Assessing the role of reserve requirements under …nancial

frictions
Carlos Montoroy
Bank for International Settlements

Preliminary draft
11 March 2011.

Abstract
The global …nancial crisis has proven that central banks must give …nancial stability
a more prominent role. This requires providing central banks with new instruments and
understanding their general equilibrium implications. This paper extends an otherwise
standard New-Keynesian model to include (i) a banking sector and an interbank market
subject to …nancial frictions in the form of collateral and liquidity contraints in a spirit
similar to Kiyotaki and Moore (2008); (ii) multiperiod credit contracts a-là Benes and Lees
(2010); and (iii) reserve requirements on deposits as a policy instrument that complements
the interest rate. The …rst two components generate an endogenous credit spread and a
credit intermediation that is subject to maturity mismatches. In this setting we evaluate
the role of reserve requirements for …nancial stability and business cycle dynamics. A
key …nding is that reserve requirements can complement monetary policy in stabilizing
the business cycle when the economy is subject to demand shocks, but not under supply
shocks.
JEL classi…cation: E50,E44, G21
Keywords: DSGE, macroprudential, monetary policy, reserve requirements on bank de-
posits, liquidity risk, maturity mismatches.

I thank Ramon Moreno, Liliana Rojas-Suarez and participants at the BIS Research Seminar for useful
comments. Research assistance by Alan Villegas is acknowledged. The views expressed in this article are those
of the author and do not necessarilly re‡ect those of the Bank for International Settlements nor those of the
International Monetary Fund.

y
O¢ ce for the Americas, Bank for International Settlements, Torre Chapultepec - Rubén Darío 281 - 1703,
Col. Bosque de Chapultepec - 11580, México DF – México; tel: +52 55 9138 0294; fax: +52 55 9138 0299;
e-mail: carlos.montoro@bis.org.

1
1 Introduction
The international …nancial crisis has made evident that a single policy objective -low and
stable in‡ation- and a single policy instrument -the interest rate- do not su¢ ce to guarantee
the stability of the …nancial system. Central bank policy objectives and the tools required to
achieve …nancial stability are being re-examined. However, the well-known "Tinbergen princi-
ple" states that at least as many instruments are required as objectives are in place. Therefore
greater emphasis on …nancial stability objectives will require central banks to complement its
main monetary policy instrument, the interest rate, with additional tools.
An instrument that has increasingly been employed in emerging market economies (EMEs)
is the reserve requirements on bank deposits. In Latin America, the central banks of Brazil,
Colombia, and Peru used them as a mechanism to contain credit in boom episodes and to ease
liquidity conditions in local markets in times of stress (Graph 1.1). Thus over the past few
years, reserves requirements on bank deposits allowed central banks to "lean-against-the wind"
during the upswing part of the cycle and bu¤er the abrupt changes in …nancial conditions
during the downswing. Furthermore, they helped maintain monetary policy focused on its
in‡ationary goals in the expansionary phase of the cycle.

Graph 1.1
Policy rates and reserve requirements
Brazil1 Colombia Peru

The vertical line marks 15 September 2008, the date on which Lehman Brothers …led for Chapter 11
bankruptcy protection.
1 E¤ective reserve requirements rates.

Sources: Bloomberg; CEIC; Central Bank of Brazil; Central Bank of Peru.

This paper re-examines the role of reserve requirements as a macroprudential tool in a

dynamic stochastic general equilibrium (DSGE) model. The paper makes important contri-
butions in capturing the macroeconomic and …nancial linkages of the economy by introducing
a banking sector, an interbank market, multi-period loan contracts, and …nancial frictions in
the form of collateral and liquidity constraints. By doing so the model captures relevant …-
nancial features of …nancial markets: retail and wholesale bank funding, …nancial maturity

2
transformation, and endogenous time varying credit spreads. Intuitively, the novel aspect is
that investment is carried out by entrepreneurs that borrow long-term from banks. To satisfy
this credit demand, banks fund themselves either from household deposits or short-term in the
interbank market. Thus, in this setting bank intermediation involves maturity transformation
risk. An additional key element is that banks´ access to interbank funds is subject to …nan-
cial frictions in the form of collateral and liquidity constraints. These frictions give rise to an
endogenous time varying credit spread. This alters signi…cantly the credit channel and the
transmission mechanism of monetary policy, and therefore the business cycle.
From a policy perspective, the contribution of this paper is to highlight that the role of
reserve requirements on bank deposits can act as a macroprudential tool that acts as a bu¤er to
smooth out the procyclicality of the …nancial system. Because of this, they have the potential
to complement the interest rate as a monetary policy tool. Simulation results show that this
is particularly the case for demand shocks but not for supply side shocks.
These two contributions must also be examined in perspective. First, over the past cou-
ple of decades the use of reserve requirements declined rapidly across the world as central
banks were increasingly managing its monetary policy through interest rates. Thus, from a
theoretical perspective its analysis, which was framed as a tax on …nancial intermediation,
gradually lost interest. By contrast, in our paper the emphasis is on reserve requirements as a
macroprudential tool to smooth out the procyclicality of the …nancial system rather than as a
monetary policy tool to achieve an in‡ation objective. Second, modern macroeconomic models
(Dynamic Stochastic General Equilibrium or DSGE models) neglected the role of …nancial
markets and the banking sector1 . Only recently, with the international crisis, have researchers
begun to examine the role of …nancial markets and its linkages with the real economy. Our
paper brings a stylised perspective that that complements the existing literature examining
the role of …nancial factors, such as the …nancial accelerator or the literature on collateral
constraints (See Bernanke, Gertler, and Gilchrist, 1999, Iiacovello, 2005 or Christiano et al,
2007a) Speci…cally, our paper adds to this literature not just in capturing credit or market
risk, but more importantly in understanding the role of liquidity risk. Furthermore, our paper
also contributes to the recent literature that examines the role of …nancial markets and how
they matter for the conduct of monetary policy (See Curdia and Woodford, 2008).
The paper is structured as follows. The next section presents a brief discussion of the
role of reserve requirements as a macroprudential tool. It also reviews the Latin American
experience with reserve requirements. Next, the DSGE model is presented. This is followed by
a brief section that analyses the role of …nancial frictions in the model. Then, we present the
calibration and simulate the model’s response to monetary policy shocks and examine the role
of reserve requirements. This is followed by a robustness analysis to understand the response
of the model to three type of shocks. A shock on monetary policy, to understand the credit
channel and transmission mechanism in the model. And a demand side and a supply side
shock. In analyzing these shocks we also explore the role of …nancial frictions in the model. A
…nal section concludes.
1
For an overview of the state of DSGE literature see Tovar (2009).

3
2 Reserve requirements on bank deposits in emerging economies
The use of reserve requirements on bank deposits as a monetary policy instrument declined
across the world over the past twenty years. This was particularly evident in advanced countries
conducting monetary policy through interest rates. In some cases, reserve requirements were
eliminated all together (e.g. Canada or New Zealand). The declining reliance on reserve
requirements had to do to a large extent with changes in the manner in which central banks
conduct monetary policy operations ( i.e. through open market operation), but also due to
…nancial innovation, and the ine¢ ciencies of taxing the banking sector vis-à-vis other …nancial
intermediaries, which put the banking sector in a clear competitive disadvantage. The role of
reserve requirements was also vili…ed as they became to be considered a de…ning element of
…nancially repressed economies (McKinnon, 1973).
In recent years the perception of this policy tool changed, especially as central banks
begun to focus their attention on systemic risk. However, a greater emphasis on systemic
risk requires a careful examination of the policy toolkit. Possibly new policy instruments
will have to be introduced or existing ones modi…ed to smooth out the …nancial cycle i.e.
reduce the procyclicality of the …nancial system. In this respect, reserve requirements on bank
deposits have features that make them suitable as a macroprudential tool (IMF (2010) and
CGFS (2010)). This is not surprising as they were originally introduced as a …nancial stability
instrument to manage liquidity risk, in a similar fashion to deposit insurance.
Reserve requirements may act as a macroprudential tool for at least three reasons (see IMF,
2010). First, in the upswing part of the business cycle they can help contain credit growth,
e¤ectively acting as a speed limit. Second, during the upswing they build up a cushion of
reserves that can be deployed during the downturn. In this manner, they can smooth out
…nancial liquidity pressures in the …nancial system during bad times, thus acting as a liquidity
bu¤er. Finally, they can be an e¤ective complement to monetary policy when the credit cycle
con‡icts with its own goals. This is best illustrated in episodes of capital surges. In such
context raising interest rates to contain aggregate demand pressures can be self-defeating, as
this may induce more capital in‡ows. Of course, under such circumstances raising reserve
requirements may be e¤ective, not only to contain aggregate demand, but also credit growth.
Certainly, these bene…ts must be weighted against its costs. Reserve requirements are a
blunt instrument and are a tax on bank intermediation. Thus their use may be more appropri-
ate when other traditional and less costly instruments (e.g. monetary policy) are insu¢ cient
to achieve …nancial or price stability (Vargas et al (2010)).
The central banks of Brazil, Colombia and Peru have actively used reserve requirements
over the past few years. The use of this instrument as a macroprudential tool in the context
of capital surges is illustrated by the case of Colombia and Peru (Graph 1, IMF (2010) and
Jara, Moreno and Tovar (2010)). Prior to the global crisis (2007-2008) the central banks of
Colombia and Peru managed reserve requirements as a prudential tool to contain pressures
on credit growth amanating from large capital in‡ows. In this context both central banks
raised marginal reserve requirements during this period. In the case of Peru, marginal reserve
requirements were also imposed on foreign currency deposits. As global …nancial conditions
deteriorated towards the end of 2008 both central banks lowered marginal reserve requirements
as a mechanism to ease liquidity conditions in the economy. The Central Bank of Brazil also

