speeches · August 20, 2003

Speech

William Poole · President

Fed Transparency: How, Not Whether GlobalInterdependenceCenter FederalReserveBankofPhiladelphia Philadelphia,Pennsylvania August21,2003 PublishedintheFederalReserveBankofSt.LouisReview,November/December2003,85(6),pp.1-8 Central bank transparency is a topic enoughtomorethanfullyexhaustthetimeavail- discussed almost as much as policy abletoday. actions themselves. Market partici- Beforeproceeding,Iwanttoemphasizethat pantshavealwayswantedtoknow theviewsIexpressherearemineanddonot the implicationsofpolicyactionsforthelikely necessarilyreflectofficialpositionsoftheFederal future course of monetary policy, but the long- ReserveSystem.Ithankmycolleaguesatthe standing practice of central bank secrecy has FederalReserveBankofSt.Louis—especially frustrated their search. In recent years, mone- RobertRasche,seniorvicepresidentanddirector tary policymakers have disclosed much more ofresearch,andDanielThornton,vicepresident than they did in the past, partly because of andeconomicadvisor—fortheirassistanceand growing interest in being more accountable and comments,butIretainfullresponsibilityfor partly because of recognition that policy actions errors. will be more effective if the market understands Myplanistoproceedbyfirstoutliningmy them better. modelofhowtheeconomyworks.Thatviewis, Discussionoftransparencyhasgonewell Ibelieve,theessentialstartingplaceforadiscus- beyondthefinancialpages.Thepastdecadehas sionoftransparency.Iwillthendiscusstwocases seennumerousprofessionalpapersontrans- inwhich,dependingonwhatviewyouhave, parencyissues.Inthisliterature,transparencyis marketparticipantsdidnotinterpretFedstate- takentomeanpublicdisclosure,andmuchofthe mentscorrectlyortheFeddidnotcommunicate discussionhascenteredonquestionssuchas: clearly.Undereitherinterpretation,therewas Howspecificshouldcentralbanksbeabouttheir somemiscommunication. policyobjectives?Shouldtheyannouncethe Iwilluse“transparency”asshorthandfor weightstheyapplytotheirinflationandoutput accuratelyconveyingaccurateinformationinclud- stabilizationobjectivesinconductingmonetary ingalltheinformationmarketparticipantsneed policy?Shouldcentralbanksdisclosetheireco- toformopinionsonmonetarypolicythatareas nomicforecasts?Shouldtranscriptsofthepolicy completeaspossible. debatebepublishedand,ifso,howsoon?Should policymakingmeetingsbetelevised? Myintenttodayisnottoreviewtheentire FUNDAMENTALS OF rangeoftransparencydebatesbutinsteadtocon- MACROECONOMIC EQUILIBRIUM centrateonissuesrelatingtotheeffectsofmone- tarypolicyinformationonmarketsandonthe Analysisofpolicycommunicationlogically effectivenessofmonetarypolicy.Icertainlydo beginswithadescriptionoftheeconomicinter- notbelievethatpoliticalaccountabilityissues actionbetweenthecentralbankandthemarkets. areunimportant,butmychosentopicislarge I’veprovidedmyviewofthisinteractiononsev- 1 MONETARYPOLICYANDINFLATION eraloccasions;hereIprovidejustenoughofa employmentstability,itisalsousefultopoint sketchofthisviewtoenablemetodiscusscom- explicitlytotheobjectiveoffinancialstability. municationissues.1 Stabilizingpolicyresponsestoseveremarket Atahighlyabstractlevel,Ibelievethatthe disruptionssuchasastockmarketcrashoraliq- appropriatemodeloftheeconomyisthatmarkets uiditycrisisfurthercontributetofosteringmaxi- behaveinanefficient,fullyinformedway.Equi- mumpossibleeconomicgrowth. libriumrequiresthatmarketparticipantsform TheFOMCimplementspolicybysettingthe accurateexpectationsaboutthebehaviorofthe intendedfederalfundsrate.Asiswellknown,a centralbank.Theeconomywillfunctionmost centralbankcannotachieveastableoutcomefor efficientlyifcentralbankpolicyhastwofeatures. theeconomyifitpegstheinterestrateataninap- First,thecentralbankmusthaveclearlyunder- propriatelevelforanylengthoftime.Thus,the stood,appropriate,andfeasibleobjectives. centralbankmustchangeitsinterestratetarget