intermediation. Financial Management, 24: 23–41.
debt restructuring. Technical report, IMF. Merton, R. C. (1974). On the pricing of corporate
Krugman, P. (1979). A model of balance-of-payments debt: the risk structure of interest rates. J. Finance, 29:
crises. J. Money Credit Banking, 11(3): 311–25. 449–70.
Krugman, P. (1998). What happened to Asia? Mimeo, Miller, G. P. (1996). Is deposit insurance inevitable? Les-
MIT. sons from Argentina. Int. Rev. Law Econ., 16: 211–32.
Krugman, P. R. and Obstfeld, M. (2006). International Minsky, H. (1992). The financial instability hypothesis.
Economics: Theory and Policy, 7th edition. Pearson. Working paper, Mimeo, Yale University.
Kupiec, P. H. and Ramirez, C. D. (2009). Bank failures Mishkin, F. S. and Eakins, S. (2011). Financial Markets
and the cost of systemic risk: Evidence from 1900– and Institutions, 6th edition. Pearson.
1930. Technical report, FDIC. Moe, T. G., Solheim, J. A., and Vale, B. (2004). The Nor-
Kyle, A. (1985). Continuous auctions and insider trad- wegian banking crisis. Technical report, Norges Bank.
ing. Econometrica, 53(6): 1315–36. Occasional Paper no. 33.
Laeven, L. and Valencia, F. (2008). Systemic banking Morris, S. and Shin, H. S. (1998). Unique equilibrium in
crises: a new database. Technical report, IMF. IMF a model of self-fulfilling currency attacks. Amer. Econ.
Working Paper. Rev., 88: 587–97.
Larosiere, J. (2009). The high-level group report on Morris, S. and Shin, H. S. (1999). Risk management with
financial supervision in the EU. Technical report. interdependent choice. Oxford Review of Economic
http://ec.europa.eu/commission barroso/president/ Policy, 15: 52–62, reprinted in Bank of England Finan-
pdf/statement 20090225 en.pdf. cial Stability Review, 7, 141–150. www.bankofengland.
Lavigne, R. (2008). Sterilized intervention in emerging- co.uk/fsr/fsr07art5.pdf.
market economies: trends, costs and risks. Discussion Moysich, A. (2000). The savings and loan crisis and its
paper 2008-4, Bank of Canada. relationship to banking. In Federal Deposit Insurance

Leeson, N. (1999). Rogue Trader. Sphere, London. Corporation, History of the 80s – Lessons for the Future.
Lehman Brothers (2008). Capital advisory group: cur- Volume I: an Examination of the Banking Crises of the 1980s
rent topics. Technical report, Lehman Brothers. and Early 1990s, Chapter 4. FDIC, Washington, DC.
Lewis, M. (2009). The man who crashed the world. Mundell, R. A. (1961). A theory of optimum currency
Vanity Fair, August. areas. Amer. Econ. Rev., 51: 657–65.
Lewis, M. (2011). The Big Short: Inside the Doomsday Murphy, D. (2009). Unravelling the Credit Crunch. CRC
Machine. Penguin. Press.
Lewis, M. K. and Mizen, P. D. (2000). Monetary Econom- Mussa, M. (2002). Argentina and the Fund: From Tri-
ics. Oxford University Press. umph to Tragedy. Institute for International Econom-
Li, D. X. (2000). On default correlation: a copula func- ics, Washington, DC.
tion approach. J. Fixed Income, 9: 43–54. Obstfeld, M. (1996). Models of currency crises with self-
Liikanen, E. (2012). High-level expert group on reform- fulfilling features. Eur. Econ. Rev., 40(3–5): 1037–47.
ing the structure of the EU banking sector. Technical Obstfeld, M. and Rogoff, K. (1996). Foundations of In-
report, European Union. ternational Macroeconomics. MIT Press.
Lowenstein, R. (2000). When Genius Failed – the Rise OECD (2002). Forty years’ experience with the OECD
and Fall of Long-Term Capital Management. Random code of liberalisation of capital movements. Techni-
House, New York. cal report.
397
Bibliography
Quagliariello, M. and Cannata, F. (2011). Basel III and Sullivan, R., Timmermann, A. and White, H. (1999).
Beyond. Riskbooks, London. Dangers of data mining: the case of calendar effects
Radelet, S. and Sachs, J. (2000). The onset of the East in stock returns. J. Econometrics, 105(1): 249–86.
Asian financial crisis. In Krugman, P., editor, Currency Taylor, J. B. (1993). Discretion versus policy rules in
Crises. University of Chicago Press. www.nber.org/ practice. Carnegie-Rochester Conference Series on Pub-
chapters/c8691. lic Policy, 39: 195–214.
Reinhart, C. M. and Rogoff, K. (2009). This Time Is Dif- The Economist (1999). South Africa’s debt: unforgivable.
ferent: Eight Centuries of Financial Folly. Princeton Uni- The Economist (US), 24 April.
versity Press. The Economist (2009). No empty threat. Credit-default
Reinhart, C. M., Rogoff, K. and Savastano, M. A. (2003). swaps are pitting firms against their own creditors.
Debt intolerance. Technical report, NBER. The Economist (US), 18 June.
Repullo, R. and Saurina, J. (2011). The countercyclical Tobin, J. (1996). A currency transactions tax, why and
capital buffer of Basel III: a critical assessment. CEMFI how? Open Economies Review, 7: 493–9.
Working Paper 1102, June. Truell, P. and Gurwin, L. (1992). False Profits: the Inside
Santos, J. A. C. (2000). Bank capital regulation in con- Story of BCCI, the World’s Most Corrupt Financial Em-
temporary banking theory: A review of the literature. pire. Houghton Mifflin, Boston, MA.
Technical Report 90, Bank for International Settle- UBS (2008). Shareholder report on UBS’s write-downs.
ments. www.bis.org/publ/work90.pdf?noframes¯1. Technical report.
Sarno, L. and Taylor, M. P. (2003). The Economics of Ex- Volcker, P. (2011). Financial reform: Unfinished busi-
change Rates. Cambridge University Press. ness. www.nybooks.com/articles/archives/2011/nov/
Schumpeter, J. (1942). Capitalism, Socialism and Democ- 24/financial-reformunfinished-business.
racy. Harper, New York. Wall Street Journal (2007). One ‘quant’ sees shakeout for
SEC and CFTC (2010). Findings regarding the market the ages – ‘10,000 years’. Wall Street Journal, 11 August.
events of May 6, 2010. Technical report. Wall Street Journal (2009a). Action on AIG unit may cost
Selgin, G. (2003). Adaptive learning and the transition taxpayers. Wall Street Journal, 13 April. online.wsj.
to fiat money. Econ. J., 113(484): 147–65. com/article/SB123957925343711995.html.
Shin, H. S. (2008). Reflections on modern bank runs: A case Wall Street Journal (2009b). Paul Volcker: think more
study of Northern Rock. Mimeo, Princeton University. boldly. Wall Street Journal, 14 December. online.wsj.
Shin, H. S. (2010). Risk and Liquidity. Clarendon Lectures com/ article/SB100014240527487048255045745863
in Finance. Oxford University Press. 30960597134.html.
Sicsic, P. (1992). Was the Franc Poincaré deliberately un- Washington Post (2004). Credit raters’ power leads to
dervalued? Explorations in Economic History, 29: 69–92. abuses, some borrowers say. Washington Post, 24
Sorkin, A. R. (2010). Too Big to Fail: Inside the Battle to November. www.washingtonpost.com/wp-dyn/arti-
Save Wall Street. Penguin. cles/A8032-2004Nov23.html.
Spanish government (2008). Las balanzas fiscales de Weisbrot, M. (2007). Ten years after: the lasting impact
las CC.AA. españolas con las AA. Públicas Centrales of the Asian financial crisis. Technical report. http://
2005. Technical report, Ministerio de Economía y www.cepr.net/documents/publications/asia_crisis_
Hacienda. 2007_08.pdf.
Stiglitz, J. (2009). Obama’s ersatz capitalism. www. Zigrand, J.-P. (2010). What do network theory and en-
nytimes.com/2009/04/01/opinion/01stiglitz.html? dogenous risk theory have to say about the effects of
r=1&scp=2&sq=stiglitz&st=cse. CCPs on systemic stability? Financial Stability Review,
Stoler, P. and Kalb, J. B. (1982). The great Vatican 14, Banque de France.
bank mystery. http://time.com/time/magazine/arti-
cle/0,9171,951806,00.html.
Sturzenegger, F. and Zettelmeyer, J. (2007). Debt
Defaults and Lessons from a Decade of Crises. MIT
Press.
398
Index
1914 crisis 2–4 agents, deposit insurance model 142–7

1987 crash and endogenous risk 52–3 agricultural depression 23
1997 Asian crisis see Asian crisis 1997 agricultural producers 205
1998 liquidity crisis 62–5 agricultural sector 20
2003–2006 mindset 330 AIB (Allied Irish Bank) 321
2007–2009 crises 325–6 AIG 9, 273, 277, 279, 289, 291, 335
Asian crisis 1997 relation to 116–17 Akerlof, G. 190n
bailouts 273–8 Alessandri, P. 249
bank capital 317–23 Allen, F. 71–2
build-up to crisis 326–30 Allied Irish Bank (AIB) 321
changing nature of banking 331–3 Allies 3, 21
collateralised debt obligations correlations 296–7 ambiguity, constructive 73, 234
crisis, 2007-2008: 333–7 Ambrosiano, Banco 129, 251
European banks in, case study 319–23 amplification mechanisms 13
Great Depression parallels with 35–7 Amsterdamsche Wisselbank 121, 123
haircuts 186 Angelini, P.L. 358
hidden and ignored risk 330–1 anti-globalism 328
liquidity policy implications 74–5 arbitrage
policy response 338 low-risk 154
regulatory failures see regulatory failures regulatory 251
subprime mortgage role 337–8 statistical 159, 333
1G currency crisis models 219–22, 224, 230 Argentina
2G currency crisis models see currency crises: models banking crises 126
capital controls 224
ABCP (asset-backed commercial paper) 141, 299–30 currency board 222
Abdelal, R. 201 currency crisis 222–4
ABS (asset backed securities) 140, 189, 299 deposit insurance 1991–1994: 148–9
accelerator effects, subprime mortgages 337 fiscal policies 222–4
Accenture 160–1 market oriented structural reforms 222
acceptances 3–4 monetary policies 223–4
accounting sovereign debt 373, 377
double entry 122 Aristotle 152
mark-to-market see marking-to-market Arnason, R. 201
accredited investors 155 Asian crisis 1997: 62, 65, 67, 73, 97–8
Acharya, V. 73 2007–2009 crises, relation to 116–17
action, prices as imperatives to 43 bailouts in crises starting in 2007 analogies 277
active prevention, financial stability 87 building up to 99–101
active risk management 54–5 capital markets 107
activity restrictions, banking regulations 241 corruption 106, 116
actual risk 53–5, 315 financial markets 106–7
adjustable pegs 206, 214 GDP 100–1, 102, 107–8, 116
Admati, A.R. 358 IMF role 98, 110, 112
Adrian, T. 14 heavy-handedness 113
advanced measurement approach, Basel II 254 as lender of last resort 113–15
adverse selection 269 moral hazard focus 112
agency problem 15 structural adjustments focus 112–13, 116
399
INDEX
Asian crisis 1997 (continued) Austria

