speeches · October 3, 2005

Speech

William Poole · President

How Predictable Is Fed Policy? UniversityofWashington Seattle,Washington October4,2005 PublishedintheFederalReserveBankofSt.LouisReview,November/December2005,87(6),pp.659-68 Day in and day out, all of us depend economicdecisionsdepend,directlyorindirectly, on a high degree of predictability in onthepredictabilityofmonetarypolicy.Monetary thenation’smonetaryarrangements. policydecisionscancreatesurprisesthataffect Consider, for example, the counter- outcomesfromhouseholddecisionsastowhat feiting problem. The problem is in fact small jobstotakeandwheretolive.Similarly,business enough that we rarely examine our paper money firmsfindthattheirdecisionsonhiringandinvest- closely. Through careful design of the currency, mentinphysicalcapitalmayturnoutwellor and careful monitoring and enforcement, the poorlydependingonthecourseofmonetary policyanditseffectsontheeconomy. Treasury and Federal Reserve together maintain Ifyoufollowthefinancialpressatall,or a highly reliable currency—its usefulness to us watchnationalTVnewsfollowingaFedpolicy is almost completely predictable. decision,youknowthatfinancialmarketshave Historically,theleastpredictableaspectof anintenseinterestinwhattheFederalReserve ourmonetarysystemhasbeenmonetarypolicy. doesordoesnotdo.Financialmarketbehavior MonetarypolicymistakesledtotheGreat providesanopportunitytostudythepredictabil- Depressionofthe1930sandtheGreatInflation ityofmonetarypolicywithsomeprecision.That ofthe1960sand1970s.Unpredictabilitycanhave ismytopictoday,organizedaroundmytitle highcosts.TheGreatInflationanditscorrection theme:HowPredictableIsFedPolicy? ledtothefailureofhundredsofsavingsandloan Beforeproceeding,Iwanttoemphasizethat associations(S&Ls).TheproblemwasthatS&Ls theviewsIexpressherearemineanddonot madelong-termmortgageloansatinterestrates necessarilyreflectofficialpositionsoftheFederal reflectingexpectationsofmodestinflation,but ReserveSystem.Ithankmycolleaguesatthe thoseexpectationsprovedtobetoolowagainand FederalReserveBankofSt.Louisfortheircom- againasinflationroseafter1965.Asinflationrose, ments.RobertRasche,seniorvicepresidentand interestratesonoutstandinglong-termfixed-rate directorofresearch,providedspecialassistance. mortgagesdidnot,creatingenormouslossesfor However,Iretainfullresponsibilityforerrors. S&Ls.Later,asinflationexpectationsbecame embedded,manyborrowedatinterestratesthat turnedouttobeunsustainablyhigh;purchasesof EVOLUTION OF FOMC farmland,forexample,financedathighinterest COMMUNICATION DURING ratesturnedouttobeextremelyunwiseasinfla- THE GREENSPAN ERA tionfellunexpectedlyintheearly1980s.Many farmersandagriculturalbanksthatlenttothem Since1989,theFOMChasadoptedmanyprac- failed. ticesthatimprovethetransparencyofitspolicy Inflationpredictabilityisthemostobvious, actions.Enhancedtransparencyisimportantfor andprobablymostimportant,consequenceof improvingtheFed’spoliticalaccountability,but predictablemonetarypolicy.Nevertheless,most fromaneconomicperspectivetransparencyis 1 MONETARYPOLICYANDINFLATION essentialifmarketsaretounderstandFedpolicy theFedoftendiscussedpolicyintermsof andthereforebemoresuccessfulinpredicting the“degreeofpressureonreserveposi- policyadjustments.Fedpolicyisimplemented tions”inthemoneymarket.Aclearfocuson throughdecisionsbytheFederalOpenMarket