DP27 Landmann Short Run Macro PDF
DP27 Landmann Short Run Macro PDF
DP27 Landmann Short Run Macro PDF
Oliver Landmann
January 2014
ISSN 1866-4113
University of Freiburg
Department of International Economic Policy
Discussion Paper Series
Editor:
Prof. Dr. Günther G. Schulze
ISSN: 1866-4113
Electronically published: 29.01.2014
Short‐Run Macro After the Crisis:
The End of the “New” Neoclassical Synthesis?
Oliver Landmann
Revised, January 2014
Abstract
The Financial Crisis of 2008, and the Great Recession in its wake, have shaken up macroeco‐
nomics. The paradigm of the “New” Neoclassical Synthesis, which seemed to provide a ro‐
bust framework of analysis for short‐run macro not long ago, fails to capture key elements of
the recent crisis. This paper reviews the current reappraisal of the paradigm in the light of
the history of macroeconomic thought. Twice in the past 80 years, a major macroeconomic
crisis led to the breakthrough of a new paradigm that was to capture the imagination of an
entire generation of macroeconomists. This time is different. Whereas the pre‐crisis consen‐
sus in the profession is broken, a sweeping transition to a single new paradigm is not in sight.
Instead, macroeconomics is in the process of loosening the methodological straightjacket of
the “New” Neoclassical Synthesis, thereby opening a door for a return to its original pur‐
pose: the study of information and coordination in a market economy.
JEL Classification: B22, B40, E10, E12, E13
Keywords: Financial Crisis, Great Recession, Macroeconomics, New Neoclassical Synthesis,
Keynesian Economics, New Classical Economics, Great Moderation
Institut für allgemeine Wirtschaftsforschung
Universität Freiburg
Platz der Alten Synagoge
D‐79085 Freiburg
Germany
Tel.: ++49(0)761‐203‐2326
Fax: ++49(0)761‐203‐2405
e‐mail: [email protected]‐freiburg.de
http://www.macro.uni‐freiburg.de/news/home
2
“The State of Macro Is Good” (Blanchard 2008)
“The State of Macro Is Not Good” (Krugman 2009)
1. Introduction1
On the eve of the financial crisis of 2008, macroeconomics was at ease and at peace with
itself. A long history of internal feuding between competing paradigms, often conducted in a
heated atmosphere, seemed to have come to an end. A state of reconciliation was finally
achieved, a “New Neoclassical Synthesis” (Goodfriend/King 1997), embodied in a macroeco‐
nomic consensus model. This model offered a blueprint for monetary policy which promised
to maintain price stability and to keep cyclical output volatility at a minimum. The prevailing
self‐confidence of macroeconomists was well captured by Lucas (2003, p. 1) who opened his
presidential address to the American Economic Association with this assessment:
Short‐run macro has fulfilled its duty, Lucas implied, so macroeconomists should redirect
their focus to the more important supply‐side issues of long‐term growth and development.
In a similar vein, Blanchard (2009) concluded that the state of macro was “good” after many
years of “enormous progress and substantial convergence”. At the time, these assessments
seemed to be vindicated by more than two decades of low inflation and moderate output
fluctuations in the advanced industrialized countries. This “Great Moderation”, as it came to
1
The author acknowledges helpful comments from participants at Wirtschaftswissenschaftliches
Seminar Ottobeuren, “Entwicklung, Stand und Perspektiven der Wirtschaftswissenschaft”, Septem‐
ber 2013.
3
be called (Bernanke 2004), was widely credited to the wise policies of central bankers who
followed the guidelines set by the New Neoclassical Synthesis model.
This happy state of affairs came to an abrupt end when the Financial Crisis of 2008 pushed
the world into its most severe slump since the Great Depression. All of a sudden, pervasive
macroeconomic instability and uncertainty was back, and so were the bitter fights among
macroeconomists about what to do about it. Although policymakers responded swiftly and
mostly with the right moves, the mainstream macroeconomic policy framework did not pre‐
pare them well for what they were facing nor was it much help in crafting the proper re‐
sponse.
As so often before, a major macroeconomic shock has come as a surprise to both policymak‐
ers and to academia. Inevitably, then, the New Neoclassical Synthesis came under attack and
macroeconomists hurried back to the drawing board to figure out what had gone wrong.
This paper attempts to give a preliminary assessment of what this soul‐searching has pro‐
duced so far. Will the New Neoclassical Synthesis be overhauled and resurrected? Or does
short‐run macroeconomics require an altogether different paradigm (Buiter 2009, De
Grauwe 2009)? And if so, how could it look?
The remainder of this paper is structured as follows: Section 2 looks back at the evolution of
macroeconomics from the Keynesian revolution to the recent New Neoclassical Synthesis.
Section 3 discusses the response of the research community to the 2008 Financial Crisis,
emphasizing two themes: the role of the financial sector in macroeconomic models and the
significance of microfoundations for the direction and scope of macroeconomic analysis.
Often in the past, times of crisis have presented a window of opportunity for the break‐
through of sweeping new paradigms. As Section 4 concludes, however, this time is different.