in individual countries 102–5 bank runs 137
lending to East Asia 99–100 euro zone and 382–5
macroeconomics 100 Great Depression 27
money and investments 100–1 autarky, deposit insurance model 143–4, 146
overconfidence 107
policy options for crisis countries 109 bad bank – good bank see good bank – bad bank
fighting 109–10 bad banking 127
foreign reserves 110–11, 116 Bagehot, Walter 266–7, 271
IMF loans 110 bail-ins 87, 272
long-term policy responses 110–11 bailouts 87, 261–3
reasons for 106 AIG 335
liquidity crisis – sudden stop 108–9 alternatives to 272–3
market factors 106–7 asset bubbles model 279–83
moral hazard 106 challenges 279
panic and contagion 106 credit facilities 273–5
performance before and after crisis 107–8 in crises starting in 2007: 273
weak fundamentals 106, 107 analysis 277
sovereign debt 100 Asian crisis analogies 277
wider lessons 115–17 European Central Bank 276, 277–8
asks 68 United Kingdom 275–6
asset-backed commercial paper (ABCP) 141, 299–301 United States 273–5
asset backed securities (ABS) 140, 189, 299 direct 267
assets equity injections 267–8, 276
banks see banks loan guarantees 269
bubbles loans 268
government guarantees causing 279–83 preference shares 268
liquidity injections stimulating 329 preferences of parties 269
model 279–83 problem assets, buying or guaranteeing 269
pro-cyclicality 330–1 euro zone 384, 386–7
cross-held 138 European Central Bank 88–9
fixed income 175, 177 governments by central banks 87–8
garbage 287, 298 lending of last resort see lending of last resort
growth 14–15 liquidity provision 268, 269–70, 273–5, 276, 277–8
lending of last resort choice 271 moral hazard 261, 266, 269, 270, 277, 278–9,
mean reverting 59 280–2
pricing, liquidity and 71–3 nature 267–72
printing money as tax on 88 politics 278–9
problem 269 successful 263–4
risk-free 47 unsuccessful 264–5
risk-weighted 256–7, 320–2 Banco Ambrosiano 129, 251
subprime 287 the Bank see Bank of England
total 320–2 Bank for International Settlements (BIS) 5, 288, 358
toxic, banks 317, 318 Bank Herstatt 129, 251
underlying, derivatives 288 bank holidays 140
unencumbered, banks 353 Bank of America 275
asymmetric copulae 297 Bank of Credit and Commerce International (BCCI)
asymmetries, information 11, 15, 164–5, 316 80, 129–30
Atkinson, P. 321 Bank of England (BoE) 4, 22, 28
Attlee, Clement 368 2007-2009 phase and 338
Australia banking supervision 81
bank capital 356 base rate 95
400
INDEX
benchmark interest rates 95 day-to-day risk 241

financial stability and 81, 338 endogenous risk 241–2
government bond holdings 87–8 governments, transfer of responsibilities to 243
lending of last resort 266–7 incentives of supervisors 242
macro-prudential regulation by 241 international 250
objectives 79 Basel see Basel Accords
origins 79 home regulators 251
Overend and Gurney crisis 265–6 host regulators 251
quantitative easing 86 regulatory arbitrage 251
reform 345 laissez-faire 240, 356
reputation risk 80 macro- and micro-prudential 240–1
bank rates 4 perverse consequences 243–4
bank runs 12, 15 principle-based 243
causes 137 pro-cyclicality 242, 247–9
circumstances of 120 reasons for 239–40
deposit insurance model 145, 146 regulatory capture 244
Great Depression 137 resource problems 245
United States 28, 139–40 smoothing the road 241–2
institutional investors 332 tick-the-box and legal approaches 242–3
market liquidity and 67 banking systems 6–7
Northern Rock 140–2 complexity 317
retail depositors 332 endogenous risk 279
as self-fulfilling prophecies 137, 145 European Union see European Union
Bankhaus Herstatt 129, 251 failures
banking good bank – bad bank resolution method 132
bad 127 Iceland 2008: 132–4
changing nature 331–3 Scandinavia 1990s 131–2
crises see banking crises US savings and loans 1980s 130–1
early 121–3 fragilities 12
European Union 347 privatisation, Iceland 133
fractional reserve see fractional reserve banking banknotes 78–9, 121
full reserve 123 bankruptcy 66, 292, 359, 361
money 121 banks
narrow 168–9 assets
originate and distribute model 188 high quality liquid 353
regulations see banking regulations sovereign bonds as liquid assets 354
shadow 10, 188, 243, 332–3 unencumbered 353
supervision 80–1 bailouts see bailouts
systems see banking systems bonuses 166–8
universal 168–9 capital
banking crises 119–21 adequacy ratios see capital adequacy ratios
bank and banking system failures see banking Basel III 348–52, 357
systems; banks calculations, Basel I and II 257
causes 126–8 capital instruments 247
costs 124–6 case study: European banks in 2007–2009 crises
in G10 economies 127 319–23
moral hazard 119, 123–4 cliff effects 249–50, 252–3
resolution 126, 132 concept of 246
banking regulations 35, 124, 133, 239 constraints 248
activity restrictions 241 criteria for 247
challenges 241–5 economic costs 358–9
crisis resolutions 241 equity 245–6
401
INDEX
banks (continued) reserves, minimum 87

European Union 357–8 risk weights 318
fragility 317, 318–19 runs see bank runs
GDP and 358–9 self-preservation 315
injections 267–8 size, politics and 8
levels 356–9 small, well-managed 16–17
maximum harmonisation principle, EU 357–8 toxic assets 317, 318
minimum 87 trading 155
national positions 356–7 see also central banks; commercial banks;
pro-cyclicality 247–9 investment banks
purposes 245 Barclays Bank 153, 275
regulations 247–50 Barings Bank 154
risk sensitivity 248–9 Barnier, Michel 293
structure optimisation 243, 317–18, 319 base rate, Bank of England 95
undermining 318–19 Basel Accords 1, 65, 241, 311
use 250 Basel I 252
complexity, incentives to maximise 317 capital adequacy ratios 256–7
conduits as 300 capital calculations 257
deleveraging 250 conduits and 300
deposit insurance 353 credit risk 252
equity 245–6, 319 issues with 252–3
Basel III 349–50 sovereign debt 255
failures 28 tier 1 capital 256
bankruptcy laws and 359 tier 2 capital 256
externalities from 240 value at risk (VaR) 253, 257–8
Great Depression 139 Basel II 253
idiosyncratic shocks 4 advanced measurement approach 254
individual banks 128–30 basic indicator approach 254
regulations preventing 120 capital adequacy ratios 256–7, 317–22
resolution 359–60 capital calculations 257
special resolution regimes 87 credit risk 254
financial innovations 243–4 criticisms 255–6
fragility 120 internal rating based (IRB) approach 253–4
funding 331–2 market discipline (Pillar III) 254
illiquidity 66, 67 market risk 254
insolvency 66 menu of approaches 253–4
large, badly run 16–17 minimum capital requirements (Pillar I) 254
large market participants 45–6 operational risk 254
lending, limits to single counterparties 86 Pillars of 254–5
leverage 13 shadow banking and 333
maximum 87 sovereign debt 255–6
liquidity, amounts 87 supervisory review process (Pillar II) 254
mergers 363 tier 1 capital 256
overnight funding 331–2 tier 2 capital 256
price takes 44–5 Basel III 74, 252,
products, sophistication 239–40 capital 348–52, 357
proprietary trading 155, 168 common equity tier 1 (CET1) 349
recapitalisation 119 costs of doing banking 358
regulations see banking regulations conservation buffers 349–50
reputation risk 319 countercyclical buffers 349, 350
reserve accounts 83 equity 349–50
402
INDEX
leverage ratio 352 Brazil

liquidity 352–6 capital controls 202
minimum capital adequacy ratios 349 Bretton Woods system 1, 35, 87, 121, 205–7, 311, 326,
tier 1 capital 349 327
tier 2 capital 349 broker–dealers 14–15
regulatory reform 343 brokers
Basel Committee for Banking Supervision (BCBS) 127, definition 172
128, 129, 131, 251–2, 342, 347 prime 56, 65
capital accords see Basel Accords Bruner, R.F. 250
Concordat 129, 252, 311 Brunnermeier, M. 70, 331
basic indicator approach, Basel II 254 bubbles 39, 53–5, 133, 278
basket pegs 214 assets see assets
Bassani, G. 359 Buchheit, L.C. 376
Baudino, P. 343 Buffett, Warren 47, 156, 288
BCBS see Basel Committee for Banking Supervision Bussiere, M. 208
BCCI (Bank of Credit and Commerce International) butterfly effects 13
80, 129–30 buying problem assets, bailouts 269
Bear Sterns 9, 164, 277, 289, 333
beggar-thy-neighbour policies 23, 209 CAC (collective action clauses) 376
Belgium Cagan, P. 92
euro zone and 382–5 Calello, P. 272
benchmarks 155 call margins 185
Benediktsdottir, S. 132 Calvi, Roberto 129
Bernanke, Ben 329, 333, 338 Camdessus, Michel 113, 114
bids 68 Canada
Bijlsma, M. 343 bank capital 356, 357
bills of exchange 122 Newfoundland confederation with 368–9
Birmingham, USA 239 capital
BIS (Bank for International Settlements) 5, 288, 358 banks see banks
Bitner, R. 190 controls see capital controls
in black 185 capital adequacy ratios (CAR) 66, 246
Black, Fisher 10–11, 69, 183 Basel I 256–7
Black Thursday 26 Basel II 256–7, 317–22
Black Wednesday 229–30 calculations 320–3
Black–Scholes–Merton option pricing model 183 leverage and 331
Black–Scholes option pricing model 49–52 minimum, Basel III 349
Bloomberg 274, 275, 278 regulations 247–50
Blundell-Wignall, A. 321 capital asset pricing model (CAPM) 47
BoE see Bank of England liquidity adjusted 73
bonds capital controls 200
covered 188 2.0: 202–3
long-term, banks 247 Argentina 224
markets 175, 176–8 Bretton Woods system 206
plain vanilla 176 developed countries moving away from 200–1
pricing 178 in developing countries 201
securitisation 189 hot money and 202, 203
bonuses 166–8, 279 impossible trinity 207
Borensztein, E. 372 pros and cons 203
Borio, C. 140, 147, 329, 350 traditional or strict (1.0) 200–2
borrowers , relationships with lenders 75 capital instruments
Bowie Bonds 188 bank capital 247
403
INDEX
capital markets definition 172