4
used average reserve requirements as a mechanism to prevent disruptions in the interbank
market following Lehman Brothers’ episode in September 2008. Speci…cally, average reserve
requirements were lowered to large and liquid banks if they extended credit to small and illiquid
banks. More recently, as capital in‡ows to emerging markets surged, the Central Bank of Peru
has actively raised marginal reserve requirements again.
From a policy perspective several questions arise. First, under liquidity pressures how
e¤ective are reserve requirements vis-à-vis interest rates? Second, how do reserve requirements
a¤ect the credit channel and the business cycle? These are the main questions that we aim to
address with the model. Given the complex nature of the model, we leave for future research
the extension to an open economy.

3 The model
The model in this section introduces three key features: retail and wholesale bank funding,
…nancial maturity transformation, and endogenous time-varying credit spreads. Speci…cally,
the banking sector in the model intermediates short-term funds from the interbank market or
from household deposits to entrepreneurs that borrow long-term to …nance investment. How-
ever, bank intermediation faces two key …nancial frictions: liquidity and collateral constraints.
This is relevant because they generate an endogenous credit spread.
The model departs from recent attempts to model …nancial markets and …nancial Frictions
in the DSGE literature. Indeed, a common approach is to model …nancial frictions follow-
ing the …nancial accelerator framework (Bernanke and Gertler,1999). This framework relies
on an optimal debt contract problem with costly state veri…cation. As a result information
asymmetries between borrowers and lenders endogenously amplify credit market developments
and generate an external …nance premium that depends inversely on the debt to net worth
ratio. The model also departs from recent e¤orts at modeling the banking sector, in the sense
that in our model banks fund themselves in the retail and wholesale market. Furthermore, we
introduce long-term loan contracts a-la Benes and Lees (2010), so that banks activity is char-
acterized by a maturity transformation problem. This implies that the credit channel takes
into account expectations about future real interest rate payments, which generates a maturity
risk and endogenously a wedge between lending rates and the interbank rate. Linking interest
rates with long-interest rates also smoothes out ‡uctuations of interest rates. This generates
a degree of interest rate persistence that has so far been modeled through a monopolistic
competitive banking sector with interest rate adjustment costs (e.g. Gerali et al, 2009).
In our model …nancial frictions also play an important role. Speci…cally, bank intermedia-
tion faces collateral and liquidity constraints. The use of collateral constraints to amplify the
credit cycle is now commonly employed (Iacovello, 2005). However, rather than relying of a
durable good asset (e.g. a durable good), our collateral is the banks …nancial assets. In that
sense our problem takes into account …nancial institutions leverage positions. Our framework
follows the recent contributions by Kiyotaki and Moore (2008) and Del Negro et al (2010). This
framework generates a demand for liquid assets. However, we depart from these approaches
by applying this problem to the …nancial sector rather than to a …nancing problem for entre-
preneurs. A key aspect is that we are able to consider the role of core funding in the model

5
and generate endogenous credit spreads that arise from the bank intermediation process.
By making the credit channel more realistic, the role of reserve requirements becomes less
trivial. The reason is that changing reserve requirements eases the constraints imposed by the
…nancial frictions, which in turn creates a more complex interaction with the real economy. In
this respect, the paper also adds to the literature that examines how credit spreads between
borrowers and lenders a¤ect the credit channel and the monetary transmission mechanism
(Curdia and Woodford, 2008).
Overall, aside of these complexities associate to the …nancial market, the economy corre-
sponds to the standard New Keynesian model with capital, as in Woodford (2003).

3.1 Households
We assume the following utility function on consumption (Cth ) and labour (Ht ) of the repre-
sentative household
1
" 1
#
X h
Ct+s 1+v
Ht+s
s
Ut = Et ; (3.1)
1 1+v
s=0
where represents the coe¢ cient of risk aversion and v captures the inverse of the elasticity of
labour supply. The optimiser consumer takes decisions subject to a standard budget constraint
which is given by
Wt Ht Bt 1 Dt 1 1 Bt Dt t Tt
Cth = + + RtD 1 + + ; (3.2)
Pt Pt Pt R t Pt Pt Pt Pt
where Wt is the nominal wage, Pt is the price of the consumption good, Bt is the end of period
nominal bond holdings, Dt is the end of period stock of deposits in the bank system, Rt is the
riskless nominal gross interest rate paid by the bonds, RtD 1 is nominal interest rate paid by the
deposits made in t 1, t is the share of the representative household on total nominal pro…ts,
and Tt are net transfers from the government. The …rst order conditions for the optimising
consumer’s problem are:
" # " #
Pt Ct+1 D Pt Ct+1
1 = Et Rt = Et Rt ; (3.3)
Pt+1 Ct Pt+1 Ct

Wt
= Cth Htv : (3.4)
Pt
Equation (3:3) is the standard Euler equation for both bonds and deposits that determines
the optimal path of consumption . At the optimum the representative consumer is indi¤erent
between consuming today or tomorrow, and consumers are indi¤erent between bonds and bank
deposits if RtD = Rt . Equation (3:4) describes the optimal labour supply decision. We assume
that labour markets are competitive and also that individuals work in each sector z 2 [0; 1].
Therefore, Ht corresponds to the aggregate labour supply:
Z 1
Ht = Ht (z)dz: (3.5)
0

6
3.2 Firms
3.2.1 Final good producers
There is a continuum of …nal good producers of mass one, indexed by f 2 [0; 1] that operate
in an environment of perfect competition. They use intermediate goods as inputs, indexed by
z 2 [0; 1] to produce …nal consumption goods using the following technology:
Z 1
"
" 1
" 1
Ytf = Yt (z) " dz ; (3.6)
0

where " is the elasticity of substitution between intermediate goods. The demand function of
each type of di¤erentiated good is obtained by aggregating the input demand of …nal good
producers:
Pt (z) "
Yt (z) = Yt ; (3.7)
Pt
where the price level is equal to the marginal cost of the …nal good producers and is given by:
Z 1
1
1 "
1 "
Pt = Pt (z) dz : (3.8)
0

and Yt represents the aggregate level of output.