Second,thecentralbankmusthaveahighlyreg- fromtimetotimetoachieveitsobjectives. ularandpredictablepolicyruleorresponsepat- Inmyabstractmodeloftheeconomy,the ternthatlinkspolicyactionstothestateofthe marketandthecentralbankhavethesameinfor- economy,includingallinformationrelevantto mationbase;neitherhasaninformationaladvan- assessingtheeconomy’sprobablefuturecourse. tage.Asnewinformationarrives,theappropriate Pushingtheideaofafullrationalexpectations interestratetoachievepolicyobjectivesmay equilibriumonestepfurther,thereshouldbea change.Givenmyassumptionthatthemarket politicalequilibriuminwhichthecentralbank andthecentralbankhavethesameinformation, pursuesobjectivesbroadlyacceptedinsociety. allplayersrespondthesamewaytothearrival Withoutbroadpoliticalsupport,monetarypolicy ofnewinformation.Thecentralbankdetermines objectivesaresubjecttochangethroughnormal theappropriatepolicyresponseknowingthat democraticprocessesandsuchchange,orthe themarketalsohasthesameinformationand prospectofit,addstouncertaintyaboutfuture understandsitsimplicationsfortheeconomy monetarypolicy. andforpolicyactions. Withregardtoobjectives,theFederalOpen Atahighlyabstractlevel,Ibelievethatthis MarketCommittee(FOMC)hasstatedrepeatedly modelaccuratelydescribesthewaytheU.S.econ- thatoneofitsobjectivesisalowandstablerate omyhasbeenworkinginrecentyears.Aswe ofinflation.AlthoughtheFOMChasnotquanti- addmoreandmoredetailtothemodel,wefind fiedthattarget,forpresentpurposesitisuseful areasinwhichtheequilibriumisnotcomplete. todiscusscommunicationsissuesasiftheFOMC Thus,myviewisthattheeconomyhasbeencon- hadannouncedaspecifictarget.Putanotherway, vergingtowardafullrationalexpectationsmacro- withregardtomarketbehaviorIbelievethatthe economicequilibrium,butisnotallthewaythere differencebetweenanexplicittargetandone asyet.Inparticular,overthepastquartercentury inferredfromFOMCdecisionsisminimaltoday therehasbeenenormousprogressinimproving andhasbeenforsomeyears. theclarityoftheFed’sobjectivesandintheFed’s TheFOMCalsohastheobjectiveofmaximum disciplineinpursuingtheobjectives.Withregard possiblestabilityofoutputandemployment. totheinflationobjective,thereisaworldofdiffer- Takentogether,lowinflationandoutputstability encebetweentoday’ssituationandthatprevailing alongtheeconomy’sgrowthpatharebelievedto inthe1970s. contributetomaximumpossibleeconomicgrowth Therehasalsobeenenormousprogressin overtime.Becauseofitsimportancetooutputand provisionofmoreaccurateandtimelyinforma- 1 WilliamPoole,“Synching,NotSinkingtheMarkets,”presentedbeforethePhiladelphiaCouncilforBusinessEconomics,FederalReserve BankofPhiladelphia,6August1999;www.stlouisfed.org/news/speeches/1999_08_06_99.html>.WilliamPooleandRobertH.Rasche, “PerfectingtheMarket’sKnowledgeofMonetaryPolicy,”JournalofFinancialServicesResearch,December2000,18(2/3),pp.255-98. 2 FedTransparency:How,NotWhether tionaboutpolicyactions.TheFOMCannounces ofeachFOMCmeeting.Beforethedecisionsof itspolicyactionsontheafternoonoftheconclu- themeetingaremadepublic,Iestimate—“guess” sionofeachregularlyscheduledmeetingand isamuchbetterword—themarketreactionto promptlyafteranyinterimmeeting.TheFedis thepolicyactionandpressreleasethataremade moreopeninmanyotherwaysaswell;forexam- publicat2:15p.m.afterthemeeting.ThenIlisten ple,theFOMCnowreleasesapolicystatement totheradiooracablenewschanneltodetermine attheconclusionofitsmeetinganddissents,if howthebondandstockmarketsrespond.Ordi- any,arealsodisclosedatthattime. narily,butnotalways,Igetthedirectionofthe MyfundamentalconceptionoftheFed’scom- marketresponsescorrect,butmyestimatesof municationchallengeistofurthertheprogress themagnitudesofthemarketreactionsareoften towardamorecompleterationalexpectations wideofthemark.Mypersonalexperienceisthat equilibrium.Putanotherway,myquestionisthis: Ifinditexceedinglydifficulttopredicthowpeo- HowmighttheFedmodifyitscommunications plewillinterpretpolicyactionsandthenuances strategysothatthemarketcanconvergeona ofthepressrelease.IsuspectthatotherFOMC