Speech

Inflation Hawk = Employment Dove CenterfortheStudyofAmericanBusiness WashingtonUniversityinSt.Louis St.Louis,Missouri July29,1999 Istart with a confession: my title this morn- • Third,intheshortrun,theremaybeeffects ing reflects personal frustration over the thatbringcriticismontheinflationhawks— publicdebateconcerningacriticallyimpor- ahigherinterestratemightincreaseunemtant monetary policy issue. The issue con- ployment,andalowerinterestratemight cerns the relationship between monetary policy reduceunemployment.But,giventhelongand unemployment. My frustration reflects the runrelationshipalreadydiscussed,analyzfact that I am often called an “inflation hawk,” ingshort-runissuesrequiresspecialcare tobesurethattheshortrunislinkedcorwith the implication, at least as it touches my rectlytothelongrun. thin skin, that I care little about unemployment. Further, the implication seems to be that those • Fourth,weneedtoexamineissuesofvariurgingtheFederalReservetokeeppushinginter- abilityaroundtheaverages.Ibelievethere est rates lower and lower are the true friends of isacompellingcasethatbothinflationand theunemployed.Irejectthisanalysiscompletely, unemploymentaremorevariablewhen theaverageinflationrateishigher.Ialso and I’ll tell you why. Low inflation, far from believethatFedpolicycancontributeto being the enemy of the unemployed, is a friend. employmentstabilitywhilekeepinginfla- Letmeoutlinethisargument,whichI’llthen tionlow,butthisbeliefismoreofaconjecdevelopinmoredetail,stepbystep. turethansomethingIcandemonstrate. • First,economistsofallstripesagreethat • Fifth,andfinally,Iinsistthattheonlyreaoverthelongrunthereisnosignificant sonablewaytothinkaboutmonetarypolicy tradeoffbetweentheaveragerateofinflation isinthecontextofuncertainty.Policyandtheaveragerateofunemployment. makershavetoplaytheodds.Afairassess- Therefore,onaverageovertime,aneasier mentoftheperformanceofpolicymakers monetarypolicydoesnotyieldlowerunem- requiresaFridayview,notaMondayploymentbutdoesyieldhigherinflation, morningviewafterthegamehasbeen allotherthingsbeingequal.Bythesame played.Inaninherentlyuncertainenvironargument,atightermonetarypolicydoes ment,policymakersshouldnotbecriticized notyieldhigherunemployment,butdoes forbeingunabletoguesswhatshocksthe yieldlowerinflation. futurewillbring,onlyformakingbadbets beforethefuturearrives. • Second,solongastheinflationrateremains low,policyactionschangingthefederal Beforeproceeding,Iwanttoemphasizethat fundsrateupordownshouldbeexpected theviewsIexpressherearemineanddonotnectohavelittleeffectonbondyieldsinthe essarilyreflectofficialpositionsoftheFederal longrun.Inparticular,ahigherfedfunds ReserveSystem.Ithankmycolleaguesatthe ratethatheadsoffincipientinflationis FederalReserveBankofSt.Louisfortheircomfriendlytobondinvestors. ments,butIretainfullcreditforerrors. 1

MONETARYPOLICYANDINFLATION THE LONG RUN higherinflationwouldnotbuylowerunemployment.Norwouldpoliciesyieldinglowerlong-run Let’sbegintheanalysiswithadiscussionof inflationcosthigherunemploymentonaverage long-runrelationships.Theissuehereisthe overthelongrun. relationshipbetweenvariablesontheaverage Inthecontextoflong-runanalysis,theoutoverasubstantialperiod,suchasadecade.The standingissuecametobethenumericalestimate calendarperiodmustbelongenough,depending ofwhatsomecalledthe“naturalrateofunemontherelationshipbeingexamined,thatadjustployment,”andwhatotherscalledthe“nonmentstotransitoryevents,orshocks,arecomplete. acceleratinginflationrateofunemployment,”or Thirtyyearsago,macroeconomistswere NAIRU.