Z 1
Yt = Ytf df: (3.9)
0

3.2.2 Intermediate goods producers

There is a continuum of intermediate good producers indexed by z 2 [0; 1]. Each …rm z pro-
duces an intermediate good using capital and labour. These …rms operate under monopolistic
competition in the intermediate goods market and demand production factors in competitive
markets. They use the following Cobb-Douglas production function:

Yt (z) = At [Kt 1 (z)] [Ht (z)]1 ; (3.10)

These …rms take as given the real wage, Wt =Pt , and the rental rate of capital, RtK . They
allocate the optimal level of capital and labour minimising costs given the technology. The
demands for both production factos are given by:

M Ct
Htd (z) = (1 ) Yt (z): (3.11)
Wt =Pt
M Ct
Ktd 1 (z) = Yt (z): (3.12)
RtK
After replacing the demand for each factor, we can obtain an expression for the real marginal
cost:
1 RtK Wt =Pt 1
M Ct (z) = ; (3.13)
At 1

7
where M Ct (z) represents the real marginal cost for each intermediate …rm. Notice that mar-
ginal costs are the same for all intermediate …rms, since technology has constant returns to
scale and factor markets are competitive, ie M Ct (z) = M Ct .
Intermediate producers set prices following a staggered pricing mechanism a la Calvo. Each
…rm faces an exogenous probability of changing prices given by (1 ). A …rm that changes
its price in period t chooses its new price Pt (z) to maximise:
1
X
s
Et t;t+s (Pt (z); Pt+s ; M Ct+s ; Yt+s ) ;
s=0

s Ct+s Pt
where t;t+s = Ct Pt+s is the stochastic discount factor. The function:
"
int (P [Pt (z) Pt M Ct ] PtP(z)
t (z); Pt ; M Ct ; Yt ) t
Yt is nominal pro…ts of the supplier of
good z with price Pt (z); where the aggregate demand and aggregate marginal costs are equal
to Yt and M Ct ; respectively. The optimal price that solves the …rm’s problem is given by
P
1
s "+1
Et t;t+s M Ct;t+s Ft+s Yt+s
Pt (z) s=0
= ; (3.14)
Pt P
1
s "
Et t;t+s Ft+s Yt+s
s=0
"
where " 1 is the price markup, Pt (z) is the optimal price level chosen by the …rm and
Pt+s
Ft+s = Pt the cumulative level of in‡ation. The optimal price solves equation (3:14) and is
determined by the average of expected future marginal costs as follows:
"1 #
Pt (z) X
= Et 't;t+s M Ct;t+s ; (3.15)
Pt
s=0

s "+1
t;t+s Ft+s Yt+s
where 't;t+s P
1 .
s "
Et t;t+s Ft+s Yt+s
s=0
Since only a fraction (1 ) of …rms changes prices every period and the remaining one
keeps its price …xed, the aggregate price level, de…ned as the price of the …nal good that
minimise the cost of the …nal goods producers, is given by the following equation:

Pt1 "
= Pt1 1
"
+ (1 ) (Pt (z))1 "
: (3.16)

Following Benigno and Woodford (2005), equations (3:14) and (3.16) can be written recur-
sively introducing the auxiliary variables N Nt and DDt :
1 "
" 1 N Nt
( t) =1 (1 ) ; (3.17)
DDt
h i
" 1
DDt = Yt (Ct ) + Et ( t+1 ) DDt+1 ; (3.18)
"
N Nt = Yt (Ct ) M Ct + Et [( t+1 ) N Nt+1 ] ; (3.19)

8
where t = Pt =Pt 1 is the gross in‡ation rate. Equation (3:17) comes from the aggregation of
individual …rms prices. The ratio N Nt =DDt represents the optimal relative price Pt (z) =Pt :
These three last equations summarise the recursive representation of the non linear Phillips
curve.

3.2.3 Capital goods producers

Capital producers transform consumption goods into capital goods and operate in perfect
competition. The capital produced is sold to the entrepreneurs at price Qt in terms of units
of consumption. The cost function is given by the following equation:

It
It 1 + (3.20)
I
Capital producers also need to consider that entrepreneurs’ demand for investment is con-
strained by the amount of credit they can receive from banks, as explained below.

3.3 Entrepreneurs
Entrepreneurs buy investment goods from the capital producers (at price Qt ) and accumulate
capital.
Kt = It + (1 ) Kt 1 (3.21)
They rent capital to intermediate goods producers at rate RtK . The rate of return of the
entrepreneurs is given by:
1
RtQ = RK + (1 ) Qt (3.22)
Qt 1 t
which is given by the rental rate of capital and the price of capital net of the depreciation rate,
all divided by the initial price of capital, Qt 1 .
Entrepreneurs …nance enterely its investment from the banks. The value of investment
each period is equal to the ‡ow of credit from the banks:
CRt
Qt It = (3.23)
Pt
where CRt is the nominal amount of of credit of each period.
We assume that entrerpreneurs take credit in a multiperiod contract with …xed repayments
as in Bennes and Lees (2010). A credit CRt taken at time t is paid back in repayments
proportional to the amount borrowed and decay at a …xed rate 2 (0; 1) : This implies the
k 1 cr
following repayment schedule: Qcr cr
t CRt , Qt CRt , :::, Qt CRt , ::: for repayments due at
cr
t + 1; t + 2; :::; t + k; respectively. Where Qt is a shadow price associated to the credit. This
form of multiperiod contract allows us to draw out the implications of long-term debt and
maturity mismatch in an analytically tractable framework.
Accordingly, the sum of all repayments due at t associated with all past loans is:
1
X
k 1
Jt 1 = Qcr
t k CRt k (3.24)
k=1

9
 Also, Jt can be written recursively as:

J t = Jt 1 + Qcr
t CRt (3.25)

The stock of loans, estimated as the present value of repayments, is equal to:

Lt = t Jt (3.26)
h i
1 1 k 1 1
where t = Et Rt + Rt Rt+1 + ::: + Rt :::Rt+k 1
+ ::: ; which can be written recursively
as:
1
t = Et (1 + t+1 ) (3.27)
Rt
Considering equation (3.25), the evolution of the stock of loans can be written as:

t cr
Lt = Lt 1 + t Qt CRt : (3.28)
t 1

The optimal condition for the entrepreneurs equalise the expected return of capital with the
real value of the repayments, given by:
0 1
1 + Qcr
Q @ t+1 A
Et Rt+1 = Et Qcr
t : (3.29)
t+1

See appendix A:2 for details on the derivation of equation (3.29).

3.4 The interbank market

To model the interbank bank market we assume that the economy is populated by a continuum
of banks b 2 (0; 1), whose objective is given by
1
X
s
max Et b ln [Ct+s (b)] (3.30)
s=0

where Ct (b) is the bankers consumption. Each period commercial banks have a random lending
opportunity. Speci…cally, this implies that the stock of loans (Lt (b)) supplied by each bank
evolves according to:
(
t
t 1
Lt 1 (b) + t Qcr
t CRt (b) with probability
Lt (b) = (3.31)
t
Lt 1 (b) with probability 1
t 1

where is the rate that loan repayments decay, as explained above. CRt (b) is the the amount
of new credit extended by each bank. We also assume that b > . Lending opportunities
are i.i.d. across time and banks. Banks lend to entrepreneurs with multiperiod contracts with
…xed repayments as explained above. The main problem these agents face is that they may
not have enough resources to take advantage of the lending opportunity. This forces them to
seek funding in the interbank market to take advantage of it.

10
 Table 3.1: Banks’balance sheet
Assets Liabilities
Loans, Lt (b) Interbank borrowing, ZtB (b)
L
Interbank lending, Zt (b) Deposits , Dt (b)
Reserves, RRt (b) net worth, Zt (b) Dt (b) + RRt (b) + Bt (b)

Access to the interbank market is going to be subject to a number of …nancial imperfections,

which are modelled, similarly as in Del Negro et al (2010), as constraints on the evolution
of the banks’balance sheets. The banks liabilities consist of the interbank borrowing, ZtB (b),
and the deposits received from households, Dt (b) :Assets consist of the claims on its own
loans, Lt (b) ;and on loans extended in the interbank market, ZtL (b) :In addition, banks must
hold some unremunerated reserves at the central bank, RRt (b). We assume that there is a
minimum amount of reserves that must be held at the central bank without any remuneration.
However the central bank may also decide to o¤er banks the possibility of holding reserves at
a discount over the rate …xed by the central bank. The bank balance sheet is then summarised
in Table 3.1, where Zt (b) = ZtL (b) + Lt (b) ZtB (b) .
The constraints on the evolution of the banks balance sheet are a collateral contraint (CC)
and a leverage constraint (LC) . The CC implies that banks can only fund in the interbank
market a fraction t of the amount of credit extended, CRt (b). The LC implies that in any
given period a bank can only borrow from the interbank market a fraction t of its claims on
its own loans, Lt (b) :
These two constraints taken together imply that the evolution of interbank borrowing is
subject to the following innequality:

ZtB (b) L (b) + CR (b) (3.32)

| t t{z1 } | t {zt }
LC CC

Note that LC and CC are similar in spirit, the former is a constraint on existing loans,
while the later applies to new loans.
The inter-temporal budget constraint of the banks can then be summarised by:

Pt Ct (b) + CRt (b) + ZtL (b) ZtB (b) Jt 1 + Xt + Rtib 1 ZtL 1 (b) ZtB 1 (b) (3.33)

In the …rst line we have the use of funds of the banks, which can be for consumption to give
new credit or for interbank lending. In the second line we have the sources of funds, which can
be from their return on loans claims, Jt 1 , and, the ‡ow of deposits from the consumers net of
the ‡ow of reserve requirements, Xt , and from repayments of the interbank loans considering
the gross interes rate Rtib . We de…ne Xt as:

Xt = Dt (b) Rt 1 Dt 1 (b) RRt (b) RRt 1 (b) (3.34)

From equation (3.34) we can see that reserve requirements a¤ect the quantity of disposable
resources that banks have to …nance entrepreneurs.

