Media Briefings

Psychology Drives Business Cycles

  • Published Date: May 2011


‘Animal spirits’ – or psychological factors – can account for about half of the business cycle
fluctuations in the United States since the 1960s. Each recession has been preceded by a
large increase in pessimism, while excessive waves of optimism reach their maximum in
the middle of expansions. These are the central findings of research by Fabio Milani,
published in the May 2011 issue of the Economic Journal.
His findings imply an active role for government policy in trying to smooth the business
cycle. He concludes that policy-makers should intervene to avoid large swings in optimism
and pessimism that are disconnected from fundamentals and which may lead to harmful
overshooting and undershooting later on.
The study is also an important step in efforts to address perceived failings of
macroeconomic theory following the financial crisis. While this work shares the same
modelling approach as previous research, it shows that psychological forces are worth
bringing back to the core of macroeconomics.
The return of psychology to economic analysis
Psychological factors drive economic recessions and expansions. Although economists
have long recognised the importance of psychology – Keynes famously referred to ‘animal
spirits’ in his General Theory in 1936 – modern macroeconomists leave these factors out of
the models they use to understand business cycles. This is probably because they are hard
to measure and hard to model with mathematical precision.
This study provides a way to introduce psychological forces into macroeconomic models
and to assess their contribution empirically. The research shows that ‘animal spirits’ do
matter: they can account for about half of the business cycle fluctuations in the United
States since the 1960s.
Beyond rational expectations: irrational waves of optimism and pessimism
Economists have long understood the importance of expectations. Earlier economists, such
as Pigou and Keynes, emphasised businessmen’s excesses of optimism and pessimism as
determinants of economic activity. Since the 1970s, however, macroeconomic models have
been almost universally based on the paradigm of rational expectations, which implies that,
on average, expectations correspond to the outcomes from the model.
This new research relaxes the assumption of rational expectations. The study considers a
popular macroeconomic model and, as a novelty, exploits data on observed expectations
about future GDP, inflation and interest rates from the Survey of Professional Forecasters.
These forecasts serve as proxies for the general state of expectations in the economy.
The research assumes that firms and households form their expectations from a nearrational
model, in which they are allowed to learn about the workings of the economy over
time, rather than having full knowledge as under rational expectations.
But expectations can depart from the forecasts implied by the learning model: at some
points, individuals may be overly optimistic – by forecasting, for example, a higher future
output than implied by their model – or overly pessimistic. These waves of over-optimism
and over-pessimism, measured as the portion of expectations that cannot be reconciled
with the learning model, define the expectation ‘shocks’ in the analysis.
Main findings
These expectation shocks have been a major determinant of business cycle fluctuations in
the United States since the 1960s. They account for roughly half of fluctuations in GDP,
while traditional supply and demand shocks account for the remaining half.
Each recession has been preceded by a large pessimism shock, while optimism shocks
reach their maximum values in the middle of expansion phases. Psychological forces are
also responsible for the persistence of macroeconomic variables: it takes several quarters
before the effects of unwarranted jumps in optimism and pessimism die off.
As a check on their interpretation, the study shows that the identified expectation shocks
are indeed correlated with published indicators of consumer and business sentiment; they
do not spuriously reflect information sets available to forecasters that are larger than those
allowed in the model.
Implications for macroeconomic research
The financial crisis has triggered a wave of criticism of the state of macroeconomic theory.
Outside observers often view macroeconomists as enamoured with the elegance of their
constructs, with less regard for the realism of their assumptions and conclusions.
A recent bestselling book by George Akerlof and Robert Shiller titled ‘Animal Spirits’ calls
for a re-examination of the benchmark assumptions in macroeconomics to incorporate
concepts from psychology. This study is an effort in that direction: while it shares the same
modelling approach as previous research, it shows that psychological forces are worth
bringing back to the core of macroeconomics.
Implications for policy
The results are also relevant for policy-makers. While rational expectations have sometimes
been at the roots, often inappropriately, of arguments about policy ineffectiveness, the
findings here definitely imply a pro-active role for government policy.
Policy-makers should intervene to avoid large swings in optimism and pessimism that are
disconnected from fundamentals and which may lead to harmful overshooting and
undershooting later on. The lessons from the housing market boom and bust have probably
been learned.
ENDS
Notes for editors: ‘Expectation Shocks and Learning as Drivers of the Business Cycle’ by
Fabio Milani is published in the May 2011 issue of the Economic Journal.
Fabio Milani is an assistant professor of economics at the University of California, Irvine.
For further information: contact Fabio Milani on +1-949-824-4519 (email:
fmilani@uci.edu; homepage: http://www.socsci.uci.edu/~fmilani/); or Romesh Vaitilingam on
+44-7768-661095 (email: romesh@vaitilingam.com).