貨幣政策是否重要 35年后的敘述方法 Does Monetary Policy Matter The Narrative Approach after 35 Years

NBERWORKINGPAPERSERIES

DOESMONETARYPOLICYMATTER?

THENARRATIVEAPPROACHAFTER35YEARS

ChristinaD.Romer

DavidH.Romer

WorkingPaper31170

/papers/w31170

NATIONALBUREAUOFECONOMICRESEARCH

1050MassachusettsAvenue

Cambridge,MA02138

April2023

ThispaperwasthebasisforChristinaRomer’sAmericanEconomicAssociationpresidentialaddress,presentedattheASSAMeetinginNewOrleansonJanuary7,2023.WearegratefultoDavidCard,WendyEdelberg,EmiNakamura,EricSwanson,JustinWolfers,andseminarparticipantsattheUniversityofCambridgeforhelpfulcommentsandsuggestions.TheviewsexpressedhereinarethoseoftheauthorsanddonotnecessarilyreflecttheviewsoftheNationalBureauofEconomicResearch.

NBERworkingpapersarecirculatedfordiscussionandcommentpurposes.Theyhavenotbeenpeer-reviewedorbeensubjecttothereviewbytheNBERBoardofDirectorsthataccompaniesofficialNBERpublications.

?2023byChristinaD.RomerandDavidH.Romer.Allrightsreserved.Shortsectionsoftext,nottoexceedtwoparagraphs,maybequotedwithoutexplicitpermissionprovidedthatfullcredit,including?notice,isgiventothesource.

DoesMonetaryPolicyMatter?TheNarrativeApproachafter35Years

ChristinaD.RomerandDavidH.Romer

NBERWorkingPaperNo.31170

April2023

JELNo.E31,E52,E58,E65,N12

ABSTRACT

Thenarrativeapproachtomacroeconomicidentificationusesqualitativesources,suchasnewspapersorgovernmentrecords,toprovideinformationthatcanhelpestablishcausalrelationships.Thispaperdiscussestherequirementsforrigorousnarrativeanalysisusingfreshresearchontheimpactofmonetarypolicyasthefocalapplication.WereadthehistoricalminutesandtranscriptsofFederalReservepolicymakingmeetingstoidentifysignificantcontractionaryandexpansionarychangesinmonetarypolicynottakeninresponsetocurrentorprospectivedevelopmentsinrealactivityfortheperiod1946to2016.Wefindthatsuchmonetaryshockshavelargeandsignificanteffectsonunemployment,output,andinflationintheexpecteddirections.Analysisofavailablepolicyrecordssuggeststhatacontractionarymonetaryshocklikelyoccurredin2022.Basedontheempiricalestimatesoftheeffectofpreviousshocks,onewouldexpectsubstantialnegativeimpactsonrealGDPandinflationin2023and2024.

ChristinaD.Romer

DepartmentofEconomics

UniversityofCalifornia,Berkeley

Berkeley,CA94720-3880

andNBER

cromer@

DavidH.Romer

DepartmentofEconomics

UniversityofCalifornia,Berkeley

Berkeley,CA94720-3880

andNBER

dromer@

Thispaperrevisitsoneofthefundamentalquestionsofmacroeconomics:Doesmonetarypolicymatter?Itisaquestionthatliesattheheartofanymodelofshort-runmacroeconomicfluctuations.Ifmonetarypolicymatters,thenitisvitaltoincludeachannelthroughwhichchangesinaggregatedemandhaverealeffects.Itisalsoaquestionthatisonceagainattheforefrontofpolicydiscussions.TheFederalReserveisinthemiddleofanaggressivetightening,inhopesofloweringinflation.Thereappeartobewidelydisparateviewsaboutthelikelyeffectsofitsactionsonrealactivity,andabouthowquicklyinflationmightfall.