11
3.5 Monetary policy
The central bank sets the risk-free interest rate, which in turn is the same paid by the banks
for the deposits because of the arbitrage condition of the households. We assume the central
bank follow a simple Taylor rule that depends on current in‡ation:

t
Rt = R : (3.35)

Also, the central bank sets the reserve requirement rate, as a fraction of the deposits
received from the households, that is:

RRt = t Dt : (3.36)

To make a comparative analysis, we assume the central bank can follow the next rules:
(
t = CRt (3.37)
+ cr CR

It can maintain …xed at some level ( ) or can either adjust it with respect to deviations of
credit with respect to its steady state value. .
Moreover, the central bank balance sheet is given by:

(RRt RRt 1) (Nt Nt 1) = 0; (3.38)

Reserve requirements correspond to a liability to the central bank, changes in the reserve
requirements need to be matched with the change in the net worth (Nt ) of the central bank2 .

3.6 Market clearing

In equilibrium labour, intermediate and …nal goods markets clear. The economy-wide resource
constraint is given by
Yt = C t + I t : (3.39)
The labour market clearing condition is given by:

Ht = Htd ; (3.40)

where the demand for labour comes from the aggregation of individual intermediate producers
in the same way as for the labour supply:
Z 1 Z 1
d d M Ct
Ht = Ht (z)dz = (1 ) Yt (z)dz (3.41)
0 Wt =Pt 0
M Ct
= (1 ) Yt t ;
Wt =Pt
2
We are not modelling explicitely the monetary base, therefore any analysis of the balance sheet of the central
bank can be done only partially. A natural extension of this model will be to include the monetary base in the
analysis.

12
 R1 "
where t = 0 PtP(z) t
dz is a measure of price dispersion.
Similarly, aggregate capital demand equals:
M Ct
Ktd 1 = Yt t (3.42)
RtK

Also, aggregate credit is equal to the proportion of banks that give credit times individual
credit:

CRt = CRt (b) (3.43)

Total consumption is equal to:
Ct = Cth + Ctb (3.44)
R
where Ctb = b Ct (b) : In equilibrium, the demand for bonds equals zero (Bt = 0) and interbank
borrowing equals its corresponding lending (ZtL = ZtB ).

4 Financial frictions in the interbank market

The interest rate adjusted by reserve requirements is de…ned by:

Rt t
Rt = > Rt ; (4.1)
1 t
where Rt includes the intermediation mark-up generated by reserve requirements. Reserve
requirements act as a tax on the banking system increasing the intermediation cost. Moreover,
the …rst order conditions for the banks determine that the interest rate paid by interbank
lending equals the reserve requirement adjusted-interest rate, that is Rtib = Rt , which is higher
than the risk-free rate. Furthermore, the …rst order conditions for the bank with a lending
opportunity determines the shadow price associated to the credit paid in multiperiod contracts:

1
X Pt+s Pt
't bL (1 t ) = Et
s
't+s+1 Qcr
t Et 't+1 Rtib t (4.2)
Pt+s+1 Pt+1
s=0
1
X Pt+s
+ Et ( )s 't+s+1 bL t+s+1 t+s Qcr
t
Pt+s+1
s=0

where 't bj for j = fL; N Lg is the Lagrange multiplier of the budget constraint for
a bank with and without a lending opportunity, respectively, and 't+1 = 't+1 bL +
(1 ) 't+1 bN L is the average Lagrange multiplier. The term on the LHS is the marginal
cost of the use of funds in terms of consumption, which equals the marginal bene…t for having
more income in the future because of larger loan stock net of the cost for interest payments
plus the expected bene…t of being able to borrow more in the interbank market because of
larger loan stock, respectively. According this condition, if situations in the interbank market
are tighter, that is lower t+1 or t , Qcrt needs to increase to maintain the equality in this

13
condition. Also, if the probability of having a lending oportunity ( ) is smaller, the expected
bene…t of being able to borrow in the interbank market next period is smaller, and Qcr t needs
to be higher. This implies a larger premium on the interest rate paid by the entrepreneurs.
After considering other optimality conditions, equation (4.2) can be simpli…ed to:

Rt
Qcr
t = (4.3)
t + t
h i h i
where t Et R1 1 + t+1 and t Et Rt t+1 t + t+1 are recursive rep-
t
resentations of the present values terms in equation (4.2). Note that from this condition, if
cr
t = 0; Qt can be lower than the risk-free rate because of the multiperiod contract. Also, the
parameter asociated to the colateral constraint t does not a¤ect the risk premium. The risk
premium of the interest rate paid by the entrepreneurs depends on the tightening conditions of
the interbank market. When the leverage constraint are tighter, that is is lower, the spread
between the interest rate paid by the entrepreneurs and the risk free rate becomes higher.
Moreover, the supply of credit comes from the budget constraint of the banks with a lending
opportunity, after considering the solution for banks’consumption. The supply of credit give
by the banks is the following:

b 1
CRt = + t Lt 1 Rt 1 t 1 Dt 1 + (1 t ) Dt ; (4.4)
(1 t) t 1

where reserve requirements a¤ect the resources available for credit. They a¤ect directly
the amount of credit supplied by the banks and it can be an additional channel that monetary
policy can work. Also, the colateral constraint a¤ects directly the supply of credit.
The equilibrium in the credit market is given by the intersection of the credit supply
(equation 4.4) and the credit demand (equation 3.29 after replacing in it equations 4.3, 3.12,
3.21 and 3.23). Credit supply is perfectly inelastic with respect the shadow price of credit
(Qcr cr
t ) while the credit demand is downward slopping with respect to Qt : Graph 4.1 shows the
e¤ects in the model of changes in the …nancial frictions and the reserve requirement rate. A
more restricted collateral constraint (lower ) reduces the supply of credit, because banks can
…nance a lower proportion of credit, leaving the demand of credit unchanged (Graph 4.1 (a)).
On the other hand, with a more restricted leverage constraint (lower ) banks can borrow a
lower fraction of its assets, leaving the supply of loans unchanged, but increasing the cost of
credit moving upwards the demand for credit (Graph 4.1 (b)). Moreover, an increase in the
reserve requirement rate (higher ) reduces the credit supply, because of the lower ‡ow of funds
available to banks, and increases the interest rate paid by the entrepreneurs, moving upwards
the demand for credit (Graphs 4.1 (c)). Notice that both a reduction in or an increase in
move upwards the credit demand in the same way a tax increases the price of goods.

14
 Graph 4.1
Credit market equilibrium: comparative statics
# # "
QtCR CRtS QtCR QtCR CRtS
CRtS

QCR QCR
t1 QCR t1
t1
QCR QCR
t0 CRtD QCR CRtD
t0
CRtD
t0

CRt1 CRt0 CRt CRt = CRt CRt

0 1
CRt1 CRt0 CRt

(a) (b) (c)

Collateral constraint Leverage constraint Reserve requirement rate

5 Simulation results
To simulate the model we …rst calibrate the parameters such that generate reasonable values
of the steady state (Table 5.1). The discount factor for households ( ) is calibrated such that
the annualised risk-free rate is 4%. In the calibration is important to consider a low parameter
for the discount factor of bankers ( b ), for having a positive steady state level of deposits
(see appendix for more details). The probability of a bank having a lending opportunity is
calibrated at 0.5. The parameters associated to the leverage constraint ( ) and the colateral
constraint ( ) are set in 0:9 and 0:8, respectively, which are equivalent to a leverage ratio of 10
and a loan-to-value ratio of 0:8. The elasticity of substitution between goods (") is calibrated
such that the mark-up of the intermediate goods producers is 15%: Moreover, we assume that
investment depreciate in 10 years, which generates a quarterly depreciation rate of 2:5%: The
deacay rate of the credit repayments is calibrated at 0:8, which imply a Macaulay’s duration
of 1:5 years. The reserve requirement rate is calibrated at 20%, a median value of the reserve
requirements in the Latin American region.

Table 5.1
Baseline calibration
= 0:99 = 1:15 = 0:25
b = 0:8 = 0:5 = 0:8
= 0:5 =1 = 0:8
= 0:9 v=1 = 0:2

According our benchmark calibration, the ratio credit over GDP in steady state (Table 5.2)
is 9:5%, which is small because of the e¤ects of …nancial frictions and reserve requirements in
the model. If both the collateral and the leverage constraint were absent together with the

15
reserve requirement, under the same calibration the credit over GDP ratio would be around
20%. The …nancial constraints and the reserve requirement make the return of capital (RQ ),
which is 1:090 in steady state, higher than the risk-free real interest rate R. Then, the stock of
capital (together with investment and credit) is smaller to generate a higher return in steady
state.

