Nobel Prize Lecture in Economics
The Study of Causal Relationships in the
Macroeconomy: The Contributions of
Christopher Sims
Eric Zivot
Robert Richards Chaired Professor of Economics
© Eric Zivot 2012
Christopher Sims Receives Nobel Prize
• Award given “for their
empirical research on
cause and effect in
the macroeconomy”
• Sargent and Sims
developed methods
for identifying and
quantifying the effects
of changes in policy
© Eric Zivot 2012
Christopher Sims at Work
© Eric Zivot 2012
Christopher Sims – Trombone Player!
© Eric Zivot 2012
Short Biography
Harvard, PhD 1968
Harvard, Assistant Professor 1968-70
Minnesota, Associate Professor 1970-1974
Minnesota, Professor 1974-1990
Yale, Henry Ford II Professor 1990-1999
– One of my PhD advisors!!!
• Princeton, Harold H. Helm '20 Professor of
Economics and Banking 1999 - present
© Eric Zivot 2012
Sims’s Main Contributions
• Statistical tests for causal relations
– “Money, Income, and Causality”, AER 1972
• Vector autoregression (VAR) modeling framework
– “Macroeconomics and Reality”, ECTA, 1980
– “Comparison of Interwar and Postwar Business
Cycles”, AER, 1980
– “Are Forecasting Models Usable for Policy Analysis?”,
Fed Quarterly Review, 1986.
• Use of Bayesian statistical methodology
– “Forecasting and conditional projection using realistic
prior distributions”. Econometric Reviews, 1984
– “Bayesian Methods for Dynamic Multivariate Models",
(with Tao Zha), IER, 1998
© Eric Zivot 2012
“Money, Income and Causality”
• Keynesian-Monetarist Debate about the causes of
business cycles
– Keynesian view: business cycles are primarily demand-driven
– Monetarist view: business cycles are primarily the result of
monetary policy mistakes
• Monetarist’s argued that growth in the stock of money
was tightly related to growth in income and they argued
that this relationship was causal - fluctuations in money
growth causing fluctuations in income.
• A statistically estimated equation with income explained
by current and past money growth implied that most of
the business cycle could be eliminated by simply making
money supply growth constant.
© Eric Zivot 2012
“Money, Income, and Causality”
• Examine the substantive question: Is there
statistical evidence that changes in the money
supply is causal in the money-income
• Develop a direct statistical test for the existence
of unidirectional causality.
– Complements the causality tests of Clive Granger,
who was also awarded the Nobel Prize for his work
on the statistical analysis causality.
© Eric Zivot 2012
Causality Testing: Framework
• If the monetarists were right in claiming that the
strong correlations of money growth with income
primarily reflected a causal influence of
monetary policy errors on income, future money
growth should not contribute to explaining
current income, once the influence of current
and past money growth on income had been
accounted for.
© Eric Zivot 2012
Causality Testing: Results
• The monetarist regressions passed the test.
Future money growth did not help predict current
• Problem: Framework only considers two
– What about interest rates, prices, unemployment,
exchange rates, etc?
– Neglecting these variables could cloud the moneyincome causality results
© Eric Zivot 2012
“Macroeconomics and Reality”
• Circa 1980, the state of the art in
macroeconometrics was to use large-scale
Keynesian-type structural models.
• These models often involved scores or even
hundreds of equations, essentially a S=D
equation for every important market, identities to
make sure things add up correctly, etc.
• But in order to estimate the parameters of these
models, the structural parameters as they are
known, you had to overcome the identification
© Eric Zivot 2012
“Macroeconomics and Reality”
• The identification problem essentially asks if its
possible to estimate the structural parameters at
• The answer, in general, is no. For example, if
every variable in the model appears in every
equation, then it won't be possible to estimate
the structural model due to simultaneity bias.
© Eric Zivot 2012
Identification of Structural Models
© Eric Zivot 2012
“Macroeconomics and Reality”
• In large models, exclusions are numerous, and
many researchers simply assumed whatever
exclusion restrictions were needed to achieve
identification, and then went on to estimate the
• Sims argued the assumptions that researchers
were imposing to achieve identification had no
theoretical basis. They were ad hoc and difficult
to defend - especially when expectations are in
the model
© Eric Zivot 2012
Sims’s VAR Modeling Framework
• Step1: Construct statistical forecasting model
(VAR) to separate expected from unexpected
• Step 2: Identify causal links and extract
fundamental shocks (VAR → SVAR)
• Step 3: Trace effects of fundamental shocks on
economy over time – Impulse response analysis
© Eric Zivot 2012
Step 1: Construct VAR
• Forecasting model for a collection of macro
variables without ad hoc identifying restrictions
• Each variable is forecast using lagged values of
itself and all other variables
– Variable = Forecast (expected) + error (unexpected)
• Short-term forecasting accuracy is typically quite
© Eric Zivot 2012
Step 2: Identify Causal Links
• Identify causal links and extract fundamental
shocks using theory-based and non-ad hoc
– Some variables react to others with a lag (e.g., M only
responds to Y with a lag)
– Long-run neutrality arguments (e.g., M has no longrun impact on Y)
• Process transforms errors in VAR equations into
fundamental shocks
© Eric Zivot 2012
Step 3: Impulse Response Analysis
Q: How do macro variables respond to unexpected
fundamental shocks ?
© Eric Zivot 2012
The Monetarist Debate Viewed Through SVARs
• Sims and others were able to show using
SVAR's that influences of monetary policy were
detectable in the data.
• But at the same time, they showed that most
movements in both money stock and interest
rates represented systematic reactions of
monetary authorities to the state of the
• Only a small part of macroeconomic fluctuations
could be attributed to erratic monetary policy.
© Eric Zivot 2012
Bayesian Methodology
• The textbook frequentist view distinguishes nonrandom, but unknown, “parameters” from
random quantities that repeatedly vary, or could
conceivably repeatedly vary.
• The Bayesian view treats everything that is not
known as random, until it is observed, after
which it becomes non-random
• Think of coin flipping experiment
© Eric Zivot 2012
Bayesian Methodology
• A Bayesian approach comfortably
accommodates uncertain prior information.
• In a large model, it allows introducing sensible
restrictions on the values of unknown
parameters, without pretending that these
restrictions are without uncertainty.
• That is, it allows introducing probability
distributions for model parameters, then allowing
the data to update or sharpen those
• It thereby avoids the need to imply unrealistic
precision in the probability distributions for
model predictions. © Eric Zivot 2012
Concluding Remarks
• Sims changed the way macroeconomists and
econometricians model the economy
• After Sims, the profession moved away from
using large scale Keynesian-style structural
macro-models and adopted the SVAR approach
• IRF analysis from SVARs provide the “stylized
facts” from which modern theory-based models
are calibrated and evaluated
• Bayesian methods have become widely
© Eric Zivot 2012

similar documents