A Tale of Two Cities

Perry Mehrling

Recently socialist economists, for the first time applying to their problem
tools of modern economic theory, found that a free market for consumer
goods, and even for labor, is probably the safest guarantee for the best
allocation of national resources to various productions (and to leisure),
provided that the amount and use of national savings are determined by
central bodies and that industries which naturally incline to a monopolistic
or semi-monopolistic organization are entrusted to public managers.
—Jacob Marschak, “Peace Economics” (1940)

It was quite a surprise, albeit a very pleasant one, to learn that my book
Fischer Black and the Revolutionary Idea of Finance (2005) had been
chosen to receive the ESHET award. I speak not from false humility, but
rather from a realistic recognition that this book, along with my previous
one, The Money Interest and the Public Interest: American Monetary

Correspondence may be addressed to Perry Mehrling, Room 2 Lehman Hall, Barnard College,
3009 Broadway, New York, NY 10027; e-mail: pgm10@columbia.edu. This paper was pre-
pared and presented originally as the Jérôme Blanqui Lecture for the May 2008 meeting of the
European Society of the History of Economic Thought, held in Prague. I would like to acknowl-
edge the helpful comments of Duncan Foley and Ken Kuttner, as well as participants in the
ESHET meeting (Prague, May 2008), especially Harald Hagemann, and participants in the
meeting of HISRECO (Lisbon, June 2008), especially Roger Backhouse, Béatrice Cherrier, and
Phil Mirowski. I would also like to acknowledge financial support from the Alliance Program
at Columbia University, which enabled a series of helpful discussions with Christian Walter and
Glenn Shafer.
History of Political Economy 42:2 DOI 10.1215/00182702-2010-001
Copyright 2010 by Duke University Press
202 History of Political Economy 42:2 (2010)

Thought, 1920–1970 (1997), represents an attempt to trace a peculiarly

American line of intellectual development, a line with roots in American
pragmatism and institutionalism. In all of this work I have been using the
story of the development of American monetary and financial thought as
a narrative thread through the quite incredible ups and downs of the twen-
tieth century, a story of war and peace as well as boom and depression.
Within that larger story, the drama of money and finance is a tale of
shifting boundaries, between public and private, between state and mar-
ket, even between the Treasury and the Federal Reserve. I have tried to
tell this story by following the effects of these shifting boundaries on the
development of monetary and financial thought inside the academy. I use
the word effects on purpose. I do not mean to deny the “performativity” of
economic theory in the sense of MacKenzie 2006; economists, especially
American economists, have always been engineers, looking to put their
ideas into practice in the world. Rather I mean to direct attention to the fact
that while the engineering impulse in economics is constant, the problems
that direct the attention of that impulse are specific to each moment in time
and arise from historical developments leading up to that time.
The biographical form that I have used in my work has thus been meant
not so much to celebrate the way that the ideas of disembedded genius
have changed the world but rather, and quite the reverse, to trace how eco-
nomic thought develops by engagement with the concrete problems faced
by each generation, an engagement that is both facilitated and obstructed
by intellectual inheritance from past generations. I focus attention on the
development of individual minds, even though my larger goal is to under-
stand the development of economic thought more generally, because the
task of understanding an individual is more delimited, and hence conceiv-
ably possible. But given that larger goal, the choice of which individual
mind to enter is clearly crucial.
Reviewers of my first book sometimes objected to my pretense to be
telling a story about American monetary thought, since I focused the story
around Allyn Young, Alvin Hansen, and Edward Shaw. They asked, “What
about Milton Friedman, James Tobin, Franco Modigliani?” Readers of my
second book similarly sometimes objected to my pretense to be telling a
story about American financial thought, since the book focuses on Fischer
Black. “What about Harry Markowitz, Bill Sharpe, Merton Miller, Franco
Modigliani, Robert Merton, Myron Scholes, and Paul Samuelson?” A con-
venient answer to this kind of question is that, on principle, I confine my
attention to people who are safely buried. I used this convenient answer
often! But the more honest answer is that I have felt the need to get some
 Mehrling / A Tale of Two Cities 203

distance from the people whom the internal processes of the academic
discipline have already anointed, in order to see more clearly the effect of
external influences on the development of thought.
I’m going to do that again today. I will be telling a story about the devel-
opment of macroeconomics, but my central figure will not be Keynes, or
Wicksell, or even Fisher or Kalecki. My story is new, and so necessarily
only a sketch. My goal today is to tell a story that looks at a period you
already know but through a different lens. My story thus complements the
standard story, such as that of Roger Backhouse and David Laidler (2004),
who use as their lens the role of Keynes’s General Theory in facilitating
transition from the diversity of interwar macroeconomics to the hegemony
of the IS-LM model in the postwar period.