aquantitativetargetendedtheambiguity. Committee(FOMC)thatsetthetarget,or • May1999:Apressstatementfollowing “intended,”federalfundsrate.Thefundsrateis theconclusionofeveryFOMCmeeting theinterestrateintheinterbankmarketonover- includesthetargetfederalfundsrateand night(one-day)loans.Thefundsrateistheful- thepolicy“bias.”Thebiasindicatedthat theCommitteewasleaningtowardan crum,oranchor,forallotherinterestratesinthe increaseordecreaseinthefundsratetarget market. buthadnotyetdecidedtoactuallychange HerearesomemilestonesofchangesinFOMC thetarget. practicesthathaveenhancedtransparency: • December1999:Initspressstatement,the • August1989:Policychangesinthetarget FOMCreplacesthepolicybiaslanguage federalfundsratearelimitedtomultiples with“balanceofrisks”languageinaneffort of25basispoints.Priorpracticeofchanging tolengthenthehorizonofitsstatementand therateinotherincrementsoftencreated provideasummaryviewofitsoutlookfor marketuncertaintyastoexactlywhatthe theeconomy. FederalReserve’sintentionwas. • January2002:ThevoteontheDirective • February1994:StartingwiththisFOMC andthenamesofdissentingmembers,if meeting,theCommitteereleasedapress any,areincludedinthepressstatement. statementdescribingitspolicyactionat Previously,thisinformationwasnotavail- theconclusionofanymeetingatwhich abletothemarketuntilthemeetingminutes theCommitteechangedthetargetfunds werereleasedfollowingthesubsequent rate.Priortothispractice,themarkethad FOMCmeeting,sixtoeightweekslater. toinferfromFedopenmarketoperations • August2003:TheFOMCintroduces whether,andhow,theFed’spolicystance “forward-looking”languageintoitspost- hadchanged.Consequently,themarketwas meetingpressstatement.1Thislanguage oftenuncertainastothecurrentsettingof suggestedtheprobabledirectionofthe Fedpolicy. targetfederalfundsrateoverthenextone ormoremeetings. • August1997:Publicacknowledgmentthat policyisformulatedintermsofatargetfor • January2005:ReleaseofminutesofFOMC meetingadvancedtothreeweeksafterthe thefundsrate.Themarkethadcometo meeting(andbeforethenextscheduled believethatthefederalfundsratewasthe FOMCmeeting). policytargetbutalluncertaintyaboutthis issuedisappearedafterthistime. Thepurposeofthesechanges,whichhave • August1997:Aquantitativetargetfederal gonealongwaytowardliftingthetraditionalveil fundsrateisincludedintheDirectiveto ofsecrecyovermonetarypolicy,istoincrease theSystemOpenMarketAccountManager transparencyofpolicy,improveaccountability, attheTradingDeskoftheFederalReserve andprovidebetterinformationtomarketpartici- BankofNewYork(the“Desk”).Previously, pantsaboutthefuturedirectionofpolicy.In 1 “Inthesecircumstances,theCommitteebelievesthatpolicyaccommodationcanbemaintainedforaconsiderableperiod.”FederalReserve PressRelease,August12,2003(www.federalreserve.gov/boarddocs/press/monetary/2003/20030812/).InthepressreleaseofJanuary28,2004, thelanguagewasmodified:“…theCommitteebelievesthatitcanbepatientinremovingitspolicyaccommodation”(www.federalreserve.gov/ boarddocs/press/monetary/2004/20040128/default.htm).Subsequently,inthepressreleaseofMay4,2004,asecondmodificationofthe languagewasintroduced:“theCommitteebelievesthatpolicyaccommodationcanberemovedatapacethatislikelytobemeasured” (www.federalreserve.gov/boarddocs/press/monetary/2004/20040504/). 