2. A Brief History of Macroeconomic Theory since Keynes
Short‐run macroeconomics, constantly propelled by the interaction of events and ideas
throughout the 20th century, has come a long way since Keynes (1936). Disregarding the in‐
complete and incoherent strands of pre‐Keynesian business cycle theory (surveyed by
Haberler 1937), we can distinguish three broad stages in the long journey of macroeconom‐
ics since the Great Depression: The first stage took macroeconomics from the Keynesian
4
Revolution to the Neoclassical Synthesis of the post‐war period. The next stage opened with
the New Classical challenge against the Neoclassical Synthesis which eventually culminated
in the Real Business Cycle Theory of the 1980s. The last part of the journey was marked by a
revival of Keynesian thought (“New Keynesian Macroeconomics”) and coincided with the
period of the Great Moderation. The term “New Neoclassical Synthesis” refers to the merger
of the key elements of New Classical and New Keynesian theory. Figure 1 summarily sketch‐
es this history whose three stages are now briefly reviewed in turn.
2.1 From the Keynesian Revolution to the Neoclassical Synthesis
As most scientific revolutions, the Keynesian revolution of the 1930s resulted from a massive
failure of the ruling paradigm to account for important observed facts ‐ in this case, the
depth and persistence of the Great Depression. However, the Keynesian revolution did not
displace the old paradigm completely, but left economics divided into two inconsistent and
competing visions of how a market economy works. What Keynes (1936) called “classical
theory”, was largely the neoclassical tradition of equilibrium price and market theory which
described the coordinating power of markets and the efficiency of market‐determined re‐
source allocation. The Keynesian income‐expenditure model, in contrast, with its core con‐
cept of an underemployment equilibrium, justified thorough skepticism with regard to the
aggregate homeostatic properties of a market economy. The controversy between “Keynes
and the Classics” (Hicks 1937) gradually died down as the post‐war economy proved much
more stable and resilient than many observers had been led to believe by somber visions of
secular stagnation (Hansen 1938).
A clever way of reconciling classical and Keynesian thought both with one another and with
the relative macroeconomic stability of the post‐war period was provided by Samuelson
(1955). He argued that a market economy, for the reasons spelled out by Keynes, cannot be
trusted with ensuring full employment on its own. But if monetary and fiscal policies are
employed to solve the problem of underemployment, they would place the economy on a
trajectory on which the principles of classical equilibrium analysis applied for the purpose of
explaining relative prices and resource allocation. This was the neoclassical synthesis. It
seemed to give an accurate and logically coherent account of macroeconomic developments
in the advanced economies during the quarter century after World War II when govern‐
ments began to use the tools of Keynesian counter‐cyclical policy actively and rapid growth
5
Interwar years:
Great Depression
Figure 1: From the Keynesian Revolution to the New Neoclassical Synthesis
prevailed with only minor cyclical fluctuations. In fact, confidence in the ability of macroeco‐
nomic policy grew increasingly stronger in this period, to the point where conventional wis‐
dom had convinced itself that the business cycle was “obsolete” (Bronfenbrenner 1969).
6
Arguably, this confidence ‐ or, as it soon turned out, overconfidence ‐ carried the seeds of
self‐destruction. Sure enough, as soon as the business cycle was thought to be vanquished, it
was back with a vengeance ‐ and the consensus of the neoclassical synthesis was gone.
2.2 The Rise of New Classical Macroeconomics
The shocks that ended the golden age of the post‐war world economy came from the supply
side. Inflationary expectations, a consequence of the gradual acceleration of inflation during
the 1960s, and the explosive rise of commodity prices destroyed the stability of the Phillips
Curve which had become a central pillar of the neoclassical synthesis. The resulting stagfla‐
tion, the simultaneous rise of inflation and unemployment, at first proved hard to accom‐
modate for the Keynesian short‐run macroeconomic model which was essentially a model of
aggregate demand.
The failure of the Keynesian paradigm to predict stagflation prepared the ground for the
New Classical counter‐revolution. Led by Robert Lucas, the New Classicals regarded the neo‐
classical synthesis as completely discredited by the breakdown of the Phillips Curve, as so
“wildly incorrect” and “fundamentally flawed” that there was no point in trying to repair it
(Lucas/Sargent 1979). What they proposed instead, was a reconstruction of macroeconom‐
ics from the first principles of general equilibrium theory, deriving all behavioral equations of
the model form individual rationality. Most importantly, Lucas (1972) extended the axiom of
rationality to the formation of expectations.
Contrary to New Classical allegations, the old model of the neoclassical synthesis was far
from hopelessly damaged. As a matter of fact, the repair work was begun almost immedi‐
ately. Augmented by an expectations‐augmented Phillips Curve and incorporating supply
shocks, the old model did quite well tracing the co‐movement of output, employment and
the price level, as conditioned by the origin of exogenous shocks (Gordon 2009). This ex‐
tended model continues to be the essence of what most intermediate‐level textbooks pre‐
sent as short‐run macroeconomic theory (Blanchard/Johnson 2013, Mankiw/Ball 2011).