rationalexpectationsequilibriumwithlesserror participantsperformsimilarexercises,thoughI thanweobservetoday? havenotaskedanyofthem. I’vesometimesthoughtIshouldkeepaformal recordofmymarketpredictions,buthavenotyet MISCOMMUNICATION— decidedtodoso.Itcouldbeasoberingexercise TWO CASES forallFOMCmemberstomaintainsucharecord. Communicationisobviouslyimperfectifthe Itisinstructivetoconsiderexamplesinwhich speaker—theFOMCinthiscase—cannotpredict communicationswerelessclearthantheymight accuratelyhowlistenerswillrespond. havebeenandtoanalyzehowsuchproblems NowconsidertwospecificexamplesofFOMC mightbeavoidedinthefuture.Communications communicationsthatIbelieveweremisread.The successesarealsoworthstudying.Thereisagrow- firstistheevolutionintheannouncementofthe ingliteraturealongtheselines,suchasanalysis “tilt”inthedirective,andthesecondisthecom- ofthemarketeffectsofthechangeinFOMCprac- municationlastMayabout“anunwelcomesub- ticeinFebruary1994toimmediatedisclosureof stantialfallininflation.”IemphasizethatI’m policydecisionsattheconclusionofFOMC meetings. offeringmypersonalinterpretationofthese Accuratecommunicationisfarmoredifficult cases;otherFOMCmembersmayhavedifferent thanitseemsatfirstglance.Completeaccuracy interpretations. requiresthatspeakerandlistenerinterpretactions Intheearly1980stheFOMCbegantovote andwordsthesameway.Inanormalconversa- onlanguagepertainingtopossiblefuturepolicy tion,individualshaveanopportunitytoclearup actions.Thislanguagewasalternativelycalled ambiguitybyraisingquestionsaboutintended the“tilt,”“bias,”or“symmetry”ofthepolicy meaning.Itispossibletoaskforclarification,or directive.Thelanguagewasgenerallyregarded askagain,beforeacting.Whatcentralbankoffi- asapplyingtopossiblepolicyactionthroughthe cials(and,ofcourse,otherofficialsaswell)say, periodendingatthenextFOMCmeeting.Histori- however,canhaveimmediatemarketimpact; cally,theFOMCdidnotreleasethislanguage marketparticipantsmayactbeforeambiguities untiltheminutesofthatmeetingwerepublished ormiscommunicationcanbecorrected.Thisfact (subsequenttothenextregularlyscheduled imposesspecialburdensoncentralbankers. FOMCmeeting).Thatmeantthatthestatement, Toillustratehowdifficultthecommunica- whenreleased,hadnoinformationvalueabout tionsprocessisinthecentralbankcontext,letme theprobabledirectionofpolicyactionsbecause relatetoyouanexerciseIgothroughattheend thestatementreferredtoaperiodalreadypast. 3 MONETARYPOLICYANDINFLATION Inanefforttobemoretransparent,theCom- Inanattempttoclarifyitscommunications, mitteedecidedinDecember1998thatitwould theFOMCestablishedasubcommitteetoreview releasethetiltlanguageimmediatelywithits bothitspolicydirectiveandthepublicannounce- policyactionattheconclusionofameetingwhen mentfollowingFOMCmeetings.Communications itexpectedtheinformationtobeparticularly practicechangedintworespects.First,theFOMC important.Theminutesofthatmeeting,released wouldissueastatementaftereverymeeting.That inlateJanuary1999,containaparagraphonthe stepeliminatedthepossibilitythatthemere Committee’sdiscussionofadisclosurepolicy.A existenceofastatementwouldbetreatedasan keypassagefromtheminutesreadsasfollows: unusualeventsignalingamajorchangeinpolicy. Nonetheless,themembersdecidedtoimple- Thesecondstepwasanew“balance-of-risks” mentthepreviouslystatedpolicyofreleasing, statementthatassessedtheoutlookforpricesta- onaninfrequentbasis,anannouncement bilityandsustainableeconomicgrowthinthe immediatelyaftercertainFOMCmeetings foreseeablefuture.DespitetheFOMC’sstated whenthestanceofmonetarypolicyremained intentionthatitsnew“balance-of-risks”wasnot unchanged.Specifically,theCommitteewould tobeinterpretedasanindicatoroffutureFOMC dosoonthoseoccasionswhenitwantedto actions,theevidencesuggeststhatitwasoneof communicatetothepublicamajorshiftinits thepiecesofinformationthatmarketanalystsdid viewsaboutthebalanceofrisksorthelikely usetoformexpectationsofalikelynear-term directionoffuturepolicy.Suchannouncements