(I’llskipovertheesotericaofpossible embroiledinaspiriteddebateastowhetherthere differencesinthemeaningsofthesetwoterms.) wasalong-runtradeoffbetweeninflationand Iftheinflationratesettleddowntoasteady5perunemployment.Ontheonesidewereeconomists centperyear,or8percentperyear,or1percent whobelievedthattheeconomycouldgenerate peryear,whatwouldbetheequilibriumrateof lowerunemploymentontheaverageifsociety unemploymentinthelabormarket?Theprofeswerewillingtoaccepthigherinflation,atleast sionconvergedonanumberintheneighborhood uptoapoint.Practicallyspeaking,thedebate of6percent,althoughsomethoughttheestimate centeredaroundthepossibilitythattheU.S.econshouldbehigherandsomelower. omyoperatingpermanentlyat,say,5percent Althoughmanyinvolvedinthisdebate annualinflationwouldgenerateloweraverage thoughttheirestimatesofthenaturalratewere unemploymentthanwouldtheeconomyoperatsecure,othersbelievedthatsuchestimatescarried ingatanannualinflationrateof,say,1percent. ahighdegreeofimprecision.Mostbelievethat Economistswhobelievedinthelong-run todaythenaturalrateislowerthanitusedtobe, tradeoffbelievedthatexperienceintheUnited althoughIdonotintendtodigintothissubject KingdomandtheUnitedStatessupportedtheir thismorning. view.Theyalsobelievedthatthatthereweresome Theimportantpointisthattheprofession theoreticalreasonsforbelievinginthetradeoff. didconvergeonthepropositionthattherateof Economistsontheothersideofthedebatebelieved unemploymentwouldnotbelowerinthelong thatmarketsarefundamentallyregulatedbywellrunifsocietyacceptedahigherrateofinflation. informed,profit-makingfirmsandwell-informed, Theupshotofthisargumentismyfirstandmost utility-maximizinghouseholds.Insuchaneconimportantpointthismorning:Theproposition omy,weshouldnotexpectthatongoinginflation thatahardlineoninflationwillyieldhigher could“buy”lowerunemployment.Thecentral reasonisthateveryoneshould,atleasteventually, unemploymentissimplynotacceptedinthe seethroughthefictionofrisingwagesandprices. economicsprofession.Exceptatthefringes, Nothing“real”shouldbeaffectedbyongoing economistswhoarepoliticallyliberalorpolitiinflation.Why,forexample,shouldanyonework callyconservativeonawidevarietyofotherissues morehoursperweek,ordelayretirement,when agreeonthisproposition. wagesandpricesarerisingat5percentperyear Ithink,infact,thatontheaverageinthelong thanwhenwagesandpricesarerisingby1percent runitispossiblethatlowerinflation,allother peryear? thingsequal,maybeassociatedwithlowerunem- Bythemid-1970s,thisissuewasdecided. ployment.Moreover,Ibelievethereisacompelling Mainstreammacroeconomistsacrossthespectrum casethatlowerinflation,allotherthingsequal, ofprofessionalopinion,exceptattheveryends, isgoodforeconomicgrowth.Amarketeconomy agreedthatneithertheorynorevidencejustified worksmoreefficiently,andfiscalandregulatory theviewthattherecouldbealong-runtradeoff distortionsarelessserious,whentheinflation betweeninflationandunemployment.Thus,a rateislowandstable.TheFederalReservecan governmentpolicythatyieldedpermanently contributetomaximumsustainableeconomic 2