However, the New Classical research agenda, due to its initial empirical success, its logical
coherence and its promise of reuniting microeconomics and macroeconomics, captured the
imagination of an entire generation of young macroeconomists in the 1970s and 1980s and
thus became the industry standard for what it meant to do serious macroeconomics. The
7
new paradigm turned out to have surprisingly powerful implications, both in disciplining the
specification of macroeconomic models and in conditioning their empirical predictions.
More importantly, it completely changed the normative tone of macroeconomics. Whereas
the neoclassical synthesis saw the government as the natural guarantor of macroeconomic
stability, the New Classicals insisted on subjecting the case for any stabilization policy to the
usual standards of welfare economics: If the government was to intervene in any way, it
must first have identified the particular market failure it wished to correct. But since the
New Classical models consisted of perfectly rational agents interacting on perfectly competi‐
tive markets, a role for an active stabilization policy was virtually ruled out by assumption.
The New Classical movement came in two waves. The first wave, initiated by Lucas (1972),
was a monetary business cycle model which attributed aggregate output movements to in‐
dividual producers confusing aggregate and relative price changes. When it became increas‐
ingly clear that this particular model was hard pressed to give a satisfactory account of the
depth and persistence of recessions, the New Classicals steered their paradigm away from
the tradition of monetary business cycle theory and turned it into a purely Real Business
Cycle (RBC) model. The RBC approach was pioneered by Kydland/Prescott (1982) and
Long/Plosser (1983). It rejected the premise, shared by both the neoclassical synthesis and
Lucas’s monetary model, that business cycles can be interpreted as mostly demand‐driven
departures from a supply‐determined path of potential output. In the RBC model, all output
movements are instead regarded as manifestations of a stochastic trend, resulting from the
optimal response of an optimizing representative individual to various unexpected exoge‐
nous shocks. The shocks are modeled as time‐series processes with an autoregressive struc‐
ture well known to the agents populating the models. In fact, the time‐series properties of
the exogenous shocks drive the dynamic behavior of RBC models to a large extent.
How good an explanation of business cycles is the RBC theory? The answer depends a lot on
what one means by an “explanation”. The New Classical methodology has changed that
meaning in a subtle way. Traditionally, explanatory power was measured by the success of
observable exogenous variables to account for the time‐series variation of some observable
endogenous variable. The RBC research agenda set itself the objective of building models
that can replicate the time‐series properties of key macroeconomics variables as fully as
possible. Similar as these two criteria sound, they do not amount to the same thing. This can
8
be seen most clearly from the way RBC theory measures technology shocks, the major driv‐
ers of its dynamics. Starting from a standard growth‐accounting framework, technology is
measured as the Solow residual in the usual way. As it happens, a growth‐accounting de‐
composition of the cyclical short‐run variation of output attributes most of this variation to
fluctuations of the Solow residual. That is, in production function terms, most of the short
run variation of output remains unexplained ‐ the residual is a “measure of our ignorance”,
as Abramovitz (1956) famously put it many years ago. The RBC methodology is to use this
measure of ignorance as an exogenous explanatory variable in its business cycle model
where ‐ surprise, surprise! ‐ it turns out to be a major driver of output fluctuations and is
indispensable for the replication of observed output movements. Simulated changes in out‐
put fit the actual changes quite closely, which RBC theory records as success. But we still do
not know what is behind the variation of the Solow residual. How much progress is this?
2.3 The New Keynesian Revival and the “New” Neoclassical Synthesis
The principle of building every macroeconomic relation on sound microeconomic founda‐
tions had, and continues to have, a lot of appeal, both intellectually and, so it appeared at
least initially, in terms of empirical explanatory power. In fact, the desirability of microeco‐
nomic foundations for macroeconomics was recognized long before the New Classicals ap‐
peared on the stage. A large body of theoretical work in macroeconomics, carried out well
within the framework of the neoclassical synthesis in the 1950s and 1960s, was mainly con‐
cerned with providing behavioral foundations to the central pillars of Keynesian theory: the
consumption function, the investment function, and the money demand function.
When inflation increased and the stability of the Phillips Curve became an issue, the micro‐
foundations literature shifted its attention to the supply side of the economy. The dramatic
implications of rational expectations in an equilibrium model of output and inflation, first
pointed out by Lucas (1972), underlined the urgency of this research. Once Fischer (1976)
and others had demonstrated that a model with a nominal rigidity, such as a contract‐
related delay in the adjustment of nominal wages or prices, has distinctly Keynesian features
even under rational expectations, the challenge was to develop a sound theoretical explana‐
tion of nominal rigidities. A “sound theory”, of course, made sure it did not depend on
money illusion or some other ad‐hoc irrationality on the part of households or firms. The
9
resulting research agenda spawned a huge literature which soon became known as “The
New Keynesian Economics” (Mankiw/Romer 1991).
