警惕货币增长与通货膨胀的关系（英）

ANoteofCautionontheRelationbetweenMoneyGrowthandInflation HelgeBerger,SuneKarlsson,andPärÖsterholm WP/23/137 IMFWorkingPapersdescriberesearchinprogressbytheauthor(s)andarepublishedtoelicitcommentsandtoencouragedebate. TheviewsexpressedinIMFWorkingPapersarethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoard,orIMFmanagement. 2023 JUN ©2023InternationalMonetaryFundWP/23/137 IMFWorkingPaper EuropeanDepartment ANoteofCautionontheRelationbetweenMoneyGrowthandInflation HelgeBerger InternationalMonetaryFundFreeUniversityBerlinCESifo SuneKarlsson SchoolofBusiness,ÖrebroUniversity PärÖsterholm◊ SchoolofBusiness,ÖrebroUniversityNationalInstituteofEconomicResearch AuthorizedfordistributionbyAlfredKammer June2023 IMFWorkingPapersdescriberesearchinprogressbytheauthor(s)andarepublishedtoelicitcommentsandtoencouragedebate.TheviewsexpressedinIMFWorkingPapersarethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoard,orIMFmanagement. ABSTRACT:WeassessthebivariaterelationbetweenmoneygrowthandinflationintheeuroareaandtheUnitedStatesusinghybridtime-varyingparameterBayesianVARmodels.Modelselectionbasedonmarginallikelihoodssuggeststhattherelationisstatisticallyunstableacrosstimeinbothregions.Theeffectofmoneygrowthoninflationweakenednotablyafterthe1980sbeforestrengtheningafter2020.Thereisevidencethatthistimevariationisrelatedtothepaceofpricechanges,aswefindthatthemaximumimpactofmoneygrowthoninflationisincreasinginthetrendlevelofinflation.Theseresultscautionagainstassertingasimple,time-invariantrelationshipwhenmodelingthejointdynamicsofmonetaryaggregatesandconsumerprices. JELClassificationNumbers: E31,E37,E47,E51 Keywords: BayesianVAR;Time-varyingparameters;Stochasticvolatility Author’sE-MailAddress: hberger@imf.org;sune.karlsson@oru.se;par.osterholm@oru.se TheviewsexpressedinthispaperarethoseoftheauthorsanddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoardorIMFmanagement.WearegratefultoItaiAgur,SaioaArmendariz,RomainDuval,NirKlein,JorgeCanalesKriljenko,ChristianeRoehler,andVinaNguyenforhelpfulcomments,andtoMorganManeelyforassistancewithdata. InternationalMonetaryFund,70019thStreetNW,Washington,DC20431,USA ÖrebroUniversity,SchoolofBusiness,70182Örebro,Sweden ◊ÖrebroUniversity,SchoolofBusiness,70182Örebro,Sweden WORKINGPAPERS ANoteofCautionontheRelationbetweenMoneyGrowthandInflation PreparedbyHelgeBerger,SuneKarlsson,andPärÖsterholm◊ InternationalMonetaryFund,70019thStreetNW,Washington,DC20431,USAe-mail:hberger@imf.org ÖrebroUniversity,SchoolofBusiness,70182Örebro,Swedene-mail:sune.karlsson@oru.se ◊ÖrebroUniversity,SchoolofBusiness,70182Örebro,Swedene-mail:par.osterholm@oru.se Contents Introduction3 DataandModel3 Results6 Conclusions10 References12 Appendix14 FIGURES 1.Euroareainflationandmoneygrowth5 2.UnitedStatesinflationandmoneygrowth5 3.Impulse-responsefunction:Effectofshockstomoneygrowthoninflationintheeuroarea7 4.Impulse-responsefunction:EffectofshockstomoneygrowthoninflationintheUnitedStates7 5.Impulse-responsefunctions:Effectofshockstomoneygrowthoninflationatdifferenttimepoints8 6.Tradeinflationandmaximuminflationresponseofshocktomoneygrowthintheeuroarea10 7.TradeinflationandmaximuminflationresponseofshocktomoneygrowthintheUnitedStates10 A1.Standarddeviationofshockstomoneygrowthintheeuroarea.14 A2.StandarddeviationofshockstomoneygrowthintheUnitedStates14 A3.Impulse-responsefunction:Effectofshockstomoneygrowthoninflationintheeuroarea15 A4.Impulse-responsefunction:EffectofshockstomoneygrowthoninflationintheUnitedStates15 TABLE 1.Logmarginallikelihoods6 1.Introduction Moneyhastypicallyplayedaminorroleintheliteratureonmonetarypolicythelastthirtyyears.1However,therecentsurgeininflation–precededbynon-standardmonetarypolicymeasureswhichhaveincreasedthesizeofcentralbanks’balancesheetsandmoneysupply–hasrevivedthedebateregardingtheroleofmoneygrowthforinflation.2Forexample,Borioetal.(2023a)arguethatundersomeconditionsforecastersofinflationcouldhavedonebetterbytakingtheinformationcontainedinthebivariaterelationbetweenthetwovariablesintoaccount. Fromaforecastingperspective,akeyissueiswhethertherelationbetweenmoneygrowthandinflationisstableovertime.Theempiricalliteratureisambiguous,withanumberofpaperssuggestingasignificant,albeitweakeningrelationshipovertime;see,forexample,DeGrauweandPolan(2005),BergerandÖsterholm(2011a;2011b),SargentandSurico(2011),DregerandWolters(2014)andGertlerandHofmann(2018).3Toinformthisdiscussion,weapplythehybridtime-varyingparameterbivariateBayesianVARframeworkofChanandEisenstat(2018)todatafromtheeuroareaandtheUnitedStatestoassessiftherelationbetweenmoneygrowthandinflationisstableortimevarying. 2.Dataandmodel OuranalysisemploysdataonCPIinflationandM3growthintheeuroareaandtheUnitedStates;seeFigures1and2.CPIinflationisgivenas𝜋�=100(𝑃𝑡⁄𝑃𝑡−4−1),where𝑃�istheconsumerpriceindexattimet;moneygrowthisgivenas𝜇�=100(𝑀𝑡⁄𝑀𝑡−4−1),where𝑀�isM3attimet.Theeuroareadatarangefrom1971Q1to2022Q4andtheUSdatafrom1961Q1to2022Q4.4 Definingthevectorof