“Doesmonetarypolicymatter?”isalsoaquestionthatisinherentlyhardtoanswer.Likesomanyempiricalquestionsineconomics,omittedvariablebiasisacentralissue.Bothmonetarypolicyactionsandrealeconomicactivityarelikelytobeinfluencedbyothervariables.Anythingthataffectsoutput—expectations,fiscalpolicy,financialstress—isalsolikelytodrivedecisionsbytheFederalReserve.AsKarekenandSolow(1963)pointedoutlongago,intheextremecasewheremonetarypolicymakersusepolicytosuccessfullycounteractotherforcesaffectingoutput,onewouldfindmonetarypolicyvariablesmovingallaroundandoutputnotchanging.Asimpleregressionofoutputonanindicatorofmonetarypolicywouldnaivelyandincorrectlyconcludethatmonetarypolicydidn’tmatter.Thatis,ifcountercyclicalmonetarypolicyactionsarecommon,theestimatedimpactofmonetarypolicywillbebiasedtowardzero.

In1989,wewroteapaperontheeffectsofmonetarypolicyusingatechniquewetermedthe“narrativeapproach”(RomerandRomer1989).Thisisanempiricaltechniquewhereonegatherssystematicevidencefromcontemporaneousqualitativesources(suchasnewspapers,governmentreports,andpolicymeetingtranscripts),andincorporatesitintostatisticalanalysis.Inthecaseofour1989study,theideawastodealwithomittedvariablebiasbyusingtheplentifulnarrativerecordonthemotivationformonetarypolicyactions.WeusedtherecordsoftheFederalReservetoidentifyasubsetofmonetaryactionsthatwerenotmotivatedbyotherfactorsaffectingoutput.Wearguedthatthebehaviorofoutputfollowingtheseshockswouldproviderelativelyunbiasedestimatesoftheimpactofmonetarypolicy.

2

Thesubtitleofthatpaper,“ANewTestintheSpiritofFriedmanandSchwartz,”capturedourintellectualdebttoMiltonFriedmanandAnnaJacobsonSchwartz.FriedmanandSchwartzpioneeredthenarrativeapproachintheirseminalwork,AMonetaryHistoryoftheUnitedStates(1963).Weliketothinkthatwehavebuiltontheirinsightsandaddedsomemodernrigortotheapproach.

Overourcareers,wehaveusedthenarrativeapproachinanumberofstudies.Forexample,weusedittoseparatetaxchangesintothosetakenforcountercyclicalreasonsorbecausespendingwasincreasing,andthosetakenformoreexogenousreasons(RomerandRomer2010).Wefoundthattheestimatedresponseofoutputtotaxchangeswasfasterandlargerwhenwelimitedtheanalysistoexogenouschanges.

Othershavealsousedthenarrativeapproachtoachievemacroeconomicidentification.JamesHamiltonusedittoarguethatmanylargeoilpriceincreasesresultedfromfactorsoutsidetheU.S.macroeconomy(Hamilton1985).Thissuggestedthatthecorrelationhefoundbetweenoilpriceshocksandrecessionsreflectedalargelycausalrelationship(Hamilton1983).ValerieRameyusednarrativesourcestoidentifywhennewsofchangesindefensespendingbecameknown(Ramey2011).Thisinformationallowedhertobetterestimatethesizeandtimingoftheeffectsofgovernmentspending.AndrewJalilusedthenarrativeapproachtoidentifyprewarfinancialcrisesintheUnitedStates,andtoisolateasubsetthatwereduetofactorsuncorrelatedwithmacroeconomicconditions—suchasmanagerialmalfeasanceorotheridiosyncraticshocks(Jalil2015).Incorporatingthisadditionalinformationallowedhimtoidentifythetruecausalimpactoffinancialcrises.

Thisaddressisanopportunitytoreflectonthenarrativeapproach,andtorevisitourfirstpaperusingit.Theheartoftheaddressinvolvesredoingour1989study.Wefeelwehavelearnedagreatdealaboutthepitfallsofnarrativeresearchandhowtodoitbetteroverthepastthree-plusdecades.InSectionI,wediscussthoselessons,andapplythemtotheidentificationofmonetarypolicyshocksintheUnitedStatessince1946.