What distinguishes New Keynesianism from the old Keynesianism of the Neoclassical Syn‐
thesis is the incorporation of most of the methodological axioms of New Classical Macroeco‐
nomics: Neutrality of money in long‐run equilibrium (a vertical Phillips Curve at natural out‐
put or unemployment), strictly forward‐looking behavior (rational expectations), an explicit
dynamic stochastic general equilibrium (DSGE) framework, and the derivation of all behav‐
ioral equations ‐ not just for the private sector, but also for the central bank ‐ from in‐
tertemporal optimality conditions. In short, a New Keynesian model looks a lot like a Real
Business Cycle model, with the addition of money and some sand in the wheels in form of a
nominal rigidity in price setting.
What made the New Keynesian model particularly popular was its ability to provide a
framework for a systematic monetary policy aiming at stable prices and output. The model
suggests a flexible inflation targeting strategy, implemented through a policy rule that has
the interest rate respond to movements in inflation and output, very much along the lines of
a Taylor rule. With its particular blend of New Classical methodology and New Keynesian
short‐run non‐neutrality of money, the model appeared to have achieved a new reconcilia‐
tion of the Classical and Keynesian traditions that fell out with each other so bitterly in the
1970s. Hence, “New Neoclassical Synthesis”.
As pointed out above, the establishment of this new paradigm as the mainstream in mone‐
tary macroeconomics coincided with the Great Moderation, a period of low inflation and
unusually stable output in the advanced economies, extending from the second half of the
1980s to the eve of the Financial Crisis in 2008. A number of reasons have been cited for the
benign macroeconomic climate of that period: improved inventory management in the pri‐
vate sector, improved macroeconomic management by policymakers, and plain luck, i.e. the
absence of major destabilizing shocks. How much of the credit is due to central banks re‐
mains controversial. There is solid evidence, however, that the macroeconomic turbulences
of the 1970s and early 1980s were associated with monetary policies which strayed widely
from the guideline of the Taylor rule (Judd/Rudebusch 1998). Not surprisingly, then, central
banks and academic macroeconomists grew increasingly confident that proper management
of the short‐term interest rate was enough to keep the macro economy on course. Once
10
again, then, just as in the late 1960s, the business cycle was believed conquered. And once
again, it did not fail to make a powerful comeback.
3. The Reappraisal of Macroeconomics After the Financial Crisis
The history of macroeconomics, as sketched above, reveals one salient parallel between
macroeconomic theory and its object, the macro‐economy: They are both subject to cycles.
They both experience times of relative calm, followed by phases of extraordinary turbulence.
And clearly, whatever one may think about the relation of macroeconomic theory to reality,
their states of calm and excitement are highly correlated. Macroeconomic turmoil was al‐
ways associated with facts and experiences which the prevailing theory found hard to ac‐
commodate. The consequences each time were excitement and controversy in the research
community. This was true in the 1930s when Keynes launched the Keynesian Revolution in
response to the inability of the contemporary mainstream to provide a plausible account for
persistent high unemployment during the Great Depression. It happened again in the 1970s
when the Neoclassical Synthesis struggled to accommodate the stagflationary shocks of the
time, which paved the way for the triumph of the “New Classical” movement. Today, the
“New Neoclassical Synthesis” pays the price for its complete neglect of the vagaries of finan‐
cial markets.
3.1 Placing the Financial Sector Back into Macroeconomics
How badly was macroeconomics shaken by the Financial Crisis of 2008? Her Majesty, the
Queen of England, embarrassed economists when she famously wondered: "If these things
were so large, how come everyone missed them?" The implied expectation that economics
should progress to the point where it can predict an event like the Financial Crisis may be
asking too much. After all, seismology was not declared a failure when it failed to pinpoint
the timing of the last major earthquake. What seismology is expected to provide, however,
are an understanding of the underlying mechanisms, an assessment of looming risks, and
indicators for early warning systems. Held against this standard, mainstream macroeconom‐
ics admittedly did a lot worse than seismology. The key elements in the causation of the Fi‐
nancial Crisis were totally absent from the paradigm of the New Neoclassical Synthesis. The
financial sector was reduced to a single money market interest rate which could be con‐
11
trolled by the central bank and acted as the sole transmitter of monetary‐policy impulses to
the goods and labor markets. There were no banks, no shadow banks, no subprime lending,
no securitization, no interconnectedness between banks, no leverage, no bubbles , no liquid‐
ity crises, no deleveraging ‐ nothing of that sort. This is not to say that there was not con‐
siderable analysis and insight about these phenomena, but whatever knowledge existed was
scattered and unconnected to the macroeconomic model of the New Neoclassical Synthesis.
The canonical macro model and financial economics simply had “no point of contact”
(Friedman 2013).
When an anthill suffers damage from a major blow, the immediate consequence is a period
of high excitement and activity as the ants scramble to rebuild and fortify the structure. This
is very much what happened in macroeconomics after the trauma of the Financial Crisis. The
architects of the New Neoclassical Synthesis quickly went to work to fill the glaring gaps in
their analysis and to add details of the financial sector to their models. The theory was ex‐
tended to analyze the proper response of monetary policy to credit spreads and to the vol‐
ume of credit supply (Curdia/Woodford 2010), to highlight the central bank balance sheet as
an instrument of monetary policy (Curdia/Woodford 2011), and to identify disruptions in
financial intermediation as a source of disturbances to economic activity (Gertler/Kiyotaki
2010), to name just a few examples. A large literature has developed along these lines. A
survey is provided by Brunnermeier et al. (2012).