A Tale of Two Equations

From an American perspective, the story of the development of macro-
economics is less about Keynes and more about the rise and fall of what
I have previously called “monetary Walrasianism” (Mehrling 1997). The
central texts of monetary Walrasianism include, among others, Franco
Modigliani’s “Liquidity Preference and the Theory of Interest and Money”
(1944), Don Patinkin’s Money, Interest, and Prices (1956), and James
Tobin’s “Liquidity Preference as Behavior towards Risk” (1958). I have
previously suggested that a key foundational figure for the monetary Wal-
rasian project was Jacob Marschak, insofar as his 1938 “Money and the
Theory of Assets” set the agenda (Mehrling 2002 n. 12). Today I want to
expand on that suggestion, both backward and forward in time.
My story is a tale of two equations. The first one is the quantity equa-
tion that Irving Fisher revived in 1911:
MV = ∑pQ. (1)
In this equation M is the quantity of money, V the velocity of money, p the
price of various goods sold, and Q the quantity of various goods sold. The
left-hand side captures the turnover of money balances, while the right-
hand side captures the exchange of goods. In 1911, Fisher was concerned
mainly with how M, V, and Q interact to determine the price level and its
fluctuation over time. The quantity theory of money, for example, uses the
quantity equation to talk about the relationship between M and the price
level. But the same framework was used subsequently, by Fisher and oth-
ers, to analyze the determination of Q and its fluctuation over time. Just
so, in my book on the development of American monetary economics,
204 History of Political Economy 42:2 (2010)

the quantity equation provides the framework for discussion by just about
everyone.
The second equation is the Euler equation that lies at the heart of both
modern finance and modern macroeconomics:
U′(Cit) = Et [δU′(Cit + 1)Rjt + 1]. (2)
In this equation Cit is the consumption of consumer i at time t, U is a func-
tion that translates consumption into utility terms, δ is the subjective dis-
count rate, and Rjt + 1 is the gross return on asset j in the period between t and
t + 1. For finance, this equation is about how asset prices depend on time
and risk preferences, the equation is called the “consumption CAPM,” and
the asset in question is typically equity or long-term bonds (Breeden 1979;
Cochrane 2001). But the same equation can be used to talk about the inter-
temporal fluctuation of income, and as such is at the core of both real busi-
ness cycle theory and its New Keynesian variants (Woodford 2003; Sargent
2008). In this application, the asset is typically capital, or a rate of interest.
I submit to you that a very large intellectual revolution is involved in
moving from the first equation to the second. The most obvious change is
a shift from money to finance, and from the quantity of money (or aggre-
gate demand) to the rate of interest as the relevant policy instrument. This
already is a huge substantive change, from the city of money to the city
of finance, from the public Board of Governors in Washington to the pri-
vate securities exchange in New York. In what follows, however, I will be
focusing attention on three far-reaching methodological changes involved
in the shift from the first equation to the second.
1. Risk. The second equation incorporates risk explicitly insofar as the
realization of the gross asset return is stochastic; hence, the expecta-
tions operator E[.].
2. Equilibrium. The second equation is a first-order condition charac-
terizing individual optimization taking price as given, but also pos-
sibly characterizing economy-wide equilibrium in an endowment
economy taking price as the equilibrating factor.
3. Time. The second equation incorporates time explicitly insofar as
the risk that is important for individual choice has to do with the
future time period t + 1 when asset returns will be realized.
In all three respects, we have clearly come a long way from Fisher 1911.
The distance we have traveled has, however, been obscured by the ten-
dency of the profession to read something like equation 2 back into Fisher,
which is to say the tendency to read modern economics as a natural exten-
 Mehrling / A Tale of Two Cities 205

sion of Fisher’s pioneering work. James Tobin (1985) more than anyone
else is responsible for this reading, but there have been plenty of historians
of thought writing in support of the story of continuity from Irving Fisher.
For my purposes, it is more important to emphasize the elements of dis-
continuity.
I have elaborated this argument in more detail elsewhere (Mehrling
2001), so a single quotation will suffice to make the point here. Here is
Fisher, writing in 1930:
While it is possible to calculate mathematically risks of a certain type
like those in games of chance or in property and life insurance where
the chances are capable of accurate measurement, most economic risks
are not so readily measured. To attempt to formulate mathematically in
any useful, complete manner the laws determining the rate of interest
under the sway of chance would be like attempting to express completely
the laws which determine the path of a projectile when affected by ran-
dom gusts of wind. Such formulas would need to be either too general
or too empirical to be of much value. (316)
This passage reveals an attitude toward the “dark forces of time and
ignorance”—the phrase is from Keynes in the General Theory—not unlike
that of Keynes himself. Like Keynes in his Treatise on Probability (1921),
and also like Frank Knight in his Risk, Uncertainty, and Profit (1921),
Fisher thought that the mathematics of risk was not an appropriate ana-
lytical framework for problems of intertemporal choice, much less inter-
temporal general equilibrium. Fisher thus explicitly rejected the line of
analytical development that leads to equation 2. But if Fisher didn’t do it,
then how did we get to equation 2? My central argument will be that mon-
etary Walrasianism served as the bridge that carried us from equation 1 to
equation 2. And the key figure in understanding monetary Walrasianism,
I suggest, is Jacob Marschak.1

Jacob Marschak
Marschak himself started with Irving Fisher and the equation of exchange,
in his PhD thesis “Die Verkehrsgleichung” (1924).2 According to his own