3

Havingredonethenarrativework,wethenincorporatethenewevidenceintoastatisticalframework.InSectionII,wediscussbothourmethodologyandourfindingsabouttheimpactofmonetarypolicyonrealeconomicindicators,suchastheunemploymentrateandrealGDP.Wefindthatacontractionarymonetarypolicyshockincreasestheunemploymentrateby1.6percentagepointsandreducesrealGDPby4.4percent,relativetotheno-shockbaseline.Theeffectsofmonetarypolicyarehighlystatisticallysignificant.Wefindthatourimprovedseriesonmonetarypolicyshocksresultsinslightlysmallerestimatesoftherealeffectsthanwefoundinour1989study,butgreaterstatisticalprecision.Wealsoexaminetheimpactofmonetaryshocksoninflation.Wefindthatcontractionarymonetarypolicyreducesinflationbyabout1.5percentagepoints,buttheeffectsdevelopslowlyandarelesspreciselyestimatedthanthoseforrealactivity.

Thoughourmainfocusisonthenarrativeapproachandestimatesoftheimpactofmonetarypolicy,inSectionIIIweconsidertheimplicationsofourfindingsforcurrentmonetarypolicy.WhatdoesournewandimprovednarrativeworkandempiricalestimationtellusaboutthelikelyoutcomeoftheFederalReserve’srecentactionstocontaininflation?

Finally,inSectionIV,weconsiderthefutureofthenarrativeapproach.Isitstillneeded,orhasitbeensupersededbyotherapproaches?And,ifitisstillneeded,doesitneedtobedonebyhumanresearchers?

I.RigorousNarrativeAnalysis

Thekeystepinnarrativeresearchistoanalyzesomenarrativesourcetogathersystematicevidencethatcanbeusedinanempiricalframeworkalongwithconventionaldata.Manyyearsago,someonereferredtowhatwedoasthe“l(fā)iterary”approach.Itwascleartheydidn’tmeanitasacompliment.Thetonewasthatnarrativeworkisinherentlylessscientificandrigorousthanconventionaldataanalysis.Wedisagreestrongly.Webelievedeeplythatnarrativeevidencecanbebothrigorousandreproducible.

4

A.FeaturesofGoodNarrativeAnalysis

Inthissection,wedescribewhatwehavelearnedaresomeoftherequirementsforrigorousnarrativeanalysis.TheserequirementsaresummarizedinTable1.Weillustratethefeatureswiththenewnarrativeresearchwehavebeendoingontheeffectsofmonetarypolicy—withaneyetowardshowinghowourworkhaschangedinlightofwhatwehavelearnedoverthepast35years.

AReliableSource.—Afundamentaldeterminantofthequalityofanynarrativeevidenceisthereliabilityofthesource.Thesourceshouldbecontemporaneousorrealtime.Otherwise,subsequenteconomicoutcomescouldaffecthowthesourcedescribesorinterpretsactions.Thenarrativesourceshouldexistforalongperiodoftimeandbereasonablyconsistentovertime.Otherwise,oneisnotgettingthesameinformationoverthefullsample.Thesourceneedstobeaccurate.Oneislookingforanarrativesourcethataccuratelyreportswhatwassaidatameeting,ortheobservationsofknowledgeable,accurateobservers.

Forourcurrentstudy,weareusingthehistoricalMinutesandTranscriptsofthemeetingsoftheFederalOpenMarketCommittee(FOMC)oftheFederalReserveSystem.

1

Theseareeitherverydetailedsummariesofthediscussionwithextensiveparaphrases,orverbatimtranscripts.Theyarecontemporaneouswiththemonetarypolicydecisions.Thereareroughly50to100pagesofmaterialpermeeting—so,witheighttotwelvemeetingsperyear(orinsomeperiodsevenmore),wearetalkingaboutalotofinformation(andreading!).Westartin1946becauseweareinterestedintheeffectsofmonetarypolicyinthepost-WorldWarIIperiod.

