Section4 FamaEMH.ppt(1).ppt

Efficient Capital Markets:,The Journal of Finance,Vol 46,No.5(Dec,1991)By Eugene F.Fama,Menu,1.The Theme2.The Main Areas of Research3.Return Predictability:Time-Varying Expected Returns4.Cross-Sectional Return Predictability5.Event Studies6.Tests for Private Information7.Conclusions,1.The Theme,Strong version of market efficiency hypothesis:information and trading costs are 0Weaker version:prices at the point where marginal benefit do not exceed MC.1st obstacle:The positive costs make the strong version hypothesis false but it could provide a good benchmark and avoid the trouble of calculating costs.2nd obstacle:The joint-hypothesis,the market efficiency is not testable but must be jointly tested with some equilibrium model(some asset pricing models),the test results will become ambiguous among both aspects,Does the joint hypothesis problem make empirical work on asset-pricing models uninteresting?,The answer is absolutely no.It has changed our views about the behavior of returns and practices of market professionalsTherefore,the purpose of this paper is that the market efficiency literature should improve our ability to describe the time-series and cross-section behavior of security returns.Although the joint-hypothesis problem exists,we cannot deny the richness and usefulness of past empirical works.,2.The Main Areas of Research,1970 review:weak-form tests;semi-strong form tests and strong-form testsNowadays the above three developed into the following respectively:(1)tests for return predictability(2)event studies(3)tests for private information,Return predictability is most detailed and the evidence of this topic is most controversial.New works believe returns are predictable from past returns,dividend yields and term structure variables.It reject the old models but head onto the joint-hypothesis problem:Does return predictability reflect rational variation through time in expected returns or irrational deviations of price from fundamental value?New tests remain suspicious but give us some new methodologies.Event studies are discussed next briefly.Its most advantage is allowing a break between market efficiency and equilibrium-pricing issues The tests for private information problem are also shadowed by the joint-hypothesis problem,3.Return Predictability:Time varying Expected Returns,Research on time-series predictability of stock returns.Recent tests consider the forecast power of D/P,E/P and term structure variables.They also examine returns for longer horizons,not only short term predictability checked by the early workIn contrast to short term returns,the predictable component of return grows to about 40%of the variance of 2-10-year returns.,3.A Past Returns,A.1 Short-Horizon ReturnsA.2 Long-Horizon ReturnsA.3 The Contrarians,A.1 Short-Horizon Returns,Pre-1970 work,tests of market efficiency assumed that expected returns are constant through time.And they found that daily,weekly and monthly returns are predictable from past returns.Fama(1965),Fisher(1966)In recent work,Lo and MacKinlay(1988)found nonsynchronous trading effect,which says that autocorrelation is stronger for small stocks.Conrad and Kaul(1988),mitigate this problem not completely by using Wed-Wed returns.,French and Roll(1986)result that stock prices are more volatile in open market,including a market inefficiency hypothesis which says the higher variance of price changes is transitory.So the pricing errors due to noise trading will be reversed eventually.,After testing the transitory effect,French and Roll found transitory portion for different samples.,A.2 Long-Horizon Returns,The early view that AR of short-horizon returns close to 0 imply insignificance is challenged by Shiller(1984)and Summers(1986),who showed that stock prices swings away from fundamental values in a highly inefficient market.If the AR1 has a slope less but close to one,the short horizon returns will have little autocorrelation and look like random walk.Stambaugh(1986)claims that long-horizon returns have strong negative AR and the swings away from fundamental values will be reversed.Fama and French(1988a)also found positive short-horizon returns AR and negative long-horizon returns AR.,A.3 The contrarians,DeBondt and Thaler(1985,1987)attack the market efficiency by unmasking irrational bubbles.They find that extreme losers will gain over 3-5years and extreme winners are just the opposite,which is attributed to market overreactionChan(1988)and Ball and Kothari(1989)argue the winner-loser results are due to failure to risk-adjust returns.,B Other forecasting Variables,The univariate tests on long-horizon returns of Fama and French(1988a)and Poterba and Summers(1988)are a statistical power failure.Less noisy proxies for expected returns should be foundB1 The evidence:D/P(Rozeff 1984,Shiller 1984)E/P(Campbell and Shiller 1988b,etc),B2 Market Efficiency,Relation between D/P and prices in an efficient market or in irrational bubbles.Fama and French(1988b):there is no evidence that low D/P signals bubbles,which is negative expected stock returns.Fama and French(1989):C and I of firms.two term-structure variables:default spread and term spread,Ferson and Harvey(1991)formally test the common expected returns hypothesis.Fama and French(1989)find that the variation in expected returns tracked by D/P and the default spread increase from low risk to high risk assets.And this reconciles with modern intertemporal asset-pricing models(Lucas,Breeden)and original consumption-smoothing stories.But the bubbles fan believes that the returns variation may just mean bubbles are correlated across assets and markets.,B3 A Caveat,The slopes are large and so have some uncertainty in forecast power.Industry-level data-dredging problem.,C Volatility Tests and Seasonals in Returns,C.1 Volatility Tests:This test is not informative about market efficiency,giving no help on the issue of whether the variation in expected returns is rational.,C2 Return seasonality,January effect,4.Cross-Sectional Return Predictability,In pre-1970 efficient markets literature,the common models are constant expected returns model and the market modelPost-1970 asset-pricing models:one-factor Sharpe-lintner-Black(SLB)model;multifactor models of Merton and Ross;consumption based CAPM group,4A SLB Model,A.1 Early