1. For the supporting arguments on risk and time, I am indebted to the work of Elton
McGoun (2007) and Christian Walter (2006), who got me started on the right track.
2. Arrow 1979 provides a biographical sketch, but see also Hagemann 2006, 2007. Marschak
was born in 1898 in Kiev, Ukraine, earned his PhD at the University of Heidelberg in 1922, and
spent the 1920s at the Kiel Institute for World Economics. His first major publication (1923)
206 History of Political Economy 42:2 (2010)

testimony (Marschak 1974, 3:3), there is a straight line of intellectual devel-

opment from that early paper to the mature statement of monetary Walra-
sianism in his 1950 paper “The Rationale of the Demand for Money and
of ‘Money Illusion.’” Thus a story of continuity from equation 1 to equa-
tion 2 works for Marschak even if not for Fisher. But that is getting ahead
of the story. To begin, it is important to emphasize that Marschak’s initial
development of the quantity framework involved integrating the theory of
money not with the theory of value—that would come later—but rather
with the theory of capital.3
The high-water mark of that early work is achieved in Marschak’s
book Kapitalbildung (Capital Accumulation), which contains the results
of an extensive research project with Walther Lederer at the Kiel Institute
for World Economics (Marschak and Lederer 1936; see Goldschmidt
1938). The publication of the book was delayed because of political devel-
opments in Germany (see Hagemann 2007), but we can see the essential
analytical framework in a paper Marschak published under the intrigu-
ing title “Econometric Parameters in a Stationary Society with Monetary
Circulation” (1934a). This publication coincided with Marschak’s depar-
ture from Germany for a position at Oxford University and hence can
be considered a kind of inaugural address for his new English-speaking
colleagues.
In this paper Marschak is interested in how the vertical organization
of industry, by subsuming much of the exchange of intermediate goods
within the boundaries of the firm, increases the efficiency of monetary cir-
culation. The whole point of the exercise is to derive an analytical frame-
work that can be taken to the data, in this case the German data for the
late 1920s. Marschak is not looking to test any particular economic the-
ory, but rather to use economic theory to suggest mathematical equations
that he can use to characterize empirical regularities in the data. That is
what Marschak thought the econometric project was all about.
In conjunction with his move to Oxford, Marschak attended the meet-
ings of the Econometric Society, held in Leiden in September–October
1933, and he wrote up the proceedings for publication. His account of his

engaged von Mises on the socialist calculation debate. In 1933 he left for Oxford University;
from 1935 to 1939 he served as director of the Oxford Institute of Statistics, from 1939 to 1942
he was at the New School for Social Research, and from 1943 to 1948 he was director of the
Cowles Commission for Research in Economics. McGuire and Radner 1972 is a Festschrift for
Marschak, and Marschak 1974 contains his collected works.
3. This intellectual project he shared with Arthur Marget (1938), among others.
 Mehrling / A Tale of Two Cities 207

own paper, “Theoretical Problems Suggested by Roosevelt’s Policy,” pro-

vides key evidence of his intellectual framework as of that date. The
important equations are as follows (Marschak 1934b, 196):
p
( )
qi = σi —i = δi ( p1, p2 . . . w . e),
w
∑pi qi = w . e = MV,
P = λ( p1 . . . pn, q1 . . . qn).
As in his earliest work, the centerpiece is the equation of exchange,
now written right to left. He has added, however, the idea that the quan-
tities q in the equation are the result of a supply σ and demand δ rela-
tionship for each good i, and he is clearly prepared to use the standard
apparatus of production functions and utility functions to help organize
thinking about the form of those supply and demand functions.4 Thus, in
effect, he has a quantity equation in center place, with a Walrasian system
appended.
At the 1933 meeting, Marschak’s main concern was to use that analyti-
cal apparatus to point out the possible inconsistency of various elements of
Roosevelt’s economic plan for the United States. Roosevelt was proposing
policies to affect certain individual prices, wages, quantities, purchasing
power, and the price level. The point of Marschak’s analytical framework
is to show that there is a logical relationship between all these elements
that has apparently not been recognized by the policy makers themselves.
This kind of contribution to policy analysis is also what Marschak thought
the econometric project was all about.
These early papers allow us to characterize Marschak’s intellectual for-
mation independent of the later work for which he is best known. In his
intellectual formation, he was a kind of proto-econometrician, pragmati-
cally using both economic theory and statistics to characterize regularities
in the data. But the purpose of these studies was always to provide the basis
for economic management, both microeconomic and macroeconomic.
In microeconomics, the big problem was monopoly and the solution was
public ownership and management, but not any further intervention into
market processes. In macroeconomics, the big problem was instability,
and the solution was monetary management. These are the problems to
which the “econometric movement”—as Kenneth Arrow (1979, 502) calls

4. A key step toward this Walrasian formulation is apparently Marschak’s 1931 habilita-
tion thesis on the elasticity of demand.
208 History of Political Economy 42:2 (2010)

it—was supposed to be providing the solution, and we must understand

Marschak’s work in that context.5
From this standpoint, we can understand Marschak’s “Money and the
Theory of Assets” (1938) not as a new departure but rather as continuous
with his previous thought. Whereas previously he had a quantity equation
framework with a Walrasian system appended, from 1938 on he would
have a Walrasian system with a money demand equation incorporated.6
Marschak assumes that utility depends on the moments x, y, and z of the
distribution of consumption yields generated by various assets, and pro-
ceeds to derive a first-order condition for the consumer optimum. If the
price of asset a is p, and the price of asset b is q, then we have the follow-
ing (p. 316):
dx dy
p Ux da + Uy da + . . .
q= , (marginal productivity equations).
dy
Ux dx + Uy +...
db db
Here we see the first pier of the bridge between equation 1 and equation 2.
The immediate impetus for Marschak’s formal approach to the prob-
lem seems to have been Hicks’s famous “Suggestion for Simplifying the
Theory of Money” (1935), which proposed to open the field of money to
the apparatus of supply and demand. Equally important, however, was the
earlier work of Hicks that urged fellow economists to overcome their fear
of using the mathematical apparatus of probability and risk. As early as
1931, Hicks had urged rejection of Frank Knight’s philosophical objec-
tions to employing the concept of risk, mainly on pragmatic grounds.
Marschak would have encountered Hicks’s views at the 1933 meeting of
the Econometric Society, if not before, as Hicks presented a paper titled
“The Application of Mathematical Methods to the Theory of Risk.” Fol-
lowing Hicks, what Marschak did in 1938 was, in a single stroke, to incor-
porate both risk and (portfolio) equilibrium into the quantity equation
framework, but not yet time.

