speeches · February 14, 2001

Speech

William Poole · President

Economic Forecasts and Monetary Policy ArkansasBusinessandEconomicSocietyandthe CentralArkansasChapteroftheRiskManagementAssociation LittleRock,Arkansas February15,2001 Of all the things that people ask me Beforeproceeding,Iwanttoemphasizethat after speeches, at meetings, at par- theviewsIexpressherearemineanddonot ties,andduringcasualconversation necessarilyreflectofficialpositionsoftheFederal over lunch, no topic arises more ReserveSystem.Ithankmycolleaguesatthe often than that of my outlook for the economy. FederalReserveBankofSt.Louisfortheircom- Strangely enough, this is a topic on which I can ments,butIretainfullresponsibilityforerrors. shed little light for anyone who is modestly well informed. My purpose today is to discuss this disconnect—why it is that people believe that I FORECAST ACCURACY have a very special economic forecasting crystal Indiscussingtheaccuracyofeconomicfore- ball and why I know that I do not. casts,I’mgoingtorelyheavilyontheBlueChip Thefirstquestionisprettyeasy.Peopleseem compendiumofforecasts.Asmanyofyouprob- toassumewithlittlethoughtthatImusthavesome ablyknow,BlueChipisamonthlynewsletter forecastingexpertise,orotherwiseIwouldn’tbe reportingforecastsfromapanelofeconomicfore- inmycurrentposition.I’llchallengethatassump- casters.Thenewsletterisdatedthetenthofevery tionbyexplainingwhyIknowIdonothavesuch month,andI’llbereferringtotheFebruary10, expertise.However,towardtheendofmyremarks 2001issue. I’llexplainhowIapplymyprofessionalexpertise TheBlueChipConsensusforecastforthe totheoutputofexperteconomicforecasters. U.S.economyisthatin2001totalrealgross I’llexploremytopicbyfirstdiscussingthe domesticproduct(GDP)willgrowby2.1percent accuracyofeconomicforecasts.Perhapsyouare overitstotalforlastyear.Itisimportanttonote alreadychucklingatthatphrase—“theaccuracy thatthisforecastreferstothetotalGDPforthe ofeconomicforecasts.”Ifso,I’mofftoagoodstart. year,andnotthegrowthfromthefourthquarter ThenI’lldigintowhyforecastsarenotveryaccu- of2000tothefourthquarterof2001,whichis rate.IthinkthatwhatIhavetosayinthisregard anothercommonwayofmeasuringrealgrowth willringtruetoyou.Ifforecastsarenotveryaccu- fortheyear.BlueChipdefinesthe“consensus” rate,howcantheFedconductmonetarypolicy? asaboutthemiddle,orthemedian,ofthe51 Indiscussingthisquestion,I’llemphasizethat forecasterssurveyed. whattheFedcandoistosetastablelong-runpath Ithinkthatforecastisreasonable,butsuppose forpolicy,yieldinglowandstableinflationon Ididnot?Suppose,hypothetically,Iweretotell theaverage.Finally,Iwillspeaktothecurrent youthatIbelievetheU.S.economywillbevery outlook,concentratingonconveyingthenature muchweakerthantheconsensusforecast.Sup- ofthemainstream,consensusforecast.Ishare pose,also,thatIweretotellyouthat,asaconse- thatforecastbecauseIamaconsumerratherthan quenceofweakGDPgrowth,corporateprofits aproducerofforecasts. aregoingtobemuchlowerthantheconsensus 1 MONETARYPOLICYANDINFLATION forecastthisyearandnextyearaswell.Moreover, Iamnotintheforecastingbusiness—Iama I’dtellyou—thisstoryisallhypothetical,remem- consumerofeconomicforecasts.WhatshouldI ber,toproveapoint—thatwiththemiserable makeofthesubstantialrangeofforecastsproduced outlookforcorporateprofitsIbelievethatthe byseriousforecasterswhomakealivingatthis stockmarketisgrosslyovervaluedatcurrentlevels. enterprise?Tome,theonlyreasonableconclusion I’msureI’dgetyourattentionbyspinning isthatthereissubstantialroomforsignificant outthisdrearyforecast.Howwouldyoureact? differencesofprofessionalopiniononthismatter. Wouldyou,forexample,pulloutyourcellphone