NEWS & VIEWS RESEARCH
reliable substitute for actual experiments is
thus becoming a reality.
Walter Thiel is at the Max-Planck-Institut
fr Kohlenforschung, 45470 Mlheim an der
Ruhr, Germany. Gerhard Hummer is at the
Max-Planck-Institut fr Biophysik,
ECONOMICS
60438 Frankfurt am Main, Germany.
e-mails: thiel@kofo.mpg.de;
gerhard.hummer@biophys.mpg.de
3. Senn, H. M. & Thiel, W. Angew. Chem. Int. Edn 48,
11981229 (2009).
4. Levitt, M. & Warshel, A. Nature 253, 694698 (1975).
5. Rosta, E., Nowotny, M., Yang, W. & Hummer, G.
J. Am. Chem. Soc. 133, 89348941 (2011).
6. Lindorff-Larsen, K., Piana, S., Dror, R. O. & Shaw, D. E.
Science 334, 517520 (2011).
7. Takamori, S. et al. Cell 127, 831846 (2006).
8. Zhao, G. et al. Nature 497, 643646 (2013).
1. Warshel, A. & Karplus, M. J. Am. Chem. Soc. 94,
56125625 (1972).
2. Warshel, A. & Levitt, M. J. Mol. Biol. 103, 227249
(1976).
SHORT-TERM UNPREDICTABILITY
EMPIRICAL FINANCIAL ECONOMICS
by John Y. Campbell
Fama showed that asset prices are extremely
difficult to predict in the short term.
The Sveriges Riksbank Prize in Economic
Sciences in Memory of Alfred Nobel was
awarded to Eugene F. Fama, Lars Peter Hansen
and Robert J. Shiller, whose empirical analysis
of asset prices has shaped our understanding
of how markets work (see figure).
Price
Predicting asset prices
Rapid movement
in response
to news
Days
amas efficient market hypothesis (EMH)
argues that competition among investors
makes the return from using information
on stock prices commensurate with the cost
of that information. Thus, if costs are zero,
prices correctly reflect all relevant information1. According to this hypothesis, if we could
easily predict that stock prices will rise tomorrow, we would all buy today, such that prices
would in fact rise today until they reflected the
information we had received. Tests by Fama in
the 1960s found that short-run returns were
mainly unpredictable, which is consistent
with a market that incorporates information
efficiently.
Fama emphasized that the EMH was not
directly testable; one can only test a joint
hypothesis of the EMH and a model detailing
the way in which expected returns are set. If, say,
small-company stocks generally outperform
large-company stocks, this might not indicate
that the pricing of small companies is inefficient, but rather that small-company stocks are
riskier and hence their investors demand high
expected returns as compensation2.
In 1981, Shiller showed that historical prices
were excessively volatile relative to their future
realized value3. This suggested that although
prices respond quickly to information, they
change for other reasons as well. Shiller interpreted this volatility as resulting from investor sentiment. Subsequent work linked excess
volatility to predictable variation in long-run
returns; short-term predictability was later
found as well.
These findings presented a serious challenge to the EMH, but Famas joint hypothesis
allows a possible explanation: time-varying
LONG-TERM PREDICTABILITY
But Shiller showed that there is greater predictability
over the longer term, and interpreted this finding as
market inefficiency resulting from investor behaviour.
Bubble and
predictable
downturn
Price
E F F I C I E N CY A N D V O L AT I L I T Y
by Christopher Polk
Crash and
predictable
recovery
Years
TESTING THEORIES
Hansens statistical techniques for testing economic
theories highlighted the attractiveness of stocks to
investors who can tolerate risk.
inancial markets continually generate
vast quantities of data on asset prices.
Fama, Shiller and Hansen have led an effort,
over almost 50 years, to use these data to
better understand the economy and investor
behaviour.
Fama observed that the return on any risky
financial asset is the sum of a required return
that a rational investor expects to earn and an
unexpected return driven by the arrival of
news. He noted that, over short time periods,
the volatility of unexpected returns is much
greater than any movement in the required
return, and hence that short-term price movements accurately reflect the news hitting the
market at each point of time.
Hansen built on Famas insight, developing
a powerful statistical method to extract from
asset returns information about key properties of the economy, such as investors average
aversion to risk, without having to model other
features of the economy that are irrelevant to
the problem at hand4,5.
Shiller pointed to data indicating that large
price swings result from the accumulation of
movements in required returns over long periods of time, and that unexpected returns reflect
not only news about the future payments that
assets will make, but also unexpected changes
in the required return6.
Together, their work has definitively shown
the value of empirical research in understanding price formation in financial markets. Fama
and Shiller have also used financial data to
construct indexes that summarize the movements of broad categories of assets, such as
groups of stocks with similar characteristics
and houses in the same city 2.
Christopher Polk is in the Department of
Finance, London School of Economics, London
WC2A 2AE, UK. John Y. Campbell is in the
Department of Economics, Harvard University,
Cambridge, Massachusetts 02138, USA.
e-mails: c.polk@lse.ac.uk;
john_campbell@harvard.edu
expected returns may be due to time-varying
risk and/or risk aversion. Understanding the
sources rational and sentiment-based of
predictable variation in returns is at the heart
of modern financial economics.
1. Fama, E. F. J. Finance 25, 383417 (1970).
2. Fama, E. F. & French, K. R. J. Financ. Econ. 33, 356
(1993).
3. Shiller, R. J. Am. Econ. Rev. 71, 421436 (1981).
4. Hansen, L. P. Econometrica 50, 10291054 (1982).
5. Hansen, L. P. & Jagannathan, R. J. Pol. Econ. 99,
225262 (1991).
6. Campbell, J. Y. & Shiller, R. J. Rev. Financ. Stud. 1,
195228 (1988).
5 D E C E M B E R 2 0 1 3 | VO L 5 0 4 | NAT U R E | 9 7
2013 Macmillan Publishers Limited. All rights reserved