2 HowPredictableIsFedPolicy? severalearlierspeechesandpaperswrittenjointly arefuturescontractsinfederalfunds.Thematur- withmembersoftheSt.LouisFedResearch ingfuturescontractissettledattheendofthe Division,IexaminedhowchangesinFedtrans- monthbasedontheaverageeffectivefederalfunds parencyhaveaffectedmarketbehavior,especially rateduringthemonth.Thus,thefederalfunds after1994.Today,Iwillreviewsomeofthosefind- futuresmarketisadirectbetontheFOMC’starget ingsandaddnewfindingsonmarketbehavior federalfundsrateinthefuture.Thenumberof overthepasttwoyears,whenthemostrecent contractstradedhaschangedovertime,ashas innovationswereintroduced.2 theleveloftradingactivityinthemarket. Overthecourseofamonth,databecome knownandthemarkettradingisbasedonthat ARE MARKETS IN SYNCH WITH informationplusexpectationsastothefederal THE FOMC? fundsrateduringtheremainingdaysofthemonth. Thereisalsotradinginthe1-month-aheadcon- OnanumberofoccasionsIhavestressed tract—forexample,inOctober2005,tradingina myviewoftheimportanceofmarketsbeing“in November2005contract.TheChicagoBoardof synch”withtheFOMC.Accumulatingevidence Tradelistsadditionalcontractsasfarasayearor hasshownthat,judgingbythereactionofthe moreout,buttradingvolumeistrivialmuch federalfundsfuturesmarket,marketparticipants beyondthe5-month-aheadcontract.Thedistant havebeenincreasinglyaccurateinpredicting contracts,becausetheytendtohavethinvolume, FOMCpolicyactionsasstepstowardmoretrans- arenotnecessarilyreliablemeasuresofmarket parencywereimplemented.Thebasicthemeof expectationsofthefederalfundsrateinthefuture. thisworkisthattheeconomywillfunctionmore However,the1-month-aheadcontractispretty efficientlyifthemarketsandtheFedareinterpret- activeandprovidesanexcellentmeasureofmarket ingincomingdatathesameway.IftheFedand expectationsforthatmonth.Inmyanalysis,Iwill themarketshavethesameviewastothepolicy focusonthe1-month-aheadcontractandcertain implicationsofnewinformation,thenthemarket othercontracts,suchasthe4-month-aheadcon- willbeabletopredictFedpolicyadjustments tract—forexample,inOctober2005,thecontract accurately. forFebruary2006. Thoseanalysesweremadepriortotheintro- NowlookatFigure1.Theperiodcovered ductionof“forward-looking”languageinthepost- startswhentradinginfederalfundsfuturescom- FOMCmeetingpressreleases,beginningin mencedinOctober1988.Thedatashownaredaily August2003.Anappropriatequestionishowthe changes(close-of-businesstoclose-of-business) “forward-looking”languagehasaffectedmarket intheyieldonthe1-month-aheadfederalfunds perceptionsoffutureFOMCpolicyactions. futurescontractondaysofscheduledFOMC Figures1and2replicateandextendthecorre- meetingsanddayswhentheFOMCchangedthe spondingfiguresfromPooleandRasche(2003; targetfundsratebetweenregularmeetings.The seefootnote2). areashadedingray,betweenplusandminus5 First,alittlebackground.Thefederalfunds basispoints,indicatesaregionthatIhavedefined markettradescontinuouslyduringthedayand asinsignificant“noise”inthismarket. theratemayfluctuateminutebyminute.Atthe InFigure1,thepointsplottedwithasquare conclusionofeachday,theFedpublishesthe showdaysonwhichtheFOMCchangedthetarget “effective”rate,whichistherateatwhichmost federalfundsrateby25basispoints.Forexample, transactionstookplace.Alsotradingcontinuously theseconddarkbluesquarefromtheleftshows duringtheday,ontheChicagoBoardofTrade, thatonthatdaythe1-month-aheadfuturescon- 2 ForareviewoftheeffectoftheearlierimprovementsinFOMCtransparencyseeW.PooleandR.H.Rasche,“TheImpactofChangesinFOMC DisclosurePracticesontheTransparencyofMonetaryPolicy:AreMarketsandtheFOMCBetter‘Synched’?”FederalReserveBankofSt.Louis Review,January/February2003,85(1),pp.1-9. 