Differentprofessionalshavedifferentviewson totellyourbrokertosellallyourcommonstock? theoutlook,andthatfactmustreflectgapsinecon- omists’knowledge.Thefactisthateconomicsis Wouldyouusethecashyouraisedascollateral notsowelldevelopedasadisciplinethatthese forlargeadditionalshortsalesofcommonstock? forecastersareallcompelledbyconfirmedknowl- Ithinknot.Ihopenot. edgetoreportforecastsclusteringwithinafew Whynot?Theanswerisprettysimple.You tenthsofapercentofeachother.Instead,what knowthateconomists’forecastsarenottobe economistscanreasonablysaythey“know”per- takenthatseriously.Whatevermightbemyfore- mitswidedifferencesofopiniononforecasts. cast,youknowthatthereareothereconomists Forecastersthemselveswillsaythatthediffering withrespectablecredentialswhoareforecasting forecastsofmanyoftheirrivalsare“reasonable.” continuingsolidgrowthafteraslowfirsthalfof Letmenowlookatanotherdimensionofthis thisyear.Justasyouwouldnotsellshortonthe rangeofforecasts.TheFebruaryBlueChipcon- basisofmyforecast,youwouldnotmortgageyour tainsatableofconsensusforecastsfor2001made housetothehilttobuystockonmarginifIwere duringthecourseof2000.TheJanuary2000con- insteadtoofferaveryupbeatforecast.Myeco- sensusfor2001wasrealgrowthof3.0percent. nomicforecast—everyeconomicforecast—must Theconsensusforecastroseoverthecourseof betreatedwithextremecaution. 2000,reachingahighof3.5percentinSeptember. IntheFebruaryBlueChip,51respondents InOctobertheconsensuswasalso3.5percent offeredforecastsfor2001and47for2002. realgrowth.InNovember,theconsensusforecast Althoughtheconsensusforecastwas2.1percent begantofall,butitslippedonlyslightly,to3.4 realGDPgrowthin2001over2000,theaverage percent.Thus,notonlydovariousforecasters ofthetop10was2.8percentgrowthandthe differabouttheoutlookatanymomentoftime, averageofthebottom10was1.3percent.The buttheychangetheirmindastimegoeson.Just mostoptimisticofthe51forecastsforrealGDP toemphasizethispoint,noteagainthatthecon- growthin2001was3.6percentandthemostpes- sensusfor2001lastNovemberwas3.4percent simisticwas0.9percent.For2002,theconsensus butnow,onlythreemonthslater,theconsensus forecastwas3.5percentrealgrowth;theaverage forthisyearhasfallento2.1percentrealgrowth. ofthetop10was4.1andforthebottom102.8 Whatshouldweconcludefromthefactthat percent.Themostoptimisticrealgrowthforecast forecastsareoftenrevisedsubstantiallyoverthe for2002was5.3percentandthemostpessimistic spanofafewmonths,asillustratedbytherecent was2.2percent. significantdownwardadjustmentintheforecast Therangeofforecastsisobviouslyquitesub- forthisyear?Foronething,certainly,onereason stantial.ThenamesintheBlueChiplistaremostly whyyououghtnottosellyourportfolio,ormort- namesthatyouwouldrecognize.Theyinclude gageyourhousetobuyadditionalstock,onthe majorbanksandothercompaniesthatmaintain basisofeconomicforecastsisthattheseforecasts economicforecastingunits.Inaddition,there differalotfromoneforecastertoanotherand arespecializedcommercialforecastingfirmsin changesignificantlywiththepassageoftime, thelist.Alloftheseforecastershaveanintense evenoveracoupleofmonths.Foranotherthing, interestingettingtheforecastright.Theydevote let’srecognizethatforecastsarenotnowand substantialtimeandresourcestotheirforecasts. neverhavebeenhighlyaccurate. 