政策制定者制定統(tǒng)一框架的歷史貨幣和金融統(tǒng)計數(shù)據(jù)之二

Historicalmonetaryandfinancialstatisticsforpolicymakers:towardsaunifiedframeworkSeptember2022onlineISBN(online)theBISThispublicationisavailableontheBISwebsite().BankforInternationalSettlements2022.Allrightsreserved.Briefexcerptsmaybereproducedortranslatedprovidedthesourceisstated.Historicalmonetaryandfinancialstatisticsforpolicymakers:towardsaunifiedframeworkudioBorioyvindEitrheimMarcFlandreauClemensJobstJanFQvigstadandRylandThomasMonetaryandEconomicDepartmenterJELclassification:C82,E42,E43,E51,N01,N10,N20,R31.KeywordsHistoricalstatisticscredithousepricesinterestratescentralbanks,financialhistory.iinancialstatisticsiFollowingtheoutbreakofthe2007–09GreatFinancialCrisis,centralbankersandpolicymakershavebecomeincreasinglyawareoftheimportanceofmacro-financialhistory.Nowadays,informedpolicymakingcallsforconstantminingofthehistoricalishasledtoextensiveeffortstocombinenationalhistoricaldatasetssoastoproduceinternationalpanels.Fewpeoplerealisehowproblematicthisassemblagereallyis.Wheretheyexist,tisticshavebeenconstructedthroughnationallyfocusedefforts.Whilethesewereoftentheproductofheroicbreakthroughs,theyhaveneverfollowedacommonplan.Theresultisthattheindividualnationalseriesthatcomprisemodernpanelsarenotonlyimperfect,butunlikeTolstoy’sunhappyfamilies,whichareallunhappyintheirownway,eachimperfectseriesisimperfectinitsownstyle.AttheinitiativeofClaudioBorio,HeadoftheBISMonetaryandEconomicDepartment,andJanFQvigstad,formerDeputyGovernoroftheCentralBankofNorway,agroupofcentralbanksmetinBaselinOctober2016andagreedtoestablishasmallnetworkonHistoricalMonetaryandFinancialStatistics(HMFS).1Buildingonlong-standingintellectualsynergiesbetweencentralbankresearchdepartmentsandacademicresearchersonmonetaryandfinancialhistory,andginternationalexpertsthenetworkhasstartedtounpacktheproblem,makingastartbyaddressingtheissuesinvolvedintheconstructionofseriesforinterestrates,creditaggregatesandhousingprices.ThepresentBISPaperisbornoutofthesetravails.Theresultisbothsoberingofourmacro-financialhistory.Itisoptimisticbecause,whilethegapsandforwardbyputtingtogetheraunifiedframework:usingitasahow-toguide,futureacrofinancialstatisticsorstatistoriansaswedubthemherewillandpolicymakerstofindasafepassagearoundthepitfallsofstatisticalhistory.AgustnCarstenser1Ithasincludedthefollowingcountries:Austria,Canada,Denmark,France,Italy,Japan,Norway,Sweden,theUnitedKingdomandtheUnitedStatesinancialinancialstatisticsfornifiedframeworkClaudioBorioyvindEitrheimMarcFlandreau,ClemensJobst,JanFQvigstadandtractInrecentyears,manycentralbankshaveengagedindataprojectsaimedatthecollectionanddocumentationofhistoricalmonetaryandfinancialstatistics(HMFS)fortheirrespectivecountries.Forthesecountries,longrunsofdataforkeymacroeconomictimeseriesarenowavailableandareincreasinglybeingusedinpolicy-orientedresearchofinteresttovariousinstitutions.Informationfromthesehistoricaldatabasesisusedtodrawparallelsbetweencurrentdevelopmentsandhistoricaleventstoshedlightontoday’spolicyissuesintheareasofmonetaryandfinancialstability.TheaimoftheBankforInternationalSettlements(BIS)HMFSprojectistoestablishanetworkbetweeninterestedcentralbanksthathavealreadyinvestedinlocalnationalHMFSdatabases,usingtheBISasahub.TheHMFSprojecthasadecentralisedstructure,witheachcentralbankprovidingtheresourcesnecessarytocontributeinformationaboutitscountry.Moreover,thenetworkrestsonapartnershipbetweencentralbanksandacademicexperts,includinganannualmeetinginwhichweexchangeviewsaboutthelateststatisticaldevelopmentsinourareaofinterestatthemembercentralbanks.ThefocusofthisundertakinghasbeenisinginterestratescreditandhousingpricesThisBISpaperisanattempttotakestockofwhatcountsas“goodpractice”andhowitshouldbeimplementedindifferentcontexts.Itisourunderstandingthatthereexistsalargeamountoftacitknowledgeamongeconomichistoriansofhowbesttoconstructhistoricaltimeseries.Attheintheirmethodsandlogic.Researchersandpolicymakersusinghistoricalstatisticsofchoices.InthisBISpaper,datasourcesandmethodsaredocumentedanddiscussedforinterestrates,creditandhousingprices,respectively.Weidentifytheanoverviewoftransversalissuesandresearch-basedevidenceonwhatmayconstitute“goodpractice”.AseparateBISPaperprovidesanoverviewofavailablehistoricaldatafortheparticipatingcountries.2AnyviewsexpressedaresolelythoseoftheauthorsandcannotbetakentorepresentthoseoftheBankforInternationalSettlements,theBankofEngland,theCentralBankoftheRepublicofAustria,theCentralBankofNorway,theUniversityofPennsylvaniaortheUniversityofVienna.vinancialstatisticsvContributorsAuthorsteamforthereportVincentBignon(BankofFranceandUniversitédeAix-Marseille)udioBorioBISyvindEitrheimCentralBankofNorwayeauUPennandWhartonSchoolrClemensJobstUniversityofViennaandCentralBankoftheRepublicofAustria)ankofEnglandJanFQvigstad(CentralBankofNorway)HMFSsteeringcommitteelaudioBorioBankforInternationalSettlementsyvindEitrheimCentralBankofNorwayeauUPennandWhartonSchoolsBankofEnglandChairHMFScountryteamsAustriaClemensJobstCanada:PatrickAlexander,ArthurBerger,RobinBrace,KimHuynh,LauraMurphy,TrevorSabeanLarrySchembrigrenncentBignonPierreSicsicDavidDemillyericoBarbielliniAmideiPaoloCroceElenaGenitoPaoloPiselliJapan:MarikoHatase,HarukoKato,MariOhnuki,ShotaroYamaneNorway?yvindEitrheim,MatsBFevolden,JanFQvigstadrJacobsonmRylandThomassMarkCarlsoniiinancialiiInthisvolumeearchingforgoodpracticeinhistoricalmonetaryandfinancialstatisticsdoverviewVincentBignon,ClaudioBorio,?yvindEitrheim,MarcFlandreau,ClemensJobst,JanlandThomastaoninterestratesiversityofPennsylvaniaClemensJobstUniversityofViennaandCentralBankoftheRepublicofAustriaataoncreditVincentBignon,BanquedeFranceandUniversitédeAix-MarseilleandreauUniversityofPennsylvaniaandWhartonSchoolofBusinessldataonrealestatepricesvindEitrheimCentralBankofNorwayClemensJobstUniversityofViennaandCentralBankoftheRepublicofAustriainancialstatisticsbleofcontents calmonetaryandfinancialstatisticsforpolicymakerstowardsaunifiedframework iiiContributors vInthisvolume viiThehistoryoftheHMFS 11.Searchingforgoodpracticeinhistoricalmonetaryandfinancialstatistics 3a.Introduction 3b.Historicalmonetaryandfinancialstatistics:thestateofplay 4c.Syntheticcountries,syntheticobjects,andthemethodologyofhistoricalmonetaryandfinancialstatistics 6BoxI.1Theneedfor“statistorians” 8dSettingtheagenda 13e.TheHMFSgoingforward 14es 2.Historicaldataoninterestrates 192.1Somegeneralconsiderations 202.2Long-terminterestrates 25BoxIIPricingterminableannuitieswhatwedidnotknowtheyknew 28BoxII.2Whennocleangovernmentbondseriesareavailable 31BoxII.3Dealingwithconversionrisk:threecases 342.3Short-termrates 352.4Conclusion 41es 3.Historicaldataoncredit 473.1Thelamppostsyndrome:pitfallsintheconstructionofhistoricalcredit 3.2Intensivemarginandrepresentativesamples:dowemeasurebankcredit BoxIIIDarkmatterinbankingacasestudy 583.3Institutionsthatarenot“banks”butactlikebanks 593.4Conclusion 62xinancialstatisticsxReferences 4.Historicaldataonrealestateprices 684.1Introductionandoverview 684.2Methods 71oxIVQualityadjustedhistoricalHPIsforNorway BoxIV.2Thehousesaletimeline–whichpricetoselect? 