令人不快的惊喜？选举与税务新闻冲击（英）

UnpleasantSurprises?ElectionsandTaxNewsShocks AntonioC.DavidandCanSever WP/23/139 IMFWorkingPapersdescriberesearchinprogressbytheauthor(s)andarepublishedtoelicitcommentsandtoencouragedebate. TheviewsexpressedinIMFWorkingPapersarethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoard,orIMFmanagement. 2023 JUN ©2023InternationalMonetaryFundWP/23/139 IMFWorkingPaper AfricanDepartment UnpleasantSurprises?ElectionsandTaxNewsShocksAntonioC.DavidandCanSever* AuthorizedfordistributionbyEdwardGemayelJune2023 IMFWorkingPapersdescriberesearchinprogressbytheauthor(s)andarepublishedtoelicitcommentsandtoencouragedebate.TheviewsexpressedinIMFWorkingPapersarethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoard,orIMFmanagement. ABSTRACT:Unanticipatedchangesintaxpolicyarelikelytohavedifferentmacroeconomiceffectscomparedtoanticipatedchangesduetoseveralmechanisms,includingfiscalforesightandpolicyuncertainty.Itisthereforeimportanttounderstandwhatdrivessuchpolicysurprises.Weexplorethenatureofunanticipatedtaxpolicychangesbyfocusingonapoliticaleconomydeterminantofthoseevents,namelythetimingofelections.Usingmonthlydatafor22advancedeconomiesandemergingmarketsovertheperiod1990-2018,weshowthatimplementationlagstendtobesignificantlylongerfortaxpolicychangeannouncementsthataremadeduringthepre-electionperiods,therebyleadingtoalowerlikelihoodof“taxnewsshocks”.Wealsofindthatimplementationlagsbecomemuchshorterfortaxpolicychangesthatareannouncedintheaftermathofelections,generatingmorefrequenttaxnewsshocks.Thispatternremainssimilarfordifferenttaxmeasuresortypesoftaxes.Thefindingsarerobusttoanumberofchecks,includingalternativedefinitionsoftaxnewsshocks,ortocontrollingforvariouseconomicandinstitutionalfactors. JELClassificationNumbers: D72,D78,P51,E02,E62,E63,H20,H30 Keywords: TaxShocks;ElectoralCycles;PoliticalEconomy Author’sE-MailAddress: AntonioC.David:adavid@imf.org;CanSever:csever@imf.org *TheviewsexpressedherearethoseoftheauthorsanddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoard,orIMFmanagement.TheauthorswouldliketothankHichamBennouna,CostasChristou,EraDabla-Norris,LucEyraud,EdwardGemayel,KouameDesireKanga,MoustaphaMbohouMama,AndualemMengistu,JaroslawWieczorekandPeterWankuruforusefuldiscussionsandfeedback.TheauthorsarealsogratefultotheparticipantsoftheIMFSeminarattheAfricanDepartmentforhelpfulsuggestions.Allerrorsareourown. WORKINGPAPERS UnpleasantSurprises?ElectionsandTaxNewsShocks PreparedbyAntonioC.DavidandCanSever1 1TheviewsexpressedherearethoseoftheauthorsanddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoard,orIMFmanagement.TheauthorswouldliketothankHichamBennouna,CostasChristou,EraDabla-Norris,LucEyraud,EdwardGemayel,KouameDesireKanga,MoustaphaMbohouMama,AndualemMengistu,JaroslawWieczorekandPeterWankuruforusefuldiscussionsandfeedback.TheauthorsarealsogratefultotheparticipantsoftheIMFSeminarattheAfricanDepartmentforhelpfulsuggestions.Allerrorsareourown. Contents 1.Introduction3 2.DataandStylizedFacts7 2.1.TaxPolicyChanges7 2.2.Elections9 2.3.OtherVariables9 2.4.StylizedFacts10 2.4.1.EventsintheSample10 2.4.2.TaxNewsShocksandElectionPeriods11 3.EmpiricalStrategy12 4.Results13 4.1.MainResults13 4.2.AlternativeDefinitionsofTaxNewsShocks15 4.3.OtherRobustnessChecks18 4.4.ExplanatoryPower21 4.5.ChangesintheTaxBaseversusTaxRates23 4.6.ChangesinConsumptionversusIncomeTaxes23 4.7.AdvancedEconomiesversusEmergingMarkets24 5.Conclusion25 References27 Appendix31 ElectionsintheSample31 ImplementationLagsandElectionPeriods32 1.Introduction Themacroeconomiceffectsoftaxpolicychangescandependonwhetherthosechangesareanticipatedbytheeconomicagentsornot.Whenthedurationbetweentheannouncementandimplementationdatesoftaxpolicychangesislonger,thatis,whenthepolicychangeis“anticipated”,agentshavemoretimetoadjusttheirbehavior(so-called“fiscalforesight”asdiscussedinLeeperetal.2012,2013),e.g.,bysubstitutingtaxableactivityintothepresentinanticipationofafuturetaxincrease.1Infact,theextantliteraturefindssignificantdifferencesintermsofthemacroeconomicimpactoftaxpolicychangeswhencomparinganticipatedandunanticipatedpolicychanges(MertensandRavn2011,2012,2014,Leeperetal.2012,2013,Ramey2016).Forinstance,inanewKeynesianDSGEmodelwithseveralstandardfrictions(includingfinanciallyconstrained,“hand-to-mouth”households),fiscalmultipliersforgovernmentspendingaremuchlower(andcanbeevennegativeonimpact),whenthereisalargedegreeoffiscalforesightrelativetothebenchmarkofnofiscalforesight(Leeperetal.2012).Incontrast,theoutputeffectsofincreasesincapitaltaxesinsuchmodelsarelargerinabsolutevalue(i.e.,morenegative)especiallyatlongerhorizons,ifthereissignificantfiscalforesight.2Itisinterestingtonotethat,whileitisclearthattheeffectsofpolicychangesdifferinthepresenceof“fiscalforesight”,itisnotevidentthatanticipatedpoliciesareinherentlywelfareenhancingwhencomparedtounanticipatedpolicies. Moreover,unanticipatedtaxpolicychanges,socalled“taxnewsshocks”,canaffectmacroeconomicoutcomesbyincreasing(policy)uncertainty.Con