This work is by no means limited to sticking financial add‐ons onto the canonical mainstream
model. Rather, some of the fundamental strategic simplifications of the model must be re‐
viewed critically. In a world of homogeneous representative agents, for example, it is hard to
rationalize a social function for financial intermediation between lenders and borrowers. Nor
could a deleveraging crisis or a Fisherian debt deflation occur in the absence of heterogene‐
ous agents who differ with regard to their spending patterns.
3.2 Realism vs. Microfoundations: The Case of the Phillips Curve
The efforts now under way to reconnect macroeconomics and financial economics promise
substantial progress. However, the reconstruction of macroeconomics is unlikely to stop
there. As Blanchard (2008) pointed out in his pre‐crisis survey, the convergence of macro‐
economics towards the canonical new‐synthesis model may have gone too far. Almost ritu‐
12
ally, the same set of optimality and equilibrium conditions appeared in paper after paper,
hardly questioned any more because they had become the defining features of serious mac‐
roeconomics. The calm of the Great Moderation did not subject this received wisdom to a
serious test. When the test came and the received wisdom was found wanting, everything
that went unquestioned for a long time was up for a critical reappraisal. Some of the critics
used strong language. Buiter (2009) flatly declared “most mainstream macroeconomic theo‐
retical innovations since the 1970s … to be self‐referential, inward‐looking distractions at
best.” (Krugman 2009) accused macroeconomics of mistaking “beauty, clad in impressive‐
looking mathematics, for truth”. Putting the hyperbole aside, a legitimate case can be made
that some of the standard assumptions of the new‐synthesis model owe their popularity
more to their theoretical elegance and analytical convenience than to their accuracy in de‐
scribing actual behavior (Caballero 2010).
A case in point is the New Keynesian Phillips Curve, originally proposed by Calvo (1983),
which is the standard model of inflation in the New Neoclassical Synthesis. This model starts
from a clever, but highly artificial and unrealistic set‐up designed to introduce a nominal ri‐
gidity into a world of perfectly forward looking rational price setters. The nominal rigidity
results from the assumption that price setters get an opportunity to adjust their prices only
at intervals that are stochastically determined and exogenous to them. An awkward sce‐
nario. But the resulting inflation equation is parsimonious and elegant, and it meets the re‐
quirement that the behavioural relations of a macroeconomic model must not be posited ad
hoc, but should be derived from the first‐order conditions of some explicit optimization
problem solved by rational individuals. Lack of realism in an assumption is not, as such, a sin
in economic modelling. But the elegance of the New Keynesian Phillips Curve comes at a
price: Its main empirical implication ‐ that there is inertia in the price level, but not in the
inflation rate ‐ is strongly contradicted by the facts (Fuhrer/Moore 1995, Mankiw 2001).
How is it, then, that the new‐synthesis model can generate the observed persistence of infla‐
tion if such persistence is not a feature of the New Keynesian Phillips Curve? Only by building
the persistence into the time‐series representations of the exogenous shocks which con‐
stantly drive the economy away from its steady‐state equilibrium path.
Thus, just as the RBC theory on which it builds, the canonical New Keynesian model lets its
exogenous shocks do most of the work in replicating the observed time‐series behavior of
13
the macro‐economy, rather than deriving these properties from its own inherent dynamic
structure. The irony here is that some simple specifications of the Phillips curve which allow
for an inertial backward‐looking momentum of inflation do a remarkably good job in describ‐
ing the inflation process, but are considered “ad hoc” because they are hard to derive theo‐
retically from a model of rational forward‐looking agents. The New Keynesian Phillips curve,
in contrast, is not considered ad hoc because it solves a straightforward maximization prob‐
lem. A clean theoretical rationalization of inflation inertia would be highly desirable, of
course. But in the meantime, one is left to wonder who is the arbiter of “ad‐hockery” in mac‐
roeconomics. Why is it ad hoc to assume downward wage rigidity or inflation inertia right
away while it is not ad hoc to assume some arbitrary exogenous frequency λ at which prices
can be revised?
3.3 Back to the Roots: Macroeconomics as the Study of Information and Coordination
On a more fundamental level, it has become increasingly clear that the desire of anchoring
aggregate relations in individual behavior, which is entirely uncontroversial as a general
principle, has led macroeconomics into an axiomatic straightjacket of rules that define what
is and is not an acceptable model. These rules were laid down early on by the New Classicals
with the commendable objective of using the reductionist neoclassical rationality principle
to discipline macroeconomic modelling. But what started as an agenda of containing arbi‐
trary discretion eventually morphed into a set of arbitrary principles itself that shut off once‐
important strands of thinking. Macroeconomics thereby lost sight of issues which, if pursued
more systematically, would have left it better prepared for the crisis it faced in 2008. Two
themes in particular moved off the radar screen of mainstream macroeconomics because
the New Classical Microfoundations agenda somehow answered them a priori: The break‐
down of coordination in a decentralized market economy and the processing of dispersed
information by households and firms.