Oneimportantvirtueofthissourceisthat,formuchofoursample,itwasconfidential.ThedetailedrecordsofFOMCmeetingswerenotintendedorexpectedtobereleasedtothepublic.Asaresult,therewaslittlereasonforpolicymakersnottospeakhonestlyandunguardedly.Startingin1993,theTranscriptsweremadepublic,butwithafive-yearlag.Thelaglikelyhelpedtokeep

1ThenarrativesourcesaredescribedinmoredetailinonlineAppendixA.TheyareallavailableontheBoardofGovernorswebsite:

/monetarypolicy/fomc_historical_year.htm

.

5

thediscussionsfrank.

2

Italsomeansthatouranalysishastoendin2016,whichisthelatestyearforwhichtheTranscriptshavebeenreleased.

Thesourceweareusingisdifferentfromthatinouroriginalstudy.Backin1989,wecouldn’tfacethethousandsofpagesofhistoricalMinutesandTranscripts.Soweinsteadmainlyusedthemuchbriefer“RecordofPolicyActionsoftheFederalOpenMarketCommittee.”These“Records”areverysimilartowhatarenowcalledthe“MinutesoftheFederalOpenMarketCommittee.”Theyarequitebrief—ontheorderofjustafewpagespermeetingearlyinthepostwarperiodtoabouttenpagespermeetinginthe1980s.Becausethese“Records”aremuchlessdetailed,theyinevitablyprovidelessinformationonthemotivationforpolicy.Moreover,becausethe“RecordofPolicyActions”wasalwaysmadepublicwithonlyashortlag,wesuspectitmaybelessforthright.Forthesereasons,ourcurrentnarrativesourcehassubstantialbenefitsovertheearlierone.

AClearSenseofWhatOneIsLookingforintheNarrativeSource.—Justasimportantasthenarrativesourceishowoneapproachesit.Thekeylessonisthatyouneedtohaveaclearideaofwhatyouarelookingforinthedocuments.Thisiswherethecreativitylies.Whatcanwelearnfromthenarrativesourcethatwouldaidinestablishingcausation?ThisiswhatmadeFriedmanandSchwartz’sMonetaryHistorysoimportant.TheirgeniuswasinknowingwhattolookforinthediariesofmonetarypolicymakersandotherearlyFederalReserverecords.

Whatarewelookingforinthenarrativerecordforthisrevisitingoftheeffectsofmonetarypolicy?Ataverybroadlevel,wearelookingfortimeswhenmonetarypolicymakerschangedmoneygrowthandinterestratesforreasonsunrelatedtocurrentorprospectiverealeconomicactivity.Thesearepolicy“shocks”inthesensethatmonetarypolicyisnotbeingdrivenbyoutputorotherfactorsaffectingoutput.Thus,thebehaviorofoutputandotherindicatorsofrealactivity

2MeadeandStasavage(2008)andHansen,McMahon,andPrat(2018)findsomechangesinthemeetingsafterparticipantslearnedthattranscriptswouldeventuallybereleased,butnoclearchangesinthefranknessofthediscussion.

6

followingsuchepisodesshouldgiverelativelyunbiasedestimatesofthecausalimpactofmonetarypolicy.

3

Theproblemwiththisbroadframingisthatitishardtospecifyconcretecriteriathatcanbeusedtodeterminewhensuchepisodesoccurred.Soinstead,welookforsomethingmorespecificinthenarrativerecord.Asinouroriginalpaper,welookfortimeswhenmonetarypolicymakersfelttheeconomywasroughlyatpotential(ornormal)output,butdecidedthattheprevailingrateofinflationwastoohigh.Policymakersthenchosetocutmoneygrowthandraiseinterestrates,realizingthattherewouldbe(oratleastcouldbe)substantialnegativeconsequencesforaggregateoutputandunemployment.Thesecriteriaaredesignedtopickouttimeswhenpolicymakersessentiallychangedtheirtastesabouttheacceptablelevelofinflation.Theyweren’tjustrespondingtoanticipatedmovementsintherealeconomyandinflation.