3 MONETARYPOLICYANDINFLATION Figure 1 Changes in Funds Futures Rate When FOMC Changed Target Funds Rate BasisPoints 50 →8/89:Allsubsequent →2/94:Allsubsequentchangesaccompanied by changesin multiplesof pressstatement at end of FOMC meeting. 40 25basispoints. 30 20 10 0 –10 –20 –30 LtBlue:Intermeeting target changes DkBlue:Target changes at scheduled meetings –40 –50 10/20/88 10/20/89 10/20/90 10/20/91 10/20/92 10/20/93 10/20/94 10/20/95 10/20/96 10/20/97 10/20/98 10/20/99 10/20/00 10/20/01 10/20/02 10/20/03 10/20/04 NOTE:(cid:4)25-basis-pointchange,(cid:2)50-basis-pointchange,(cid:1)75-basis-pointchange,(cid:3)change thatmarketwasnotimmediatelyawarehadoccurred(Poole,Rasche,andThornton,2002). tractrosebyabout0.12percentagepoints,or12 targetrateby75basispoints.Ifyoulookcare- basispoints.Thus,onthatday,themarkethad fullyatthedarkbluetriangles,youcanseethat predictedabouthalfoftheactualchangeof25 thefuturesmarketdidaprettygoodjobofpredict- basispointsinthetargetfundsrate.Lookingfar- ingchangesof50basispoints.Thelargesterror, thertotherightinFigure1,youcanseethatmost in2002,wasabout–0.18,or18basispoints.One ofthedarkbluesquaresfallwithinthegrayband. wayofinterpretingthaterroristhatthemarket ThatmeansthatondaysofscheduledFOMC wasbettingonadeclineinthetargetrateof25 meetings,mostofthetimethemarketcorrectly basispointsandputtingsomeprobabilityona predictedtheCommittee’sactionofchangingthe declineof50basispoints.Lookingatallthedark targetfundsrateby25basispoints.Remember bluepoints,themarkethasdoneaprettygood thatthedatashowthemarket’stradingtheday jobofpredictingratechanges,especiallyafter oftheFOMCmeeting.TheCommittee’sdecision February1994. isnotgenerallypredictedaccuratelyweeksor NowlookatthelightbluepointsinFigure1. monthsinadvance. AlsoinFigure1,pointsplottedwithadark TheseshowdaysonwhichtheFOMCtookpolicy bluetriangleindicatedaysofscheduledmeetings actionsbetweenregularlyscheduledmeetings. whentheFOMCchangedthetargetfundsrateby Aswiththedarkbluepoints,squaresindicate 50basispoints.Thereisonepoint,November15, federalfundstargetchangesof25basispoints 1994,plottedwithadarkbluediamond,showing andtrianglesindicatechangesof50basispoints. thesinglecaseinwhichtheFOMCchangedthe Althoughtherearefewlightbluepointsafter 4 HowPredictableIsFedPolicy? Figure 2 Changes in Funds Futures Rate on FOMC Meeting Dates With No Change to Target Funds Rate BasisPoints 50 →2/94:Allsubsequent changes accompanied by pressstatement at end of FOMC meeting. 40 30 20 10 0 –10 –20 –30 –40 –50 10/21/88 10/21/89 10/21/90 10/21/91 10/21/92 10/21/93 10/21/94 10/21/95 10/21/96 10/21/97 10/21/98 10/21/99 10/21/00 10/21/01 10/21/02 10/21/03 February1994,itisclearthatFOMCpolicyactions targetchange,havegeneratedlittleifanynews ondaysotherthanscheduledmeetingsarenot inthefederalfundsfuturesmarket.Suchdecisions wellpredicted.Thatresultishardlysurprising, havebeenwellanticipatedbymarketparticipants. giventhatsuchmeetingsarenotannouncedin AsyouhavebeenlookingatFigures1and2 advance.Theveryexistenceofsuchameeting, youhaveprobablywonderedwhythepointsat