Asian crisis 1997: 107 financial stability and 164–6
Great Depression 26 Central Powers 3, 20, 21
CAPM see capital asset pricing model CET1 (common equity tier 1), Basel III 349
CAR see capital adequacy ratios Chaebol 105
Carr, S.D. 250 Chicago Board of Trade 152
carry trades 157–9, 202 Chile
Carville, James 175 capital controls 202
Case–Shiller index 55 China
cash capital controls 201, 212
flows 187 exporting deflation 328
opportunity costs of holding 72 hyperinflation 93
cash-in-the-market pricing 71–3 renminbi as possible reserve currency 212
Castello, Francesch 167, 239 sovereign debt 377
catastrophe bonds 297 United States and, currency wars 210
CBT (computer-based trading) 160 Churchill, Winston S. 22
CCP see central counterparties Citigroup 68–9, 274–5
CDO see collateralised debt obligations City of London 2–3, 7
CDS see credit default swaps clearing 171–2
central banks 77–8 clearing houses 166, 172, 290, 291
bailing out 87 cliff effects, bank capital 249–50, 252–3
European Central Bank 88–9 CMG (crisis management groups) 360
governments 87–8 co-insurance 140
banking supervision 80–1 CoCo (contingent convertibles) 351–2
challenges for central banking 90 collateral liabilities 75
conflicts between financial stability and monetary collateralised debt obligations (CDO) 16
policy 90–1 2007–2009 crises correlations 296–7
hyperinflation 92–3 calculations 303–7
independence 91 as catastrophe bonds 297
losing control of money 91–3 correlated defaults 295–6, 305–7
exchange rate interventions 197–9 credit default swaps and, interlocking exposures 316
financial stability and 79, 80–1, 86–7, 90–1, 344–5 credit ratings 293
first generation currency crisis model 219–22 examples 294–5, 303–5
independence 90, 91, 328 fractional reserve banking and 298–7
interest rates 82–3, 91, 94–5 idiosyncratic risk and 298–9
lending of last resort 270–2 issues with 296–9
liquidity providers 73–4, 316 nature 293
monetary policy 80–1, 82 optimisation 297–8
conflicts between financial stability and 90–1 rating agencies and 297–8, 299
interest rates 82–3 squared 298
open-market operations 82, 83, 85 synthetic 298
quantitative easing 82, 84–6 systemic risk and 298–9
reserve requirements 82, 84 toxic assets 318
monetary stability 79, 80 tranches 293–6, 297–8, 305
origins 78–9 collective action clauses (CAC) 376
regulatory reform 344–5 colleges of supervisors 347–8
reputation risk 80 commercial banks 168–9
South Korea 105 asset growth and leverage growth 14–15
Thailand 102–3 common equity tier 1 (CET1), Basel III 349
central counterparties (CCP) 86 competition, systemic risk and 10
credit default swaps and 290 competitive devaluations 23–4, 209
404
INDEX
complex products 287 Coval, J. 297

complexity coverage ratio, liquidity see liquidity
banking systems 317 covered bonds 188
banks’ incentives to maximise 317 CRA see credit rating agencies
financial instruments 287 crawling pegs 214
pro-cyclicality, adding to 331 Credit-Anstalt 27
regulatory failures 317 credit crunches 247
supervisory approach to 317 credit default swaps (CDS) 16, 134,
computer-based trading (CBT) 160 153, 289
conduits 299–301, 332, 337 AIG and 335
confidence 1, 15, 327 benefits 290
conflicts of interest, credit rating agencies 181–2 collateralised debt obligations and, interlocking
Congo see Democratic Republic of Congo; exposures 316
Republic of Congo criticism 289
conservation buffers, Basel III 349–50 example 302–3
constructive ambiguity 73, 234 mechanics of 301–3
contagion moral hazard 292
Asian crisis 1997: 106 naked 292–3
carry trades as source of 158–9 nature 289
sovereign debt repayment reason 372 netting out 291
contingent convertibles (CoCo) 351–2 network vulnerabilities 290–2
contingent liabilities 75 payment flows 290
continuous linked settlement platforms 129 perverse incentives 292
contractionary monetary policy 82 premiums 302
contracts for difference 242 risk
convergence trades 56, 58 individual 290
convexity 184 systemic 291–2
cooperation, international 347–8 on sovereign debt 292–3
Copeland, L. 224, 230, 264 credit events, CDSs 289, 302–3
corporate bonds, rating agencies 299 credit facilities, bailouts 273–5
corporates credit markets 175–6
asset growth and leverage growth 14–15 credit models 176, 183–5
correct exchange rates 196 credit rating agencies 179–83
correlated defaults haircuts 185–6
collateralised debt obligations 295–6, 305–7 margins 185–6
correlations 48, 331, 333 market for credit 176–9
corruption 106, 116, 127–8, 133 marking-to-market 187
costs securitisation 188–90
banking crises 124–6 credit models 176, 183–5
of doing banking, Basel III 358 credit rating agencies (CRA) 175
of exchange rate changes 225 collateralised debt obligations and
liquidity coverage ratio 354 297–8, 299
opportunity 72 conflicts of interest 181–2
countercyclical buffers, Basel III 349, 350 critics 181
counterparties 151, 171 European sovereign debt crisis 182
central see central counterparties issues with 181–3
hedge funds 155 legal protection 183
single, bank lending limits to 86 mistakes by 299
counterparty credit risk 165, 185 oligopoly 182–3
counterparty risk 172 opinions on creditworthiness 179
credit default swaps 290 quality of ratings 182
405
INDEX
credit ratings 134, 175, 179–80 success and failure 380–1

collateralised debt obligations 293 transfer unions 380
maturities and 180–1 currency wars 209–10
quality of 182 Cyprus
ratings shopping 182 banking system 6–7
spreads and 180–1 euro zone and 382–5
yields and 180–1 sovereign debt crisis 387
credit risk 178–9
Basel I 252 danger, liquidity ratios 356
Basel II 254 dangerous financial instruments 285–307
counterparty 165, 185 Danielsson, J. 48, 70, 167, 201, 255, 258, 313, 377
default probabilities 179 Darling, Alistair 338
recovery 179 data mining 298
regulation 74 Davis, E.P. 62, 64
credit spreads 184 DAX 154
Credit Suisse 275, 321 day-to-day risk
creditors banking regulations 241
credit default swaps and 290–2 risk management failures targeting 311
creditworthiness 75, 179 targeting 311–13
crisis management groups (CMG) 360 de Ramon, S. 358
crisis resolutions dead weight losses 88, 243
banking regulations 241 debasement of money 121
Crockett, Andrew 9, 54, 240, 330 debt
cross-border loans 3, 4 evolution, European sovereign debt crisis
cross-held assets 138 background 382–4
currencies, reserve see reserve currencies sovereign see sovereign debt
currency boards 213, 222 subordinated see subordinated debt
currency crises 208–9 debt securities 62
models 217–18 decisions, deposit insurance model 145–6
Argentinian crisis 222–4 defaults
European crisis, 1992–1993: 228–31 correlated see correlated defaults
first-generation (1G) models 219–22, 224, 230 costs analysis, sovereign debt repayment reason 372
global games currency crisis model 231–4 immunity, sovereign debt repayment reason 372–3
second-generation (2G) models 224–8, 229, 230, probabilities 179, 184–5
234, 264, 374 sovereign debt see sovereign debt
currency markets 193–5 deficits, persistent 219
capital controls 200–3, 207 deflation 22, 24, 25, 27, 84–5, 204–5
crises see currency crises deleveraging 30, 62, 250
exchange rates see exchange rates delinquency rates 55
fixed or floating 195–7 delta and dynamic replication 49–52
interventions 196, 197–9 demand
overvaluation perils 208–9 downward-sloping 43
reserve currencies 210–12 functions 45–6
sterilisation 198–9 upward-sloping 43–6
undervaluation and ‘currency wars’ 209–10 demand deposits 12, 120
currency mismatches 107, 110 Democratic Republic of Congo
currency reserves 116, 329 sovereign debt enforcement 375
currency transactions Denmark
spot, Tobin tax 169 bank capital 356
currency unions 213 Faroe Islands sovereign debt crisis 377–9
Europe, history 379–8 dependence 48
406
INDEX
non-linear see non-linear dependence Dodd–Frank Act, US 166, 168

deposit insurance 35, 131, 138, 353 Dodd–Frank orderly liquidation authority 359
government schemes 146–7 dollar, US see United States
modelling 142–4 dollarisation 214–15
fractional reserve banking systems and domestic bond markets 177–7
equilibrium decisions 144–6 domestic sovereign debt 369–71
regulatory response 146–7 dominant strategies equilibrium, deposit insurance
Northern Rock 140–2 model 146
privately funded schemes 146–7 Dominican republic
pros and cons 147 banking crisis 127–8
Argentina 1991–1994: 148–9 sovereign debt enforcement 375
moral hazard 147 double entry accounting 122
wholesale markets impact 147–8 downward-sloping demand 43
United States 140 dual role of prices see prices
deposits, demand 12, 120 Duffie, D. 165, 293, 297
depressions see Great Depression Dutch disease 7
depth 69–70 Dybvig, P. 138, 142
deregulation 131–2 dynamic instability 206
banking crises causes 127 dynamic trading strategies 48–53
regulatory failures and 311
derivatives 16 early agents, deposit insurance model 142–7
meaning 288 early banking 121–3
markets 67 East Asian crisis see Asian crisis 1997
over-the-counter 67, 288 EBA (European Banking Authority) 346, 347
pricing 313 ECB see European Central Bank
subprime mortgages linked 337 econometrics 298
volume of trading 288–9 economic costs of capital 358–9
desired exchange rates 224–5 economic development
destabilising nature of stability 9–10 sovereign debt repayment reason 372–3
Deutsche Bank 275, 321 economy, costs on, banking crises 125–6
devaluations 28, 224, 225 ECU (European Currency Unit) 228, 381
competitive 23–4, 209 effective dates, CDSs 302
internal 22, 209 efficient market hypothesis (EMH) 329–30
developed countries Egypt
capital controls, moving away from 200–1 sovereign debt enforcement 375
developed markets Elgin, G. 121
sterilisation in 199 Elliot Capital Management 223, 376
developing countries Elliot, G. 265
capital controls 201 Emanuel, Rahm 342
Dexia 275 emergency measures in crises, TBTF institutions 363
Diamond, D. 138, 142 emerging markets, sterilisation in 199
Diamond–Dybvig deposit insurance model 142–7 EMH (efficient market hypothesis) 329–30
Dinallo, Eric 289 empirical predictions 54–5
direct bailouts see bailouts EMS (European Monetary System) 381
dirty floats 214 endogenous prices 45–6
disclosed reserves, Basel Accords 256 endogenous risk 39–41
discount markets 3 1987 crash and 52–3
discount rates 94 actual risk 53–5
Disyatat, P. 329 banking crises 119
diversification 331, 337, 362–3 banking regulations 241–2
Dodd, D. 151, 156 banking systems 279
407
INDEX
endogenous risk (continued) European sovereign debt crisis 74