Table 5.2
Steady state - baseline calibration
CR=P L=P
Y = 0:095 Y = 0:654
I
Y = 0:095 Qcr = 0:290
C
Y = 0:905 R = 1:010
D=P
Y = 0:361 RQ = 1:090

5.1 Reserve requirements and the steady state

Reserve requirements generate a distortion in steady state, because it increases the cost of credit
(equation 4.3) as a tax in the banking system. However, in this model reserve requirements
do not a¤ect the supply of credit in steady state (equation 4.4) because it doesn’t a¤ect the
‡ow of funds of banks (X) in steady state. Since the steady state is an stationary equilibrium,
deposits in real terms are constant and so there is no change in the level of reserve requirements
in steady state3 .
Graph 5.1 shows the e¤ects in the steady state of the model of considering di¤erent values
of the reserve requirement rate. The red line corresponds to the baseline calibration. A
higher reserve requirement increases exponentially the interest rate paid by credit (panel d),
which reduces the level of GDP (panel b). Moreover, it also changes the composition of GDP,
reducing investment (and credit) while consumption increases. That is, the higher cost of
…nancing crowds-out investment.
3
However, these result will change if we consider growth in the model. Reserve requirement will bring another
distortion through the credit supply.

16
 Graph 5.1
Steady state of the model1
Credit / GDP GDP
1 1

0.9
0.8
0.8

0.6 0.7

0.6
0.4
0.5

0.2 0.4
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
τ (a) τ (b)

Consumption / GDP Credit interest rate

1.12 3

1.1
2.5
1.08

1.06 2

1.04
1.5
1.02

1 1
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
τ (c) τ (d)
1 Steady state values normalised to the case of zero reserve requirements (eg value= 1 when = 0)

5.2 E¤ects of reserve requirements as a monetary policy instrument

The central bank can use actively reserve requirements as a monetary policy instrument in
addition to the interest rate. When doing this, it a¤ects the transmission mechanism of the
interest rate and the macroeconomic response to di¤erent shocks.

Monetary policy transmision mechanism:

To analyse how the transmission mechanism of monetary policy changes, we analyse …rst
how is the response of an exogenous change in the monetary policy interest rate. Graph 5.2
shows the impulse responses to an iid shock to the Taylor rule (equation 3.35) when the reserve
requirement rate is maintained …xed. Under the baseline calibration (red line), an increase in
the interest rate has the traditional e¤ects: it contracts household consumption and increases
the interest rate paid by credit. Moreover, it reduces both investment and credit with a lag.

17
There is also a small increase in investment and credit in the …rst period, generated by the
initial reduction in the price of investment.
The duration of the multiperiod credit contract has also implications for the transmission
mechanism of monetary policy. A higher parameter implies a slower decay of the repayments
and a higher duration of the loan. When increasing from 0:8 to 0:95 (the blue line), the
impact of an interest rate shock on both investment and credit is reduced. This is because an
incraese in the interest rate reduces the loans stock (as the present value of the payments gets
smaller), reducing also the supply of credit because of the leverage constraint. As the duration
of the loans gets higher, the e¤ect of a transitory increase of the interest rate today is smaller,
and reduces less the supply of credit. Then, monetary policy looses power.
On the other hand, a tighter leverage constraint in steady state increases the impact on
both investment and credit. A reduction in from 0:9 to 0:3 (the green line), makes the e¤ect
of the interest rate on the supply of credit more important. However, also note that the e¤ect
(pass-through) on the credit interest rate is unchanged.

Reserve requirement transmission mechanism:

In graph 5.3 we analyse the e¤ect of an exogenous increase in the reserve requirement
rate of 10% with some persistence (red line). This increases the credit interest rate, which
reduces credit and investment. Moreover, the reduction in GDP also decreases household
consumption. As in‡ation decreases, the interest rate also falls. However, after one period
there is some recovery in investment and GDP generated by the decrease in the price of
investment. Furthermore, a tighter leverage constraint (blue line) reduces the impact of reserves
requirements on credit. This e¤ect is opposite to the one on the interest rate shock.
How is a¤ected the transmission mechanism of monetary policy is a¤ected when
we use reserve requirements?
We compare the di¤erent responses to shocks maintaining …xed the reserve requirement
rate with those where the reserve requirement rate follows a feedback rule. Since an objective
of macropudential policy is to moderate the …nancial cycle, in order to limit the likelihood of
…nancial stress, we consider a policy rule for the reserve requirements rate that responds to
deviations of credit from its steady state (equation 3.37).
Graphs 5.4 shows the impulse responses to a shock in the aggregate demand equation
(3.39). Given the size and persistence calibrated for this shock, it implies a implies a maximum
increase in the reserve requirement rate of 15%. Also, the increase in credit is around 1/3 than
in the counterfactual case without the use of reserve requirement. Moreover, the transmission
mechanism of the interest rate is improved. The interest rate has to increase a half than in
the counterfactual case and in‡ation is controlled better.
However, it is not necessarily the case that using a feedback rule for the reserve requirement
rate always improves the power of the interest rate. It will depend on the source of the shock.
For instance, graph 5.5 shows the impulse responses to a productivity shock. In this case
both credit and investment increases, and in‡ation decreases. A reserve requirement rule on
credit moderates credit growth, however since this generates a negative output gap (aggregate
demand grow less than aggregate supply), in‡ation decreases more and the interest rate should
have a higher response to control.

18
 How about the response to shocks that a¤ect the …nancial constraints?
Our framework allows also analysing what happens when the …nancial conditions change.
For example, in the case of a …nancial crisis …nancial conditions get tighter, which in our setup
will imply an increase in the risk premium paid by credit. In our model we have two types of
…nancial constraints, the leverage constraint and the collateral constraint. As shown in graph
5.6, an exogenous reduction in the fraction t that a bank can borrow from its own loans imply
an increase in the cost of investment, reducing both credit and GDP. The endogenous response
of the interest rate partially o¤set this e¤ect, however even when the credit interest rate is
is reduced, it takes time for credit to recover. On the other hand, an endogenous reduction
of 15% in the reserve requirement rate (blue line) complements the response of the interest
rate, making credit and other macroeconomic variables more stable. Moreover, the outcome of
using a reserve requirement rate feedback rule is similar in the case of an exogenous reduction
in the fraction t that banks can fund from the credit (graph 5.7 ).

19
 Graph 5.2
Impulse responses to a monetary policy shock1
Credit Investment Investment price
0.5 0.5 0.5

0 0 0

-0.5 Baseline -0.5 -0.5

-1 higher λ -1 -1
lower φ
-1.5 -1.5 -1.5
2 4 6 8 2 4 6 8 2 4 6 8

GDP Household Consumption Credit interest rate

0.2 0.2 1.5

0 0 1

-0.2 -0.2 0.5

-0.4 -0.4 0

-0.6 -0.6 -0.5

2 4 6 8 2 4 6 8 2 4 6 8

Policy interest rate Inflation

0.6 0.4

0.4 0.2

0.2 0

0 -0.2

-0.2 -0.4
2 4 6 8 2 4 6 8
1 Shock iid with 1% standard deviation. Sensitivity analysis with =0.95 and =0.3.

Graph 5.3
Impulse responses to an increase in the reserves rate1
Credit Investment Investment price
0.4 0.4 0.02
Baseline
0.2 lower φ 0.2
0
0 0
-0.02
-0.2 -0.2

-0.4 -0.4 -0.04

2 4 6 8 2 4 6 8 2 4 6 8

GDP Household Consumption Credit interest rate

0.01 0.01 0.06

0 0.04
0
-0.01 0.02
-0.01
-0.02 0

-0.03 -0.02 -0.02

2 4 6 8 2 4 6 8 2 4 6 8

Policy interest rate Inflation Reserves rate

0.02 0.01 0.2

0.01 0.005 0.15

0 0 0.1

-0.01 -0.005 0.05

-0.02 -0.01 0
2 4 6 8 2 4 6 8 2 4 6 8
1 Shock with autocorrelation coe¢ cient of 0.5 and standard deviation of 0.1.

Sensitivity analysis with =0.3.