aswellasthechangeinthetargetfundsrate, theendoftheperiod,startingwithAugust2003, necessarilytakesthemarketbysurprise. fallalmostpreciselyonzero,indicatingnoerror Figure2showsdayswhentheFOMCmetand atallinpredictingFOMCdecisions.Startingwith madeapolicydecisionnottochangethetarget thestatementissuedafteritsmeetingofAugust12, federalfundsrate.Ascanbeseeninthefigure, 2003,theFOMChasincluded“forward-looking” thereweretimesbefore1998whenthemarket languagethathasfacilitatednearlyperfectmarket putsomeprobabilityonaratechange,andwhen forecastsoftheFOMCdecisionatitsnextmeeting. theFOMCdidnotchangethefundsratetarget Initially,thelanguageindicatedthat“policy thatmeantthatthefuturesmarketadjustedto accommodationcanbemaintainedforaconsid- reflectthatoutcome.However,mostofthetime erableperiod.”Thatlanguagesuggestedthatthe thepointsfallinthegrayband,meaningthatthe FOMCwouldnotchangethetargetfundsrateat futuresmarketdidnotchangebymuchbecause itsnextmeeting.AfterthemeetingofJanuary28, themarkethadcorrectlypredictedthattheFOMC 2004,thelanguagewasmodifiedtosaythat“the wouldnotchangethetargetrate. Committeebelievesthatitcanbepatientinremov- Wecansummarizetheseresultsasfollows: ingitspolicyaccommodation.”Themarketread ParticularlysinceFebruary1994,policydecisions thatlanguageassuggestingthattheperiodofan takenatregularlyscheduledFOMCmeetings, unchangedtargetfundsratewascomingtoanend, whetherornotthey’veinvolvedafederalfunds butwasnotyetover. 5 MONETARYPOLICYANDINFLATION Figure 3 Changes in Funds Futures Rate When FOMC Changed Target Funds Rate Futures Yield Changes forMonthAfterNext FOMC (basis points) 50 40 30 20 10 0 –10 –20 –30 DkBlue:January 1999–June 2003 –40 LtBlue:August 2003–September2005 –50 –50 –40 –30 –20 –10 0 10 20 30 40 50 ChangesinYieldsonNext-Month Futures Contract (basispoints) Finally,thestatementissuedafteritsmeeting ofthe25-basis-pointincreaseinthetargetfunds ofMay4,2004,saidthat“theCommitteebelieves rate. thatpolicyaccommodationcanberemovedata BobRascheandI,inourearlierwork,found pacethatislikelytobemeasured.”Thatlanguage thatthefuturesmarket,althoughpredictingFOMC thenappearedineverystatementthroughthe decisionsquiteaccuratelyadayinadvance,usu- mostrecentstatementissuedafterthemeeting allydidnotpredictaccuratelyseveralmonthsin onSeptember20,2005.Themarketcametoread advance.Butnowwehaveanewquestionto thelanguageasindicatingthattheFOMCwould explore:IfthecontentoftheFOMCpressreleases raisethetargetfundsrateby25basispointsat sinceAugust2003issignalingthepolicydecision itsfollowingmeeting.TheFOMCinfactdidso atthenextFOMCmeetingsoclearly,howwellare ateverymeetingthroughthemostrecentone, marketsnowpredictingthemoredistanttrajectory andthemarketpredictionerrorswerenegligible. ofthetargetfederalfundsrate? Inparticular,noneofthepolicyactionstakensince Addressingthisquestionisabitmorecompli- theintroductionofthislanguage—thatpolicy catedthanitmightseem.TheFOMCandmarket accommodationcanberemovedatameasured participantsunderstandthatmonetarypolicy pace—hasgeneratedanylarge(greaterthan5 cannotbelockeddownlonginadvancebecause basispoints)changeintheyieldonthe1-month- anunpredictableeconomiceventmightmakea aheadfundsfuturescontractonthedayofthe policyadjustmenthighlydesirable.SinceJune FOMCaction.Oneachoftheseoccasionsthe 2004,thepressreleasehasclearlyindicatedthat, futuresmarkethasmadeanalmostperfectforecast whiletheCommitteeintendstoproceedatameas- 6 