Success:expected returns are a positive linear function of market betaRolls criticism is too strong for the proxies of market portfolio.,A2 Anomalies,Empirical attacks on the SLB models Basu(1977,1983)Banz(1981):stocks size explainsOne argument:the beta are noisy and the anomalies are correlated with the true betaBall(1978):use multifactor models to explain the SLB anomaliesOther work shows there is spillover among the SLB anomalies.,How many of the SLB anomalies have roles in expected returns?,Fama and French(1991):E/P,leverage and book-to-market equity weaken but do not absorb the relation between size and expected returns.In truth,the premier SLB anomaly is the weak role of market beta in the cross-section of returns.Stambaugh(1982)finds that:there is positive relation between expected returns and beta across bonds and stocks;the relation between expected returns of stocks and beta is weak.,A3 Market Efficiency:A4 The Bottom Line:SLB model is a sharp benchmark and passes the test of practical usefulness.,B Multifactor Models,Ross 1976:APTLehman and Modest(1988)test 15 factors,leaving an unexplained size effect like SLB model.Tests of APT leads to the problems of factors number.The limitation of factor analysis:it does not give answer to how these factors relate to uncertainties about consumption and portfolio opportunities.Chen,Roll and Ross(1986):address the economic implication question(flexibility),C Consumption-Based Asset-Pricing Models,Rubinsten 1976,Lucas 1978,Breeden 1979:the interaction between optimal consumption and portfolio decisions leads to a positive linear relation between the expected returns and betas.Tests use tastes or joint distribution of consumption growth and returns.,Empirical work tests its time series and cross-section predictions,using GMM(Hansen and Singletion 1982).But tests of the consumption model failUnconditional tests of Equity-premium puzzle(Mehra and Prescott1985):must have high risk aversion to explain the spread,for one example habit formation(Constantinides 1990)But high risk aversion may be true,D Where Do We Stand,D1 The Bad News:Rejections of SLB models are common and the consumption-based model fares worse than SLB models.The multifactor mode seems better,but using ex post data for variables.D2 The good News:Rejections are not clean.The three models are not exclusive.,5 Event Studies,Fama,Fisher,Jensen and Roll(1969).Event study is useful in corporate finance.A.Some of the Main Results:Unexpected changes in dividends are associated with stock price changes of the same sign,in contrast to MM theory.New issues of common stock are bad news for stock prices.Corporate-control transactions like mergers and tender offers,B Market Efficiency,When the stock price response to an event is large and concentrated in a few days,calculating abnormal returns has little implications.Quick adjustment is consistent with efficiency.However,the dispersion of returns increases around information events,and event studies on average adjustment of prices to information cannot tell us the variance.,When the response of prices to information is slow,event studies have joint-hypothesis problem.Some event studies suggest that stock prices do not respond quickly to specific information.,6 Tests for Private Information,1970 view:only two cases of market inefficiencyAfter 1970:The profitability of insider trading is now in detail;Some security analysts have information not reflected in stock prices;Some professional investment managers have access to private information.,6A Insider Trading,Jaffe(1974)finds that the market does not react quickly to public information about insider trading,8 months afterSeyhun(1986):insider buying is more important in small firms.,6B Security analysis,The value line rankings are another anomalyStickel(1985)uses event-study methods to show that there is an announcement effect in rank changes that more clearly implies that value line has information no reflected in prices.The information in Value Line rank changes is also stronger for small stocks.,C Professional Portfolio Management,Jensen(1968,1969):for the 1945-1964 period,returns in funds are on average 1%per year below the market line,and then concludes that mutual-fund managers do not have private information.But recent studies differ.Henriksson(1984)finds that average returns to fund investors are trivially different from the Sharpe-Lintner market line.Ippolito(1989)finds that fund returns are 0.83%per year above the market line and concludes his results to the“noisy rational expectation”(Grossman and Stigitz 1980),Ippolitos mutual-fund evidence is not confirmed by tests on pension plans.How to reconcile the opposite results for mutual funds and pension funds?Elton,Gruber,Das and Hklarka(1991);,7 Conclusions,Event studies:it can give a clear picture of the speed of adjustment of prices to informationPrivate Information:insiders and outsiders;Changes in value lines rankings of firms lead to permanent changes in stock prices.The investors studied are most mutual fund managers.The limitations of three popular models,Return predictability,Controversy about market efficiency centers largely on this domainPositive short-horizon return AR and negative long-horizon return AR,but the latter has less precision.Other variables used:dividend yields,E/P,default spreads,term spreads,etc.,Real economy equilibrium perspective:rational variation in expected returns is caused either by shocks to tastes for consumption or by technology shocks.Two tasks in the future:The 1st task;If the above is true,we can profit from the work on variation in expected returns across bonds and stocks,across international markets and work on factors related to the cross-section of expected returns,The 2nd task is to establish links between expected returns and business conditions,like corporate financial index or growth rates of output.,Some preparations:Event study Analysis,Seven steps:Event definition;Selection criteria;Normal and abnormal returns;Estimation procedure;Testing procedure;Empirical results;Interpretation and conclusions.,CAPMAPTCCAPM,ICAPM,SDF,The general equilibrium and the no arbitrage equilibrium,An important perspective into the difference between economics and finance researchTheoretical background relationship:one example:interest rate term structure models,