2 EconomicForecastsandMonetaryPolicy Thereisasubstantialprofessionalliterature WHY FORECASTS GO ASTRAY ontheaccuracyofeconomicforecasts.Todiscuss Wheredoforecasterrorscomefrom?Whydo thisliteratureinanydetailwouldrequirethatI forecastsgoastray?Understandingtheanswers getintoamind-numbingexpositionofexactly tothesequestionshelpsustobetterappreciate whatisbeingforecast—forexample,theprelimi- whateconomicforecastsmean. naryorfinallyrevisedGDPestimate—andother Letmebeginwithananalogy.Supposeyou suchmatters.Butletmegivetheflavorofthis weretoaskaprofessorofchemistryatyourlocal research.Everyforecastoughttohaveastandard universitywhatwillhappeninthechemistrylab errorattachedtoit.Whendiscussingannualaver- nextSeptember27intheafternoon.Theprofessor agedata,whichIhavebeendoingsofartoday, mightlookabitpuzzledandthengoontoexplain thestandarderrorforarealGDPforecastisin that,iftheexperimentconcernstheeffectof theneighborhoodof1percentagepoint.Thatis, pressureontheboilingpointofwater,thenthe whenwesaythattheforecastfor2001is2.1per- answerisverywell-known.But,Iinsist,Iwanta centrealgrowth,whatwemeanisthatthereis simpleanswer:Whatisgoingtohappen?Don’t aboutatwo-thirdsprobabilitythatrealgrowth confuseme,professor,withmorequestions. Myquestiontothechemistryprofessor,in willbe2.1percentplusorminus1.0percent. fact,iseitherfoolishormeaningless.Chemistry Thatmeans,ofcourse,thatthereisaone-third canprovidegoodpredictionswhentheexperi- probabilitythattheactualoutcomewillbeeither mentisknown.Withoutknowledgeofwhat higherorlowerthantherangefrom1.1to3.1 experimentistoberun,it’simpossibletoknow percentrealgrowth. whattheoutcomeisgoingtobe. Wealsoneedtoattachanerrorbandtothe Economicforecastersfacethesameproblem. inflationforecast.For2001,theBlueChipcon- Thedisciplineofeconomicsprovidessubstantial sensusfortheconsumerpriceindexisthatits insightsaboutwhatwillhappenifsomespecified annualaveragewillbe2.6percentabove2000, eventtakesplace.But,notknowingwhatevents witharangefromhightolowamongthepanel mighttakeplace—whateconomicexperiments offorecastersof1.8to3.5percent.Basedonmy mightberun—wehaveaveryweakbasisformak- readingoftheprofessionalliteratureonthismat- ingaforecast.Incurrentcircumstances,weare ter,Ithinkthatthereasonablestandarderrorto beingaskedtoofferaforecasteventhoughwedo attachtotheyear-aheadinflationforecastisprob- notknowwhatthenatureofpossibletaxlegisla- ablyabitsmallerthanfortheyear-aheadrealGDP tionwillbe,whateventsmightoccurabroad, forecast,butperhapsnotallthatmuchsmaller. whattheresolutionoftheCaliforniaelectricity situationwillbe,andsoforth.Clearly,weface Letmesummarizethisdiscussionbriefly thousandsofpossibleeventsthatcouldderail beforecontinuing.Economicforecastsaresubject evenaverysensibleforecast.So,onereasonfore- toaninherentrangeoferror.IntheUnitedStates, castsgoastrayisthatthingshappenthatthefore- myassessmentoftheevidenceisthateveryreal casterhadnotforeseenandinmostcasescould GDPforecastshouldhaveanerrorbandaround notforesee. itofaboutplusorminus1percentagepointfora Forecastsalsogoastraybecauseeconomists forecastmadeatthebeginningoftheyear.For havenotbeenabletopindownaccuratelyhow inflation,theerrorbandshouldbeonlyalittle theeconomywillrespondtoparticularevents. smaller.Evenwiththeselargeerrorbands,out- Ofcourse,thelikelyeffectsofsomeeventsare comesoutsidetheforecastrangewillbeobserved betterunderstoodthanothers.Thatis,evenafter fairlyoften.Failuretounderstandthelimitsof someeventsoccur,weareoftenuncertainabout economicforecastingcanonlyproduceproblems theirlikelyeffects.Incurrentcircumstances,for foranyoneactingonthebasisoftheforecasts. example,wedonothavesolidpredictionsofthe 3 MONETARYPOLICYANDINFLATION likelyeffectsofthedeclineinthestockmarket year.Forecastsareavaluableinputtoeconomic afterlastMarch,orofthemilddepreciationofthe policy,providedconsumersofforecastslikeme