policyaction.2Myperception,however,isthat wouldnotbemadeaftereverychangeinthe symmetryofthedirective,butonlywhenit thebalance-of-riskslanguagedidnotcometo seemedimportantforthepublictobeaware haveacompletelysettledmeaninginthemarket. ofanimportantshiftinthemembers’views. Formysecondexample,considerthestate- mentfollowingtheFOMCmeetinglastMaythat AttheconclusionofthemeetinginMay1999, referredto“anunwelcomesubstantialfallin theFOMCforthefirsttimereleasedastatement inflation.”Insubsequentcommentaryinthefinan- thatincludedthe“tilt”inthepolicydirective. cialpress,thisstatementwasinterpretedtomean Theformalstatementreferredto“thefederalfunds oneormoreofthefollowingthings:(i)Acutin operatingobjectiveduringtheintermeeting theintendedfundsrateattheJune2003meeting period.”ManymembersoftheFOMCbelieved waslikely.(ii)Anyincreaseintheintendedfunds thatthemarketoverreactedtotheMaytiltstate- mentandtosubsequenttiltstatementsaswell. ratewithinthenextyearwashighlyunlikely.(iii) Thestatementsdidattractconsiderableattention, TheFOMCwouldimplement“unconventional andmarketanalystsbegantospeculateabout monetarypolicyactions”suchasaggressively changesintheintendedfundsrateatfuture purchasinglong-termTreasurybonds.Interpre- FOMCmeetingsbasedonthetilt,orsymmetry, tation(iii)gainedforceandamajorrallyinlong- announcedbytheFOMC. termTreasurymarketsensued,drivingthe10-year Themarketreactiontothestatementreleased Treasuryratetoamore-than-40-yearlowof3.13 immediatelyaftertheMay1999FOMCmeeting percent. shouldnot,perhaps,havebeenasurprisetothe Speakingstrictlyformyself,Ibelievethere Committee.TheCommitteehadsaid,afterall,in aretwoimportantpointsthatthestatementof itsminutesoftheDecember1998meetingthatit May6triedtocommunicatethatdidn’treally wouldmakesuchanannouncement“whenit comeacross.First,inflationhasnowrecededto wantedtocommunicatetothepublicamajor alevelwhereforthefirsttimein40yearsinflation shiftinitsviews...” risksaresymmetric:Fromthecurrentinflation 2 RobertH.RascheandDanielL.Thornton,“TheFOMC’sBalance-of-RisksStatementandMarketExpectationsofPolicyActions,”Federal ReserveBankofSt.LouisReview,September/October2002,84(5),pp.37-50. 4 FedTransparency:How,NotWhether rateneithersustainedincreasesnorsustained DISCLOSURE STRATEGY decreasesaredesirable.Second,inthewordsof Givenmyemphasisontheeconomicpurpose myFOMCcolleagueGovernorBernanke,“FOMC ofdisclosure,Iseenoroomformerelysatisfying behaviorandrhetorichavesuggestedtomany curiosityaboutwhatgoesoninFOMCmeetings. observersthattheCommitteedoeshavean Thegeneralnatureofwhatgoesoninmeetings implicitpreferredrangeforinflation.Mostrele- caneasilybeinferredbyreadingmeetingtran- vanthere,thebottomofthatpreferredrange scripts,whicharereleasedwithafive-yearlag. clearlyseemstobeavaluegreaterthanzeromeas- Theappropriatecommunicationsgoal,inthecon- uredinflation,atleast1percentorso.”3Onsev- textofhowtheeconomyfunctions,shouldbeto eraloccasionsinthepastIhavestatedthatmy minimizemarketuncertaintyaboutmonetary preferredinflationtargetiszeroinflation,properly policy.Itisimportanttoemphasizethatuncer- measured.SinceIbelievethatthemajorprice taintyaboutfuturemonetarypolicyactionscan- indicesemployedtodayaresubjecttosome notbeeliminatedbecausethoseactionsdepend upwardbiasandmeasurementerror,thegoal criticallyoninformationthatcannotitselfbepre- “zeroinflation,properlymeasured”translates dicted.Whatneedstobeminimized,therefore, intoalow,positivemeasuredrateofinflation.In isuncertaintyaboutcentralbankresponsesto myjudgment,1percentmeasuredinflationfor newinformation. theconsumptionpriceindexisintheneighbor- I’mgoingtoconcentratemydiscussionon hoodofpricestabilityasIdefineit. thepolicystatementissuedattheconclusionof Tome,though,anannouncementthatinflation eachFOMCmeeting,butsomeofmycomments isnowdowntoanappropriatelong-runtarget havebroaderapplicability.Thecommunication shouldnotbyitselfhaveledtoasharpdecline attheconclusionofeachFOMCmeetingisacriti- calonebecausemarketparticipantsareprimed