HowPredictableIsFedPolicy? uredpace,itsintentionisconditionedonfuture notbeensurprises,norhavetheygeneratedrevi- informationaboutthestateoftheeconomy.3A sionstomarketexpectationsofthepolicyaction secondquestionconcernstheinformationthatis attheimmediatefutureFOMCmeeting. generatingadjustmentsofmarketexpectations. Ifweexamineallthedailydata,andnotjust OntheverticalaxisofFigure3arechanges datafordaysofFOMCmeetings,fromthebegin- inthefuturesyieldforthecontractofthemonth ningofJuly2003throughmid-September2005, subsequenttothenextFOMCmeeting.These therewereonlytwodayswhentheyieldonthe changesareplottedonthedatewhenanFOMC 1-month-aheadfundsfuturescontractchanged policyactionwasannounced.Becausethereare by5basispointsormore.Thosetwodayswere eightscheduledmeetingsperyear,sometimes June15,2004,andSeptember1,2005.Onthe therearemeetingsinadjacentmonthsandsome- firstofthesedaysChairmanGreenspantestified timesnot,whichmeansthatthefuturescontract atthehearingsforhisrenominationasChairman studiedissometimestwomonthsaheadandsome- oftheBoardofGovernors.OnthatdaytheJuly timesthreemonthsahead.Anexamplemayhelp 2004futuresyielddecreasedby5basispoints.On inunderstandingwhatdatawereused.Inthecase theotherday,September1,2005,theOctober ofthemeetingonMay15,2001,themarketknew 2005futuresyielddecreasedby6basispointsin thattherewasascheduledmeetingonJune26-27, theaftermathofHurricaneKatrina.Thus,since 2001.Thus,hereweexaminedthechangeon July2003thereissimplynotmuchinformation May15intheJuly2001futurescontract—the2- inthe1-month-aheadcontractonhowmarket month-aheadcontract.However,thenextsched- expectationsarereactingtonews.Thisfindingis uledmeetingfollowingtheoneonJune26-27,was insharpcontrasttoearlierPoole-Raschefindings, onAugust21;therewasnomeetingscheduled whichIdonothavetimetodiscussinanydetail forJuly.Thus,therelevantchangeinthefederal here,thateconomicnewssuchasemployment fundsfuturesmarketonJune27wasforthe reportsoftentriggeredsignificantchangesinthe September2001contract—the3-month-ahead 1-month-aheadfuturescontract. contract. However,overthepasttwoyears,the4-month- InFigure3,onthehorizontalaxisarechanges aheadfuturescontractismoreinformative. intheyieldonthe1-month-aheadfuturescon- BetweenJune2003andmid-September2005 tract,andontheverticalaxisarechangesinthe therewere24occasionswhentheyieldonthis appropriate2-or3-month-aheadcontractasjust contractchangedby5ormorebasispoints.One explained.Thepointsplottedindarkblueare wastheoccasionoftheannouncementofa25- fortheperiodfromthebeginningof1999through basis-pointincreaseinthetargetfundsrateon June2003.Duringthisperiod,yieldsonthetwo June30,2004.Onthatdatetheyieldonthe futurescontractsessentiallymovedone-for-one October2004fundsfuturescontractdecreased asindicatedbythetightscatterofthepoints by8basispoints.Atfirstglancethismarketreac- aroundthe45-degreeline.Thatis,ifthe1-month- tionmightsuggestsomeconfusionaboutFOMC aheadcontractchangedby30basispoints,soalso intentions. didtheappropriate2-or3-month-aheadcontract. Infact,twopiecesofinformationbecame Thepointsplottedinlightbluearefromthesuc- availablewiththeFOMCpressreleaseonthat cessionof25-basis-pointmovesduringtheperiod day.First,theincreaseof25basispointsinthe ofthe“measuredpace”language.Allbutoneof targetfundsratewastheinitialpolicyaction thesepointsfallintothegrayboxthatdesignates undertheforward-lookinglanguageofremoving changesoflessthan5basispointsinabsolute policyaccommodation“atapacethatislikelyto value.Hencethechangesinthetargetfundsrate bemeasured.”Onthepreviousday,theOctober duringthemeasuredpaceregimehavegenerally 2004fundsfuturescontracthadclosedat1.90. 