dollaragainsttheEuroinrecentmonths,orof understandtheirlimitations. theapparentdownturnintheJapaneseeconomy, WhereIthinkmyprofessionalexpertise orofnumerousotherthingsthathavehappened mattersisinreadingandassessingtheworkof orseemlikelytohavehappened.Knowledgejust economicforecasters.IknowwhyIthinksome isn’tallthatcomplete.Wecanpokefunatecon- forecasts—almostalwaystheonesatthe omistsorlaughattheirplight.Infact,mostofus extremes—are“offthewall.”Tounderstandwhy economistsdoboth.Still,attheendoftheday, forecastshavethecharacteristicstheydorequires wehavetotakeaccountsomehowofthefactthat extensiveknowledgeofbusinesscycledynamics knowledgeisincomplete.Afterall,wearetalk- andlong-runeconomicgrowthprocesses.This ingaboutadeadlyseriousbusinessbecausethe understandingisanimportantinputtomaking consequencesofpolicymistakes,asweallknow, monetarypolicydecisions. canbeveryunfortunate. Asyoumayknow,twiceeachyeareach Anotherreasonforecastsgoastrayisthat memberoftheFederalOpenMarketCommittee behaviordependsimportantlyonexpectations (FOMC),theFed’smainmonetarypolicybody, aboutthefuture.Thedecisionspeoplemakeabout submitsaneconomicforecastinpreparationfor changingjobs,aboutsavingandspending,and theFederalReserve’sMonetaryPolicyReportto thatfirmsmakeabouthiringworkersandinvest- theCongressandtheChairman’scongressional ingcapitaldependcriticallyontheirexpectations testimonyonthatReport.ChairmanGreenspan aboutthefuture,andinsomecasesthefairly submittedsuchareportandtestifiedjusttwo distantfuture.Economistsunderstandsomething daysago.Althoughtheforecastsofindividual abouthowexpectationsareformed,butthereare FOMCmembersarenotmadepublic,theMonetary hugegapsinwhatweknow.Isuspectthatfor PolicyReporthasatableshowingtherangeand theindefinitefutureeconomistswillbefaced centraltendencyofthoseforecasts.AttheSt.Louis withuncertaintiesaboutsuddenshiftsofpublic Fed,weprepareourforecastbystudyingwhat opinion,ornon-shiftswhenchangesinopinion professionalforecasters,bothinsideandoutside mightseemlogical. theFed,aresayingandexaminingwhetherwe I’vediscussedthesourcesofforecasterrors mighthavesomereasontodifferfromtheconsen- notbecauseIbelievethatwehaveanyrealistic susforecast.Bycomparingthecentraltendency chanceofwipingthemawaybutratherbecause oftheFOMCforecastswiththeBlueChip theyaresoobviouslyrelevantwhenwespenda Consensus,forexample,itisevidentthatthe fewminutesthinkingaboutthem.Understanding FOMCforecastsareinthemainstreamofprofes- whyforecastsgoastrayisanimportantpartof sionalforecastersingeneral.Nooneshouldbe understandingwhatvalueforecastshave. surprisedatthisoutcome;theprofessionalfore- castingfraternitybasesitsworkonanaccumulated bodyofforecastingtechniquesbroadlyshared IMPLICATIONS FOR MONETARY amongeconomistswhospecializeinthisarea. POLICY GivenwhatI’vesaidsofar,though,Ihope Whataretheimplicationsoflimitedforecast youagreethatI’dbecrazytomakemypolicy accuracyformonetarypolicy?Theplacetostartis positiondependonwhetheraforecastisafew withtheobservationthatprofessionalforecasters tenthsofapercenthigherorlower.NorshouldI doinfactforecastmoreaccuratelythansimple paygreatattentiontoargumentsamongforecasters alternativesonemightconsider.Forexample, aboutafewtenthsofapercentrealgrowthor forecastersonaveragebeatnaiveforecastssuch inflation.Giventherecordofforecasterrors,a asthatgrowthwillbeconstantorthesameaslast forecastofrealgrowthof2percentisbasically 4 EconomicForecastsandMonetaryPolicy indistinguishablefromoneof2percentor1per- ion.Marketinterestratesrespondsensitivelyto cent.Quitefrankly,itisludicroustopresenta