5. Arrow’s impression that Marschak’s contribution was largely synthetic tends to miss
Marschak’s preexisting capacious analytical framework within which it was possible to find
room for narrower contributions of many different kinds. Similarly, his impression that Mar-
schak changed his economics to fit the changing environments within which he worked tends
to miss the continuity of his thought.
6. The early history of money demand theory, a missing chapter in Mirowski and Hands
2006, remains to be written. Those interested in tracing Marschak’s own development on this
point will want to consider, in addition to the endpoints Marschak 1938 and Marschak 1950,
Makower and Marschak 1938 and the intermediate steps (Marschak 1943, 1949).
 Mehrling / A Tale of Two Cities 209

It is important to emphasize that Marschak apparently saw this emer-

gent monetary Walrasian construct as completely compatible with the
work of Keynes. For himself, he still preferred to think and talk in terms
of money and purchasing power, rather than aggregate demand, but
more for the sake of analytical clarity than anything else (see for exam-
ple Marschak 1942b). Knut Wicksell, after all, had shown how an excess
of investment over saving could be understood as the same thing as an
excess supply of money (Marschak 1941b). In this sense there is a straight
line from Marschak’s original (1934a, 1934b) quantity-theoretic econo-
metric project to the neo-Walrasian econometric project mooted in Mar-
schak 1942a and Marschak and Andrews 1944, the papers that are usually
treated as Marschak’s contribution to the Cowles econometric modeling
effort.7
It is important also to emphasize that Marschak apparently saw this
monetary Walrasian construct as completely compatible with the work
of the American institutionalists. Marschak’s “Methods in Economics”
(1941a), which reviews a discussion of the work of the institutionalist Fred-
erick C. Mills, is clearly a precursor to Tjalling Koopmans’s famous “Mea-
surement without Theory” (1947), but with much more sympathy for the
institutionalists. (See also Marschak [1951] 1974, his “Comment on Mitch-
ell.”) Like the institutionalists, Marschak’s project was to a large extent
about “social engineering,” about understanding the world in order to
make it better. For this goal, he insisted, it is vital to avoid the “ravages of
methodology” (1941a, 441) and to use pragmatically whatever tools are
available to work toward the goal.
In his “Cross Section of Business Cycle Discussion” (1945) Marschak
was even more pedagogically explicit, urging young folk “not to despise
God’s gift of equations,” which can be used to integrate knowledge com-
ing from economic theory and knowledge coming from statistical mea-
surement, in the manner of Tinbergen 1939.8 Keynes famously urged his
audience to dream of a possible future world when economists would be
about as important as dentists. Marschak held out such a humble role as
a possible life choice for his own students in the here and now. As it hap-
pened, many found such a role attractive, and the rise of monetary Wal-
rasianism followed the career trajectory of those students.

7. It follows that Marschak 1934a and 1934b should be viewed as part of the prehistory of
econometrics that the standard story (Morgan 1990) treats as beginning with Trygve Haavelmo.
8. Marschak and Lange [1940] 1995 provides a defense of Tinbergen that Keynes blocked
from publication in the Economic Journal.
210 History of Political Economy 42:2 (2010)

The Second Generation

The main themes of Marschak’s monetary Walrasianism were thus all in
place by 1945, if not earlier. What remained to do was to shore up the ana-
lytical foundations of the project on the one hand, and to develop the pol-
icy applications on the other. A key step showing the way forward on the
first front was von Neumann and Morgenstern’s work to show how ratio-
nal choice under risk could be understood to imply the existence of an
implicit von Neumann–Morgenstern utility construct (Marschak 1946).
Marschak himself took the opportunity to reformulate his 1938 paper as
Marschak 1950, which therefore qualifies as a second pier in the bridge
between our two equations. But he was content to leave the rest of the work
to the broader econometric movement. For this purpose, no one was more
important than the young Kenneth Arrow.
As early as 1941, Arrow had written his master’s thesis on stochastic
processes under Abraham Wald and Harold Hotelling at Columbia Uni-
versity. His comfort with the mathematics of probability subsequently
proved valuable for the war effort, most notably in his paper “On the Use
of Winds in Flight Planning” (1949), which was only published after the
war.9 In this paper, Arrow solved exactly the kind of problem that Irving
Fisher in 1930 had held out as obviously insoluble; Fisher’s projectile
becomes Arrow’s airplane. If Fisher was wrong about that problem, then
maybe the problem of intertemporal choice under uncertainty was not so
obviously insoluble either. We can understand Arrow’s subsequent work as
an attempt to explore just such a possibility, first with the microeconom-
ics of choice under risk (1951), then with the general equilibrium problem
(1953). Here we have the third pier in our bridge, because here we finally
have an explicit treatment of time.
Hicks had already introduced in Value and Capital (1939) the idea of
extending the well-known static model of general equilibrium to incorpo-
rate both risk and time, by the simple stratagem of treating commodities
at different times and in different states of the world each as a distinct com-
modity with its own distinct price. In 1953 Arrow showed that a complete
set of state-contingent securities markets can play exactly the same role as
a complete set of commodity futures markets, in the sense of implement-
ing the same allocation.
Arrow 1953 can thus be read as a more formal foundational version of
Marschak 1950. Money itself is behind the scenes as a security that pays