Therequirementsthatpolicymakersweretakingactionsandwerewillingtoacceptrealeconomicconsequencesaredesignedtoensurethatpolicymakerswerenotjustengagedinritualhandwringing.Intermsofthesimpleaggregatesupply/aggregatedemandframework,wearelookingfortimeswhentheFederalReservedeliberatelyshiftedtheaggregatedemandcurveback.Ifmonetarypolicyhasrealeffects,outputshouldfallaftersuchactions.Thatis,thesearecontractionarymonetarypolicyshocks.

Focusingonaveryspecifictypeofpolicymovementunrelatedtocurrentorprospectiveeconomicdevelopmentscomeswithbothacostandabenefit.Thecostisthatitmeanswewillnotbeusingsomeepisodesthatarelegitimateforestimatingthecausaloutputeffectsofmonetarypolicy.Thiswouldlikelydecreasetheprecisionofourestimates,butshouldnotintroducebias.Thebenefitisthatmakesourcriteriaandidentificationproceduresmuchmoreconcrete,andso

3Importantly,weusetheterm“shock”tomeanmovementsinmonetarypolicyunrelatedtocurrentorprospectiverealeconomicactivity.Theyneedn’tbeunanticipatedwhentheyoccur.Thatis,iftheFederalReservetelegraphsachangeinpolicy,itcanstillbeusedtoestimatetheeffectsofpolicyaslongasitisnotdrivenbycurrentorprospectiverealeconomicactivity.Similarly,thechangeinpolicyneedn’tbeadeparturefromtheFederalReserve’susualbehavior.Ifmonetarypolicymakersrespondintheirusualwaytosomethingunrelatedtocurrentorprospectiverealoutput,thatstillconstitutesashockbyourdefinition.

7

lesspronetoerrors.

Animportantextensionthatwedointhenewworkistobroadenthecriteriatoincludeexpansionarymonetarypolicyshocks.Inparticular,wenowalsolookfortimeswhenpolicymakersbelievedthattheywereatastablelevelofeconomicactivity,buttookactionstolowertheunemploymentrate—andwerewillingtoacceptadverseconsequencesforinflation.Thatis,welookfortimeswhenpolicymakersweredeliberatelyshiftingtheaggregatedemandcurveoutbecauseofachangeintheirviewoftheacceptableordesirablelevelofunemployment.Ifmonetarypolicyhasrealeffects,outputshouldrisefollowingsuchactions.

ApproachtheNarrativeSourceDispassionatelyandConsistently.—Todothenarrativeanalysisrigorously,oneneedstoapproachthequalitativesourcedispassionatelyandconsistently.Researchersneedtoresistthetemptationtoseewhattheywanttoseeinthenarrativesource.Wefindthathavingmultiplereadersforthesamedocumentprovidesavaluablecounterbalancetothistemptation.Likewise,onehastoreadtheentiresourcefrombeginningtoend—notlookharderforinformationinsomeperiodsthanothers.

DocumenttheNarrativeEvidenceCarefully.—Wehavealsofoundthatwritingdetaileddescriptionsofouranalysisofthenarrativeevidenceisacrucialpartofthework.Thedocumentationprocessclarifiesourownthinkingandsometimesmakesuschangeourmindsabouttheclassificationofanepisode.And,itmakesiteasierforsubsequentresearcherstocheckourwork.Onechangethatwehavenoticedinourresearchisthatthesenarrativeappendixeshavebeengettinglongerandlonger.

B.ExamplesoftheNarrativeAnalysisofMonetaryShocks

Toprovideafeelforhowweapplyourcriteria,herewedescribesomeoftheevidencefortwoepisodesweidentifyasmonetaryshocks.Themuchmoredetaileddescriptionofthenarrativeevidenceforthesetwoshocks,aswellastheothersweidentify,iscontainedinonlineAppendixA.

ContractionaryShockinDecember1988.—December1988isanexampleofa

8

4AllcitationstotheTranscriptsaretoU.S.BoardofGovernorsoftheFederalReserveSystem(1976–2016).

contractionarymonetarypolicyshock.Itisthelatestcontractionaryshockinoursample(which,asnotedabove,endsin2016).

