外文原文Dividendpolicy:areview1.IntroductionExplainingdividendpolicyhasbeenoneofthemostdifficultchallengesfacingfinancialeconomists.Despitedecadesofstudy,wehaveyettocompletelyunderstandthefactorsthatinfluencedividendpolicyandthemannerinwhichthesefactorsinteract.AllenandMichaely(1995)concludethat‘‘muchmoreempiricalandtheoreticalresearchonthesubjectofdividendsisrequiredbeforeaconsensuscanbereached’’(p.833).ThefactthatamajortextbooksuchasBrealeyandMyers(2002)listsdividendsasoneofthetenimportantunsolvedproblemsinfinancereinforcesthisconclusion.ThefirstempiricalstudyofdividendpolicywasprovidedbyLintner(1956),whosurveyedcorporatemanagerstounderstandhowtheyarrivedatthedividendpolicy.Lintnerfoundthatanexistingdividendrateformsabenchmarkforthemanagement.Companies’managementusuallydisplayedastrongreluctancetoreducedividends.Lintneropinedthatmanagersusuallyhavereasonablydefinitivetargetpayoutratios.Overtheyears,dividendsareincreasedslowlyataparticularspeedofadjustment,sothattheactualpayoutratiomovesclosertothetargetpayoutratio.2.DividendirrelevanceandtaxclientelesWhileLintner(1956)providedthestylisticdescriptionofdividends,thewatershedinthetheoreticalmodellingofdividendswasalmostsurelytheclassicpaperMillerandModigliani(1961),whichfirstproposeddividendirrelevance.Essentially,theirmodelisaone-periodmodelundercertainty.Givenafirm’sinvestmentprogram,thedividendpolicyofthefirmisirrelevanttothefirmvalue,sinceahigherdividendwouldnecessitatemoresaleofstocktoraisefinancesfortheinvestmentprogram.Thecrucialassumptionhereisthatthefuturemarketvaluewillremainunaffectedbycurrentdividends.Theargumentrestsontheassumptionsthattheinvestmentprogramisdeterminedindependentlyandthateverystockholderearnsthesamereturn(i.e.thediscountrateremainsconstant).MillerandModigliani’sdividend-irrelevanceargumentiselegant,butthisdoesnotexplainwhycompanies,thepublic,investmentanalystsaresointerestedindividendannouncements.Clearly,theobservedinterestindividendannouncementmustberelatedtosomeviolationoftheMillerandModiglianiassumptions.MillerandModigliani,whileformulatingtheirfamousdividendirrelevancepropositions,observedthatinthepresenceoftaxation,investorswillformclienteleswithspecificpreferencesforparticularlevelsofdividendyields.Thisspecificpreferencefordividendsmaybedetermined,interalia,bythemarginaltaxratesfacedbytheinvestor.Alteringthedividendlevel,accordingtoMillerandModigliani,leadsonlytoachangeintheclienteleofshareholdersforthefirm.Partofthedividendpuzzlearisesfromthefactthatdividendsaretypicallytaxedatahigherratecomparedtotheincomefromcapitalgains.Thishascertainlybeenhistoricallytruealthoughinrecentyearswehavenoticedamovetoeliminate/reducetaxondividends.Weshould,therefore,expectinvestorstoprefercashfromcapitalgainsovercashfromdividends.MillerandScholes(1978)provideaningeniousschemetoconvertdividendincometocapitalgainsincome.RecentlyAllenetal.(2000)haveadvancedatheorybasedontheclienteleparadigmtoexplainwhysomefirmspaydividendsandothersrepurchaseshares.AvariantoftheclienteletheoryhasalsobeenadvancedbyBaker(2004)wheretheypositthatdividendpaymentsareinresponsetodemandsfrominvestorsfordividends.3.InformationalasymmetryandsignallingmodelsDeviationsfromtheMillerandModigliani(1961)dividendirrelevancepropositionisobtainableonlywhentheassumptionsunderlyingthesettingofMillerandModiglianiareviolated.Thetax-cliente...