bubbles 53–5 financial stability responsibilities 81
conclusion 59 government bond holdings 89
dual role of prices 43–6 lending of last resort 271–2
dynamic trading strategies 48–53 main refinancing operations fixed rate 94
feedback loops 63, 72, 185 marginal lending facility 94–5
harmonised behaviour 70 monetary policy 81
LTCM crisis of 1998: 56–9 operational independence 89
Millennium Bridge 41–3 ownership 89
perceived risk 53–5 reforms 346
pro-cyclicality 330–1 zombie banks and 128
risk 46–8 European Currency Unit (ECU) 228, 381
risk management failures and 311 European Monetary System (EMS) 381
enforcement of sovereign debt see European Securities and Markets Authority
sovereign debt (ESMA) 182, 346
engineering, financial 298 European sovereign bonds
Enria, Andrea 357 default probabilities 184–5
Epstein, G. 32 European sovereign debt 371
equilibria European sovereign debt crisis 35
decisions, deposit insurance background 377
modelling 144–6 debt evolution 382–4
fundamental 206 European currency arrangements since Bretton
multiple see multiple equilibria Woods 381–2
equity Faroese crisis 1992: 377–9
bank capital 245–6 individual countries 384–7
banks 319 causes 387–8
Basel III 349–50 central bank liquidity injections 73–4
conduits 299 common currency impact 37
injections, bailouts 267–8, 276 credit rating agencies and 182
markets 176 global impact 388–9
structured investment vehicles 300 Great Depression and 37
tangible, Basel Accords 256 resolution difficulties 388
tranches, CDOs 293–4 zombie banks 128
ERM (exchange rate mechanism) 23, 228–31 European Systemic Risk Board (ESRB) 81, 346
Ervin, W. 272 European Union (EU)
ESMA (European Securities and Markets Authority) bank capital 357–8
182, 346 banking sector reforms 169
ESRB (European Systemic Risk Board) 81, 346 banking supervision 81, 347
Estonia banking system size 6–7
euro zone and 382–3, 385 banking union 347
euro see European Union common currency 37
euro zone see European Union currency arrangements since Bretton
eurodollar market 243–4 Woods 381–2
euroisation 214–15 euro 381
European banking union 347 currency union 379–82
European Banking Authority (EBA) 346, 347 European crisis, 1992–1993 implications
European banks in 2007–2009 crises 319–23 for 231
European Central Bank (ECB) 6, 381 as reserve currency 212
2007-2009 phase and 338 sovereign debt crises cause 387–8
bailouts 88–9, 276, 277–8 euro zone 23, 381–2
benchmark interest rates 94–5 bailouts 384, 386–7
408
INDEX
debt evolution 382–4 monetary unions 213

individual countries 384–7 single currency pegs 213
members’ leaving not allowed 388 target zones or bands 214, 228
European crisis, 1992–1993: 228–31 exchange-traded commodities 152
European Currency Unit (ECU) 228, 381 exchanges 171
European Monetary System (EMS) 381 exit consents, sovereign debt 376
Exchange Rate Mechanism 23, 228–31 exogenous prices 44–5
financial stability 81 exogenous risk 39–40, 59
financial transactions tax 170 exogenous shocks 3, 70–1
gross domestic product 382–4 expansionary monetary policy 82
Maastricht debt criteria 381–2 export-led growth 19
macro-prudential policymaking, lack of 346 external sovereign debt 369–71
macro-prudential regulation by 241 externalities
member government bail-outs, European Central from bank failures 240
Bank 88–9 banking 239
micro-prudential regulation by member firesale see firesale externalities
states 241 negative 120
monetary policies 81 extreme risk
of member states 231 definition 312
naked CDSs, ban on 292–3 risk management failures not targeting 311
regulatory reforms and 346 targeting failure 311–13
single rulebook 347, 357
Solvency II 343 failures
sovereign debt crisis see European sovereign debt of banks see banks
crisis in risk management and regulations
transfer union 347, 380, 388 309–10
zombie banks 128 capital and the crisis 317–23
evergreening 128 regulatory failures see regulatory failures
excessive risk taking 311 Fannie Mae 188
exchange rate mechanism (ERM) 23, 228–31 Faroe Islands 377–9
exchange rates 203–4 fat tails 313–14
adjustable pegs 214 FDIC (Federal Deposit Insurance Corporation)
basket pegs 214 140, 359
Bretton Woods system 205–7 Fed see Federal Reserve System
cost of changes 225 Fed funds interest rates 82, 94
crawling pegs 214 Federal Deposit Insurance Corporation (FDIC) 140,
currency boards 213 359
currency unions 213 Federal Home Loan Bank System, US 130
desired 224–5 Federal Reserve System (Fed) 22, 28, 32
dirty floats 214 benchmark interest rates 94
dollarisation 214–15 credit facilities in bailouts 273–5
euroisation 214–15 government bond holdings 87–8
fixed 195–7, 204 Great Depression 140
1G currency crisis models 219–22 liquidity injections 328–9
2G currency crisis models 224–5 liquidity provision in bailouts 269–70, 273–5
floating 195–7 origins 79
1G currency crisis models 220 principle-based approach to regulation 243
free floats 214 reform 345
gold standard 204–5 support for foreign banks 168
impossible trinity 207 target federal funds rate 94
managed floats 214 federalism, fiscal 224
409
INDEX
feedback Bank of England responsibilities 81

carry trades 157–8 central banks and see central banks
loops central counterparties and 164–6
endogenous risk 63, 72, 158 conflicts between monetary policy and 90–1
positive 133 definition 86
vicious 58, 64, 70–1 European Central Bank responsibilities 81
virtuous 64 European Union 81
negative 83 financial transactions tax 170
Ferguson, T. 32 marking-to-market and 187
fiat money 121 neglect 328
fiats 83 passive prevention 86–7
financial crises regulation 163
1931: 27–8 resolution 87
definition 261 Financial Stability Board (FSB) 345–6, 359, 360
financial engineering 298 Financial Stability Forum (FSF) 341, 345
financial fundamentals Financial Stability Oversight Council
Indonesia 104 (FSOC), USA 346
financial innovations 155–6, 243–4 financial transactions tax (FTT) 169–70
financial instability hypothesis 9–10 Finland
financial institutions banking crisis 126, 131–2
interconnectedness 315–16 euro zone and 382–3, 385
risk taking 9 firesale effect 315–16
systemic risk creation roles 9–10 firesale externalities 13–14, 55, 185
financial instruments firesale prices 66, 132
complexity 287 first-generation (1G) currency crisis models 219–22,
dangerous 285–307 224, 230
market liquidity and 67 first globalism 1, 99, 327
see also derivatives; collateralised debt obligations; First World War 2–4, 19
credit default swaps Great Depression causal event 20–2
financial intermediation reparations 21–2
deposit insurance model 145 fiscal costs, banking crises 125–6
financial liberalisation fiscal federalism 224
banking crises causes 127 fiscal policies, Argentina 222–4
financial markets Fitch 179, 299
Asian crisis 1997: 106–7 fixed exchange rates see exchange rates
Great Depression 26–7 fixed income assets 175, 177
parallels between Great Depression and 2007–2009 flights to quality 62–3
crises 36 floating exchange rates see exchange rates
Financial Policy Committee (FPC) 346 flood insurance 10
financial regulations see regulations Flood, R.P. 219
financial returns 313, 314 forbearance, regulatory 272–3
financial sector contribution to systemic risk 8 foreign exchange (FX)
Financial Services Authority (FSA) carry trade 157
2007-2009 phase and 338 rates see exchange rates
banking supervision 81 see also entries beginning with currency
bonuses 167 foreign loans, Sterling 3
micro-prudential regulation by 241 foreign reserves
Northern Rock 140 Asian crisis 1997: 110–11, 116
reform 345 foresight, perfect 219, 222
financial stability Foresight Group, UK 7
active prevention 87 FPC (Financial Policy Committee) 346
410
INDEX
fractional reserve banking 67, 78, 120, 121 G-SIB (global systemically important banks) 349,
bank runs 137, 139 350–1, 362, 364
collateralised debt obligations and 298–7 G10 economies
deposit insurance modelling 144–6 banking crises in 127
systemic risk and 11–13 G20 341, 342, 345–6
fragilities gains from trade, deposit insurance model 144
bank capital 317, 318–19 Gale, D. 71–2
banking systems 12 gaming 354–5
banks 120 garbage assets 287, 298
France Garber, P.M. 219
bank capital 356 GATT (General Agreement on Tariffs and Trade) 35
euro zone and 382–5 Gaussian copula 295, 297
sovereign debt defaults 369 GDP see gross domestic product
United Kingdom and, currency wars 209–10 General Agreement on Tariffs and Trade (GATT) 35
Frankel, J.A. 213 Germany
Franz Ferdinand, Archduke 3 bank capital 356
fraud bank runs 137
BCCI 129–30 euro zone and 382–5, 386
Iceland banking system crisis 133 European Exchange Rate Mechanism 228–9
free floats 214 Great Depression 27
free riding 35 hyperinflation 92–3
free trade 35 interest rates 229
Friedman, Milton 32, 85–6 mark as reserve currency 228
Friedman’s helicopter 85–6 monetary policies 228–9
FSA see Financial Services Authority reunification 229
FSB (Financial Stability Board) 345–6, 359, 360 sovereign bonds 184–5
FSF (Financial Stability Forum) 341, 345 sovereign debt enforcement 375
FSOC (Financial Stability Oversight Council), USA 346 Giffen goods 43
FT100 index 154 Glass–Steagall Act, USA 29, 168, 241
FTT (financial transactions tax) 169–70 Glass–Steagall Deposit Insurance Act 140
full reserve banking 123 global games currency crisis model 231–4
the Fund see International Monetary Fund global imbalances 21, 329
fundamental equilibrium 206 global impact of European sovereign debt crisis 388–9
fundamental uncertainties 233–4 global savings gluts 329
fundamentals 217 global systemically important banks (G-SIB) 349,
2G currency crisis models 227 350–1, 362, 364
financial, Indonesia 104 globalism 326–7
macroeconomic, Indonesia 104 first 1, 99, 327
underlying, prices reflecting 43 history forgotten 327
weak, Asian crisis 1997: 106, 107 second 327
funding banks 331–2 systemic risk 327
funding liquidity 353 Washington consensus and anti-globalism 328
Asian crisis 1999: 108 GMAC 274
crises and 67 gold 47, 121
market liquidity and, interdependencies between gold standard 22, 27, 121
70–1 adjustment mechanism 205
meaning 65–7 deflation and 204–5
risk 65, 300 exchange rate regime 204–5
scarceness 72 Great Depression and 27–8, 33, 37
futures 152 impossible trinity 207
FX see foreign exchange lending of last resort 267
411
INDEX
gold standard (continued) gold standard and 27–8, 33, 37