Formuchof1987and1988,policymakerswereworriedthatinflationwouldpickupiftheydidn’ttighten.Inresponse,theymadesmallmovestowardrestraint.Theseinitialmovesdon’tcountasacontractionaryshockbyourcriteria.Policymakersweren’ttryingtoshifttheaggregatedemandcurvebackfromastablelevel;rathertheywerejusttryingtoholditsteadyagainstotherforcestendingtoshiftitout.

InMay1988,policymakersstartedtosaythatthecurrentlevelofinflationwasunacceptable,notjustthattheywantedtopreventinflationfromrisingfurther.Forexample,onesaid:“Intermsofourowninflationrate…wehavebeenstalledataratethatIthinkistoohighformostofus”(W.LeeHoskins,Transcript,May17,1988,p.5).

4

Anothersaid:“whateverislikelytohappenonthewageandpriceside,itdoesn’tseemtomethatthere’sgoingtobeanydecelerationnextyearunlessweact.Ithinkitistimeforsomefurtheraction”(GaryStern,pp.4–5).However,anumberofothermemberswerelessclearthattheywantedtoliterallyreduceinflationandwerehesitanttoincurthepossibleoutputconsequences.Aconcernaboutrisinginflation,ratherthanitslevel,continuedthroughthesummerandfallof1988.Forexample,attheAugust1988meeting,onemembersaid:“thebottomlineisthatweareinaterritoryofacceleratinginflationandwehavetoresistthatgrowth”(EdwardBoehne,Transcript,August16,1988,p.17).

AdesiretoreduceinflationandawillingnesstoacceptoutputconsequencesbecamemuchmorewidespreadattheDecember1988meeting.Forexample,onemembersaid:“Ithinkthejobbeforeusistocontaintheinflationandtoslowthiseconomydown”(Forrestal,Transcript,December13–14,1988,p.56).Anothersaidthat,whilethefragilityofthefinancialsystemwasarealconcern,“wecan’tdesignmonetarypolicytoavoidanydifficultiesinvarioussectors.…

9

[W]e’vegottofocusoninflationandifsomethinggoeswrongthenyoucanaddressthoseproblems”(H.RobertHeller,p.51).Thestaffpredicatedtheirforecastonagoalofreducinginflation.Theywarnedthat“ifitistheaimoftheCommitteenotmerelytoholdthelineoninflationbut,rather,torestoreadownwardtrendby1990,thenitmaybenecessarytoruntheriskofsomefinancialstressandeconomicweakness”(MichaelPrell,Transcript,PresentationMaterials,December13–14,1988,p.1).FOMCmembersneverthelessagreedtoembarkonasignificanttightening.

Thisepisodecountsasacontractionarymonetarypolicyshockbecause,atastablelevelofgrowthandunemployment,policymakersdecidedthatthecurrentlevelofinflationwasunacceptableandtookactionstoreduceit.Andtheyclearlyunderstoodandacceptedthattherecouldbesubstantialadverseconsequencesforoutputandunemployment.

ExpansionaryShockinJanuary1972.—January1972isanexampleofanexpansionarymonetarypolicyshock.Indeed,itistheonlyexpansionaryshockwefind.

Followingtherecoveryfromthemildrecessionof1969–70,therewasasenseamongFOMCmembersandthestaffthatunemploymenthadstabilizedatanelevatedlevel.Forexample,inNovember1971,theFederalReservestaffprojectedthatunemploymentwouldfallonlyslightly(from6percentto5.3percent)in1972(JosephZeisel,MemorandumofDiscussion,November16,1971,p.40).

5

ByDecember1971,therewasagrowingconsensusthattheprevailinglevelofunemploymentwasunacceptablyhigh,andthattheFederalReserveneededtotakeexpansionaryactions.Forexample,onemember“believedtheappropriateposturefortheSystematthispoint

5AllcitationstotheMemorandaofDiscussionaretoU.S.BoardofGovernorsoftheFederalReserveSystem(1967–1976).In1967,followingthepassageoftheFreedomofInformationAct,theFederalReserveseparatedoutasmallamountofthematerialinthehistoricalMinutesandcalledthemthe“MinutesofActions.”Theremainingmaterial,whichappearstocontainessentiallyallofwhatwaspreviouslyinthehistoricalMinutes,wasputinanewdocumentcalledtheMemorandumofDiscussion.LikethehistoricalMinutes,theMemorandaofDiscussionarewritteninthethirdpersonandcontainextensiveparaphrasesofthediscussion.

