Novy-Marx-Robert-2013-The-other-side-of-value--The

Theothersideofvalue:Thegrossprofitabilitypremium$RobertNovy-MarxSimonGraduateSchoolofBusiness,UniversityofRochester,500JosephC.WilsonBlvd.,Box270100,Rochester,NY14627,UnitedStatesarticleinfoArticlehistory:Received15December2011Receivedinrevisedform29May2012Accepted6June2012Availableonline26January2013JELclassification:G12Keywords:ProfitabilityValuepremiumFactormodelsAssetpricingQualityinvestingabstractProfitability,measuredbygrossprofits-to-assets,hasroughlythesamepowerasbook-to-marketpredictingthecrosssectionofaveragereturns.Profitablefirmsgeneratesignificantlyhigherreturnsthanunprofitablefirms,despitehavingsignificantlyhighervaluationratios.Controllingforprofitabilityalsodramaticallyincreasestheperformanceofvaluestrategies,especiallyamongthelargest,mostliquidstocks.Theseresultsaredifficulttoreconcilewithpopularexplanationsofthevaluepremium,asprofitablefirmsarelesspronetodistress,havelongercashflowdurations,andhavelowerlevelsofoperatingleverage.Controllingforgrossprofitabilityexplainsmostearningsrelatedanomaliesandawiderangeofseeminglyunrelatedprofitabletradingstrategies.&2013ElsevierB.V.Allrightsreserved.1.IntroductionProfitability,asmeasuredbytheratioofafirm’sgrossprofits(revenuesminuscostofgoodssold)toitsassets,hasroughlythesamepowerasbook-to-market(B/M)predictingthecrosssectionofaveragereturns.Grossprofits-to-assetsisalsocomplimentarytobook-to-mar-ket,contributingeconomicallysignificantinformationabovethatcontainedinvaluations,evenamongthelargest,mostliquidstocks.Theseconclusionsdifferfromthoseofearlierstudies.Forexample,whileFamaandFrench(2006)findsthatearningshasexplanatorypowerinFamaandMacBeth(1973)crosssectionregressions,FamaandFrench(2008,p.1663)findsthat‘‘profitabilitysortsproducetheweakestaveragehedgeportfolioreturns’’amongthestrategiestheyconsiderand‘‘donotprovidemuchbasisfortheconclusionthat,withcontrolsformarketcapandB/M,thereisapositiverelationbetweenaveragereturnsandprofitability.’’Grossprofit-abilityhasfarmorepowerthanearnings,however,pre-dictingthecrosssectionofreturns.Strategiesbasedongrossprofitabilitygeneratevalue-likeaverageexcessreturns,eventhoughtheyaregrowthstrategiesthatprovideanexcellenthedgeforvalue.Thetwostrategiessharemuchincommonphilosophi-cally,despitebeinghighlydissimilarinbothcharacter-isticsandcovariances.Whiletraditionalvaluestrategiesfinancetheacquisitionofinexpensiveassetsbysellingexpensiveassets,profitabilitystrategiesexploitadifferentdimensionofvalue,financingtheacquisitionofproduc-tiveassetsbysellingunproductiveassets.Becausethetwoeffectsarecloselyrelated,itisusefultoanalyzeprofitabilityinthecontextofvalue.Valuestrategiesholdfirmswithinexpensiveassetsandshortfirmswithexpensiveassets.Whenafirm’smarketvalueislowrelativetoitsbookvalue,thenastockpurchaseracquiresarelativelylargequantityofbookassetsforeachdollarspentonthefirm.Whenafirm’smarketpriceishighrelativetoitsbookvaluetheoppositeContentslistsavailableatSciVerseScienceDirectjournalhomepage:://dx.doi.org/10.1016/j.jfineco.2013.01.003$IthankGeneFama,AndreaFrazzini,TobyMoskowitz,theanonymousreferee,andparticipantsoftheNationalBureauofEconomicResearchAssetPricingandQGroupmeetings.E-mailaddress:[e-mail]nomics108(2013)1–28istrue.ValuestrategieswerefirstadvocatedbyGrahamandDodd(1934),andtheirprofitabilityhasbeenshowncountlesstimessince.Previousworkarguesthattheprofitabilityofvaluestrategiesismechanical.Firmsforwhichinvestorsrequirehighratesofreturn(i.e.,riskyfirms)arepricedlowerand,consequently,havehigherbook-to-marketsthanfirmsforwhichinvestorsrequirelowerreturns.Becausevaluationratioshelpidentifyvariationinexpectedreturns,withhigherbook-to-marketsindicatinghigherrequiredrates,valuefirmsgeneratehigheraveragereturnsthangrowthfirms(Ball,1978;Berk,1995).Whilethisargumentisconsistentwithrisk-basedpricing,itworksjustaswellifvariationinexpectedreturnsisdrivenbybehavioralforces.Lakonishok,Shleifer,andVishny(1994)arguethatlowbook-to-marketstocksareonaverageoverpriced,whiletheoppositeistrueforhighbook-to-marketstocks,andthatbuyingvaluestocksandsellinggrowthstocksrepresentsacrudebuteffectivemethodforexploitingmisvaluationsinthecrosssection.Similarargumentssuggestthatfirmswithproductiveassetsshouldyieldhigheraveragereturnsthanfirmswithunproductiveassets.Productivefirmsthatinvestorsdemandhighaveragereturnstoholdshouldbepricedsimilarlytolessproductivefirmsforwhichinvestorsdemandlowerreturns.Variationinproductivityinthiswayhelpsidentifyvariationininvestors’requiredratesofreturn.Becauseproductivityhelpsidentifythisvariation,withhigherprofitabilityindicatinghigherrequiredrates,profitablefirmsgeneratehigheraveragereturnsthanunprofitablefirms.Again,theargumentisconsistentwith,butnotpredicatedon,rationalpricing.Consistentwiththesepredictions,portfoliossortedongrossprofits-to-assetsexhibitlargevariationinaveragereturns,especiallyinsortsthatcontrolforbook-to-market.Moreprofitablefirmsearnsignificantlyhigheraveragereturnsthanunprofitab