Arguably, macroeconomics once started out as the study of coordination failures in large
systems of interconnected decentralized markets (Leijonhufvud 1981). In particular, mass
unemployment was explained by Keynes (1936) as a failure of the market economy to coor‐
dinate the transaction plans of quantity‐constrained households and firms. Whether or not
this theme was adequately captured by the IS‐LM apparatus of the neoclassical synthesis is
debatable. But there is no question that macroeconomics completely abandoned the analy‐
14
sis of coordination failures when it adopted the New Classical research agenda (Laidler 2009,
Leijonhufvud 2009a, Spahn 2009). Once Robert Lucas had convinced the research commu‐
nity that stable macroeconomic relations could only be obtained if one started with optimiz‐
ing agents interacting through cleared markets in a competitive economy, coordination fail‐
ures were defined away. Departures from first‐best outcomes could still occur, but only be‐
cause incomplete information prevented agents from reading the market signals correctly.
Thus, while transaction plans may have been based on a mistaken interpretation of informa‐
tion available to households and firms, they were still perfectly coordinated. In such a set‐
up, the government had no business intervening in the machinery of coordination. If there
was any role for public policy, it was to avoid becoming a source of macroeconomic noise
itself.
Real Business Cycle theory led macroeconomics even further down that road. It eliminated
markets and prices by reducing macroeconomics to the study of a rational, but lonely Robin‐
son Crusoe who faces shocks to his productivity. In a Robinson Crusoe economy, coordina‐
tion cannot become an issue by definition. Moreover, that same Robinson, while living in a
stochastic world with uncertainty, fully understands the particular shocks to which he is ex‐
posed and he also understands the stochastic nature of the system that generates these
shocks. In this way, any possible problems associated with the processing of information
have been moved off the radar screen of macroeconomics as well. The Robinson Crusoe
model is about as far as one can go in distancing oneself from the study of coordination fail‐
ures. The lonely nerd who is perfectly informed about the relevant frequency distributions
over an infinite time horizon and who understands and strictly observes his intertemporal
budget constraint is a poor starting point if one is out to understand a world of burst bub‐
bles, failed Ponzi games and mass defaults as typically produced by a financial crisis.
The New Keynesians, eager to retain as much of the New Classical Microfoundations as pos‐
sible in order to evade the criticism that had brought down the old neoclassical synthesis,
have largely bought into the Robinson‐Crusoe caricature of the macro economy. The New
Neoclassical Synthesis thus differs from the RBC model only marginally ‐ mainly by pouring
some sand into the wheels of Robinson’s optimization machinery in the form of a nominal
rigidity which delays the response of the price level to a monetary disturbance. The resulting
short‐run non‐neutrality of money is what gives the New Neoclassical Synthesis that
15
“Keynesian” flavour which is needed to get a role for money and monetary policy in a theory
of business cycles. By its very construction, this “stability‐with‐frictions” paradigm cannot
address the more severe instability brought about by a genuine breakdown of coordination
(Leijonhufvud 2009a).
What is more, a whole range of coordination issues is resolved a priori by the rational‐
expectations hypothesis which, as a central tenet of New Classical thought, was also em‐
braced by the New Neoclassical Synthesis. The rationality of rational expectations refers to
the efficient use of information and it implies not only an astonishing information‐processing
capability of households and firms, but also a high degree of implicit coordination among
them. As agents are assumed to form expectations in line with the model they inhabit, they
are assumed to have the full picture given by the model and, therefore, to share the same
picture. On the face of it, this is a remarkable behavioural proposition. At the very least, it
makes for a curious contrast with the fierce debates raging among professional economists
about the right macroeconomic model.
In the aftermath of the 2008 Financial Crisis, the standards of what passes as respectable,
microfounded macroeconomics have begun to shift away from the axiomatic postulates of
the New Classical School towards an approach which places more emphasis on behavioural
realism (Akerlof/Shiller 2009). In particular, ways are explored to move beyond rational ex‐
pectations towards a paradigm which takes into account learning and which allows for
agents struggling to process new and confusing information (De Grauwe 2012,
Frydman/Phelps 2013, Woodford 2013). Most of the ideas that are developed along these
lines are not entirely new. Concepts like herding, animal spirits, or multiple equilibria have
been developed and explored to various degrees in the past, but they have subsequently
been pushed to the fringes of mainstream macroeconomics because they did not fit into the
self‐imposed straightjacket of the New Neoclassical Synthesis.
4. Conclusion: This Time Is Different
This paper has reviewed the evolution of macroeconomics from Keynes to the New Neoclas‐
sical Synthesis and sketched the reappraisal of the state of knowledge that has begun after
the Financial Crisis of 2008. The story of macroeconomics was told in section 2 as a sequence
16
of crises, paradigms, paradigm wars, and attempted syntheses between conflicting para‐
digms:
♦ The Great Depression of the 1930s paved the way for the breakthrough of the
Keynesian paradigm, but left the field divided between microeconomics and macro‐
economics.
♦ The Neoclassical Synthesis was an attempt to reconcile “Mr. Keynes and the Classics”
(Hicks 1937), but did not bring unity methodologically.