9. Here is another example of the spillover of the “protocols of war” into economics
(Klein, forthcoming).
 Mehrling / A Tale of Two Cities 211

in every state. Like the earlier papers of Marschak, Arrow’s intertempo-

ral general equilibrium model was supposed to be the analytical founda-
tion on which more policy-oriented work could be built. The Fed-MIT-
Penn econometric model (Ando and Modigliani 1969) can thus be seen as
the ultimate realization of Marschak’s vision, as can also Tobin’s “Gen-
eral Equilibrium Approach to Monetary Theory” (1969). The high-water
mark of monetary Walrasianism was also the high-water mark of Amer-
ican Keynesianism.
This sketch of the rise of monetary Walrasianism begs the question
why it ultimately fell. What happened? An article by Robert Lucas, “An
Equilibrium Model of the Business Cycle” (1975), will serve to set the
scene. Lucas takes as his text a Tobin-style monetary growth model, which
he finds inadequate in ways that have become well known: “On the one
hand, it is easy to postulate agents and market institutions which ignore or
foolishly waste information: the result is a theory which seriously under-
states agents’ abilities to vary their decision rules with changes in the
environment (such as, for example, the theory underlying the major econo-
metric forecasting models [to wit, the FMP model])” (1138). For our pur-
poses the important bit is the sentence that follows: “It is equally easy to
postulate ‘efficient’ securities markets which rapidly transmit all informa-
tion to all traders: the result is a static general equilibrium model.”
Here Lucas seems to have in mind a model like that of Arrow 1953,
which, notwithstanding the time and state subscripts on the commodities,
he interprets as hopelessly tied to a static and certain world. Prices in the
model are established once and for all at the start of time, and time involves
nothing more than selecting which contingent branch to follow at each
fork. The problems of risk and time have thus not in fact been addressed
by the intertemporal general equilibrium model, as Arrow himself would
agree. “The existence theorem for general intertemporal equilibrium can
be taken as a proof that perfect foresight is at least a consistent theory”
(Arrow 1978, 159). The analytical foundations of monetary Walrasianism
turned out not to be so secure as had been thought.
Lucas’s reference to “‘efficient’ securities markets,” which conflates a
literature in finance (Fama 1965, 1970) with the literature on intertempo-
ral general equilibrium in economics, inadvertently points the way to the
future. Already proving foundational for the emerging new field of modern
finance, the idea of efficient markets would become the foundation also for
a new direction in macroeconomics. One of the ironies of history is that it
was Paul Samuelson—a Keynesian but never a monetary Walrasian—who
got the ball rolling to clarify the idea of efficient markets by connecting it
212 History of Political Economy 42:2 (2010)

to a particular kind of stochastic process known as a martingale (Samuel-

son 1965, 1973, 1976, 1984).
Samuelson has often told the story about how his thinking on these
matters was sparked by a postcard from Jimmie Savage at the University
of Chicago that drew his attention to the early work of Louis Bachelier.10
Bachelier had assumed that security prices follow a Brownian motion and
had derived various formulae from that assumption. Samuelson’s con-
tribution was to show that efficiency meant that security prices, suitably
adjusted, would follow a martingale, which is to say that expected future
price, suitably discounted, is equal to the present price. The key to Samuel-
son’s proof was to tease out the implications of the idea that the expected
value of a pure speculation must be zero.
In all of these papers, Samuelson worked with an exogenously fixed
discount rate, and he was only ever thinking about asset prices, not at all
about macroeconomic fluctuations. In a more general economic model,
however, the discount rate should be endogenous and should fluctuate
with the economy as a whole. That is the model of Lucas’s “Asset Prices
in an Exchange Economy” (1978), in which he derives the Euler equation
that has come to orient modern macroeconomics.11 He writes it like this
(p. 1434):

( ) ( )
U′ ∑ yj pi (y) = ß ∫ U′ ∑ y′j (y′i + pi (y′))dF (y′, y).
j j

Lucas himself did not go so far as to treat this equation as the center-
piece of a real business cycle theory—he confined himself to an exchange
economy—but his students did. This is the sense in which modern mac-
roeconomics comes from the theory of efficient markets.
Indeed, in a sense the fundamental problems of finance and macro-
economics are very much the same: both seek mechanisms for controlling
the dark forces of time and ignorance, and for that purpose both adopt

10. Two decades earlier, Arrow (1941) had included a discussion of Bachelier in his master’s
thesis: “Bachelier used probabilities continuously changing in time to treat the theory of specu-
lation on exchanges. Both simple operations and options are considered. The mathematical
expectation of the speculation is taken to be 0.” Thus, Arrow was studying martingales uptown
at Columbia University at exactly the same time that he was attending Marschak’s seminar
downtown at the New School for Social Research.
11. See also LeRoy 1973. Both Lucas and LeRoy conclude from their analysis that there is
no reason to suppose that the martingale property will necessarily be a feature of efficient asset
market prices. In other words, Samuelson’s results were a special limiting case. The martingale
property was subsequently rehabilitated in finance by shifting attention to risk-neutral pricing
under a “martingale equivalent” probability measure (Harrison and Kreps 1979).
 Mehrling / A Tale of Two Cities 213