currentconditionsanddomuchofthestabiliza- forecasttothetenthofpercentagepointwhen tionwork.TheFedsetsastablelong-runenvi- thediscussionisintendedforanon-professional ronmentwithinwhichthemarketscanmake audience.Doingsoprovidesacompletelymis- short-runadjustmentsinresponsetocurrent leadingpictureofwhataforecastmightmean. developments. I’veusedtenthsofpercentagepointtodayindis- Letmenowreturntomyopeningparagraph cussingtheBlueChipforecastsbecausethatis whereInotedtheintenseinterestpeopleshow howBlueChipreportstheforecasts.Butwhen- inmyviewsabouttheeconomicoutlook.People everIdiscusstheoutlook,Italkintermsofrough doseemtoexpectthatIwillbeabletoshedsome ranges,orforecastsroundedofftothenearesthalf lightonwheretheeconomymaybegoingand percentagepoint.Thus,todayIwouldnotspeak somelightonfutureFederalReservepolicy.Ihope of2.1percentrealgrowthbutofabout2percent IhaveconvincedyouthatIdonothaveanygems realgrowth.AndIwouldhopetogetacrossthe ofunconventionalwisdomtoofferonwherethe pointthatwhatthatforecastreallymeansisgrowth mostlikelyinthe1to3percentrange.Iamnot economyisgoing.Iacceptthejudgmentofexpert sayingthatitmakesnodifferencewhethergrowth forecasters,justasIdoexpertlawyers.Iquestion is1percentor3percent,butonlythatatthistime andprobetobetterunderstandexpertadvice,but wemustbepreparedtodealwiththisrangeof donottrytoconstructfromscratcheithermyown uncertainty. economicforecastormyownlegalanalysis. Evenifnear-termforecastswerenotsubject NorcanIprovideanyhelpinforecastingthe toconsiderableuncertainty,itisimportantto Fed’snextpolicyadjustment.Theissuehereis recognizethatmonetarypolicyadjustmentscan- thateconomicconditionsdosometimeschange nothavepreciseandquickeffectsontheeconomy. rapidly,andthejoboftheFOMCistocollateall TherehasbeenlotsoftalkaboutwhethertheFed thescrapsofinformationuptothetimeofa canengineera“softlanding”fortheeconomy.In meeting.Idonotevenmakefinalmyownposition earlyDecember,Iexperiencedatruesoftlanding; atanFOMCmeetinguntilIquiteliterallyget aftertouringtheBoeingF-18assemblyplantin there.Ofcourse,Iamaccumulatinginformation St.Louis,IhadashortsessionontheF-18simu- lator.TheinstructorhelpedmetolandanF-18on allalong,butIknowfromexperiencethatnew anaircraftcarrierdeck.Thatmaneuverisroutine informationandnewstaffevaluationofexisting intheNavy,butIcanassureyouthattheFederal informationcanandshouldaffectmythinking. Reservecannotlandtheeconomyonacarrier Forexample,IreadcarefullytheFedstaffeco- deckintheeconomicoceanweface.Instead,the nomicforecast,whichisavailableonlyacouple Fed’sjobistokeeptheeconomyonasoundlong- ofdaysbeforetheFOMCmeeting. runtrack;wecannotavoidmostshort-runeco- Nooneshouldhavegreatconfidenceinwhat nomicfluctuations,butwecanhelptoprevent actiontheFOMCwilltakefourweeks,forexam- thoseinevitablejoltsfrombecomingcumulative ple,aheadofameeting.Theworldisjusttoofull andpushingtheeconomyfaroffitslong-run ofsurprisesforthat.Aswecomeuptothedayof growthtrack. themeetingitself,themarketandtheFedwill Fedpolicyhasbeenextremelysuccessfulin havedigestedalloftheinformationavailable, recentyearsinmaintainingconfidencethatthe andinrecentyearshavemostoftencometoa rateofinflationwillremainlowandstableover aperiodofyears.Thisstablelong-runoutlookis commonjudgmentabouttheappropriatepolicy whatenablesthemarketstoworkthroughneces- action.Thisisasitshouldbegiventheuncer- saryshort-runadjustmentsinconstructivefash- taintiesoftheworldweface. 