inthe10-yearbondrate.WhatIthinkhappened toreacttonewsofapolicyactionanditsration- wasthatthemarket,seeingthattheintended ale.Thestatementisnecessarilyshortanditsets federalfundsratewasdownto1percent,thought thestageforFOMCmemberstoprovidesubse- thattheFedwasrunningoutofroomtoimple- quent,morethoroughdiscussionsofpolicy.I’ll mentpolicythroughsettingatargetfederalfunds concentrateontwoaspectsofthestatement.The rate.IftheFedweretoswitchtosettingatarget firstistheextenttowhichthestatementshould forlong-terminterestrates,thensuchapolicy provideaforecastinsomeformoffuturepolicy wouldreduceoreliminateforatimedownside actions,andthesecondisthestructureofthe priceriskonlong-termTreasurybonds.That statementitself. wouldjustifybiddingthe10-yearbondpriceup Givenmyrationalexpectationsmacroeco- (theratedown),becausethepriceriskwould nomicmodelandmydesiretocreateamore becomeone-sided—bondpricescouldgoupbut completeequilibrium—anequilibriuminwhich notdown,oratleastnotdownbyverymuch expectationserrorsareminimized—thecentral verysoon.Overtime,however,themarketcame communicationsissueistoexplaintothemarket tobelievethattheFOMCwasnotcontemplating thenatureofthepolicyrulethatdetermines theneedforanunconventionalpolicyinthenear hownewinformationfeedsintopolicyactions term,andbondpricesfell.Indeed,bondyields designedtoachieveascloselyaspossiblethe backeduptoalevelabovewheretheyhadbeen centralbank’spolicyobjectives.Achievingclar- justbeforetheMayFOMCmeeting. itywithrespecttopolicyobjectivesisactually 3 BenS.Bernanke,“AnUnwelcomeFallinInflation?”presentedbeforetheEconomicsRoundtable,UniversityofCalifornia,SanDiego,23 July2003,inLaJolla,California;<http://www.federalreserve.gov/boarddocs/speeches/2003/20030723/default.htm>. 5 MONETARYPOLICYANDINFLATION quitesimplecomparedwithexplainingthenature actionsarealmostentirelycontingentonthe ofthepolicyrule. arrivalofnewinformation.Forthatreason,I Thefundamentalproblemisthatthereisno believethatanFOMCforecast,ortilt,towarda policyrulebywhichwecancalculatetheappro- specificfuturepolicyactionismorelikelytobe priatepolicyactionfromobserveddata.Thereis misleadingtothemarketthanhelpful.Itistrue insteadaregularitytopolicyofthesort“youknow thatattheconclusionofameetingIhaveasense itwhenyouseeit.” oftheprobabilitiesofvariousfuturepolicy Sometimesweobserveastrikingchangein actions,andIsuspectthatotherFOMCmembers someparticularvariable,suchastheunemploy- thinkaboutthepolicyprocessthesameway.I mentrate,thatallbutdemandsapolicyresponse. mightbelieve,forexample,thatnewinformation Mostofthetime,though,policyactionsflowfrom wouldbeveryunlikelytoleadmetowanttoraise anaccumulationofdata,mostofwhichpointin theintendedfundsrateatthenextmeetingbut thesamedirection.Itjustisnoteasytodescribe might,incombinationwithinformationalready “youknowitwhenyouseeit.”Iwouldbeabso- known,makethecaseforcuttingtheintended lutelydelightedifresearcherscouldprovidea rate.AndImightassignaprobabilitytoafuture quantifiedpolicyrule,atleastasabasecase. cutof0.5,or0.3,orsomeothervalue.Buteven TherulesuggestedbyJohnTaylorishelpful,but inthissituationIwouldnotwanttoruleoutan veryincomplete.Ithinkitunfortunatethatwe increaseintheintendedrate,forIcancertainly havenotseenintheprofessionalliteraturean imaginenewinformationthatwouldcompelan evolutionofapolicyrulethatbuildssubstantially increase. ontheworkbegunbyTaylor.Buttheproblemis Question:CouldtheFOMCasapractical averydifficultone;foronething,itisnecessary matterdecideontheprobabilityandconveythat fromtimetotimetodiscountchangesinanimpor- probabilityaccuratelytothemarket?Myownview tanteconomicvariablebecauseofsuspected isthatonlyrarelycouldtheFOMCagreeonwhat anomaliesinthestatisticsthemselves. theprobabilityshouldbe,andeventhenitwould Thus,wehavetolivewiththeunfortunate beextremelydifficulttoconveythatprobability factthatthemonetarypolicyworldisoneof“I’ll tothemarket.Moreover,iftheprobabilityishigh, knowitwhenIseeit.”Weneedtokeepthatfact