mechanics of 204 implications for future policy 34, 35
survival 204 European sovereign debt crisis 37
United Kingdom 27–8 parallels with 2007–2009 crises 35–7
United States 33 schadenfreude 26–7
Goldman Sachs 274–5, 314, 333 stock market crash 26
good bank – bad bank United States
banking system failures resolution method 132, build-up 22–3
263, 265 crisis 28–9
good equilibrium, deposit insurance model 145, 146 double dip 29
Goodhart, C. 80, 250, 251, 270, 271 economy impact 25
Goodhart’s Law 332 financial regulation impact 29
Gorter, J. 343 great moderation 10, 62, 313, 330
Gosh, A.R. 213 Greece
government bonds euro zone and 382–5
Bank of England holdings 87–8 hyperinflation 93
European Central Bank holdings 89 sovereign debt crisis 386, 388
Fed holdings 87–8 Greenspan, Alan 62, 278, 313, 328
government debt Greenspan puts 269, 328–9
United States 211–12 gross domestic product (GDP)
governments 1929 rankings 21
banking regulations transfer of Argentina 223
responsibilities to 243 Asian crisis 1997: 100–1, 102, 107–8, 116
deposit insurance schemes 146–7 bank capital impact 358
guarantees for bank obligations 264 banking crises effects on 125–6
causing asset bubbles 279–83 capital and 358–9
policies 10 central bank interest rates and 83
risk-free 184 European Union 382–4
systemic risk creation roles 10–11 global 389
TBTF institutions and 362, 363 Ireland 263–4
Graham, B. 151, 156 stability trade-off and 358–9
Great Depression 1, 19–20 Sweden 263
bank failures 139 growth
bank runs 28, 137, 139–40 assets 14–15
build-up 20–2 export-led 19
agricultural depression 23 leverage 14–15
bad monetary policy in UK 22 guaranteeing problem assets, bailouts 269
competitive devaluations 23–4 guarantees
deflation 24, 25 deposit insurance model 146
United States 22–3 by governments for bank obligations see
capital markets 26 governments
causes 29 Guidotti–Greenspan rule 107
collapse in trade 30, 336 Gurwin, L. 129
global leadership lacking 34–5
monetary policy role 30–4 haircuts 185–6
national interests, focus on 34–5 Haiti
Smoot–Hawley Act 30, 36 sovereign debt enforcement 375
tariffs 29–30 Haldane, A.G. 249
trade restrictions 29–30 Hale, D. 368
financial crisis 1931: 27–8 hands-off approach 7
financial markets 26–7 Hanke, S.H. 93
412
INDEX
Hannover Re 181 illiquid stocks 73

Harman, Jeremiah 271 illiquidity
HBOS 275, 276 banks 66, 67
hedge funds 56, 65, 70, 148, 155, 333 insolvency and, distinction between 66
hedging 67, 288 ILOLR (international lender of last resort) 277
markets 151–2 imbalances, global 21, 329
short selling 162 IMF see International Monetary Fund
herding 167 immediacy 69
Herring, R.J. 129 impact function, prices 45–6
Herstatt 129, 251 impossible trinity 207
HFT (high-frequency trading) 159–62 in black 185
hidden risk 330–1 in red 185
high-frequency trading (HFT) 159–62 incentives
high-powered money see M0 to become TBTF institutions 361–2
high productivity 62 perverse 11, 16–17, 292
high quality liquid assets 353 regulatory failures 317
historical values 187 supervisors of banking regulations 242
home regulators 251 incidence of attack 234
Homer, S. 369 incompetence 133
Hong Kong independence, central banks 90, 91, 328
Asian crisis 1997: 97, 102 Independent Commission on Banking,
Honohan, P. 126 UK 7, 169
Hoover, Herbert 27 India
Hoover Moratorium 27 sovereign debt defaults 369
host regulators 251 indifference curves, deposit insurance
hot money, capital controls and 202, 203 model 144
house prices 55 individual risk, credit default swaps 290
households, asset growth and leverage growth 14–15 Indonesia
HSBC 168, 169, 321, 333 Asian crisis 1997: 97, 102, 104, 106, 108, 113
Hume, David 204 banking crises 126
Hungary ineffectiveness, regulations 311
bank runs 137 inflation 328
Great Depression 27 hyperinflation 22, 73, 92–3
hyperinflation 93 sovereign debt reduction through 370–1
hybrid capital instruments 318–19 unexpected, central banks creating 87
hybrid instruments United Kingdom 22
banks 247 United States 211
Basel Accords 256 information asymmetries 11, 15, 164–5, 316
hyperinflation 22, 73, 92–3 initial margins 185, 187
innovations, financial 155–6, 243–4
Iceland insolvency
banking system failure 2008: 132–4, 242 illiquidity and, distinction between 66
capital controls 201–2 instability
idiosyncratic risk dynamic 206
collateralised debt obligations and 298–9 financial instability hypothesis 9–10
too big to fail institutions and 362–3 stability breeding 309
idiosyncratic shocks 4 Institute of International Finance (IIF) 342, 358, 359
ignored risk 330–1 institutional investors, bank runs 332
IIF (Institute of International Finance) 342, 358, 359 institutionalisation 155
IKB 301, 333 institutions, sovereign debt enforcement 375
Iksil, Bruno 153 instruments, hybrid see hybrid instruments
413
INDEX
insurance interventions, currency markets 196, 197–9

credit default swaps see credit default swaps intolerance, sovereign debt 373–4
deposit see deposit insurance investment banks 168–9
insurable interest 292 securitisation and 189
moral hazard and 124 investments
portfolio 49, 52 Asian crisis 1997: 100–1
regulations 343 domestic savings-based 101
social, optimal, deposit insurance model 143–4 funds from short-term borrowings from
interbank exposures 75 international capital markets 101
interconnectedness 3, 327 portfolio, hot money 202
interdependence 11, 15–16 investors
interest rates accredited 155
central banks 82–3, 91, 94–5 institutional, bank runs 332
discount rates 94 risk-averse short-term 329
Fed funds 82, 94 risk neutral 184
Germany 229 strike 70
increases, build-up to Great Depression 22 IOSCO (International Organisation of Securities
prime 94 Commissions) 347
risk-free 82, 94 IRB (internal rating based) approach, Basel II 253–4
short 82, 94 Ireland
target 82, 94 bailout 2008: 264–5
uncovered interest rate parity see uncovered interest banking system 6–7
rate parity euro zone and 382–5
interests, conflicts of, credit rating agencies 181–2 sovereign debt crisis 386
internal devaluations 22, 209 irrational exuberance 278
internal rating based (IRB) approach, Basel II 253–4 Italy
international banking regulations see banking euro zone and 382–5
regulations European Exchange Rate Mechanism 230
international bond markets 177–8 sovereign bonds 184–5
international cooperation 347–8 sovereign debt crisis 387
international institutions, regulatory reform 345–6 Ito, T. 107, 108
international lender of last resort (ILOLR) 277
International Monetary Fund (IMF) 5, 35, 64, 87 Japan
Argentinian crisis 222–3 banking crises 127
Asian crisis 1997 role see Asian crisis 1997 sovereign debt 373–4
Bretton Woods 206 zombie banks 128
capital controls 201–2, 203 jawboning 105
financial crises database 124–6 JP Morgan 274–5, 333
heavy-handedness 113 JP Morgan Chase 152–3, 315
as lender of last resort 113–15
moral hazard 112 Kalb, J.B. 129
packages 98 Kasdeki, J.-R. 375
Indonesia 104, 113 Kaufman, G.G. 10
South Korea 105, 113 Keating, C. 167
Thailand 103, 113 Kerviel, Jérôme 154, 167
reform and 346 Keynes, John Maynard 22, 27, 40, 84, 90, 93, 169
sovereign debt restructuring 375 Kindleberger, C. 30–1
structural adjustments 112–13, 116, 328 King, Mervyn 338
Washington consensus champion 328 Klingebiel, D. 126
International Organisation of Securities Commissions Kohn, M. 239
(IOSCO) 347 Kovacevich, R.M. 147
414
INDEX
Krueger, A.O. 375 stigma effect 271

Krugman model of governments guaranteeing terms of lending 270
borrowers 147 Lenin, Vladimir 93
Krugman, P. 219, 279, 281 leverage
Kupiec, P.H. 120 banks see banks
Kwok, A.K.F. 93 capital adequacy ratios and 331
Kyle, A. 69 constraints, prices 43–6
growth 14–15
Laeven, L. 124–6, 208 haircuts and 186
Lagarde, Christine 292 increases 63
large market participants, banks as 45–6 margins and 186
late agents, deposit insurance model 142–7 ratios, Basel III 352
latent variables, risks as 318 structured investment vehicles 300
Latin America 327 leveraged traders 70
sovereign debt defaults 369, 372 Lewis, M. 335
Lavigne, R. 199 Li, D.X. 295
LCR (liquidity coverage ratio) 74, 353–5 liabilities, collateral and contingent 75
leadership, global, lacking during Great Depression Liberia
34–5 sovereign debt enforcement 375
Leeson, Nick 154 LIBOR (London Interbank Offered Rate)
legal approaches to banking regulations 242–3 price-fixing scandal 153
legal protection, credit rating agencies 183 Liikanen, Erkki 169
legal steps, sovereign debt enforcement 375–6 limit order markets 68–9
legal tender 121 limited liability corporations 124
Lehman Brothers liquidation
bailouts and 273, 276, 277 companies 371
credit default swaps contributing to failure 289, of Newfoundland 369
290, 291 liquidity 48, 61–2, 352–3
effect of failure 334–5 1998 crisis 62–5
inaction costs and 126 asset pricing and 71–3
information asymmetry and 164 asymmetric information and 316
quant event and 314 automated systems and 162
resolution regimes and 359–60 banks 87
risk–return trade off and 9 Basel III 352–6
UK banks, failure effects on 276 complex financial models ignoring 287
lemons problems 190 coverage ratios (LCR) 74, 353–5
lenders crises
relationships with borrowers 75 Asian crisis 1997: 108–9
lending global 2008: 335–6
banks see banks funding see funding liquidity
to East Asia 99–100 injections 73–4, 328–9
lending of last resort (LOLR) 4, 35, 79, 268 market see market liquidity
asset choice 271 meaning 65
Bank of England 266–7 funding liquidity 65–6
European Central Bank 271–2 illiquidity and insolvency, distinction between 66
historical origins 265–7 limit order markets 68–9
IMF 113–15 market liquidity 65–6, 67
motivation 270 models 70
parallels between Great Depression and 2007–2009 interdependencies between market and funding
crises 36 liquidity 70–1
principles 266–7 liquidity and asset pricing 71–3
415
INDEX
liquidity (continued) Luxembourg