InflationHawk=EmploymentDove growthbypursuingpoliciesleadingtopricesta- ableeconomicgrowth.Economistscananddo bility,orverylowandstableinflationifyou argueaboutthemagnitudesoftheseemployment prefertostatethegoalthisway. effects,butbroadlyagreeontheirdirection. Overthelongrun,however,ratesofunem- Finally,maintaininglowinflationontheaverage ploymentandeconomicgrowtharedetermined willbebond-investor-friendly. primarilybynonmonetaryconditions.Itiseasy todemonstratethispointbycomparingU.S.and Eurozoneexperience.Bothareashavehadsimilar THE SHORT RUN inflationratesoverthelast15yearsorso,but Ifthereisnolong-runtradeoffsuchthat unemploymentinmajorcontinentalEuropean unemploymentislowerwheninflationishigher, countrieshasbeenfarhigherthanintheUnited thenperhapsthematteratissueconcernsashort- States.France,forexample,hashadanunemployruntradeoff.Whatmightbeinvolvedhere? mentrateabove9percenteveryyearsince1984. Thereareseveraldifferentwaystoexamine Althoughmyfocusthismorningisonmonetheshort-runlogic.Suppose,forexample,we tarypolicyandunemployment,Ishouldcomment considerahypotheticalexperimentinwhicha brieflyabouttheimplicationsofmyargument centralbankhasbeenpursuingapolicyfora forinterestrates.Overthelongrun,theaverage longtimethatyields2percentannualinflation. levelofnominalinterestratesisdeterminedpri- Nowsupposethecentralbankchangesitspolicy marilybytheaveragerateofinflation.Thehigher toonethatyieldsalong-runrateofinflationof therateofinflationisontheaverage,thehigher 2.5percentperyear.Duringthetransitiontothe averageinterestrateswillbe.Asuccessfulmonehigherrate,themoreexpansionarymonetary tarypolicythatmaintainspricestability,therefore, policymightforatimebiddowntheunemployisnotonlyemployment-friendlybutalsointerestmentrate.Butgiventhefactthatthereisnolongrate-friendly.Idonotknowwhatpolicyactions willbeappropriatelaterthisyear,butIwantto runtradeoff,thedeclineinunemploymentwill emphasizethatitissimplynottruethatincreases betemporary.Thenetresultisthattheeasier inthefundsratecauselastingincreasesinbond monetarypolicyyieldsatemporaryreduction rates.Forexample,theFedraisedthefederalfunds ofunemploymentandapermanentincreaseof ratesignificantlyoverthecourseof1994,butby inflation. December1995,the30-yearTreasurybondrate AnotherversionofthisargumentgoessomeendeduplowerthanithadbeeninJanuary1994. thinglikethis:Supposeafavorabledisturbance— Fromreadingthefinancialpressoverthelast suchasatechnologicalchangethatraisestherate coupleofweeks,itseemsthatbondinvestorsare ofproductivitygrowth—impactstheeconomy.If jitteryoverthepossibilitythattheFedmightraise thecentralbankmaintainsmonetarypolicyasis, thefederalfundsratelaterthisyear.Investors perhapstheinflationratedropsfrom2percent oughtinsteadtobereassuredthattheFedwill,if peryearto1.5percentperyear.Asomewhatmore necessary—andIemphasizetheif—raiseratesto expansionarypolicy,ontheotherhand,might keepinflationlow.Bondinvestorshavefarmore keeptheinflationrateat2percentperyearbut tofearfromasustainedriseintherateofinflation yieldatemporarybonus:reducingtheunemploythantheydofromtheupsanddownsofthefed- mentrateforatime.Inthiscase,theunchanged eralfundsraterequiredtokeeptheeconomyon monetarypolicy“buys”apermanentlylower anevenkeel. inflationrateinexchangeforanunchangedrate Tosummarize,higherinflationdoesnotbuy ofunemployment;thealternativemonetarypolicy lowerunemployment;ifanything,pricestability keepstheinflationrateunchangedbutyieldsa yieldsthelowestpossibleunemploymentrate, temporarilylowerunemploymentrate.Thesetwo allotherthingsbeingequal.Moreover,pricesta- casesareessentiallythesameexceptthattheend bilitycontributestoachievingmaximumsustain- resultisadifferentinflationrate.Thecasesarethe 3