itseemstomethatinmostcasesitwouldmake inmindwhendesigningcommunicationspolicy. moresensetosimplytakethepolicyactionat Explainingapolicyaction—elucidatingthe thecurrentmeetingratherthanbroadcastitas considerationsthatledtheFOMCtodecideto likelyatthenextmeeting. adjusttheintendedfundsrate,ortoleaveit Theold“tilt”languagecausedproblems,I unchanged—isworthwhile.Overtime,theaccu- think,preciselybecausedifferentFOMCmembers mulationofsuchexplanationshelpsthemarket, haddifferentinterpretationsofwhatprobabilities andperhapstheFOMCitself,tounderstandwhat attachedtowhatwords.AndIthinkthemarket thepolicyregularitiesare.Itisalsoimportantto understandthatmany—perhapsmost—policy viewwas,atleastsometimes,thatiftheFOMC actionshaveprecedentvalue.IftheFOMCtakes chosetochangethebias,itmustbedoingsoto actionAincircumstancesX,thenexttimecir- announceasignificantprobabilityoffuturepol- cumstancesXarisetheFOMCshouldalsotake icyaction.Ithinksomeobserversalsotendto actionA,orhavegoodreasonnottodoso.One reactasfollows:Iftheprobabilityishigh,why oftheadvantagesofpublicdisclosureoftherea- shouldn’ttheFOMCactnow?Iftheprobability sonsforpolicyactionsisthattherequiredexpla- islow,whytalkaboutit?Iftheprobabilityisina nationforcestheFOMCtothinkthroughwhatit middlerange,willdisclosingthetilthelpthe isdoingandwhy. markettopricesecuritiesmoreefficiently—that Discussingfuturepolicyactionsisadifferent is,moreinlinewiththetruelikelihoodoffuture matter.Inmyviewoftheworld,futurepolicy policyaction? 6 FedTransparency:How,NotWhether Furthermore,thetiltlanguagewassometimes thatnewinformationwillarrivepointingtoward usedinanefforttoreducethenumberofdissents futurepolicyactions.Understandingpolicyobjec- intheFOMC.Inthiscase,thelanguagemayhave tivesandthepolicyrule,themarketcanputitself providedinaccurateinformation,becausethe insidetheFed’sheadandmakethesameguesses majoritymaynothavebelievedthattherewas theFedcanmake. anysignificantprobabilityoffuturepolicyaction IfinformationontheFed’sthinkingaboutits inthedirectionindicated. futurepolicydirectionisnotredundant,then Anotherproblemisthatofactingconsistently thatfactalonedoesnotnecessarilycallforthe withguidanceabouttheprobabledirectionof Fedtoforecastfuturepolicyactions.Theissue futurepolicy.Sometimesnewinformationarrives formeisquitedifferent.Ifthemarketdoesn’tsee thatisclearlycompellinginthedirectionofnot whatIsee,whynot?Whatisthemarketmissing, actinginaccordancewiththeguidance.Afore- andwhatdowemakeofthefact?Perhapsthe castofapolicyaction,madebeforethenewinfor- bettercoursewouldbetodisclosetheunderlying mationarose,maythenhavecreatedadilemma informationthemarketisapparentlymissing,or foracentralbank.Thecentralbanktheneither callattentiontoinformationthemarketisneglect- breakswhatthemarketregardsasacommitment ing.Thattomeisabetterstrategythanhintingat orlivesuptothecommitmentatthecostof anunconditionalpolicydirection,becausethe ignoringnewinformationcallingforadifferent essenceofwhatthemarketneedstoknowisnot policyaction.However,moreofteninformation theintendedfederalfundsrateinsixweeks.What willbeindecisive;onceguidanceisannounced, themarketneedstoknowisthepolicyresponse theburdenofprooftendstoshifttowardshowing functionbywhichthecentralbankactsinacon- whytheforecastedactionisinadvisable,whereas sistentwayovertimeandonethatisefficientin withoutguidancetheburdenofprooftends fosteringsuccessinachievingpolicyobjectives. towardjustifyinganaction. Thisdiscussionassumesthatthemarketis Allinall,then,I’vecometotheviewthat missingsomething.But,couldtheproblembe FOMClanguageforecastingfuturepolicyactions thatthemarketseessomethingIdonot?Howcan isprobablycounterproductiveinmostcircum- IbesosurethatIknowtheappropriatedirection stances.Idonot,however,ruleoutthedesirability forfuturepolicyactions?IfitistheFedthatis offorecastingfuturepolicyinsomecasesgiven missingsomethingandnotthemarket,thendis- thattherationalexpectationsmodelfromwhich