net stable funding ratio (NSFR) 74, 355 banking system 6–7
Northern Rock 141 euro zone and 382–3, 385
opportunity costs of holding 72
parallels between Great Depression and 2007–2009 M0 11–12, 30–1, 67, 84
crises 36 M1 11–12, 31, 67, 84
policy implications 73 M2 12, 30–1, 67
2007–2009 crises 74–5 M3 12
central banks’ liquidity injections 73–4 Maastricht debt criteria 381–2
macro-prudential policies 74–5 Maastricht Treaty 373
moral hazard problems 74 Macrae, R. 166
regulations 74 macro-prudential
premiums 73 policies
providers of first resort 74 European Union lack of policymaking 346
provision in bailouts see bailouts liquidity policy implications 74–5
ratios regulations
coverage ratio (LCR) 74, 353–5 banking 240–1
danger 356 reform 343
net stable funding ratio (NSFR) 74, 355 tools
regulatory reform 352–6 capital controls 203
risk macroeconomics
conduits 300–1 Asian crisis 1997: 100
maximisation 332 fundamentals, Indonesia 104
regulations 74 Madoff, Bernie 147, 154
structured investment vehicles 300–1 main refinancing operations fixed rate, ECB 94
traps 84–6 maintenance margins 185, 187
liquidity adjusted capital asset pricing model 73 Malaysia
Litan, R.E. 129 Asian crisis 1997: 97, 102, 108
living wills 87, 241, 361 capital controls 201, 203
Lloyds TSB 276 Malta
loan guarantees, bailouts 269 banking system 6
loan to value ratios 86 euro zone and 382–3
loans managed floats 214
bailouts 268 marginal lending facility, ECB 94–5
cross-border 3, 4 marginal products 280
foreign, Sterling 3 marginal rates of substitution, deposit insurance
maturities 12 model 144
securitised 4 marginal rates of transformation, deposit insurance
LOLR see lending of last resort model 144
London Interbank Offered margins
Rate see LIBOR calls 70–1, 185
London Stock Exchange 3–4 increases, central counterparties 165–6
London whale 153, 315 initial 185, 187
long-term bonds, banks 247 leverage and 186
loss given default 179 maintenance 185, 187
loss spirals 71 requirements 63, 70
losses, dead weight 88, 243 spirals 71
low-risk arbitrage 154 mark-to-market accounting see marking-to-market
Lowenstein, R. 56 market discipline (Pillar III), Basel II 254
LR (leverage ratio), Basel III 352 market factors, Asian crisis 1997: 106–7
LTCM 56–9, 63, 64, 65, 152 market liquidity
416
INDEX
funding liquidity and, interdependencies between minimum capital adequacy ratios, Basel III 349
70–1 minimum capital requirements (Pillar I), Basel II 254
increasing importance 67 Minsky, Hyman 9, 40, 330
meaning 65–6 modern portfolio theory (MPT) 47
price impacts 68 Modigliani–Miller analysis 358
risk 65 Moe, T.G. 132, 263
terminology 69–70 momentum strategies 53
market makers 68, 172 momentum trading 159
market orders 68 monetary aggregates 11
market oriented structural reforms, Argentina 222 monetary base see M0
market participants, trading risk 154–5 monetary institutions, credibility 224
market risk monetary policies
Basel II 254 Argentina 223–4
regulation 74 build up to crisis 328–9
markets central banks see central banks
derivatives 67 conflicts between financial stability and 90–1
discount 3 contractionary 82
efficient 329–30 European Central Bank 81
limit order 68 European Union see European Union
markets’ fear gauge see VIX expansionary 82
marking-to-magic 187 Germany 228–9
marking-to-market 10, 187, 287 Great Depression 30–4
marking-to-model 187 impossible trinity 207
Marsh, D. 381 neutral 82
Marton, William McChesney Jr 77 primacy 328
maturities UK, Great Depression 22
credit ratings and 180–1 United States 22–3
loans 12 monetary stability, central banks 79, 80
mismatches 99, 107, 110, 332 monetary unions 213, 381
maximum harmonisation principle, EU 357–8 money
maximum leverage, banks 87 Asian crisis 1997: 100–1
McCarthy, Sir Callum 140–1 debasement 121
mean reversion 57 definition 121
mean reverting assets 59 fiat 121
mean–variance diagrams 47 high-powered see M0
Medici bank 121–3 hot, capital controls and 202, 203
Mellon, Andrew 26 legal tender 121
merger arbitrage 57 losing control, central banking challenges 91–3
mergers, banks 363 multiplier 12
Merrill Lynch 275 narrow see M1
Merton, R. 10, 56, 183 parallels between Great Depression and 2007–2009
Mexico crises 36
banking crises 126 printing 4, 87–8
sovereign debt 373, 377 price to pay 92
mezzanine notes, structured investment vehicles 300 as tax on assets 88
mezzanine tranches, CDOs 293–4 quantity theory of 204
microeconomic theory 43 supply
micro-prudential regulations 240–1, 345 contraction 30–2
military sovereign debt enforcement 375 multiplier 67
Millennium Bridge 41–3 reductions 67
Miller, G.P. 148 types 11–12
417
INDEX
money market mutual funds 148 New York Federal Reserve Bank (NYFed) 22, 58, 63,
Moody’s 179–81, 299 79, 333, 358
moral hazard 10 Newfoundland 368–9
Asian crisis 1997: 106, 112 Nigeria
asset bubbles, government guarantees causing sovereign debt enforcement 375
279–83 non-linear dependence 48, 287, 297, 313–15, 331, 333
bailouts see bailouts normal distribution 313, 314
banking crises 119, 123–4 normality 313–15
credit default swaps 292 Northern Rock 28, 137, 140–2, 188, 190, 275, 332
definition 123 Norway
deposit insurance 147 banking crisis 127, 131–2
government guarantees causing asset bubbles notional principals, CDSs 302
279–83 NSFR (net stable funding ratio) 74, 355
insurance and 124 NYFed see New York Federal Reserve Bank
liquidity injections increasing 329
liquidity policy implications 74 Obama, Barack 269
risk taking and 124 Obstfeld, M. 219
securitisation 190 off-balance sheet risk 332
Morgan, John Pierpont 250 offers 68
Morgan Stanley 274 oligopolies, credit rating agencies 182–3
Morris, S. 158, 231 on the margin 19
mortgage originators 189 ‘one minute to midnight’ 272
mortgage-related products 55 ongoing crisis, 2007-2009 phase 325–6
motivation, lending of last resort 270 build-up to crisis 326–30
Moysich, A. 131 changing nature of banking 331–3
MPT (modern portfolio theory) 47 crisis, 2007-2008: 333–7
multiple equilibria 225–8 hidden and ignored risk 330–1
deposit insurance model 145 policy response 338
Mundell, R.A. 380 subprime mortgage role 337–8
Mussa, Michael 224 ongoing developments in financial regulation 341–65
open market operations 32
naked credit default swaps 292–3 central banks’ monetary policy 82, 83, 85
naked short selling 163 liquidity provision in bailouts 269
narrow banking 168–9 operational risk, Basel II 254
narrow money see M1 opportunity costs 72
national banking champions 362 optimal risk sharing, deposit insurance model 145
national discretion, liquidity coverage ratio 354 optimal social insurance, deposit insurance model
national institutions, regulatory reform 344–5 143–4
national interests, focus on, Great Depression 34–5 optimisation
national positions, capital 356–7 bank capital structure 317–18, 319
negative externalities 120 collateralised debt obligations 297–8
negative feedback 83 deposit insurance model 144
net stable funding ratio (NSFR) 74, 355 option pricing theory 183
Netherlands optionality 288
bank capital 356 options 67
euro zone and 382–5 put 49–52
netting 172, 289 short 57
netting out 291 orders 171
network effects 3 originate and distribute model of banking 188
network vulnerabilities, credit default swaps 290–2 Oskarsson, H. 377
neutral monetary policy 82 OTC see over-the-counter
418
INDEX
output loss, banking crises 126 politics

over-the-counter (OTC) bailouts 278–9
derivatives 67, 288 bank size and 8
markets 177 Ponzi schemes 153–4, 155
trading 171 portfolio insurance 49, 52
overconfidence 167 portfolio investments, hot money 202
Asian crisis 1997: 107 portfolio theory 47, 313
Overend and Gurney 167, 265–7 Portugal
overnight funding 331–2 euro zone and 382–5
overproduction 23 sovereign debt crisis 386–7
overvaluation perils 208–9 positive feedback loops 133
power, political, banks 8
packages, IMF see International Monetary Fund PPP (purchasing power parity) 219, 222
Pangloss value 281–2 predictions, empirical 54–5
panic and contagion, Asian crisis 1997: 106 preference shares
Panizza, U. 372 bailouts 268
Papandreou, George 386 banks 247
Pareto efficiency, deposit insurance model 145 preferred stock, Basel Accords 256
Paris Club 375 premiums
parties 171 credit default swaps 302
see also counterparties liquidity 73
passive prevention, financial stability 86–7 price takers, banks as 44–5
payment flows, credit default swaps 290 prices
Pedersen, L.H. 70, 73 dual role 43–6
pension funds 329 endogenous 45–6
perceived risk 53–5, 58, 315, 337 exogenous 44–5
perfect foresight 219, 222 firesale 66, 132
persistent deficits 219 impact function 45–6
Peru as imperatives to action 43
sovereign debt enforcement 375–6 leverage constraints and upward-sloping demand
perverse consequences of banking regulations 243–4 43–6
perverse incentives 11, 16–17, 292 market liquidity impacts 68
petrodollars 327 reflecting underlying fundamentals 43
Pfandbriefe 188 reserve currencies and 211
Philippines supply and demand functions 46
Asian crisis 1997: 102 pricing
plain vanilla bonds 176 assets, liquidity and 71–3
Plaza Accord 208 bonds 178
policies cash-in-the-market 71–3
2007–2009 phase of ongoing crisis response 338 primary markets 177
Asian crisis 1997 see Asian crisis 1997 prime brokers 56, 65
governments 10 prime interest rates 94
liquidity see liquidity Prince, Charles 278
macro-prudential see macro-prudential policies principle-based regulations 243
monetary see monetary policies printing money see money
speculation issues see speculation private sector bail-outs, European Central Bank 89
trading see trading privately funded deposit insurance schemes 146–7
policymakers, liquidity coverage ratio dilemma 354 privatisation
policymaking, macro-prudential, EU lack of 346 banking systems, Iceland 133
political pain 224 problem assets, buying or guaranteeing,
political power, banks 8 bailouts 269
419
INDEX
pro-cyclicality regulations 64–5, 237–9