10

wasoneofdoingwhatitcouldwiththepolicyinstrumentsatitsdisposaltofosterandencourageeconomicexpansion”(J.DeweyDaane,MemorandumofDiscussion,December14,1971,p.60).

InamoveverysimilartooneusedbyFederalReserveChairmanPaulVolckerin1979,ChairmanArthurBurnscalledaspecialmeetingoftheFOMConJanuary11,1972“becausehehadbecomeseriouslyconcernedaboutthepresentstanceofmonetarypolicy”(MemorandumofDiscussion,January11,1972,p.4).Hesaidthat“unlesstheaggregatesnowbegantogrowatadequaterateshewould…feelthattheremightbesomevalidityinachargethattheSystemwasnotsupportingthepoliciesoftheAdministrationandCongress”(p.62).InanotherparalleltoVolcker’sactionsin1979,Burnsalsoputontheagendaaproposaltoadoptreservetargeting.And,hewasveryclearthat“therecouldbeafurtherreductionininterestrates,possiblyofsignificantdimensions”(pp.63–64).

AnumberofFOMCmembersfearedthattheswitchtorapidreservegrowthandlowerinterestrateswouldgenerateinflation.Forexample,onesaidthat“theCommitteeshouldconsiderwhetherstimulatingtheeconomytogreaterheightsintheshortrunwouldnotinvolveacostintheformofaresurgenceofinflationarypressureslateron”(PhilipColdwell,MemorandumofDiscussion,January11,1972,p.71).Anothersaidthat“itshouldberecognizedthatthebattleagainstinflationwasnotyetover,andthatundulyaggressivepolicyactionswouldinvolvetheriskofrekindlinginflationaryexpectations”(JamesRobertson,p.90).Nevertheless,despitethreedissentingvotes,amajorityofthecommitteeagreedtoradicallymoreexpansionarypolicy.

Thisepisodemeetsourcriteriaforanexpansionarymonetarypolicyshock.Fromastablelevel,theFOMCdecidedtotakeactiontolowerunemploymentbecauseitfeltthecurrentlevelwasunacceptable.Policymakersweredeliberatelyshiftingouttheaggregatedemandcurve,despitewidespreadconcernthatitwouldbeinflationary.

C.ResultsoftheNarrativeAnalysis

Theoutcomeofallthisreadingandanalysisistheidentificationoftendatesofmonetary

11

policyshocksovertheperiod1946–2016.Ninearecontractionaryshocksandoneisexpansionary.ThesedatesareshowninTable2.

Interestingly,althoughweexpandedoursamplebyalmost30years,wefoundnomonetaryshocksbetween1988and2016.Atsomelevel,thisisnotterriblysurprising.Inflation,untilveryrecently,hadbeenlowandsteadyforalmost30years.And,Americanmonetarypolicyhasbeenwidelyviewedasdeftandcarefullycalibrated.Itwastobeexpectedthattheremightbenothingtocallaclear-cutanti-inflationarymonetarypolicyshocksincethelate1980s.Obviously,however,thatmayhavechangedinthepastyear—wheninflationonceagainrosesubstantiallyandmonetarypolicymakershavetakenstrongactionstoreduceit.WediscusstherecentexperienceinSectionIII.

Partofthemotivationforrevisitingthistopicandnarrativeworkwastoseehowwellwehaddoneoriginally.Thisisatestbothofouryoungerselvesandourmorelimitednarrativesource.Forthemostpart,wewereprettyhappywithouroriginalfindings.Inadditiontothenewdatesofmonetaryshocks,Table2alsoshowsthedateswesetoriginally.

HYP