closingapolicytiltwillclearlybemisleading,or Iamreasoningisclearlyanabstraction.WhatI theoddsarethatitwillturnouttobemisleading. thinkweneedtodoistoanalyzethecircum- Historically,theFOMC(andothercentral stancesunderwhichtheabstractmodelprovides banks)wenttogreatlengthstoavoidproviding misleadingguidancewithrespecttocommuni- guidanceaboutfuturepolicydirection.Indeed, cationsstrategy. oneoftheargumentsthattheFedusedinthe Itistruethatpolicyworksinpartbychanging defenseofsecrecyintheMerrillcasein1975was expectationsandthereforethetermstructureof thattheimmediatereleaseoftheinformationin interestrates;thatisthebasicargumentfavoring thedirectiveorinFOMCdeliberationswould disclosureoffuturepolicydirection.However, produceexpectationsthatwoulddestabilize thecruxofthematteristhis:Ifthemarketfully financialmarkets.Thatargumentisincomplete understandsthepolicyrule,orpolicyregularity, atbest.Somedisclosuresclearlystabilizerather andhasthesameinformationtheFOMChas, thandestabilizemarkets;secrecycancreateincor- thenanFOMCforecastoffuturepolicydirection rectmarketguessesaboutwhattheFedisdoing. isuselessinformationbecauseitisredundant.If OnesuchcasearoseonThanksgivingeve1989, themarketandtheFedhavethesameinformation, whentheOpenMarketDeskintervenedtosupply thenthemarketcandeterminetheprobabilities reservesfortechnicalreasons.Atthistimethere 7 MONETARYPOLICYANDINFLATION wasnoannouncementoftheintendedfundsrate. isnottruethatlotsofwordsequalslotsofdisclo- Theinterventionwaswidelyinterpretedbymar- sureandgreatertransparency. ketparticipantsasasignalthattheFOMChad Becausethemarketrespondsimmediatelyto reduceditstargetforthefederalfundsratefrom policyactionsandstatements,itisimportant around8.50toabout8.25percent.Ittookseveral thattheFOMCnotfinditselfinthepositionof tradingdaysbeforethemarketsortedoutthecon- havingtoclarifyitsstatementstocorrectmisin- fusion.Onthisoccasionsecrecyproducedunnec- terpretations;explainingtheexplanationcanadd essaryvolatilityinfinancialmarkets.Numerous touncertaintyandraisequestionsaboutfuture otherexamplesprovideconvincingevidence,in policystatements,whichmarketparticipants myview,that,ingeneral,disclosureofactual mightcometoexpecttobeclarifiedorinterpreted. policyactionsisstabilizingratherthandestabi- Thebestwaytoavoidtheseproblemsistonarrow lizing.Butitisnotappropriatetogeneralizefrom therangeofphrasesusedinthestatement. thevalueofimmediatedisclosureofpolicyactions AsIexplainedearlier,myviewisthatthe todisclosureof“everything.” statementshouldconcentrateonexplainingthe Todiscusstheformatofthepolicystatement policyactionanditsrationale,andnothintat attheconclusionofeachFOMCmeeting,I’llstart futurepolicyactions.Giveninformationavail- withanobservation.Supposethestatementis ableatthetimeofameeting,Ibelievethatthe confinedtoonepage.Witheventhosefewwords, standingassumptionshouldbethatthepolicy therichnessoflanguageandtheimportanceof actionatthemeetingisexpectedtopositionthe wordorderinconveyingmeaningyieldtheresult stanceofpolicyappropriately.Thepurposeofthe thatthestatementcontainsanenormousrangeof statementshouldbetoexplainwhythepolicy possibilities.Themultiplicityofpossiblemean- action,orlackofaction,haspositionedpolicy ingsismadeevenlargersinceeachstatementis appropriatelygiventheinformationavailable. readagainstthebackdropofthepreviousone. Asamatteroflogic,thecurrentbalance-of- Thus,whatisrelevantisnotjustwordchoiceand riskslanguagecreatessomeambiguity.Ifrisks orderbutchangesfromthepreviousstatement. areassessedasunbalanced,whywaspolicynot Asanaside,theimportanceofstatement adjustedfurthertocreateabalancegoingforward? changescanmakeitdifficulttoimprovethe Apossibleansweristhatanunbalancedrisk statementovertime.Toavoidmisinterpretation assessmentforeshadowsfuturepolicyaction. ofchanges,itisbestifachangedapproachor Butthe“tilt”interpretationofanunbalanced formatcanbeannouncedinadvancesothatthe riskassessmentseemsatoddswiththerationale