♦ The Stagflation of the 1970 cast doubt on the Neoclassical Synthesis and helped the
New Classical Revival win the day. Microfoundations‐cum‐Rational‐Expectations be‐
came the new industry standard.
♦ The New Neoclassical Synthesis was an attempt to press the Keynesian notions of
nominal rigidity and short‐run non‐neutrality of money into the New Classical mold.
This paradigm provided an analytical foundation for a monetary‐policy doctrine
which was followed by numerous central banks and widely credited for the Great
Moderation.
♦ Yet another macroeconomic crisis, the Financial Crisis of 2008, shattered the Great
Moderation and discredited the New Neoclassical Synthesis.
But this time is different in one important respect. This time, no single new revolutionary
paradigm was waiting in the wings to topple the ruling paradigm and to take its place. Ra‐
ther, there is a wealth of ideas for reshaping macroeconomics. Some of them can be ac‐
commodated within the general framework of the New Neoclassical Synthesis, others re‐
quire a more radical departure from traditional thinking.
It would be premature to consign the New Neoclassical Synthesis to the scrap heap of aban‐
doned economic ideas. After all, it has worked remarkably well for a long time during the
Great Moderation. But the calm of that period lulled macroeconomists into a complacency
which prevented them from noticing how much they had narrowed down their subject, both
in scope and methodology. The experience of the Financial Crisis of 2008 and the subse‐
quent slump of the world economy suggests at least two general lessons for macroeconom‐
ics:
17
First, the failure of the New Neoclassical Synthesis to capture any of the mechanisms that
led to the Financial Crisis revealed a fatal blind spot, namely the almost total neglect of the
financial sector. More fundamentally, macroeconomic theory failed to see that the Great
Moderation fostered complacency and encouraged risk taking, thereby undermining its own
foundations as well as those of the underlying macroeconomic paradigm. One could inter‐
pret this failure in a broad sense as a vindication of the Lucas Critique (Lucas 1976). But even
more, it vindicated the analysis of heterodox thinkers such as Hyman Minsky (1986) who had
developed a theory of a cycle of stability and instability in capitalist economies, based on
endogenous shifts in attitudes towards risk.
Second, the market system appears to obey quite different laws of motion in tranquil peri‐
ods like the Great Moderation and in times of crisis. This observation calls for a generaliza‐
tion of macroeconomic theory to allow for non‐linearities: Small shocks that push the econ‐
omy not too far away from its steady‐state equilibrium may entail a different, more benign
response than large, once‐in‐a‐lifetime shocks. Leijonhufvud (1973, 2009b) has evoked the
notion of a corridor around the long‐run equilibrium which divides the impulse‐response
space into compartments with distinctly different response patterns. The relevance of this
general idea became evident when interest rates across the world hit the zero lower bound
after 2008 and the profession was forced to relearn the macroeconomics of the liquidity
trap. Even within the broad logic of the New Neoclassical Synthesis, it is possible to derive
results that appear paradoxical when compared to what one would expect under normal
conditions when monetary policy has traction (Eggertson 2010). A noteworthy and widely
discussed example is the multiplier effect of fiscal policy.
Thus, like the Great Depression of the 1930s, the Financial Crisis of 2008 has shattered the
prevailing consensus in the profession. But unlike the Great Depression or the stagflation of
the 1970s, it has not produced a sweeping transition to a single new paradigm. When histo‐
rians of economic thought will look back to how the 2008 Financial Crisis has reshaped mac‐
roeconomics, they may well conclude that the field has come out of the crisis less self‐
confident and less unified, but with a richer agenda and a broader scope as it struggled to
escape from the analytical straightjacket of the New Neoclassical Synthesis.
18
References
Abramovitz, Moses (1956), Resource and Output Trends in the US Since 1870, American Eco‐
nomic Review, 46, 5‐23.
Akerlof, George and Robert Shiller (2009), Animal Spirits, Princeton and Oxford.
Bernanke, Ben (2004), The Great Moderation, Remarks at the meetings of the Eastern Eco‐
nomic Association, Washington, DC, February 20.
Blanchard, Olivier (2009), The State of Macro, Annual Review of Economics, 1(1), 209‐228.
Blanchard, Olivier and David Johnson (2013), Macroeconomics, Boston et al.
Bronfenbrenner, Martin (1969), Is the Business Cycle Obsolete? New York.
Brunnermeier, Markus, Thomas Eisenbach and Yuliy Sannikov (2012), Macroeconomics with
Financial Frictions, NBER Working Paper No. 18102, May.
Buiter, Willem (2009), The unfortunate uselessness of most ‘state of the art’ academic
monetary economics, http://blogs.ft.com/maverecon, March 3.
Caballero, Ricardo (2010), Macroeconomics After the Crisis: Time to Deal with the Pretense‐
to‐Knowledge Syndrome, Journal of Economic Perspectives, 24, 85‐102.
Calvo, Guillermo (1983), Staggered Prices in a Utility‐Maximizing Framework”, Journal of
Monetary Economics, 12, 383‐398.
