Journal of Economic Perspectives—Volume17,Number1—Winter2003—Pages83–104 From Efficient Markets Theory to Behavioral Finance
Robert J.Shiller
服装行业特点A cademicfinance has evolved a long way from the days when the efficient
markets theory was widely considered to be proved beyond doubt.Behav-
ioralfinance—that is,finance from a broader social science perspective including psychology and sociology—is now one of the most vital rearch pro-grams,and it stands in sharp contradiction to much of efficient markets theory.
The efficient markets theory reached its height of dominance in academic circles around the1970s.At that time,the rational expectations revolution in economic theory was in itsfirst blush of enthusiasm,a fresh new idea that occupied the center of attention.The idea that speculative ast prices such as stock prices always incorporate the best information about fundamental values and that prices change only becau of good,nsible information meshed very well with theoret-ical trends of the time.Prominentfinance models of the1970s related speculative ast prices to economic fundamentals,
using rational expectations to tie together finance and the entire economy in one elegant theory.For example,Robert Merton published“An Intertemporal Capital Ast Pricing Model”in1973,which showed how to generalize the capital ast pricing model to a comprehensive intertemporal general equilibrium model.Robert Lucas published“Ast Prices in an Exchange Economy”in1978,which showed that in a rational expectations general equilibrium,rational ast prices may have a forecastable element that is related to the forecastability of consumption.Douglas Breeden published his theory of“consumption betas”in1979,where a stock’s beta(which measures the nsitivity of its return compared to some index)was determined by the correlation
y Robert J.Shiller is the Stanley B.Resor Professor of Economics and also affiliated with the Cowles Foundation and the International Center for Finance,Yale University,New Haven, Connecticut.He is a Rearch Associate at the National Bureau of Economic Rearch, Cambridge,Massachutts.His e-mail address is͗robert.shiller@yale.edu͘.
84Journal of Economic Perspectives
of the stock’s return with per capita consumption.The were exciting theoretical advances at the time.In1973,thefirst edition of Burton Malkiel’s acclaimed book, A Random Walk Down Wall Street,appeared,which conveyed this excitement to a wider audience.
In the decade of the1970s,I was a graduate student writing a Ph.D.disrtation on rational expectations models and an assistant and associate professor,and I was mostly caught up in the excitement of the time.One could easily wish that the models were true descriptions of the world around us,for it would then be a wonderful advance for our profession.We would have powerful tools to study and to quantify thefinancial world around us.
Wishful thinking can dominate much of the work of a profession for a decade, but not indefinitely.The1970s already saw the beginnings of some disquiet over the models and a tendency to push them somewhat aside in favor of a more eclectic way of thinking aboutfinancial markets and the economy.Browsing today again throughfinance journals from the1970s,one es some beginnings of reports of anomalies that didn’t em likely to square with the efficient markets theory,even if they were not prented as significant evidence against the theory. For example,Eugene Fama’s1970article,“Efficient Capital Markets:A Review of Empirical Work,”while highly enthusiastic in its conclusions for market efficiency, did report some anomalies like slight rial dependencies in stock market returns, though with the tone of pointing out how small the anomalies were.
The1980s and Excess Volatility
From my perspective,the1980s were a time of important academic discussion of the consistency of the efficient markets model for the aggregate stock market with econometric evidence about the time ries properties of prices,dividends and earnings.Of particular concern was whether the stocks show excess volatility relative to what would be predicted by the efficient markets model.
The anomalies that had been discovered might be considered at worst small departures from the fundamental truth of market efficiency,but if most of the volatility in the stock market was unexplained,it would call into question the basic underpinnings of the entire efficient markets theory.The anomaly reprented by the notion of excess volatility ems to be much more troubling for efficiency markets theory than some otherfinancial anomalies,such as the January effect or the day-of-the-week effect.1The volatility anomaly is much deeper than tho reprented by price stickiness or tatonnement or even by exchange-rate overshoot-ing.The evidence regarding excess volatility ems,to some obrvers at least,to imply that changes in prices occur for no fundamental reason at all,that they occur becau of such things as“sunspots”or“animal spirits”or just mass psychology.