5 MONETARYPOLICYANDINFLATION THE ECONOMIC AND POLICY tionexpectationsoverthelongerrunindicate thathouseholdsandfirmscontinuetobelieve OUTLOOK TODAY thatinflationwillremainintherangeofrecent AsIalreadynoted,theconsensusoutlookis years. about2percentrealgrowthandabout2percent I’llsummarizebyemphasizingthreepoints. CPIinflationin2001,measuredbytheannual First,theoutlookI’veoutlinedreflectsthemiddle averagefor2001overtheannualaveragefor2000. oftherangeofbestprofessionaljudgment.That Lookingatquarterlydata,theBlueChipConsensus istherangewhereitisprudentformeasapolicy- forthefirstquarteris1percentrealgrowthand makertohangmyhat.Second,theoutlookis thenabout2percentforthesecondquarter. subjecttouncertaintyandwillmostlikelybe Thereafter,BlueChipseestheeconomygradu- revisedincomingmonthsasnewinformation allypickingupspeedtoagrowthrateofabout3 arrives.Third,becausenewinformationwillin percentinthefourthquarter.Growthisexpected alllikelihoodchangetheoutlookinsomerespects, tocontinueataboutthatrateinto2002. monetarypolicycannotbelockedintoagiven Thisoutlookshouldberegardedasvery pathfortheprincipalpolicyinstrument—the heartening.Forecastersbelievethattheeconomy’s targetfederalfundsratesetbytheFOMC.Every- slowerfirsthalfwillbefollowedbygrowththat oneshouldunderstandthatthepolicyforecasts ismoreorlessconsistentwiththeeconomy’s forupcomingFOMCmeetingscurrentlybuiltinto estimatedlong-termgrowthpotential.Theesti- themarketmaywellneedtoberevised,either mateofgrowthpotentialreflectstheunderlying upordown.Nopolicyactionsareever“baked demographicsofthelaborforce,whichpointto intothecake”longinadvanceofFOMCmeetings. growthinlaborinputofabout1percentperyear Idonotwanttofinishbyleavingtheimpres- andestimatedtrendproductivitygrowthinthe sionthatFOMCdecisionsaredrivenbyshort-run ballparkof2.5percent.Theproductivitygrowth data.TheFedhasacommitmenttolowandstable figuremustberegardedwithconsiderableuncer- inflationasitsmostfundamentalpolicygoalover tainty,asproductivityprocessesarenotvery thelongrun.Short-runpolicyadjustmentsmust wellunderstood.WhatIthinkwecansayabout alwaysbeconsistentwiththatfundamentalgoal. productivitygrowthisthateverypassingyear Thispointisespeciallyimportantinthecurrent bringsadditionalconvincingevidencethatU.S. environment,whereunderstandableconcerns productivitygrowthisnowmeasurablyhigher aboutthecurrentstateoftheeconomymaytend thanitwasinthe1970sand1980s.Theissueis toobscureconsiderationoflong-runpolicyobjec- howmuchhigherand,moreimportantly,how tives.Ourjobistofilterouttheshort-runnoise monetarypolicyshoulddealwiththerangeof inthedata,aboutwhichwecandonothingany- uncertaintyoverproductivitygrowth. way,andmaintainastablepolicyenvironment Mostimportantly,forme,isthatthroughall focusedonkeepinginflationlowinthelongrun. thesecurrenteconomicadjustmentsanduncer- Thesearetheconditionsinwhichindividuals, taintiestheinflationrateremainswellcontrolled. firms,andmarketscanmakenecessaryshort-run Inenergymarkets,futurespricesindicatethat adjustmentsinanefficientmanner. thebestguessisthatenergypriceswillbetrend- IhopeI’vegivenyouaperspectiveonhow ingloweroverthenextseveralyears.Current economicforecastsfitintomyconceptionofthe informationcontinuestosuggestthatfirmsbelieve Fed’spolicyprocess. theyhaveverylittlepricingpower.Dataoninfla- 6

Cite this document

APA

William Poole (2001, February 14). Speech. Speeches, Federal Reserve. https://whenthefedspeaks.com/doc/speech_20010215_poole

BibTeX

@misc{wtfs_speech_20010215_poole,
  author = {William Poole},
  title = {Speech},
  year = {2001},
  month = {Feb},
  howpublished = {Speeches, Federal Reserve},
  url = {https://whenthefedspeaks.com/doc/speech_20010215_poole},
  note = {Retrieved via When the Fed Speaks corpus}
}