20
 Graph 5.4
Impulse responses to a demand shock1
Credit Investment Investment price
0.4 0.4 0
Baseline
0.3 RR rule on credit 0.3 -0.05

0.2 0.2 -0.1

0.1 0.1 -0.15

0 0 -0.2
2 4 6 8 2 4 6 8 2 4 6 8

GDP Household Consumption Credit interest rate

0.04 0.02 0.2

0.03 0 0.15

0.02 -0.02 0.1

0.01 -0.04 0.05

0 -0.06 0
2 4 6 8 2 4 6 8 2 4 6 8

Policy interest rate Inflation Reserves rate

0.06 0.03 0.2

0.15
0.04 0.02
0.1
0.02 0.01
0.05

0 0 0
2 4 6 8 2 4 6 8 2 4 6 8
1 Shock with autocorrelation coe¢ cient of 0.8 and standard deviation of 0.05.

Graph 5.5
Impulse reponses to a productivity shock1
Credit Investment Investment price
0.2 0.15 0.8
Baseline
0.15 RR rule on credit 0.1 0.6

0.1 0.05 0.4

0.05 0 0.2

0 -0.05 0
2 4 6 8 2 4 6 8 2 4 6 8

GDP Household Consumption Credit interest rate

0.2 0.4 0

0.15 0.3 -0.1

0.1 0.2 -0.2

0.05 0.1 -0.3

0 0 -0.4
2 4 6 8 2 4 6 8 2 4 6 8

Policy interest rate Inflation Reserves rate

0 0 0.15

-0.01 0.1
-0.02
-0.02 0.05
-0.04
-0.03 0

-0.06 -0.04 -0.05

2 4 6 8 2 4 6 8 2 4 6 8
1 Shock with autocorrelation coe¢ cient of 0.9 and standard deviation of 0.2.

21
 Graph 5.6
Impulse responses to a shock to the LC ( )1
Credit Investment Investment price
0.2 0.2 0.06

0 0 0.04

-0.2 -0.2 0.02

-0.4 Baseline -0.4 0

RR rule on credit
-0.6 -0.6 -0.02
2 4 6 8 2 4 6 8 2 4 6 8

GDP Household Consumption Credit interest rate

0 0.03 0.02

0.02 0
-0.01
0.01 -0.02
-0.02
0 -0.04

-0.03 -0.01 -0.06

2 4 6 8 2 4 6 8 2 4 6 8

Policy interest rate Inflation Reserves rate

0.01 0.01 0.05

0 0
0
-0.01 -0.05
-0.01
-0.02 -0.1

-0.03 -0.02 -0.15

2 4 6 8 2 4 6 8 2 4 6 8
1 Shock with autocorrelation coe¢ cient of 0.5 and standard deviation of 0.1.

Graph 5.7
Impulse responses to a shock to the colateral constraint ( )1
Credit Investment Investment price
0.4 0.4 0.06
Baseline
0.2 RR rule on credit 0.2
0.04
0 0
0.02
-0.2 -0.2

-0.4 -0.4 0
2 4 6 8 2 4 6 8 2 4 6 8

GDP Household Consumption Credit interest rate

0.01 0.02 0.02

0
0 0.01
-0.02
-0.01 0
-0.04

-0.02 -0.01 -0.06

2 4 6 8 2 4 6 8 2 4 6 8

Policy interest rate -3 Inflation Reserves rate

x 10
0.01 5 0

0 -0.05
0
-5 -0.1
-0.01
-10 -0.15

-0.02 -15 -0.2

2 4 6 8 2 4 6 8 2 4 6 8
1 Shock with autocorrelation coe¢ cient of 0.5 and standard deviation of 0.25.

22
6 Conclusions
This paper has presented a DSGE model with a banking sector, an interbank market, multi-
period loan contracts and …nancial …rctions in the form of collateral and liquidity contraints.
As a results banks become engaged in a maturity transfomation activity in which they are
able to fund themselves from retail deposits or the wholesale market and lend long-term to
entrepreneurs. Collateral and liquidity constraints introduce accelerator type features that
generate procyclical credit cycles and introduce a source of liquidity shortages in the economy.
In this setting we examine the role of reserve requiremnts, not as a monetary policy tool
but as a macroprudential tool to handle liquidity problems in the economy. In this setting
we examine how the credit channel and the monetary transmission mechanism change. A key
…nding is that reserve requirements can complement monetary policy in stabilizing the business
cycle when the economy is subject to demand shocks, but not under supply shocks.

23
References
[1] Benes, Jaromir and Kirdan Lees (2010), "Multi-period …xed-rate loans, housing and mon-
etary policy in small open economies", forthcoming International Journal of Central Bank-
ing.
[2] Bernanke, B, M Gertler and S Gilchrist (1999). “The Financial Accelerator in a Quanti-
tative Business Cycle Framework” in J Taylor and M Woodford (eds) Handbook of
[3] Macroeconomics. Volume 1C, North Holland.
[4] Betancourt, Y, N Rodríguez, C Varela and H Vargas (2010): “E¤ects of reserve require-
ments in an in‡ation targeting regime: the case of Colombia”, mimeo, Bank of the Re-
public (Colombia).
[5] Christiano, L., Eichenbaum, M. and Evans, C. (2005): “Nominal Rigidities and the Dy-
namic E¤ects of a Shock to Monetary Policy,”Journal of Political Economy 113,pp. 1-45.
[6] Chrisitiano, L, R Motto and M Rostagno (2007a). “Financial Factors in Business Cycles”,
Mimeo, European Central Bank.
[7] Curdia, Vasco and Michael Woodford (2008). “Credit frictions and optimal monetary
policy”, mimeo Columbia University.
[8] Del Negro, M., G. Eggertsson, A. Ferrero and N. Kiyotaki (2009), “The Great Escape? A
Quantitative Evaluation of the Fed’s Non-Standard Policies”, mimeo.
[9] Gerali, A, Stefano Neri, Luca Sessa, and Federico Signoretti (2009), “Credit and banking
in a DSGE model”, mimeo Bank of Italy.
[10] Iacovello, M (2005). “House Prices, Borrowing Constraints and Monetary Policy in the
Business Cycle”, American Economic Review, vol 95, pp 739–764.
[11] International Monetary Fund (2010): Regional Economic Outlook: Cool in the North,
Warming in the South. Western Hemisphere Department.
[12] Jara, A and C Tovar (2008): “Monetary and …nancial stability implications of capital
‡ows in Latin America and the Caribbean”, BIS Papers, no 48. Moreno, R (2006): “The
changing nature of risks facing banks”, BIS Papers, no 28, pp 67-98.
[13] Kiyotaki, N. and J. Moore (2008), “Liquidity, Business Cycles, and Monetary Policy”,
mimeo.
[14] Smets, F and R Wouters (2003). “An Estimated Stochastic Dynamic General Equilibrium
Model of the Euro Area”, Journal of the European Economic Association, vol 1, no 5, pp
1123–1175.ia).
[15] Tovar, Camilo E. (2009). DSGE Models and Central Banks. Economics: The Open-Access,
Open-Assessment E-Journal, Vol. 3, 2009-16. doi:10.5018/economics-ejournal.ja.2009-16,
http://dx.doi.org/10.5018/economics-ejournal.ja.2009-16

24
A Appendix: Some important derivations
A.1 The entrepreneurs problem
A.1.1 The multiperiod credit contract problem
Entrepreneurs take credit in a multiperiod contract with …xed repayments as in Bennes and
Lees (2010). A credit CRt in nominal values taken at time t is paid back in repayments
proportional to the amount borrowed and decay at a …xed rate i 2 (0; 1) : This implies the
k 1 cr
following repayment schedule: Qcr cr
t CRt , Qt CRt , :::, Qt CRt , ::: for repayments due at
t + 1; t + 2; :::; t + k; ::: respectively. Where Qcr
t is a shadow price associated to the credit. This
form of multiperiod contract allows us to draw out the implications of long-term debt and
maturity mismatch in an analytically tractable framework.
Accordingly, the sum of all repayments due at t associated with all past loans is:

Jt 1 = Qcr cr
t 1 CRt 1 + Qt 2 CRt 2 + ::: + k 1
Qcr
t k CRt k + :::
1
X
k 1 cr
= Qt k CRt k (A-1)
k=1

Also, Jt can be written recursively as:

Jt = Qcr cr
t CRt + Qt 1 CRt 1 + ::: + k
Qcr
t k CRt k + :::
= Qcr
t CRt + Qcr
t 1 CRt 1 + ::: + k 1
Qcr
t k CRt k + :::
= Qcr
t CRt + Jt 1 (A-2)

which is equal to equation (3.25) in the main text. The stock of loans, estimated as the present
value of repayments due at t + 1; t + 2; :::is equal to:
1
Lt = Et Qcr cr
t CRt + Qt 1 CRt 1 + ::: + k
Qcr
t k CRt k + :::
Rt+1
1
+ Et Qcr cr
t CRt + Qt 1 CRt 1 + ::: +
k cr
Qt k CRt k + :::
Rt+1 Rt+2
1
+ 2 Et Qcr cr
t CRt + Qt 1 CRt 1 + ::: +
k cr
Qt k CRt k + :::
Rt+1 Rt+2 Rt+3
+:::

This can be rewritten as:

Lt = t Qcr cr
t CRt + Qt 1 CRt 1 + ::: + k
Qcr
t k CRt k + :::

where
1 1 1
t = Et + + 2 + :::
Rt+1 Rt+1 Rt+2 Rt+1 Rt+2 Rt+3
1 1 1
= Et 1+ + + :::
Rt+1 Rt+2 Rt+2 Rt+3

25
which after using the de…nition t+1 and the law for iterated expectations, is equal to:
1
t = Et [1 + t+1 ] (A-3)
Rt+1

t is a discount factor that depends on future interest rates!!