814.3Historicalcontextandprimarysources 814.4Conclusionsandlessonsgoingforward 86References umesinthisseries 1inancial1ThehistoryoftheHMFSThehistoricalmonetaryandfinancialstatistics(HMFS)projectisanetworkofcentralbanksandacademicexpertshostedattheBankforInternationalSettlements(BIS)focusedonpolicy-relevantlong-runhistoricalmonetaryandfinancialdata.ItoperatesundertheaegisoftheIrvingFisherCommittee(IFC).Thenetwork’sdeparturepointisarecognitionthathistoricaldataislessaboutcollectionthanproduction.Productionofdatarequiresathoroughunderstandingofhistoricalmarkets,instrumentsandinstitutions,andasoundmethodologicalframeworktobleacrosstimeandspaceTheambitionofHMFSistoprovideaforuminwhichpolicymakers,statisticians,economistsanderapproachestohistoricaldataandencouragemutuallearning.ctlikethisisboundtohavemanyoriginsInthepasttheconstructionofhistoricalmonetaryandfinancialstatisticshasbeenasharedendeavouramongacademicsandcentralbanks.Inrecentyears,manycentralbankshaveengagedinimportantdataprojectsaimedattheproduction(collectionanddocumentation)oficsfortheirrespectivecountriesUpuntilnowinasystematicfashion,ject.Twomilestonesexplainhowthesenationaleffortsconverged.First,in2006,aconferencewasorganisedinViennaattheCentralBankoftheRepublicofAustriabyMarcFlandreau(thenaprofessoratSciencesPo,Paris)andEdiHochreiter(thenoftheAustriancentralbank),towhichrepresentativesofasignificantnumberofcentralbankswereinvitedandparticipated.AspecialsessionconsideredthepossibilityoflaunchinganinitiativethatwouldhavebeenverycloselyrelatedtoHMFS.productionofhighqualityhistoricaldatawasaglobalpublicgoodandthat,absentproperdesignincentives,ittendedtobeinshortsupply.Whilenationalsolutionshadbeenfound,therewasaneedforinternationalarrangements.Aspolicymakers,centralbankshadshownakeeninterestindevelopingthetoolsthatwouldenablethemtoremedythatspecificmarketfailure,sincetheywouldbetheprimaryusersoftheresultingdata.Becauseoftheirpublicmandate,asthemanyexamplesindicate,centralbankscouldtakesomeresponsibilityasproducersandrepositoriesforhistoricalmonetaryandtheaegisofaninternationalorganisationsuchastheBIS.Inparallel,theCentralBankofNorway(which?yvindEitrheimhadrepresentedattheViennaconference),launchedanambitiousprojectinanticipationofthecelebrationofitsbicentenary10yearslater(in2016).ThisprojectwasledbyJansmanagedby?yvindEitrheim.Oneconcern,whichtheleadersofthisprojectexpressedfromtheprojectandthenationaldataitproducedwouldalsobeofinterestfromaninternationalpointofviewandforthatreason,compatiblewithotherdataTothatendtheprojectcoordinatorsappointedMichaelBordoandMarcFlandreauasadvisors.Theproject,finalisedin2016,ledtothe2inancialstatistics2constructionofnewhistoricalseriesandtoseveralsignificantpublicationsfocusedonNorway.jectusheredinthelaunchofHMFSAttheinitiativeofClaudioBorioattheBISandJanQvigstad,agroupofcentralbanksthathadalreadyinvestedinlocalHMFSdatabases(severalofwhichhadparticipatedinthe2006banksofAustria,Canada,Denmark,France,Italy,Japan,Norway,Sweden,theUnitedinformationandtoidentifyandimplementbestpracticeapproaches.ThreesubsequentmeetingstookplaceinBaselinOctober2017,2018and2019.Thepresentvolumeisbornoutoftheseworkingsessions.In2022,thegroupstartedoperatingundertheaegisoftheBISIrvingFisherCommittee,whichbringstogethertheheadsofstatisticsatoverninetycentralbanks.MarcFlandreau(UPennandWhartonSchool)supervisedthefinalisationofthevolumewiththehelpofClemensJobst(UniversityofViennaandtheCentralBankofioBISyvindEitrheimCentralBankofNorway),MarcFlandreauandRylandThomas(BankofEngland),aswellasVincentBignon(BankofFranceandUniversitéAix-Marseille),ClemensJobstandJanFherpeoplewhohelpedwiththeprojectatdifferentpoints,areidentifiedinthetext.3inancial31.SearchingforgoodpracticeinhistoricalmonetaryandfinancialstatisticsVincentBignon,ClaudioBorio,?yvindEitrheim,MarcFlandreau,ClemensJobst,JanFQvigstadandRylandThomasa.Introductionycommunitytogivecloserattentiontohistory.Aswassubsequentlyrealised,themannerinwhichthesecuritisationofmortgageserredbyconsideringtoonarrowahistoricaltimeframeforthecomputationofmortgagedefaultrisksledtoaseriousunderestimateofpotentiallosses,and,asaresult,ofthetrueexposureoffinancialinvestorsandmarkets.Thishasbeenwidelyinterpretedasareminderoftheneedtoconsiderhistorymoreseriouslyandhasledtoahealthydemandfor“strongerhistory”andmorerigorousproductionofhistoricaldata.However,ashistorianshavelongknown,neitherthestudyofhistorynorthecollectionofretrospective“data”areinthemselvesasafeguardagainstmisguidedinference.Theupshotisthat,ontheonehand,historicalinvestigationwillneverproduceOntheotherhand,historicalinvestigationis,moreoftenthannot,theonlywayinwhichcertainquestionscanbeaddressed.Thestudyofhistory,ratherthanbeingawaytoestablishtangiblefacts,isamannerofconstructingthem.Constructiontakesplaceinaconstantbackandforthbetweensourcematerialandtheinvestigator’sconceptualandmethodologicalchoicesConsciouslyintegratingthisprocessofconstructionintotheprocessofresearchisthebestwayeconomichistorianscanofthesubprimecrisis).Thisiswherethemaincontributionofeconomichistorycanbefound:byassistingtheknowledgeprocess,itcanincreasetheoddsthat“history”willbeusedratherthanmisused(Eichengreen(2012,2015)).Ifavenerabletraditionistobesummoned,thecurrentrelevanceofhistorytomacro-financialpolicymakingmaybeseenasderivingfromthedata-focusedhistoricalmethodologyarticulatedinthe18thcenturybyJosephPriestley.Priestley’s“Newchartofhistory”,publishedaspartofhisLecturesonhistoryandgeneralpolicy,reflectedhisradicalempiricalagendaandhadanenduringlegacyin19thcenturyfinanceandfinancialpolicy(Priestley(1788)).ThroughtheintermediaryofFrancistofhistory”wassubsequentlyrevisedandupdated,andnowprovidedaglobalchronologyofnationsandinternationalconflicts(Baily(1812)).ContemporarieshadalreadynoticedthatBritishinterestrateswerehigherinwarthaninpeacetimeandachronologyofconflictwasvaluabletodeterminethesaferateofinteresttobeusedingovernmentcalculations.ItwasalsousefultoprivateactorsengagedinvitaltiessuchascomputingthepresentvalueoflifeinsurancepoliciesarbitrageetcSubsequently,theexperienceoftheworkingclassesduringtheindustrialisationphaseinthe19thcenturyledtomanyinvestigationsintothestandardandcostofliving,andthedevelopmentofofficialwageandpricestatistics(Bowley(1937)).Theneedtofinancetwoworldwarsinthe20thcenturyacceleratedthedevelopmentof4inancialstatistics4hyperinflationsofthe1920s(Cagan(1956))andthemonetaryinterpretationoftheGreatDepressionbyFriedmanandSchwartz(1963)provokedrevivedinterestinmeasurementofthemoneysupplyandthedevelopmentofmonetaryaggregates,asubsequentlandmarkwasCapieandWebber(1985).3Withthisinmind,thisvolumeprovidesareflectiononthepossibilityofconstructingahistoricalscorecardformonetaryandfinancialtimeseries.Asaresult,whiletheultimategoalisindeedtheproductionofusabletimeseries(andmanyaregivenintheonlinecountryappendices),thisvolumeismorefundamentallyhanexerciseinvolvesandarticulatingamethodology,aswellasanorganisationalframeworkinwhichthemethodologycanbeembedded.Inparticular,itreviewsthevarioussolutionstothedifficultiesencounteredaswellastheextenttowhichthesesolutionscanberealisticallyimplemented–andatwhatcost.Accordingly,therestofthisintroductorychapterreviewsthekeyideasthatunderpintheHMFSproject.Inparticular,itoutlinestheepistemologicalproblemsparticularjuncture.Italsoreviewsthehistoryoftheprojectanddiscussestheorganisationaltemplatethathasbeenadopted.bHistoricalmonetaryandfinancialstatistics:thestateofplayonehand,country-specificinitiativeshavefocusedonthecollectionofawiderangeofmacroeconomicandfinancialtimeseriesforaparticularcountryorgeographicalarea.Ontheotherhand,therehavebeencompilationsofcross-countrydatasets,comparisons.historicalstatisticspublishedbytheBankofJapan(BankofJapan(1966))andsuccessiveeditionsofHistoricalstatisticsoftheUnitedStates,culminatingintheteretalForBritainandIrelandBrianMitchellsvolumesonBritishhistoricalstatisticswithvariouscollaboratorsalsostandout.TheywerethemselvesmadepossiblebylandmarkachievementssuchasDeaneandCole(1967).4SubsequentcasesincludetheworkoftheScandinaviancentralbanks(ieDanmarksNationalbank,CentralBankofNorwayandSverigesRiksbank)andtheBankofItaly.Theselatterprojectsprovideaccesstomultiplespreadsheetstogethernds“Amillenniumofmacroeconomicdata”spreadsheet,whichhasproducedasingledatasetwithaccessto(multiple)historicalseriesandspliced(headline)dataupto3Bordo(1986).Seeegthecollectionofcross-countrystudiesfocusedonthelong-termbehaviourofvelocity(BordoandJonung(1987,1990)).4SeeMitchellandDeane(1962),MitchellandJones(1971)andMitchell(1988).5SeeAbildgren(2017),Edvinssonetal(2010,2014)andEitrheimetal(2004,2007).ForItaly,thepioneeringcontributionwasDeMattia(1968);itwasextendedbythehistoricalseriesoftheBankofItaly(CollanaStoricadellaBancad’Italia–www.bancaditalia.it/pubblicazioni/collana-storica/index.html),publishedinseveralseparatevolumes(Cotulaetal(2003)).Detailsonthedocumentationanddataareavailableonline:www.bancaditalia.it/statistiche/tematiche/stat-storiche/index.html.5inancial5ofEnglandwasmorereliantonreferencestooriginaldatasourcesforthenstructiondetailsTheseprojectsoffervaluablematerialbecauseofthewaytheylinkdata,datasources,andtheinstitutionalandeconomicsignificanceofalternativeseries.Bydesignhowever,country-specificapproachestendtobeidiosyncratic.TheBankofEnglandandDanmarksNationalbankhaveprimarilyreliedonindividualsatthosecentralbankswitharesearchinterestindevelopinghistoricalstatistics.TheBankofItaly,CentralBankofNorwayandSverigesRiksbankhavetypicallycommissionedperiodic,large-scaleprojectsinvolvingcollaborationbetweencentralbankandacademicresearchers.Internationalcomparabilityisnotusuallywithintheirremit.Earlyillustrationsofcross-countryinitiativesincludethepioneeringworkoninternationalhistoricalnationalstatisticsbyAngusMaddison,7thebroad-baseddatasetsonEuropeanhistoricalstatistics(Mitchell(1975)),anddatasetsforinternationalinterestrates(HomerandSylla(2005)).Inturn,suchefforts–especiallythoseconcerninglong-runpricesandincomedynamics–gobacktopreviousattemptsaimedatassemblinglargecross-countrydatasets.AlandmarkthatgeneratedmanysubsequentspilloverswastheInternationalScientificCommitteeonPriceHistorycreatedinbyWilliamBeveridgeandEdwinFrancisGay,theresultofafive-yeargrantbytheRockefellerFoundation.However,exceptfortheworkofHomerandSylla,andthefiscalseriesinMitchell,thesedatasetsprovidelimitedmacro-financialinformation.Morerecently,thefinancialglobalisationofthe1990sand2000sandnewfoundconcernsaboutlong-rundebtdynamicshaveboostedinterestinhistoricalprecedentsandledresearcherstocomplementdataonrealvariableswithdataonfinance.Apioneeringeffortatdocumentingfinancialdatafortheso-calledfirstglobalisationwave(1880–1913)istheGlobalFinancedatabaseassembledbydfromexistingsourcesbutmoreimportantlyitisanentirelynewattempttoorganiseadatabasethattimeseriesforgovernmentdebtandinterestratesforsovereigncountriesand15colonies,fromprimaryandsecondaryarchivalsources.TheGlobalFinancedatabasestandsoutbyenablingafocusoncomparabilityandthedevelopmentofauniformcross-countrymethodology.Itwasthefirst6SeeThomasandDimsdale(2017).Thespreadsheet(availableatwww.bankofengland.co.uk/-/media/boe/files/statistics/research-datasets/a-millennium-of-macroeconomic-data-for-the-uk.xls)isacollectionofofficialandresearchdataorganisedintoannual,quarterly,monthly,weeklyanddailysections.Eachsectioncontainsa“headlineseries”sheetcontainingasetofcontinuoustimeseriesthatlinkvarioushistoricalcomponentstogether,plusmoredetailedworksheets,whichshowhowthesetimeseriesarecalculatedtogetherwithalternativehistoricalseriesthatcanbeused.Asageneralprinciple,theannualseriesaretypicallybasedonseparateandmorereliablesourcesthanhigher-frequencyseriesandcoverawiderrangeofdata.Thecalculationsunderlyingjoined-upseriesaremadeclearinthespreadsheet,allowinguserstocreatetheirownseriesbasedondifferentassumptions.Atpresent,usersneedtoconsultunderlyingsourcesfordetailedinformationontheconstructionofthecomponenthistoricaldataseriesused,althoughadditionaldocumentationofthemethodsandchoicesmadeiscurrentlyunderway.7Maddison(2006).ThedataaremaintainedandrevisedbytheMaddisonProject,hostedbytheUniversityofGroningenandavailablehere:www.rug.nl/ggdc/historicaldevelopment/maddison/.Arelatedprojectforthepost-1950periodisthePennWorldTable,onthemostrecentversionseeFeenstraetal(2015).8SeeFlandreauandZumer(2004)fordetail.Thedataareavailableat/database/global-finance/.SeeAccominottietal(2011)foradiscussionofthecolonialdata,whichwereaddedsubsequently.6inancialstatistics6timeublicdebtgoingfarbackintimeItsdiscussionofgovernmentborrowingitsdatahavebeencollatedinsubsequentwork(forinstanceReinhartandRogoff(2009),Eichengreenetal(2019)),norealattentionwaspaidtothemethodologicalpitfallsitidentified,andinfactthemorerecentcontributionssplicetogetherdisjointedmethodologies.9Followingthesubprimecrisis,theGlobalFinancedatabasebecamethebasisoftheMacrohistoryLabdatabase(Jordàetal(2016,2017)).TheMacrohistoryLabcollatedmacro-financialhistoricalseriesfromsecond-handsourcesfor14countriesgoingbackto1870.10Thesehistoricaleffortsalsohadparallelswithmorefocuseddevelopmentsandinterestbystatisticiansengagingwithcontemporaryissues,andwhosoughttoexpandthedatahorizon.EarlyanalyticalworkincludesavolumebasedonthejointconferenceonRealEstateIndicatorsandFinancialStabilityorganisedbytheInternationalSettlementsBISinWashington,DC,in2003(BIS(2005)).ItsimportancewasdemonstratedbytheGreatFinancialCrisisof2008–09,andledtheBIStostartpublishingadetaileddatasetonpropertyprices.TheBISdatabasecurrentlycoversresidentialpropertypricesfor58advancedandemergingeconomieswithastartingdateof1970.AnotherinternationalhousepricedatabaseworthmentioningismaintainedbytheFederalReserveBankofDallasanddocumentedinMacketal(2011).Theyhavecollecteddatafromnationalpublicsourcesandthedatabasecontainsquarterlydatafor22countriesgoingbackto1975.11c.Syntheticcountrie