The efficient markets model can be stated as asrting that the price P t of a 1A good discussion of the major anomalies,and the evidence for them,is in Siegel(2002).
share(or of a portfolio of shares reprenting an index)equals the mathematical expectation,conditional on all information available at the time,of the prent value P*t of actual subquent dividends accruing to that share(or portfolio of shares).P*t is not known at time t and has to be forecasted.Efficient markets say that price equals the optimal forecast of it.
Different forms of the efficient markets model differ in the choice of the discount rate in the prent value,but the general efficient markets model can be written just as P tϭE t P*t,where E t refers to mathematical expectation conditional on public information available at time t.This equation asrts that any surprising movements in the stock market must have at their origin some new information about the fundamental value P*t.
It follows from the efficient markets model that P*tϭP tϩU t,where U t is a forecast error.The forecast error U t must be uncorrelated with any information variable available at time t,otherwi the forecast would not be optimal;it would not be taking into account all information.Since the price P t itlf is information at time t,P t and U t must be uncorrelated with each other.Since the variance of the sum of two uncorrelated variables is the sum of their variances,it follows that the variance of P*t must equal the variance of P t plus the variance of U t,and hence, since the variance of U t cannot be negative,that the variance of P*t must be greater than or equal to that of P t.
六级什么时候出成绩
Thus,the fundamental principle of optimal forecasting is that the forecast must be less variable than the variable forecasted.Any forecaster who forecast consistently varies through time more than the variable forecasted is making a rious error,becau then high forecasts would themlves tend to indicate forecast positive errors,and low forecasts indicate negative errors.The maximum possible variance of the forecast is the variance of the variable forecasted,and this can occur only if the forecaster has perfect foresight and the forecasts correlate perfectly with the variable forecasted.
If one computes for each year since1871the prent value subquent to that year of the real dividends paid on the Standard&Poor’s Composite Stock Price Index,discounted by a constant real discount rate equal to the geometric average real return1871–2002on the same Standard&Poor Index,onefinds that the prent value,if plotted through time,behaves remarkably like a stable trend.2In contrast,the Standard&Poor’s Composite Stock Price Index gyrates wildly up and down around this trend.Figure1illustrates the patterns.
How,then,can we take it as received doctrine that,according to the simplest efficient markets theory,the stock price reprents the optimal forecast of this prent value,the price responding only to objective information about it?I argued in Shiller(1981),as did also Stephen LeRoy and Richard Porter(1981),that the stability of the prent value through time suggests that there is excess volatilit
y in
2The prent value,constant discount rate,is computed for each year t as p*
const,t ϭ¥ϭtϩ1
ϱ(Ϫt)D,
翻译英语句子
whereis a constant discount factor,and D t is the real dividend at time t.An assumption was made about real dividends after2002.See note to Figure1.
Robert J.Shiller85
the aggregate stock market,relative to the prent value implied by the ef ficient markets model.Our work launched a remarkable amount of controversy,from which I will recall here just a few highlights.