quite similar conventions of modeling the dark forces as if they followed a

well-defined stochastic process. The problem of both is therefore concep-
tualized as a problem of risk control. Finance is concerned with individ-
ual control of portfolio risk; macroeconomics is concerned with social
control of aggregate risk. Foundations having been established, subse-
quent developments in both finance and macroeconomics have been driven
by the fact that, empirically speaking, the Euler equation around which
both fields organize themselves is simply an embarrassment (Lettau and
Ludvigson 2009).
In finance, the practical response to empirical failure was to write the
Euler equation as
1 = Et [Mit + 1 Rjt + 1], (3)
where M is a stochastic discount factor treated as a free variable that must
fit the cross-section pattern of asset returns.12 It follows from 3 that
0 = Et [Mit + 1(Rjt + 1 – Rj′t + 1) ].
John Cochrane (2001) emphasizes that such a form is especially useful for
practitioners whose interest focuses on relative asset prices. In this formu-
lation, we simply take the time and risk premia implied by the data on some
asset prices and use them to calculate the price of some other assets. We do
not ask where the premia come from, although we might well develop more
or less elaborate statistical models to help us characterize them.
In macroeconomics, the practical response to empirical failure was
essentially the reverse, to abandon asset pricing consequences in order to
focus on macroeconomic variables. To achieve this, the asset in the Euler
equation has generally been treated as simply an interest rate, and the Euler
equation has been interpreted as the IS curve in a larger macroeconomic
model. Because the gross return in the Euler equation is a real return, while
the interest rates we see in the real world are nominal, there is implicitly
an expectation of price inflation involved (the AS curve), as well as an
expectation of nominal interest rates (the Taylor rule, which replaces the
LM curve of the Tobin-era models).
Michael Woodford (2001, 232) writes the IS curve version of the Euler
equation as
yt = Et yt + 1 − σ(it − Et πt + 1) + gt.

12. Under the martingale equivalent measure, equation 3 can be written as 1 = E Mt [Rt + 1] or
Pt = E Mt [Pt + 1].
214 History of Political Economy 42:2 (2010)

In this equation, y is log GDP, i is the nominal interest rate, and π is the
rate of inflation, so the difference in parentheses is the prospective real
rate of interest, and g is an exogenous disturbance (introduced in part to
mop up the empirical failure of the Euler equation). We are encouraged
to think about this equation as a linearized version of the intertemporal
Euler equation, where the parameter σ captures the intertemporal rate of
substitution. Woodford emphasizes that such a form is especially useful
for practitioners whose interest focuses on monetary policy, which is
understood here to be about setting the optimal parameters of the rule
driving the nominal rate of interest.

Conclusion
The life and work of Jacob Marschak make clear that the econometric
movement was, at least in part and at its inception, about building the ana-
lytical capacity to implement a kind of market socialism. For that pur-
pose the economy was viewed as a kind of multidimensional stochastic
process whose fluctuations had proven to be unacceptably violent. Eco-
nomic policy intervention was obviously required, but effective inter-
vention would require much more detailed knowledge of the underlying
stochastic process. Therefore, philosophical objections to the use of the
mathematics of risk for economic problems had to be put aside. A proba-
bilistic risk model, although likely a poor approximation of reality, could
not be a worse approximation than a certainty model, and anyway rational
thought about pressing economic problems could only be helped by bring-
ing out into the open the implicit assumptions underlying trained intuition.
The goal was simply to use any and all resources to defeat the dark forces
of time and ignorance.
Of course, the government is not the only one trying to defeat the dark
forces, as Lucas (1976) pointed out in his famous critique. Households
and firms too can be presumed to be steering toward their own chosen
targets in the face of their own stochastic challenges, and it is not obvi-
ous a priori that governmental intervention offers anything more than an
additional stochastic challenge for individual agents to overcome. Indeed,
in this respect, the whole development of modern finance can be under-
stood as offering an image of one sector of the economy where the inter-
action of individual portfolio allocation decisions arguably achieves an
efficient result without governmental intervention of any kind. This strong
 Mehrling / A Tale of Two Cities 215

result clearly rests on the maintained hypothesis that the dark forces can
be adequately modeled as a well-behaved stochastic process, which is cer-
tainly questionable (Arrow 1981).
In both cases, both the project of the econometric movement and the
project of modern finance, the ultimate objective was to develop practical
methods of risk control. For that purpose, again in both cases, it seemed
acceptable to abstract from the more intractable features of the problem in
order to make some progress. In effect, finance and macroeconomics have
both adopted specific conventions for treating the problem of the dark
forces. There is nothing in principle wrong with that, nor indeed anything
particularly new.13
What is more worrisome is the fact that finance and macroeconomics
have adopted different conventions, and conventions moreover that are
deeply inconsistent with one another, and this despite their mutual embrace
of the very same Euler equation. When we remember that these inconsis-
tent conventions are now deeply embedded in the institutional structures
underlying our dual economy of risk control, we see the possibility that
economic events can develop in ways that make it impossible for both con-
ventions simultaneously to adapt smoothly to a changing underlying risk
environment.

References
Ando, Albert, and Franco Modigliani. 1969. Econometric Analysis of Stabilization
Policies. American Economic Review 59.2:296–314.
Arrow, Kenneth J. 1941. Stochastic Processes. Unpublished thesis, Master of Arts
in the Faculty of Pure Science, Columbia University.
———. 1949. On the Use of Winds in Flight Planning. Journal of Meteorology
6:150–59.
———. 1951. Alternative Approaches to the Theory of Choice in Risk-Taking Situa-
tions. Econometrica 19.3:404–37.
———. 1953. The Role of Securities in the Optimal Allocation of Risk-Bearing. In
Econometrie, 41–47. Paris: Centre Nationale de la Recherche Scientifique.