Also, after using the de…nition for Jt , the stock of loans can be written as:

Lt = t Jt (A-4)

Replacing Jt = Lt = t from this equation in the recursive representation of Jt , the evolution of

loans can be written as:
Lt Lt 1
= Qcr
t CRt +
t t 1
cr t
Lt = t Qt CRt + Lt 1 (A-5)
t 1

A.1.2 The optimal condition of the entrepreneurs

The entrepreneurs’pro…ts are given by:
ent
t = Pt RtK Kt 1 Qt It + CRt Jt 1 (A-6)

where Jt = Qcr
t CRt + Jt 1: From this, solving for CRt :
(1 L)
CRt = Jt
Qcr
t

where L is the lag operator. Then, pro…ts become:

ent (1 L)
t = Pt RtK Kt 1 Qt It + Jt Jt 1 (A-7)
Qcr
t
(1 L)
we calculate the present disconted value of the pro…ts considering Qcr as the discount
t
factor:

ent (1 L) ent (1 L)2 ent

t + t+1 + t+2 + :::
Qcr
t Qcr cr
t Qt+1
(1 L)
= Pt RtK Kt 1 Qt It + K
Pt+1 Rt+1 Kt Qt+1 It+1
Qcr
t
(1 L)2 K
+ cr Pt+2 Rt+2 Kt+1 Qt+2 It+2 + ::: Jt 1
Qcr
t Qt+1

The problem for this entrepreneur can be written recursively as

(1 L)
V (Kt 1) = max Pt RtK Kt 1 Qt It + Et V (Kt ) (A-8)
Kt ;It Qcr
t

26
 which equals:

1
1+ V (Kt 1) = max Pt RtK Kt 1 Qt It + Et V (Kt ) ;
Qcr
t Kt ;It Qcr
t
1 1=Qcrt
V (Kt 1) = max Pt RtK Kt 1 Qt It + Et V (Kt ) ;
Kt ;It 1 + Qcr 1 + Qcr
t t

or

1 1=Qcr
t
V (Kt 1) = max Pt RtK Kt 1 Qt It + Et V [It + (1 ) Kt 1] : (A-9)
It 1 + Qcr 1 + Qcr
t t

The …rst order condition of the Bellman equation (A-9) with respect to It is:

1 1=Qcr
t
Pt Qt + Et V 0 (Kt ) = 0: (A-10)
1 + Qcr 1 + Qcr
t t

To calculate Et V 0 (Kt ) in (A-10) we need to calculate …rst V 0 (Kt 1) in (A-9) using the
envelope theorem:

1 1=Qcr
t
V 0 (Kt 1) = Pt RtK + (1 ) Et V 0 (Kt ) : (A-11)
1 + Qcr 1 + Qcr
t t

We also use the …rst order condition (A-10) to simplify equation (A-11):
1 1
V 0 (Kt 1) = Pt RtK + (1 ) Pt Qt ; (A-12)
1 + Qcr 1 + Qcr
t t

then we evaluate equation (A-12) in t+1 and take expectations:

2 3
1 1
Et V 0 (Kt ) = Et 4 K
Pt+1 Rt+1 + (1 ) Pt+1 Qt+1 5 ; (A-13)
1 + Qcr 1 + Qcr
t+1 t+1

and replace this in the …rst order condition (A-10):

0 1 " #
1 + Qcr K + (1
Rt+1 ) Qt+1
cr @ t+1 A
Et Qt = Et : (A-14)
t+1 Qt

This condition (3.29) in the main text.

27
A.2 Solving for the banks problem:
The budget constraint of the bank is:

Pt Ct (b) + CRt (b) + ZtL (b) ZtB (b) Jt ib

1 (b) + Xt (b) + Rt 1 ZtL 1 (b) ZtB 1 (b) (A-15)

1
Jt (b) = Lt (b) (A-16)
t

t cr
Lt (b) = Lt 1 (b) + t Qt CRt (b) (A-17)
t 1
After considering these three equations, the budget constraint in t and in t + 1 becomes:

Pt Ct (b) + CRt (b) + ZtL (b) ZtB (b) (A-18)

1
Lt 1 (b) + Xt (b) + Rtib 1 ZtL 1 (b) ZtB 1 (b)
t 1

where Lt 1 (b) = t 1
t 2
Lt 2 (b) + t 1 Qcr
t 1 CRt 1 (b) : We claim that a bank with a lending
oportunity, the constraint (3.32) binds, and for the others not (see proof).
The budget constraint for a bank with lending opportunity, after replacing ZtB (b) =
t Lt 1 (b) + t CRt (b), becomes:

Pt Ct bL + (1 t ) CRt b
L
t Lt 1 (b)
1
Lt 1 (b) + Xt bL + Rtib 1 ZtL 1 (b) ZtB 1 (b) (A-19)
t 1

Similarly, the budget constraint for a bank without lending opportunity (that is
CRt (b) = 0):
1
Pt Ct bN L + ZtL bN L Lt 1 (b) + Xt bN L + Rtib 1 ZtL 1 (b) ZtB 1 (b)
t 1

where Xt (b) = (1 t ) Dt (b) (Rt 1 t 1 ) Dt 1 (b) :

Considering the above, the Lagrangian for a lending bank is:

L = ln Ct bL 't bL BCt bL + Et ln Ct+1 bL 't+1 bL BCt+1 bL jbL

t
+ (1 ) Et ln Ct+1 bN L 't+1 bN L BCt+1 bN L jbL
t + ::: (A-20)

Similarly, the Lagrangian for a non-lending bank:

L = ln Ct bN L 't bN L BCt bN L + Et ln Ct+1 bL 't+1 bL BCt+1 bL jbNL

(A-21)
t
+ (1 ) Et ln Ct+1 bN L 't+1 bN L BCt+1 bN L jbN
t
L
+ :::

28
where we have de…ned the following notation for the budget constraints in real terms as:
1
Pt Ct bL + (1 bL t ) CRt
t Lt 1 (b) Lt 1 (b) Xt bL 1
BCt bL t 1
L
Zt 1 bL Rtib 1 B
Zt 1 b L Pt
8 h i 9
< Pt+1 Ct+1 bL + (1 t+1 ) CRt+1 b
L
t+1
t
L t 1 (b) + t Q cr CR bL
t t
= 1
BCt+1 bL jbL
t
h i t 1

: 1 t
Lt 1 (b) + t Qcr
t CRt b
L Xt+1 + Rtib B t Lt 1 (b) + t CRt b
L ; Pt
t t 1
( h i )
1
Pt+1 Ct+1 bN L + Zt+1L bN L t
L t 1 (b) + t Q cr CR bL
t 1
BCt+1 bN L jbL
t
t t 1 t
Xt+1 bN L + Rtib B t Lt 1 (b) + t CRt b
L Pt+1

and
1 1
BCt bN L Pt Ct bN L + ZtL bN L Lt 1 (b) Xt bN L Rtib 1 ZtL 1 (b) ZtB 1 (b)
t 1 Pt
8 h i 9
< Pt+1 Ct+1 bL + (1 t+1 ) CRt+1 b
L B
t+1
t
L t 1 (b) = 1
BCt+1 bL jbN
t
L h i t 1

: 1 t
Lt 1 (b) Xt+1 bL Rtib ZtL bL ; Pt+1
t t 1
( h i )
1
Pt+1 Ct+1 bN L + Zt+1L bN L t
Lt 1 (b) 1
BCt+1 bN L jbN
t
L t t 1
Xt+1 bN L Rtib ZtL bN L Pt+1