just so soThe principal issue regarding our original work on excess volatility was in regard to thinking about the stationarity of dividends and stock prices.My own work in the early 1980s had followed a tradition in the finance literature of
Figure 1
Real Stock Prices and Prent Values of Subquent Real Dividends
(annual data )
taber
200020201980196019401920PDV, Constant Discount Rate
PDV, Interest Rates
freeloop
PDV, Consumption
Real Stock Price (S&P 500)
19001880186010
100100010000
P r i c e Notes:The heaviest line is the Standard &Poor 500Index for January of year shown.The less-heavy line is the prent value for each year of subquent real dividends accruing to the index discounted by the geometric-average real return for the entire sample,6.61percent.Dividends after 2002were assumed equal to the 2002dividend times 1.25(to correct for recent lower dividend payout)
了不起的盖茨比英文简介
and growing at the geometric-average historical growth rate for dividends,1.11percent.The thin line is the prent value for each year of subquent real dividends discounted by one-year interest rates plus a risk premium equal to the geometric average real return on the market minus the geometric average real one-year interest rate.The dashed line is the prent value for each year of subquent real dividends discounted by marginal rates of substitution in consumption for a reprentative individual with a coef ficient of relative risk aversion of 3who consumes the real per capita nondurable and rvice consumption from the U.S.National Income and Product Accounts.Real values were computed from nominal values by dividing by the consumer price index (CPI-U since 1913,linked to the Warren and Pearson producer price index before 1913)and rescaling to January 2003ϭ100.Some of the very latest obrvations of underlying ries were estimated bad on data available as of this writing;for example,the consumer price index for January 2003was estimated bad on data from previous months.Source data are available on ͗ale.edu/ϳshiller ͘,and the further descriptions of some of the data are in Shiller (1989).See also footnotes 1,5and 6.
86Journal of Economic Perspectives
From Efficient Markets Theory to Behavioral Finance87 assuming that dividendsfluctuated around a known trend.3However,one might also argue,as do Marsh and Merton(1986),that dividends need not
stay clo to a trend and that even if earnings followed a trend,share issuance or repurcha could make dividends depart from a trend indefinitely.In addition,if business managers u dividends to provide a smoothedflow of payouts from their busi-ness,then the stock prices might be expected to shift more rapidly than divi-dends.Marsh and Merton argued that such dividend smoothing could make stock prices unstationary in such a way that infinite samples prices appear more volatile than the prent values.
Thus,the challenge became how to construct a test for expected volatility that modeled dividends and stock prices in a more general way.As such tests were developed,they tended to confirm the overall hypothesis that stock prices had more volatility than an efficient markets hypothesis could explain.For example, West(1988)derived an inequality that the variance of innovations(that is,surpris) in stock prices must be less than or equal to the variance of the innovations in the forecasted prent value of dividends bad on a subt of information available to the market.This inequality is quite general:it holds even when dividends and stock prices have infinite variances so long as the variance of the innovation(the unexpected change)in the isfinite.Using long-term annual data on stock prices, West found that the variance of innovations in stock prices was four to20times its theoretical upper bound.4John Campbell and I(1988)recast the time ries model in ter
resultfrom
ms of a cointegrated model of real prices and real dividends,while also relaxing other assumptions about the time ries,and again found evidence of excess volatility.5Campbell(1991)provided a variance decomposition for stock returns that indicated that most of the variability of the aggregate stock market conveyed information about future returns,rather than about future dividends.
Another contested issue regarding the early work on excess volatility ques-tioned the assumption of the early work that the efficient markets model was best conveyed through an expected prent value model in which the real discount rate is constant through time.The assumption of a constant discount rate over time can only be considered afirst step,for the theory suggests more complex relationships. 3It should be pointed out that dividend payouts as a fraction of earnings have shown a gradual downtrend over the period since1871and that dividend payouts have increasingly been substituted by share repurchas.Net share repurchas reached approximately1percent of shares outstanding by the late1990s.However,share repurchas do not invalidate the theoretical model that stock prices should equal the prent value of dividends.See Cole,Helwege and Laster(1996).
shuttlebus4In more technical terms,this argument is over whether dividends could be viewed as a stationary ries. The discussion was often phrad in terms of the“unit root”property of the time ries,where a unit root refers to notion that when a variable is regresd on its own lags,the characteristic equati
on of the difference equation has a root on the unit circle.West(1988)can be viewed as a way of addressing the unit root issue.In our1988paper,Campbell and I handled nonstationarity by using a vector autore-gressive model including the log dividend-price ratio and the change in log dividends as elements.
5Barsky and De Long(1993),however,later showed that if one assumes that real dividends must be twice differenced to induce stationarity(so that dividends are even more unstationary in the n that dividend growth rates,not just levels,are unstationary),then the efficient markets model looks rather more consistent with the data.
>3166是什么意思