MONETARYPOLICYANDINFLATION samebecausethemarginalinfluenceofmonetary ofinflationishigher.Therecessionsof1973-75 policyperse,ignoringotherthingsgoingonsuch and1981-82,whichwerequitesevere,werea asthechangeinproductivitygrowth,isthesame directconsequenceofrisinginflationandrising inbothexamples. inflationexpectations.Between1965and1985, IhavenodoubtthatexcessiveFedtightening unemploymentmayhaveaveragedaround6perwouldraisetheunemploymentrate.Tousean cent,butitswungbetween3.5and9.7percent absurdexample,iftheFedweretoraisethefederal onanannualaveragebasis.Ifwecomparetwo fundsratefromitscurrentlevelof5percentto10 economies,bothwiththesameaverageunempercenttomorrowandkeepitthere,theunem- ploymentrate,everyonewillagreethattheeconploymentratewouldrisesubstantially.Some omywiththemorestableunemploymentrateis peopleseemtoreasonthatifabigincreasein preferabletotheonewithawidelyfluctuating thefundsratewouldcauseabigincreaseinthe unemploymentrate. unemploymentrate,thenasmallincreaseinthe Thecorrelationbetweenafluctuatinginflation fundsratewouldlogicallycauseasmallincrease rateandafluctuatingunemploymentrateisfar intheunemploymentrate.Butthatargumentis fromaccidental.Uncertaintybybusinessesand flawed.Wheneconomicconditionsareshifting, householdsabouttheinflationrateleadsthemto astheyalmostalwaysare,smalladjustmentsin makedecisionsthatturnout,willy-nilly,tobe thefederalfundsrate,onewayortheother,may mistakeninthelightofsubsequentevents.Itis berequiredtomaintainpricestabilityandprevent simplyverydifficulttoplanefficientlywhen thedevelopmentofinflationordeflation,either pricesintheaggregatearechangingrapidly.Someofwhichwoulddestabilizeemployment. timesthemistakesareononeside,andsometimes Tosummarizetheshort-runargument,under ontheotherside.Thenetresultofanunstable someconditionsitistruethatamoreexpansioninflationenvironmentisthatswingsinthebusiarymonetarypolicymightyieldtemporarilylower nesscyclebecomeunavoidable.Therefore,my unemployment,andalessexpansionarymonetary personalconvictionisthatthecentralbankmust policymightyieldtemporarilyhigherunemploynottakechanceswiththeinflationrate;ifinflation ment.TheremaybetimeswhentheFederal getsawayfromus,arecessionwillnotbefar Reservewillhavetomakeanexplicitchoice behind.ItisforthisreasonthatIconsidermy aboutwhethertopursue,orpermit,thesetemstanceasaso-calledinflationhawktobethatof poraryeffects.Itisalsotruethatamajorpolicy afriendoftheunemployed,orpotentiallyunemmistakecouldquicklyhavemajoreffectson ployed,shouldtherebearecession. unemployment.Theproblemistofigureouthow Myconjectureisthattimelypolicyactions toavoidmistakesbymakingtimelyadjustments canhelptostabilizeemploymentwithoutaffecting inthefederalfundsratetokeeptheeconomyon eitherthelong-runaveragerateofunemployment asmooth,noninflationarytrack. oraveragerateofinflation.Forexample,Ibelieve thattimelyFedactionlastfallhelpedtoprevent financialmarketdisturbancesfollowingthe LOW INFLATION AND Russiandefaultfromfeedingintotherealecon- EMPLOYMENT STABILITY omy.Still,Ibelievethatweknowsolittleabout Ibelievethatthetraditionalanalysisofboth theseissuesonasystematicbasisthatthepropothelongrunandtheshortruniscorrectand sitionneedstobeofferedasaconjecture.And important.Economistshavealsonotedanother becauseourknowledgeissolimited,wehaveto importantpointthatbearsdirectlyontoday’s becarefulaspolicymakersthateffortstocushion policyissues—thehighertheaveragerateof short-runshocksdonotturnintopolicyactions inflation,themorevolatiletheeconomyseems thatinadvertentlyaremoreexpansionaryorcontobe.Thevariabilityofbothinflationandemploy- tractionarythanisconsistentwiththeparamount mentisstatisticallyhigherwhentheaveragerate long-runobjectiveofpricestability. 4