The …rst order conditions for the lending bank are:

1
Ct bL : = 't bL (A-22)
Ct (bL )
1
X Pt+s Pt
CRt bL : 't bL (1 t ) = Et
s
't+s+1 Qcr
t Et 't+1 ib
R(A-23)
t t
Pt+s+1 Pt+1
s=0
1
X Pt+s
+ Et ( )s 't+s+1 bL t+s+1 t+s Qcr
t
Pt+s+1
s=0
" #
1 't+1 bL + (1 ) 't+1 bN L
Dt bL : 't bL (1 t) = Et (Rt t) (A-24)
Pt Pt+1

1
X
L s Pt+s Pt
't b (1 t) = Et 't+s+1 Qcr
t Et 't+1 Rtib t (A-25)
Pt+s+1 Pt+1
s=0
1
X Pt+s
+ Et ( )s 't+s+1 bL t+s+1 t+s Qcr
t
Pt+s+1
s=0

29
Similarly, the …rst order conditions for the non-lending bank are:
1
Ct bN L : = 't bN L (A-26)
Ct (bN L )
Rtib Rtib
ZtL bN L : 't bN L = Et 't+1 bL + (1 ) Et 't+1 bN L (A-27)
Pt+1 Pt+1
" #
1 't+1 bL + (1 ) 't+1 bN L
Dt bN L : 't bN L (1 t) = Et (Rt t) (A-28)
Pt Pt+1

From (A-27) and (A-28), we have:

Rt t
Rtib = > Rt ; (A-29)
1 t

the interest rate charged by interbank loans is higher than the risk free rate that pay for
the deposits. The reserve requirement rate generates a mark-up on that rate (it works as a
tax).
Condition (A-23) determines the shadow price associated to the credit paid in multiperiod
contracts. Using condition (A-24) and (A-29) to simplify (A-23), and representing recursively
the in…nite sum terms, we get:

Rt
Qcr
t = (A-30)
t + t
h i h i
1
where t Et R 1 + t+1 and t Et R t+1 t + t+1 are recursive repre-
t t
sentations of the present values terms in equation (A-23).
Also, note from conditions (A-24) and (A-28) that 't bL = 't bN L , which imply that
Ct bL = Ct bN L from conditions (A-22) and (A-26). There is perfect risk sharing between
bankers, the demand for deposits will equalise the consumption for both type of banks
The budget constraints for the banks with and without lending opportunity, are respec-
tively:
1
Pt Ct bL + (1 t ) CRt bL = + B
t Lt 1 (b) + Xt (b) + Rtib 1 ZtL 1 (b) ZtB(A-31)
1 (b)
t 1
1
Pt Ct bN L + ZtL bN L = Lt 1 (b) + Xt (b) + Rtib 1 ZtL 1 (b) ZtB 1 (b) (A-32)
t 1

Because of the assumption on logarithmic utility, total consumption of banks is given

1
Pt Ct bL = (1 b) + B
t Lt 1 (b) Rt 1 t 1 Dt 1 (b) + Rtib 1 ZtL 1 (b) ZtB 1(A-33)
(b)
t 1
1
Pt Ct bN L = (1 b) Lt 1 (b) Rt 1 t 1 Dt 1 (b) + Rtib 1 ZtL 1 (b) ZtB 1 (b) (A-34)
t 1

Then, credit is de…ned from the budget constraint of those that lend:

30
 1
CRt bL (1 t) = b + t Lt 1 Rt 1 t 1 Dt 1 (b) + (1 t ) Dt (A-35)
t 1

31
A.3 Aggregation
The Household’s budget constraint:

Pt Cth = Wt Ht + Rt 1 Dt 1 Dt + t + Tt (A-36)
Intermediate producer’s pro…ts:

int Wt Ht
t = Pt Yt RtK Kt 1 (A-37)
Pt

Capital producer‘s pro…ts:

cap
t = Pt (Qt 1) It (A-38)
Entrepreneur‘s pro…ts:
ent
t = Pt RtK Kt 1 Qt It + CRt Jt 1 (A-39)

Total pro…ts are:

int cap ent

t = t + t + t
= Pt Yt Wt Ht Pt It + CRt Jt 1 (A-40)

After replacing total pro…ts in the household‘s budget constraint, we obtain:

Pt Cth = Pt Yt Pt It + CRt Jt 1 + Rt 1 Dt 1 Dt + Tt (A-41)

The banker’s budget constraint is:

Pt Ctb + CRt = Jt 1 + Dt Rt 1 Dt 1 RRt + RRt 1 (A-42)

The central bank’s budget constraint is:

Tt = (Tt Tt 1) = RRt RRt 1 (A-43)

Summing up this three constraints, we have:

Yt = Cth + Ctb + It (A-44)

which equation (3.39) in the main text.

32
B Appendix: The set of equations
B.1 The non-linear set of equations
B.1.1 Aggregate demand
Total demand [Yt ]:
Yt = Ct + It : (B.1-1)
Total Consumption [Ct ]:
Ct = Cth + Ctb : (B.1-2)
Consumption of households [Cth ]:
" ! #
h
Ct+1
1
1 = Et Rt : (B.1-3)
t+1 Cth

Consumption of bankers [Ctb ]:

1 Lt 1 Dt 1 1
Ctb = (1 b) + t Rt 1 t 1 ; (B.1-4)
t 1 Pt 1 Pt 1 t
where Xt =Pt is de…ned below.
Investment [It ] :
CRt
Qt It = : (B.1-5)
Pt

B.1.2 Aggregate supply

Phillips curve [ t] :
1 "
" 1 Nt
( t) = 1 (1 ) ; (B.1-6)
Dt
h i
" 1
Dt = Yt (Ct ) + Et ( t+1 ) Dt+1 ;
"
Nt = Yt (Ct ) M Ct + Et [( t+1 ) Nt+1 ] :

Marginal costs [M Ct ]
1
1 RtK Wt =Pt
M Ct = : (B.1-7)
At 1

B.1.3 Labour market

Labour demand [Ht ]:
1
1 Wt =Pt
Ht = Yt t: (B.1-8)
1 M Ct
Labour supply [ W
Pt ]:
t

Wt
= Cth Htv : (B.1-9)
Pt

33
B.1.4 Capital market
Capital demand [RtK ]
1
1 RtK
Kt 1 = Yt t: (B.1-10)
M Ct
Capital supply [Kt ]:
Kt = It + (1 ) Kt 1: (B.1-11)
Return of capital [RtQ ]:
1
RtQ = RtK + (1 ) Qt : (B.1-12)
Qt 1

Arbitrage condition for capital [Qt ]:

0 1
1+ Qcr
Q @ A:
Et Rt+1 = Et Qcr
t
t+1
(B.1-13)
t+1

B.1.5 Interbank market

Arbitrage condition between CRt bL and Dt bL : [Qcr
t ]

1 = Qcr
t ( t + t ): (B.1-14)
Evolution (supply) of Loans [ L
Pt ]:
t

Lt cr CRt t Lt 1 1
= t Qt + : (B.1-15)
Pt Pt t 1 Pt 1 t

Supply of credit (from balance of banks) [ CR

Pt ]:
t

CRt 1 Lt 1 Dt 1 1 Dt
(1 t) = b Rt 1 t 1 + (1 t) : (B.1-16)
Pt t 1 Pt 1 Pt 1 t Pt

Flow of deposits net of the ‡ow of reserves [ X

Pt ]
t

Xt Dt Dt 1 1 RRt RRt 1 1
= Rt 1 : (B.1-17)
Pt Pt Pt 1 t Pt Pt 1 t

Deposits supply [ D
Pt ], from the budget constraint of the households
t

Wt Ht Dt 1 1 Dt t
Cth = + Rt 1 + : (B.1-18)
Pt Pt 1 t Pt Pt

34
B.1.6 Monetary policy
Taylor rule
t
Rt = R : (B.1-19)

Reserve requirements rule

RRt = t Dt : (B.1-20)

and 8
>
>
< h cr i
Qcr Q 1
= Rt = R (B.1-21)
t
t
>
>
: t 2

B.1.7 Other equations

1
t = Et 1+ t+1 : (B.1-22)
Rt
1
t = Et 1+ t+1 (B.1-23)
Rt

t = Et t+1 + t+1 (B.1-24)

Rt
Rt t
Rt = (B.1-25)
1 t

t Wt CRt 1 Lt 1 1
= Yt Ht It + ; (B.1-26)
Pt Pt Pt t 1 Pt 1 t

35