Curdia, Vasco and Michael Woodford (2010), Credit Spreads and Monetary Policy, Journal of
Money, Credit, and Banking, 42 (supplement 1), 3‐35..
Curdia, Vasco and Michael Woodford (2011), The Central‐Bank Balance Sheet as an Instru‐
ment of Monetary Policy, Journal of Monetary Economics, 58, 54‐79.
De Grauwe, Paul (2009), Top‐Down versus Bottom‐Up Macroeconomics, Paper presented at
the CESifo Conference “What’s Wrong With Macroeconomics?”, Munich, November 6‐7.
De Grauwe, Paul (2012), Booms and Busts: New Keynesian and Behavioural Explanations, in
Robert Solow and Jean‐Philippe Touffut (eds.), What’s Right with Macroeconomics?
Cheltenham.
Eggertsson, Gauti (2010), What Fiscal Policy Is Effective at Zero Interest Rates? NBER Macro‐
economics Annual, 59‐112.
Fischer, Stanley (1977), Long Term Contracts, Rational Expectations, and the Optimal Money
Supply Rule, Journal of Political Economy, 85, 191‐205.
Friedman, Benjamin (2013), The Simple Analytics of Monetary Policy: A Post‐Crises Ap‐
proach, Harvard University, January.
19
Frydman, Roman and Edmund Phelps, eds. (2013), Rethinking Expectations, Princeton.
Fuhrer, Jeffrey and George Moore (1995), Inflation Persistence, Quarterly Journal of Econom‐
ics, 110, 127‐160.
Gertler, Mark and Nobuhiro Kiyotaki (2010), Financial Intermediation and Credit Policy in
Business Cycle Analysis, in Benjamin Friedman and Michael Woodford (eds.), Handbook
of Monetary Economics, Vol. 3, 547‐599.
Goodfriend, Marvin and Robert King (1997), The new Neoclassical Synthesis and the Role of
Monetary Policy. In Ben Bernanke and Julio Rotemberg (eds.), Macroeconomics Annual
12, 231‐283.
Gordon, Robert (2009), Is Modern Macro or 1978‐era Macro More Relevant to the Under‐
standing of the Current Economic Crisis?, Paper presented to International Colloquium
on the History of Economic Thought, Sao Paulo, Brazil, August 3.
Haberler, Gottfried (1937), Prosperity and Depression, Geneva.
Hansen, Alvin (1938), Full Recovery or Stagnation?, New York.
Hicks, John (1937), Mr. Keynes and the Classics, A Suggested Interpretation, Econometrica, 5,
147‐159.
Judd, John and Glenn Rudebusch (1998), Taylor’s Rule and the Fed: 1979‐1997, Federal Re‐
serve Bank of San Francisco Economic Review, 3, 3‐16.
Keynes, John Maynard (1936), The General Theory of Employment, Interest, and Money.
London.
Krugman, Paul (2009), How Did Economists Get It So Wrong? New York Times Magazine,
September 6.
Kydland, Finn und Edward Prescott (1982), Time to Build and Aggregate Fluctuations,
Econometrica, 50, 1345‐1370.
Laidler, David (2009), Lucas, Keynes, and the Crisis, Journal of the History of Economic
Thought, 32, 39‐62.
Leijonhudvud, Axel (1973), Effective Demand Failures, Swedish Economic Journal, March,
reprinted in Leijonhufvud (1981), 103 129.
Leijonhudvud, Axel (1981), Information and Coordination, New York and Oxford.
Leijonhudvud, Axel (2009a), Macroeconomics and the Crisis: A Personal Appraisal, CEPR Pol‐
icy Insight 41, London.
Leijonhudvud, Axel (2009b), Out of the Corridor: Keynes and the Crisis, Cambridge Journal of
Economics, 33, 741‐757..
20
Long, John und Charles Plosser (1983), Real Business Cycles, Journal of Political Economy, 91,
39‐69.
Lucas, Robert E. (1972), Expectations and the Neutrality of Money, Journal of Economic The‐
ory, 4, 103‐124.
Lucas, Robert E. (1976), Econometric Policy Evaluation: A Critique, Carnegie‐Rochester Con‐
ference Series on Public Policy, 1, 19‐46.
Lucas, Robert E. and Thomas Sargent (1979), After Keynesian Economics, Federal Reserve
Bank of Minneapolis Quarterly Review, 3, 1‐16.
Mankiw, N. Gregory (2001), The Inexorable and Mysterious Tradeoff Between Inflation and
Unemployment, Economic Journal, 111, C45‐C61.
Mankiw, N. Gregory and Laurence Ball (2011), Macroeconomics and the Financial System,
New York.
Mankiw, N. Gregory and David Romer, eds. (1991), The New Keynesian Economics, Cam‐
bridge.
Minsky, Hyman (1986), Stabilizing the Unstable Economy, New Haven and London.
Spahn, Peter (2009), The New Keynesian Microfoundations of Macroeconomics, Jahrbuch für
Wirtschaftswissenschaften, 60, 181‐203.
Woodford, Michael (2013), Macroeconomic Analysis without the Rational Expectations Hy‐
pothesis, NBER Working Paper No. 19368, August.