13. See Keynes 1936, chap. 12: “In practice we have tacitly agreed, as a rule, to fall back
on what is, in truth a convention. The essence of this convention . . . lies in assuming that the
existing state of affairs will continue indefinitely. . . . We know from extensive experience
that this is most unlikely. . . . Nevertheless the above conventional method of calculation will
be compatible with a considerable measure of continuity and stability in our affairs, so long
as we can rely on the maintenance of the convention.”
216 History of Political Economy 42:2 (2010)

———. 1978. The Future and the Present in Economic Life. Economic Inquiry
16.2:157–69.
———. 1979. Jacob Marschak. In International Encyclopedia of the Social Sciences,
Biographical Supplement, edited by David L. Sills, 500–507. New York: Free
Press.
———. 1981. Futures Markets: Some Theoretical Perspectives. Journal of Futures
Markets 1.2:107–15.
Backhouse, Roger E., and David Laidler. 2004. What Was Lost with IS-LM. HOPE
36 (supplement): 25–56.
Breeden, Douglas T. 1979. An Intertemporal Asset Pricing Model with Stochastic
Consumption and Investment Opportunities. Journal of Financial Economics
7:265–96.
Cochrane, John H. 2001. Asset Pricing. Princeton: Princeton University Press.
Fama, Eugene F. 1965. The Behavior of Stock-Market Prices. Journal of Business
38.1:34–105.
———. 1970. Efficient Capital Markets: A Review of Theory and Empirical Work.
Journal of Finance 25.2:383–417.
Fisher, Irving. 1911. The Purchasing Power of Money. New York: Macmillan.
———. 1930. The Theory of Interest. New York: Macmillan.
Goldschmidt, R. W. 1938. Review of Kapitalbildung. Journal of Political Economy
46.4:585–87.
Hagemann, Harald. 2006. Jacob Marschak (1898–1977). In The Biographical Dic-
tionary of American Economists, edited by R. B. Emmett, 2:596–603. New York:
Thoemmes.
———. 2007. German-Speaking Economists in British Exile, 1933–1945. Banca
Nazionale del Lavoro Quarterly Review 60.242:323–63.
Harrison, J. Michael, and David M. Kreps. 1979. Martingales and Arbitrage in Mul-
tiperiod Securities Markets. Journal of Economic Theory 20:381–408.
Hicks, J. R. 1931. The Theory of Uncertainty and Profit. Economica 32 (May):
170–89.
———. 1935. A Suggestion for Simplifying the Theory of Money. Economica, n.s.,
2 (February): 1–19.
———. 1939. Value and Capital. Oxford: Clarendon Press.
Keynes, John Maynard. 1921. A Treatise on Probability. London: Macmillan.
———. 1936. General Theory of Employment, Interest, and Money. London: Mac-
millan.
Klein, Judy. Forthcoming. Protocols of War: The Mathematical Nexus of Econom-
ics, Statistics, and Control Engineering, 1940–1960.
Knight, Frank H. 1921. Risk, Uncertainty, and Profit. Boston: Houghton Mifflin.
Koopmans, Tjalling. 1947. Measurement without Theory. Review of Economic Sta-
tistics 29.3:161–72.
LeRoy, Stephen F. 1973. Risk Aversion and the Martingale Property of Stock Prices.
International Economic Review 14.2:436–46.
 Mehrling / A Tale of Two Cities 217

Lettau, Martin, and Sydney Ludvigson. 2009. Euler Equation Errors. Review of
Economic Dynamics 12:255–83.
Lucas, Robert E. 1975. An Equilibrium Model of the Business Cycle. Journal of
Political Economy 83.6:1113–44.
———. 1976. Econometric Policy Evaluation: A Critique. In The Phillips Curve and
Labor Markets, edited by Karl Brunner and Allan H. Meltzer, 19–46. Amster-
dam: North-Holland.
———. 1978. Asset Prices in an Exchange Economy. Econometrica 46.6:1429–45.
MacKenzie, Donald. 2006. An Engine, Not a Camera: How Financial Models
Shape Markets. Cambridge: MIT Press.
Makower, H., and Jacob Marschak. 1938. Assets, Prices, and Monetary Theory.
Economica, n.s., 5 (August): 261–88.
Marget, Arthur W. 1938. The Theory of Prices: A Re-examination of the Central
Problems of Monetary Theory. New York: Prentice Hall.
Marschak, Jacob. 1923. Wirtschaftsrechnung und Gemeinwirtschaft. Archiv für
Sozialwissenschaft und Sozialpolitik 51:488–500.
———. 1924. Die Verkehrsgleichung. Archiv für Sozialwissenschaft 52:344–83.
———. 1931. Elastizität der Nachfrage. Tubingen: J. C. B. Mohr.
———. 1934a. Econometric Parameters in a Stationary Society with Monetary Cir-
culation. Translated by M. Krost. Econometrica 2.1:91–111.
———. 1934b. The Meeting of the Econometric Society in Leyden, September–
October, 1933. Econometrica 2.2:187–203.
———. 1938. Money and the Theory of Assets. Econometrica 6.4:311–25.
———. 1940. Peace Economics. Social Research 7.1–4:280–98.
———. 1941a. A Discussion on Methods in Economics. Journal of Political Econ-
omy 49.3:441–48.
———. 1941b. Wicksell’s Two Interest Rates. Social Research 8.1–4:469–78.
———. 1942a. Economic Interdependence and Statistical Analysis. In Studies in Math-
ematical Economics and Econometrics in Memory of Henry Schultz, edited by
Oskar Lange, Francis McIntyre, and Theodore O. Yntema, 135–50. Chicago: Uni-
versity of Chicago Press.
———. 1942b. Identity and Stability in Economics: A Survey. Econometrica
10.1:61–74.
———. 1943. Money Illusion and Demand Analysis. Review of Economic Statistics
25.1:40–48.
———. 1945. A Cross Section of Business Cycle Discussion. American Economic
Review 35.3:368–81.
———. 1946. Neumann’s and Morgenstern’s New Approach to Static Economics.
Journal of Political Economy 54.2:97–115.
———. 1949. Role of Liquidity under Complete and Incomplete Information. Amer-
ican Economic Review 39.3:182–95.
———. 1950. The Rationale of the Demand for Money and of “Money Illusion.”
Metroeconomica 2.2:71–100.
218 History of Political Economy 42:2 (2010)