bank capital 247–9 bank capital 247–50
banking regulations 242, 247–9 bank failures, preventing 120
endogenous risk 330–1 banking see banking regulations
systemic risk 10, 11, 13–15 financial stability 163
productivity, high 62 hedge funds 155
products ineffectiveness 311
banks, sophistication 239–40 international financial regulations: Basel see Basel
complex 287 Accords
marginal 280 liquidity policy implications 74
prompt corrective action 65, 87 ongoing developments in financial regulation
prop desks 333 341–65
proprietary trading, banks 155, 168 principle-based 243
proprietary trading desks 70 reform see regulatory reform
protection buyers, CDSs 289, 301–2 targeting day-to-day risk not extreme risk 311
protection sellers, CDSs 289, 301–2 tick-the-box 243
prudential regulations 65 value-at-risk 257–8
versus systemic regulations 315–16 regulatory arbitrage 251
publicly observable agents, deposit insurance regulatory capture 130–1, 244, 279
model 143 regulatory failures 310
purchasing power parity (PPP) 219, 222 complexity, incentives and resources 317
put options 49–52 deregulation and 311
focus of risk management and
QE see quantitative easing regulations 311–13
quality normality and non-linear dependence
of credit ratings 182 313–15
flights to 62–3 prudential versus systemic regulations 315–16
quant crisis 2007: 70, 71, 313–14, 333 regulatory forbearance 272–3
quantitative easing (QE) 4, 28, 36, 371 regulatory reform 341–2
Bank of England 86 Basel III see Basel Accords
central banks’ monetary policy 82, 84–6 capital levels 356–9
liquidity provision in bailouts 269 central banks 344–5
United Kingdom 87 different initiatives in conflict 343
United States 87 European banking union 347
quantity theory of money 204 international cooperation and colleges of
supervisors 347–8
rainmakers 167 international institutions 345–6
Ramirez, C.D. 120 liquidity 352–6
rating agencies see credit rating agencies multiple agendas 342–3
ratings shopping 182 recovery and resolution 359–61
recapitalisation, banks 119 national institutions 344–5
Reconstruction Finance Corporation (RFC) 28, 36, new systemic risk institutions 346–7
140, 271 supervisors 344–5
recovery 179, 361 too big to fail institutions 361–4
in red 185 Reinhart, C.M. 203, 327, 373
Reding, Viviane 182 relative value trades 56
reduced form credit models 183 repatriation of funds 3
reference entities, CDSs 289, 302 repayment of sovereign debt, reasons see sovereign
reference instruments, CDSs 301 debt
reference obligations, CDSs 302 Republic of Congo (Congo Brazzaville)
Regulation Q, USA 243–4 sovereign debt enforcement 375
420
INDEX
Republic of Ireland see Ireland risk–return trade-off 8

Repullo, R. 350 risks
reputation risk 271 actual see actual risk
banks 319 assessments, complex products 287
central banks 80 counterparty see counterparty risk
sovereign debt repayment reason 372 credit see credit risk
reserve accounts, banks 83 day-to-day see day-to-day risk
reserve banks see central banks endogenous see endogenous risk
reserve currencies 197, 210–12 exogenous see exogenous risk
Chinese renminbi 212 extreme see extreme risk
euro 212 funding liquidity 65, 300
Germany mark 228 hidden 330–1
prices and 211 idiosyncratic see idiosyncratic risk
US dollar status, threats to 211–12 ignored 330–1
reserve requirements 12 individual, credit default swaps 290
central banks’ monetary policy 82, 84 as latent variables 318
reserves liquidity see liquidity
banks, minimum 87 management see risk management
currency 116, 329 market see market risk
disclosed, Basel Accords 256 market liquidity 65
revaluation, Basel Accords 256 maximisation, yield seeking as 156–7
undisclosed, Basel Accords 256 meaning 46–7
resiliency 69 operational, Basel II 254
resolution perceived see perceived risk
bank failures 359–60 reputation see reputation risk
of banking crises 126, 132 sensitivity
financial stability 87 bank capital 248–9
living wills 361 lack, Basel I 252
regimes 359–60 pro-cyclicality in banking 350
resources settlement 129
banking regulations problems 245 systematic 5
regulatory failures 317 systemic see systemic risk
restructuring sovereign debt 375, 376 tail see extreme risk
retail depositors, bank runs 332 taking
returns, financial 313, 314 bonuses and 166–8
reunification of Germany 229 excessive 311
revaluation reserves, Basel Accords 256 financial institutions 9
RFC (Reconstruction Finance Corporation) 28, 36, moral hazard and 124
140, 271 regulatory reform and 343
risk-averse short-term investors 329 target levels 311–12
risk-free assets 47 trading and see trading
risk-free governments 184 undetected 330
risk-free interest rates 82, 94 weights 318
risk management roaring twenties 19
active 54–5 Rogoff, K. 203, 219, 327, 373
failures 309–10 rogue traders 154
capital and the crisis 317–23 Roosevelt, Franklin Delano 28, 30, 33, 140
targeting day-to-day risk not extreme risk 311 Rothschild, Nathan 159–60
focus 311–13 Royal Bank of Scotland 168, 275, 276, 321
risk neutral investors 184 rules, trading 156
risk-weighted assets (RWA) 256–7, 320–2 runs on banks see bank runs
421
INDEX
Russia short interest rates 82, 94

1998 crisis 62, 65 short options 57
sovereign debt 375 short selling 36, 162–3
RWA (risk-weighted assets) 256–7, 320–2 shorting see short selling
SIC (Special Investigation Commission), Iceland 132,
S&L see savings and loans 242
S&P (Standard & Poor) 179–81 Sicsic, P. 209
S&P-500 154 SIFI (systematically important financial institutions)
Santander 321 350–1
Saurina, J. 350 silver 121
savings and loans (S&L) Singapore
banking system failure 1980s, US 130–1 Asian crisis 1997: 97, 102
savings gluts 329 Singer, Paul 375–6
scandals, trading 152–4 single currency pegs 213
Scandinavia 1990s banking system failure 131–2 single rulebook, EU 347, 357
schadenfreude 26–7 SIV (structured investment vehicles) 16, 300–1, 332, 337
Schoenmaker, D. 80 Six Flags 292
Scholes, Myron 56, 183 Slovakia
Schumpeter, Joseph 5 euro zone and 382–3, 385
Schwartz, A. 32 Slovenia
SEC see Securities and Exchange Commission euro zone and 382–3, 385
second-generation (2G) currency crisis models see Smoot–Hawley Act, USA 30, 36
currency crises: models smoothing the road, banking regulations 241–2
second globalism 327 SocGen 154, 321
Second World War 1, 29, 90, 368, 371 social insurance, optimal, deposit insurance model
secondary markets 177 143–4
Securities and Exchange Commission (SEC) 29, 182 Société Générale 154, 321
legal approach to regulation 243 Solvency II, EU 343
reform 345 sophistication, bank products 239–40
regulatory capture 244 Sorokin, A.R. 360
securitisation 55, 67, 140, 188–90 Sotheby’s 160–1
securitised loans 4 South Africa
seigniorage 87 sovereign debt 372
selection, adverse 269 South Korea
self-fulfilling prophecies, bank runs as 137, 145 Asian crisis 1997: 97, 102, 105, 108, 113
self-preservation, banks 315 banking crises 126
Selgin, G. 121 capital controls 202–3
sell-on-loss 52 sovereign bonds as liquid assets 354
Senior Supervisors Group 348 sovereign debt
sensitivity, risk see risk Basel I 255
settlement methods, CDSs 302–3 Basel II 255–6
settlement risk 129 credit default swaps on 292–3
settlements 171–2 crises 10, 367–8
shadow banking 10, 188, 243, 332–3 Asian 1997: 100
shadow exchange rates 220–1 European see European sovereign debt crisis
shares see preference shares; stocks Faroe Islands 377–9
Shin, H.S. 14, 50, 140, 158, 231 Newfoundland 368–9
shocks triggers 374
exogenous 3, 70–1 defaults 369
idiosyncratic 4 corporate default differences 371
systematic 4 domestic 369–71
422
INDEX
enforcement speculative attacks 208, 217, 219–22

encouraging agreement of lenders 376 European crisis, 1992–1993: 229–30
extreme legal steps 375–6 first-generation (1G) models 222
history, lessons from 376–7 global games currency crisis
inducing restructuring, carrots and sticks 376 model 231–4
institutions 375 second-generation (2G) models 227–8
military enforcement 375 sponsors, SPVs 188–9
vulture funds 375–6 spot currency transactions, Tobin tax 169
European 371 spreads 62–3
exit consents 376 credit 184
external 369–71 credit ratings and 180–1
intolerance 373–4 SPV see special purpose vehicles
nature 369 square-root-of-time rule 258
problems 328 stability
reduction through inflation 370–1 breeding instability 309
repayment reasons destabilising nature of 9–10, 330
contagion 372 financial see financial stability
default costs analysis 372 monetary, central banks 79, 80
default immunity 372–3 trade-off, GDP and 358–9
economic development 372–3 stagflation 90
reputation risk 372 stamp duty 169–70
restructuring 375, 376 Standard & Poor (S&P) 179–81
risk-free 255–6 Standard Chartered 348
vulnerabilities 374 standard deviation 314
vulture funds 223 statistical arbitrage 159, 333
Spain Steiglitz, Joseph 269
banking crises 127 sterilisation, currency markets 198–9
euro zone and 382–5 Sterling 3
European Exchange Rate Mechanism 230 stigma effect, lending of last resort 271
sovereign debt 377 stock markets
crisis 387 booms 22
defaults 369 crash, Great Depression 26
special purpose vehicles (SPV) 188–90 Flash Crash, USA 160–2
collateralised debt obligations 293 indices 154
conduits 299 stocks
special resolution regimes, failed banks 87 illiquid 73
Special Investigation Commission (SIC), Iceland 132, preferred, Basel Accords 256
242 Stoler, P. 129
speculation 151–2 strategic uncertainties 233–4
activities see trading: activities stress tests on European banks 347
bank proprietary trading 155 strict capital controls 200–2
credit default swaps 292 Strong, Benjamin 23, 32
currency markets 196 structural adjustments
financial transactions tax 169–70 Asian crisis 1997: 112–13, 116
overvaluation 208 IMF see International Monetary Fund
policy issues 163–4 structural credit models 183
bonuses 166–8 structured investment vehicles (SIV) 16, 300–1, 332,
central counterparties 164–6 337
financial transactions tax 169–70 subordinated debt
universal and narrow banking 168–9 banks 247
stabilising, gold standard 204 Basel Accords 256
423
INDEX
subprime systemic risk 1–2