3 “Nonetheless,theCommitteewillrespondtochangesineconomicprospectsasneededtofulfillitsobligationtomaintainpricestability.” FederalReservePressRelease,June30,2004(www.federalreserve.gov/boarddocs/press/monetary/2004/20040630/default.htm). 7 MONETARYPOLICYANDINFLATION Figure 4 Large 4-Month-Ahead Funds Futures Changes and Employment Surprises Four-Month-AheadYieldChange(basispoints) 20 15 10 5 0 –5 –10 Dk Blue:January 1999–June 2003 –15 LtBlue:August 2003–September 2005 –20 –400 –300 –200 –100 0 100 200 300 400 ActualPayroll EmploymentMinusSurveyPredictedChange SincetherewereonlytwoscheduledFOMCmeet- September,butonlyabouta0.3probabilityofa ingsduringtheJuly-October2004period,this movelargerthan50basispointsduringthat yieldimpliedamarketexpectationofacumulative period.Consistently,aftertheannouncement, 75-basis-pointmoveandroughlya0.6probability theprobabilityofa50-basis-pointmoveatthe ofanadditional25-basis-pointmoveoverthree AugustFOMCmeetingimpliedbytheAugust FOMCmeetingsinJunethroughSeptember.The contractfelltoaround0.25.Theprobabilityofa AugustandSeptemberfuturescontractyields 50-basis-pointmoveattheSeptemberFOMC werequiteconsistent.TheAugustcontractyield meetingimpliedbytheSeptembercontract,con- impliedaprobabilityofabout0.4ofa50-basis- ditionalona25-basis-pointmoveinAugust,also pointmoveattheAugustFOMCmeetingcondi- felltoaround0.25. tionalupona25-basis-pointincreaseattheJune Theotherinformationthatbecameavailable meeting.TheSeptembercontractyieldimplieda intheJune2004pressreleasewasintroduction probabilityofabout0.5ofa50-basis-pointmove ofthequalificationtothemeasuredpacelanguage attheSeptemberFOMCmeetingconditionalupon that“theCommitteewillrespondtochangesin a25-basis-pointincreaseatboththeJuneand economicprospectsasneededtofulfillitsobli- AugustFOMCmeetings. gationtomaintainpricestability.”OnJuly2,2004, Therevelationofa25-basis-pointmoveatthe theinitialestimateofpayrollemploymentgrowth June2004meetingsolidifiedmarketimpressions forJune2004wasonly112,000newjobs,com- that“measuredpace”meantasuccessionof25- paredwithsurveypredictionsontheorderof basis-pointincreasesinthetargetfundsrate.At 250,000.Afterthisnews,asecondlargedownward thecloseoftradingaftertheJune30policyaction, adjustmenttotheOctoberfuturesyieldoccurred, theOctoberfuturesyieldof1.83impliedtwo reducingtheexpectationofanypolicyactionin additionalincreasesof25basispointsthrough excessof25basispointsateitherofthenexttwo 8 HowPredictableIsFedPolicy? Figure 5 Small 4-Month-Ahead Funds Futures Changes and Employment Surprises Four-Month-AheadYieldChange(basispoints) 20 15 10 5 0 –5 –10 DkBlue:January 1999–June 2003 –15 LtBlue:August 2003–September 2005 –20 –400 –300 –200 –100 0 100 200 300 400 ActualPayroll EmploymentMinusSurveyPredictedChange FOMCmeetingstozero.TheyieldsontheAugust ampleempiricalinvestigationinmanymarkets andSeptembercontractsalsoimpliedapproxi- indicatethattheappropriateconceptisthe matelyzeroprobabilityofanythingotherthan employmentsurprise,measuredbythedifference 25-basis-pointmovesatthetwoforthcoming