changeinapproachisclearlyseparatedfroma forsubstitutingthebalance-of-riskslanguagefor changeinpolicy. theprevioustiltlanguage.Whatwouldbeclearer, Ifthestatementistoconveypolicyintent Ithink,wouldbetousethebalance-of-riskslan- accuratelyandwithminimalambiguity—surely guagetoexplainthatinformationsincethepre- desirablecharacteristicsintermsofminimizing viousmeetingindicatedthatriskswerebecoming expectationalerrorsinthemarket—thenthe unbalancedinaparticularway,andforthatrea- numberofpossiblemeaningsmustbenarrowed sontheFOMCadjustedtheintendedfederalfunds insomeway.OnewaywouldbefortheFOMCto rate. chooseamongarelativelyfewstandardphrases, Separatinggrowthrisksfrominflationrisks, atleastinlanguageprovidingasummarystate- asintheMaystatement,makesalotofsense. mentofthepolicystance. Whenemploymentchangeandinflationdataare Somewillregardthisapproachasproviding plottedinascatterdiagram,allfourquadrants “boilerplate”languagewithlittlerealmeaning. containlotsofobservations.Sometimesemploy- Myownjudgmentisthatitisbettertoprovide mentandinflationrisetogether,orfalltogether. boilerplatewithclearmeaningthanrichlanguage However,justaboutasoftenthetwovariables withamultiplicityofpossiblemeanings.Itjust moveinoppositedirections.Becauseallfour 8 FedTransparency:How,NotWhether quadrantsarepopulated,asummarypolicyjudg- I’vetriedtopresentaframeworkforthinking menthastobecommunicatedindicatingthe abouthowtoimprovetransparency.Istartwith FOMC’sweightingoftherisks.Itisrelatively aviewoftheworldbasedonastandardrational easytoexplainthatapolicytighteningwasocca- expectationsmacroeconomicmodel.Anefficient sionedbyarisingriskofhigherinflationand equilibriumrequiresthatthemarketsunderstand strongeremploymentgrowth;butwhenemploy- whatthecentralbankisdoing.Thecommunica- mentgrowthandinflationareheadedinopposite tionschallengeforthecentralbankistoexplain directions,thesummarypolicylanguageneedsto morethoroughlyandcompletelywhatitisdoing. indicatethattheFOMCacted,ordidn’t,because Thatmeans,aboveall,explaininghownewinfor- itgavemoreweighttotheinflationriskthanthe mationfeedsintopolicyactions.Ihavealotof employmentrisk,orviceversa.Theissueisnot, skepticismaboutforecastingfuturepolicyactions bytheway,thatinflationriskismoreorless becausetheyproperlyflowfromnewinformation importantthanemploymentrisk,butratherthat thatisnotitselfpredictable. currentinformationmightsuggestthatrecent Accuratecommunicationrequiressettled employmentchanges,say,aretransitory. meaningsforwords.Foranygivenword,wecan Thisdiscussionmakesclearthataminimally consultadictionaryandweusuallydiscoverthat accuratesummarystatementexplainingamone- eachEnglishwordhasseveralmeanings,which tarypolicyactionisstillprettycomplicated.The canbequitedifferent.Thereisnodictionaryin FOMCmustweighinflationrisks,employment whichwecanlookuptheseveralmeaningsofa risks,andformajudgmentbalancingorweight- paragraph.Themeaningofapolicystatement— ingthetworisks.Beyondthat,fromtimetotime preferablyonlyone—mustbeestablishedbythe specialfactorswillintrude,suchasthetragic centralbank,throughconsistentpracticeover eventsof9/11orunusualliquiditycrises. timeandthroughmoreextendeddiscussionof whatthelanguagemeans. Ithinkitfairtosaythatsystematicstudyof CONCLUDING OBSERVATIONS thehowoftransparencyisinitsinfancy,andI hopethatmyremarksheresparkotherstoanalyze Transparencyisaworthygoal,butimproving theseissues. transparencyishardwork.Mythinkingisstill evolving,butonethingIknowisthatthemoreI considertheissuetheharderitseems. 9

Cite this document

APA

William Poole (2003, August 20). Speech. Speeches, Federal Reserve. https://whenthefedspeaks.com/doc/speech_20030821_poole

BibTeX

@misc{wtfs_speech_20030821_poole,
  author = {William Poole},
  title = {Speech},
  year = {2003},
  month = {Aug},
  howpublished = {Speeches, Federal Reserve},
  url = {https://whenthefedspeaks.com/doc/speech_20030821_poole},
  note = {Retrieved via When the Fed Speaks corpus}
}