assets 287 1914 crisis 2–4
market 296–7, 298 bank capital and 359
mortgages bonuses and 167
2007-2009 phase of ongoing crisis role 337–8 collateralised debt obligations and 298–9
crisis, US 188 competition and 10
substitution, marginal rates, deposit insurance model 144 computer-based trading 160
successful bailouts 263–4 concept 4–9
sudden stop creation 8, 9
Asian crisis 1997: 108–9 financial institutions’ roles 9–10
hot money 202 government roles 10–11
Suharto, President of Indonesia 104, 112, 113, 114 credit default swaps 291–2
Sullivan, R. 159 definition 5
super senior tranches, CDOs 293–4 elimination 8–9
superpower status financial sector contribution to 8
United States gaining, United Kingdom losing 21 fundamental origins 11
supervision fractional reserve banking 11-13
banking 80–1 information asymmetry 15
see also banking regulations; regulations interdependence 15–16
supervisors perverse incentives 16–17
colleges of 347–8 procyclicality 13–15
incentives 242 globalism 327
regulatory reform 344–5 ignored 315–16
TBTF institutions and 362 ignoring 8
supervisory review process (Pillar II), Basel II 254 new institutions 346–7
supply functions, prices 46 notions of 5
swaps 67 regulatory reform 343, 358
credit default see credit default swaps systematic risk contrasted 5
Sweden too big to fail institutions and 362–3
bailout 1992: 263–4 vulnerability 6–8
bank capital 356
banking crisis 126, 127, 131–2 TA (total assets) 320–2
Swedish Riksbank 78 tail risk see extreme risk
Swiss National Bank (SNB) 34 Taiwan
Switzerland 21 Asian crisis 1997: 97, 102
bank capital 356 taking risks see risk: taking
banking crises 127 tangible equity, Basel Accords 256
Sylla, R. 369 target bands 214
synthetic collateralised debt obligations 298 target interest rates 82, 94
system events 5, 6 target levels, risk 311–12
systematic risk target zones 214, 228
systemic risk contrasted 5 tariffs, Great Depression 29–30
systematic shocks 4 TARP (Troubled Asset Relief Program) 36, 269, 273
systematically important financial institutions (SIFI) Taylor, J.B. 82
350–1 Taylor rule 82–3
systemic crises 1, 5 TBTF institutions see too-big to fail institutions
bailouts 261 technical trading 159
bank capital and 358 termination dates, CDSs 302
liquidity 74 Thailand
systemic regulations versus prudential regulations Asian crisis 1997: 97, 102–3, 108, 113
315–16 capital controls 202
424
INDEX
tick-the-box approaches to banking regulations 242–3 strategies 48–53, 156

tier 1 and tier 2 capital terminology 171–2
Basel I and II 256 traditional capital controls (1.0) 200–2
Basel III 349 tranches, collateralised debt obligations 293–6,
tightness 69–70 297–8, 305
Tobin, James 169–70 transfer unions 88, 347, 380, 388
Tobin tax 169–70 transformation, marginal rates of, deposit insurance
too-big to fail (TBTF) institutions 4, 361 model 144
actions 363 transparency 187
central counterparties 165 Trapanese, M. 359
deposit insurance and 148 treasuries, reform 345
emergency measures in crises 363 the Treasury, UK 81
governments and 362, 363 trend following 156
incentives to become 361–2 Triffin dilemma 296
problem not improving 363–4 Triffin, Robert 296
regulatory reform and 343 trigger selling 333
supervisors and 362 triggers, sovereign debt crises 374
view until 2007: 362–3 tripartite regulatory system, UK 81
total assets (TA) 320–2 Troubled Asset Relief Program (TARP) 36, 269, 273
toxic assets, banks 317, 318 Truell, P. 129
trade Turkey
collapse, Great Depression 30 banking crises 126
dates, CDSs 302 sovereign debt enforcement 375
free 35
gains from, deposit insurance model 144 UBS 321
parallels between Great Depression and UIP (uncovered interest rate parity) 197, 219, 222, 225
2007–2009 crises 36 UK see United Kingdom
restrictions, Great Depression 29–30 uncertainties 233–4
surpluses 22 uncovered interest rate parity (UIP) 197, 219, 222, 225
traders, rogue 154 underlying assets, derivatives 288
trading 151–2 underlying fundamentals, prices reflecting 43
activities 156 undermining bank capital 318–19
carry trades 157–9 undervaluation and ‘currency wars’ 209–10
high-frequency trading 159–62 undetected risk 330
short selling 162–3 undisclosed reserves, Basel Accords 256
technical trading 159 unencumbered assets, banks 353
value investing 156 unexpected inflation, central banks creating 87
yield seeking as risk maximisation 156–7 United Kingdom
banks 155 bad monetary policy 22
computer-based 160 bailouts 275–6
policy issues 163–4 bank capital 356, 357
bonuses 166–8 banking crises 127
central counterparties 164–6 banking system 7
financial transactions tax 169–70 European Exchange Rate Mechanism 230
universal and narrow banking 168–9 France and, currency wars 209–10
risk and 154 gold standard 27–8
financial innovation 155–6 Great Depression 27–8
hedge funds 155 quantitative easing 87
market participants 154–5 sovereign debt defaults 369
rules 52–3, 156 superpower status, losing 21
scandals and abuse 152–4 tripartite regulatory system 81
425
INDEX
United States Venezuela

bailouts 36, 273–5 banking crisis 126, 127–8
bank capital 356, 357 sovereign debt enforcement 375
bank runs 137, 139–40 vicious feedback loops 58, 64, 70–1
banking crises 126, 127 Vickers Report 169
capital controls 200 Vinier, David 314
China and, currency wars 210 virtuous feedback loops 64
crisis 28–9 VIX 57, 333, 336
deposit insurance 140 volatility 57, 58
dollar as reserve currency 197, 205–6, 211–12 clusters 313–14
Federal Home Loan Bank System 130 Volcker, Paul 156, 168, 330
Flash Crash 160–2 Volcker Rule 168
gold standard 33 vulnerabilities
government debt 211–12 network, credit default swaps 290–2
Great Depression sovereign debt 374
build up 22–3 systematic risk 6–8
crisis 28–9 vulture funds 223, 375–6
double dip 29
economy impact 25 Wachovia 275
financial regulation impact 29 Wall Street Bubble 22
house prices 55 Wall Street collapse 25, 26
inflation 211 wars 10
monetary policies 22–3 Washington consensus 98, 99, 203, 328
money supply contraction 30–2 Watkins, C. 166
prompt corrective action 87 weak fundamentals
quantitative easing 87 Asian crisis 1997: 106, 107
Regulation Q 243–4 Wells Fargo 274–5
savings and loans banking system failure 1980s 130–1 wholesale markets
subprime mortgage crisis 188 deposit insurance impact 147–8
superpower status, gaining 21 Northern Rock and 141
universal banking 168–9 World Bank 35, 87, 124–6
unsuccessful bailouts 264–5 World Trade Organization (WTO) 35
upper tier 1 capital, Basel Accords 256 World War, First see First World War
upward-sloping demand 43–6 wrong rate argument 196
USA see United States WTO (World Trade Organization) 35
usury 121 WWI see First World War
utility, deposit insurance model 143–4, 145–6 WWII see Second World War
V-shaped crises 107–9 yield seeking as risk maximisation 156–7

Valencia, F. 124–6, 208 yields, credit ratings and 180–1
valuations, complex products 287 Yugoslavia
value-at-risk (VaR) 47, 312 hyperinflation 93
Basel Accords 253, 257–8
credit exposures, inappropriate use for 331 Zhu, H. 165
flaws 314–15 Zigrand, J.-P. 165
value investing 156 Zimbabwe
VaR see value-at-risk hyperinflation 92–3
vega 57 zombie banks 128, 272, 278
426

Global Financial Systems Stability and Risk by Danielsson

Uploaded by

Copyright:

Available Formats

Global Financial Systems Stability and Risk by Danielsson

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Global Financial Systems Stability and Risk by Danielsson

Uploaded by

Copyright:

Available Formats

Global Financial Systems

Stability and Risk

Under what circumstances have we achieved financial stability?

Jon Danielsson is Reader in Finance and member of the Financial

First published 2013 (print and electronic)

© Jon Danielsson 2013 (print and electronic)

ISBN: 978-0-273-77466-2 (print)

British Library Cataloguing-in-Publication Data

Library of Congress Cataloging-in-Publication Data

Print edition typeset in 9.25/13.5 Stone Humanist ITC Std by 75

NOTE THAT ANY PAGE CROSS-REFERENCES REFER TO THE PRINT EDITION

2 The Great Depression, 1929–1933 19

5 The central bank 77

6 The Asian crisis of 1997 and the IMF 97

7 Banking crises 119

8 Bank runs and deposit insurance 137

9 Trading and speculation 151

9.4 Policy issues 163

10 Credit markets 175

11 Currency markets 193

12 Currency crisis models 217

13 Financial regulations 237

15 Dangerous financial instruments 285

16 Failures in risk management and regulations

17 The ongoing crisis: 2007–2009 phase 325

18 Ongoing developments in financial regulation 341

18.5 Recovery and resolution 359

19 Sovereign debt crises 367

Errata and online chapter

The author gratefully acknowledges the support of the Economic and

We are grateful to the following for permission to reproduce copyright material:

Getty Images: Cate Gillon / Staff 173, Agus Lolong 147.

Cover images: Front: Getty Images

Links to other chapters

Readings for this chapter

Notation specific to this chapter

1.1 Case study: The 1914 crisis

Unfortunately, that sophistication created particular vulnerabilities, such as interconnect-

1.2 The concept of systemic risk

1.2.1 Notions of systemic risk

A more comprehensive definition is:

1.2.2 Who is vulnerable?

0% 500% 1000% 1500% 2000%

The United Kingdom

1995 1997 1999 2001 2003 2005 2007 2009

Figure 1.3 UK output index, 1995 Q1 = 100

Banks, bank size and politics

1.2.3 Should we eliminate systemic risk?

1.3 Who creates systemic risk?

1.3.1 Role of financial institutions

1.3.2 Role of the government

1.4 Fundamental origins of systemic risk

1.4.1 Fractional reserve banking and systemic risk

Example 1.1 Fractional reserve banking

$750 10% 30% 50%

Definition 1.2 Externality Externality occurs where cost or benefit accrues to

Financial dif ficulties

Figure 1.6 Firesales and acceleration effects

−1.5% 0.0% 1.5% −25% 0% 25%

Asset growth and leverage growth

16 Failures in risk management and regulations