InflationHawk=EmploymentDove PLAYING THE ODDS contributiontomaximumsustainableeconomic growthistopursuemonetarypolicyfocusedon Anessentialfeatureofpolicymakingunder pricestability,tomaintainanefficientandreliable uncertaintyisthatthepolicymakermustplay paymentssystem,andtofosterasafebanking theodds.Ofcourse,wewanttostacktheoddsas system.Thecentralbankisnotresponsiblefor muchinourfavoraspossible,buthavingdone otherkeydriversofeconomicgrowth.These everythingwecantoimprovetheodds,wemust include,innoparticularorder,theeducation stillmakesomebetsbasedonprobabilities.This system,themaintenanceofpublicsafety,the viewofthepolicyproblemisnotpeculiarto efficientallocationofresourcesthroughgovernmonetarypolicy,butprevailsinallareasofpublic mentaltaxandspendingprograms,theculture andprivatedecisionmakingwhereactionstoday ofentrepreneurship,inventionandtechnical mustbetakeninthecontextofuncertaintiesabout progress,andsoforth. thefuture,andevenaboutthecurrentsituation. Supposethecentralbankweresuccessfulin OneofthefrustrationsIfeelasIreadallthe achievinglowandsteadyinflationthrougha policyadvicedirectedtheFed’swayisthatsome controlmechanismbasedonregulatingofthe commentatorsfailtoanalyzewhichrisksare amountofmoneyintheeconomy.Theninterest worthtakingandhowtocalculatetheodds.A rates—whichbyassumptionarenotdirectlyconcomprehensivepolicyanalysisrequiresthatwe trolledbythecentralbank—wouldfluctuateup considerallpossiblescenariosandtheirprobabilanddownasvariouseventsandshocksoccurred. ities.Thenextstepistoconsidervariouspolicy Infact,theFederalReservedoesusethefederal optionsandthelikelyeconomicoutcomesunder fundsrateasitsshort-runpolicyinstrument.A eachofthescenarios.That’showIapproachpolsuccessfulpolicyrequires,therefore,thattheFed icymaking.Ofcourse,Ialsotakeaccountofthe adjustthefundsrateupanddownasnecessary factthatpolicyisnotetchedinstonebutcanbe tobeconsistentwithpricestability.Whatactions adjustedovertime,thatpolicyactionsyieldpolicy arenecessaryisneverperfectlyclear,andmistakes effectswithalag,andsoforth. arealwayspossible.Butitsimplycannotbethe Itiseasytoseethattheexercisebecomes casethateveryfederalfundsrateincreaseis complexprettyquickly,butthatiswhatmustbe worker-unfriendly.Ifthatwerethecase,thenwe facedinanycomprehensivepolicyanalysis.I shouldexpectapolicyyieldingpricestability myselfamconvincedthatasignificantincrease andfullemploymenttobeaccompaniedbya ininflation,especiallyifaccompaniedbygrowing permanentlydecliningtrendtointerestrates,a concernsinthemarketplaceaboutthefuturerate nonsensicalexpectation. ofinflation,wouldadverselyaffecteconomicout- Finally,apolicyyieldingpricestabilitywill comes.Itisimportanttounderstandthatpolicy alsoyieldaflattrendaroundwhichthefederal dependsonboththedesiredoutcomeandthe fundsratewillfluctuate.Thelongbondyield, flowofdataabouttheeconomythataffectsthe whichreflectstheaverageofexpectedshortrates probabilitiesofvariousscenarios.Itistherefore overthelifeofthebond,shouldfluctuaterelanotjustunwisebutimpossibleformetoforecast tivelylittleinthisenvironment.TimelyFed myownpositionatthenextFOMCmeeting, actiontoraisethefundsratewhenrequiredto becausethatpositionissubjecttorevisionbased maintainpricestabilityis,therefore,notonly onnewinformation. worker-friendlybutalsobond-investor-friendly. I’vesaidmypiece,andhopethatfromnowon thepresswillrefertomeasan“unemployment ENDPIECE dove.”ButIdoubtit! I’llfinishbydrawingthethreadsofthisargumenttogether,butinawaythatisabitdifferent thanmyargumentabove.TheFederalReserve’s 5