———. [1951] 1974. Comment on Mitchell. In Economic Information, Decision,

and Prediction, 3:359–70. Dordrecht: D. Reidel.
———. 1974. Economic Information, Decision, and Prediction. 3 vols. Dordrecht:
D. Reidel.
Marschak, Jacob, and William H. Andrews Jr. 1944. Random Simultaneous Equa-
tions and the Theory of Production. Econometrica 12.3–4:143–205.
Marschak, Jacob, and Oskar Lange. [1940] 1995. Mr. Keynes on the Statistical Ver-
ification of Business Cycle Theories. In Foundations of Econometric Analysis,
edited by David F. Hendry and Mary S. Morgan, 390–98. Cambridge: Cambridge
University Press.
Marschak, Jacob, and Walther Lederer. 1936. Kapitalbildung. London: Hodge.
McGoun, Elton. 2007. The History of Quantitative Risk Measurement. In vol. 2 of
Pioneers of Financial Economics, edited by G. Poitras and F. Jovanovic, 192–211.
Northampton, Mass.: Edward Elgar.
McGuire, C. B., and Roy Radner, eds. 1972. Decision and Organization: A Volume
in Honor of Jacob Marschak. Amsterdam: North-Holland.
Mehrling, Perry. 1997. The Money Interest and the Public Interest: American Mon-
etary Thought, 1920–1970. Harvard Economic Studies 162. Cambridge: Harvard
University Press.
———. 2001. Love and Death: The Wealth of Irving Fisher. In Research in the His-
tory of Economic Thought and Methodology, edited by Warren J. Samuels and
Jeff E. Biddle, 47–61. Amsterdam: Elsevier Science.
———. 2002. Don Patinkin and the Origins of Postwar Monetary Orthodoxy. Euro-
pean Journal of the History of Economic Thought 9.2:161–85.
———. 2005. Fischer Black and the Revolutionary Idea of Finance. Hoboken, N.J.:
Wiley.
Mirowski, Philip, and D. Wade Hands, eds. 2006. Agreement on Demand: Con-
sumer Theory in the Twentieth Century. Supplemental issue to vol. 38 of HOPE.
Durham, N.C.: Duke University Press.
Modigliani, Franco. 1944. Liquidity Preference and the Theory of Interest and Money.
Econometrica 12.1:45–88.
Morgan, Mary S. 1990. The History of Econometric Ideas: Historical Perspectives
on Modern Economics. Cambridge: Cambridge University Press.
Patinkin, Don. 1956. Money, Interest, and Prices: An Integration of Monetary and
Value Theory. Evanston, Ill.: Row, Peterson.
Samuelson, Paul A. 1965. Proof That Properly Anticipated Prices Fluctuate Ran-
domly. Industrial Management Review 6 (Spring): 41–49.
———. 1973. Proof That Properly Discounted Present Values of Assets Vibrate Ran-
domly. Bell Journal of Economics and Management Science 4.2:369–74.
———. 1976. Is Real-World Price a Tale Told by the Idiot of Chance? Review of Eco-
nomics and Statistics 58.1:120–23.
———. 1984. Stochastic Land Valuation: Total Return as Martingale Implying Price
Changes—A Negatively Correlated Walk. In Collected Scientific Papers, vol. 5,
essay 327. Cambridge: MIT Press.
 Mehrling / A Tale of Two Cities 219

Sargent, Thomas J. 2008. Evolution and Intelligent Design. American Economic

Review 98.1:5–37.
Tinbergen, Jan. 1939. Statistical Testing of Business Cycle Theories. 2 vols. New
York: Agatha Press.
Tobin, James. 1958. Liquidity Preference as Behavior towards Risk. Review of Eco-
nomic Studies 25 (February): 65–86.
———. 1969. A General Equilibrium Approach to Monetary Theory. Journal of
Money, Credit, and Banking 1.1:15–29.
———. 1985. Neoclassical Theory in America: J. B. Clark and Fisher. American
Economic Review 75.6:28–38.
Walter, Christian. 2006. Les martingales sur les marchés financiers: Une conven-
tion stochastique? Revue de synthèse 127.2:379–91.
Woodford, Michael. 2001. The Taylor Rule and Optimal Monetary Policy. Ameri-
can Economic Review 91.2:232–37.
———. 2003. Interest and Prices: Foundations of a Theory of Monetary Policy.
Princeton: Princeton University Press.