CHAPTER11TheEfficientMarketHypothesisCHAPTER11TheEfficientMarketMauriceKendall(1953)foundnopredictablepatterninstockprices.Pricesareaslikelytogoupastogodownonanyparticularday.Howdoweexplainrandomstockpricechanges?EfficientMarketHypothesis(EMH)MauriceKendall(1953)foundnEfficientMarketHypothesis(EMH)EMHsaysstockpricesalreadyreflectallavailableinformationAforecastaboutfavorablefutureperformanceleadstofavorablecurrentperformance,asmarketparticipantsrushtotradeonnewinformation.Result:Priceschangeuntilexpectedreturnsareexactlycommensuratewithrisk.EfficientMarketHypothesis(EEfficientMarketHypothesis(EMH)Newinformationisunpredictable;ifitcouldbepredicted,thenthepredictionwouldbepartoftoday’sinformation.Stockpricesthatchangeinresponsetonew(unpredictable)informationalsomustmoveunpredictably.Stockpricechangesfollowarandomwalk.EfficientMarketHypothesis(EFigure11.1CumulativeAbnormalReturnsBeforeTakeoverAttempts:TargetCompaniesFigure11.1CumulativeAbnormaFigure11.2StockPriceReactiontoCNBCReportsFigure11.2StockPriceReactiInformation:ThemostpreciouscommodityonWallStreetStrongcompetitionassurespricesreflectinformation.Information-gatheringismotivatedbydesireforhigherinvestmentreturns.Themarginalreturnonresearchactivitymaybesosmallthatonlymanagersofthelargestportfolioswillfindthemworthpursuing.EMHandCompetitionInformation:ThemostpreciousWeakSemi-strongStrongVersionsoftheEMHWeakVersionsoftheEMHTechnicalAnalysis-usingpricesandvolumeinformationtopredictfuturepricesSuccessdependsonasluggishresponseofstockpricestofundamentalsupply-and-demandfactors.WeakformefficiencyRelativestrengthResistancelevelsTypesofStockAnalysisTechnicalAnalysis-usingpriTypesofStockAnalysisFundamentalAnalysis-usingeconomicandaccountinginformationtopredictstockpricesTrytofindfirmsthatarebetterthaneveryoneelse’sestimate.Trytofindpoorlyrunfirmsthatarenotasbadasthemarketthinks.SemistrongformefficiencyandfundamentalanalysisTypesofStockAnalysisFundameActiveManagementAnexpensivestrategySuitableonlyforverylargeportfoliosPassiveManagement:NoattempttooutsmartthemarketAcceptEMHIndexFundsandETFsVerylowcostsActiveorPassiveManagementActiveManagementActiveorPasEvenifthemarketisefficientaroleexistsforportfoliomanagement:DiversificationAppropriaterisklevelTaxconsiderationsMarketEfficiency&

PortfolioManagementEvenifthemarketisefficienResourceAllocationIfmarketswereinefficient,resourceswouldbesystematicallymisallocated.Firmwithovervaluedsecuritiescanraisecapitaltoocheaply.Firmwithundervaluedsecuritiesmayhavetopassupprofitableopportunitiesbecausecostofcapitalistoohigh.Efficientmarket≠perfectforesightmarketResourceAllocationIfmarketsEmpiricalfinancialresearchenablesustoassesstheimpactofaparticulareventonafirm’sstockprice.Theabnormalreturnduetotheeventisthedifferencebetweenthestock’sactualreturnandaproxyforthestock’sreturnintheabsenceoftheevent.EventStudiesEmpiricalfinancialresearcheReturnsareadjustedtodetermineiftheyareabnormal.MarketModelapproach:a.rt=a+brmt+et

(ExpectedReturn)b.ExcessReturn=

(Actual-Expected)

et=rt-(a

+brMt)HowTestsAreStructuredReturnsareadjustedtodetermMagnitudeIssueOnlymanagersoflargeportfolioscanearnenoughtradingprofitstomaketheexploitationofminormispricingworththeeffort.SelectionBiasIssueOnlyunsuccessfulinvestmentschemesaremadepublic;goodschemesremainprivate.LuckyEventIssueAreMarketsEfficient?MagnitudeIssueAreMarketsEffWeak-FormTestsReturnsovertheShortHorizonMomentum:GoodorbadrecentperformancecontinuesovershorttointermediatetimehorizonsReturnsoverLongHorizonsEpisodesofovershootingfollowedbycorrectionWeak-FormTestsReturnsoverthPredictorsofBroadMarketReturnsFamaandFrenchAggregatereturnsarehigherwithhigherdividendratiosCampbellandShillerEarningsyieldcanpredictmarketreturnsKeimandStambaughBondspreadscanpredictmarketreturnsPredictorsofBroadMarketRetP/EEffectSmallFirmEffect(JanuaryEffect)NeglectedFirmEffectandLiquidityEffectsBook-to-MarketRatiosPost-EarningsAnnouncementPriceDriftSemistrongTests:AnomaliesP/EEffectSemistrongTests:AnFigure11.3AverageAnnualReturnfor10Size-BasedPortfolios,1926–2008Figure11.3AverageAnnualRetFigure11.4AverageReturnasaFunctionofBook-To-MarketRatio,1926–2008Figure11.4AverageReturnasFigure11.5CumulativeAbnormalReturnsinResponsetoEarningsAnnouncementsFigure11.5CumulativeAbnormaStrong-FormTests:

InsideInformationTheabilityofinsiderstotradeprofitabilityintheirownstockhasbeendocumentedinstudiesbyJaffe,Seyhun,Givoly,andPalmonSECrequiresallinsiderstoregistertheirtradingactivityStrong-FormTests:

InsideInInterpretingtheAnomaliesThemostpuzzlinganomaliesareprice-earnings,small-firm,market-to-book,momentum,andlong-termreversal.FamaandFrencharguethattheseeffectscanbeexplainedbyriskpremiums.Lakonishok,Shleifer,andVishneyarguethattheseeffectsareevidenceofinefficientmarkets.InterpretingtheAnomaliesTheFigure11.6ReturnstoStylePortfolioasaPredictorofGDPGrowthFigure11.6ReturnstoStylePInterpretingtheEvidenceAnomaliesordatamining?Someanomalieshavedisappeared.Book-to-market,size,andmomentummayberealanomalies.InterpretingtheEvidenceAnomaInterpretingtheEvidenceBubblesandmarketefficiencyPricesappeartodifferfromintrinsicvalues.RapidrunupfollowedbycrashBubblesaredifficulttopredictandexploit.InterpretingtheEvidenceBubblStockMarketAnalystsSomeanalystsmayaddvalue,but:DifficulttoseparateeffectsofnewinformationfromchangesininvestordemandFindingsmayleadtoinvestingstrategiesthataretooexpensivetoexploitStockMarketAnalystsSomeanalMutualFundPerformanceTheconventionalperformancebenchmarktodayisafour-factormodel,whichemploys:thethreeFama-Frenchfactors(thereturnonthemarketindex,andreturnstoportfoliosbasedonsizeandbook-to-marketratio)plusamomentumfactor(aportfolioconstructedbasedonprior-yearstockreturn).MutualFundPerformanceTheconFigure11.7EstimatesofIndividualMutualFundAlphas,1993-2007Figure11.7EstimatesofIndivConsis