betweentheinitialestimateofthechangeinpay- FOMCmeetings. rollemploymentandsurveypredictionsofthe Oftheremaining23“largechanges”since employmentchange.Figure4presentsascatter June2003,12(52percent)occurredondayswhen plotofchangesintheyieldonthe4-month-ahead theemploymentdatawerereleased.Three“large futurescontractandemploymentsurprises.The changes”occurredintheimmediateaftermathof 12pointsplottedinlightblueindicatethefutures HurricaneKatrina(August30–September1,2005.) marketreactiontotheemploymentprediction Eightchangesoccurredondayswheneconomic errors.SinceIconsideronly“largechanges”in datawerereleased,butinnoneofthesecaseswere thefuturesyield,therearenoobservationsin theremultiplereleasesofanysinglestatistical thegrayshadedareaof5basispoints. series.4Hence,sincetheintroductionof“forward- Withoneexception,allofthelightbluepoints looking”language,marketexpectationsappear fallinthefirstandthirdquadrantsofthegraph tohavereactedsystematicallyonlytoemployment andareroughlyconsistentwithalinearrelation- data. shipbetweenthechangesinthefuturesyield Aretheobservedmarketreactionssuggestive andtheemploymentsurveypredictionerror. thatmarketsexpectasystematicreactionbythe Apparently,marketexpectationsofthefuture FOMCtoemploymentdata?Economictheoryand levelofthefundsratesincetheFOMCintroduced 4 Thereleasesthatoccurredontheseeightdateswerewholesaleinventories(9-Jun-04),retailsales(14-Jun-04),CPI(15-Jun-04),personal income(28-Jun-04),industrialproduction(15-Apr-05),leadingindicators(21-Apr-05),ISM(6-Sep-05),andproductivity(7-Sep-05). 9 MONETARYPOLICYANDINFLATION “forward-looking”languageintothepressrelease mentsurprisesdidnotgenerateanyappreciable areadjustedinthesamedirectionastheemploy- reactioninthefundsfuturesyield. mentsurprise.Thisfindingisconsistentwithan Myconclusionfromtheseobservationsis understandingthattheshort-runstrategyofthe thatmarketsentimenthascoalescedaroundthe FOMCinvolvesadjustmentsinthefuturepathof viewthatnewsaboutemploymentgrowthisa significantinfluenceonthepathofthetargetfunds thetargetfundsratewhenincomingdatasuggest rateintheforeseeablefuture. thattheriskstoemploymentgrowthhavechanged. I’llfinishwithtwoobservations.First,my ThepointsplottedindarkblueinFigure4 emphasisonmarketreactionstoemployment indicate“largechanges”inthe4-month-ahead surprisesdoesnotmeanthatthemarketignores fundfuturesrateondaysthatthepayrollemploy- inflation.Whathashappenedinrecentyearsis mentdatawerereleasedbetweenJanuary1999 thatcoreinflation—inflationexcludingeffects andtheintroductionof“forward-looking”lan- offoodandenergy—simplyhasnotgenerated guageinAugust2003.Thereare19suchevents significantsurprises.TheFedhasemphasized ofatotalof131“largechange”eventsduringthis thatitfocusesoncoreinflationsothatmonetary period(15percent).Againalmostallofthepoints policydoesnotreacttoenergyandfoodprices, lieinthefirstandthirdquadrantsofthegraph. whichtendtobehighlyvolatile.Ihavenodoubt Also,theredoesnotappeartobeanysystematic thatboththeFOMCandthemarketwouldrespond differencebetweenthescattersofpointsbefore tosurprisesincoreinflationthatseemedlikely andaftermid-2003.Thus,atthehorizonofthe tobepersistentandtoindicateadeveloping 4-month-aheadfuturescontract,theadjustment inflationproblem. ofmarketexpectationstonewsaboutpayroll Second,recentchangesinthefederalfunds employmentdoesnotappeartohavebeensys- futuresrateinresponsetorapidlychanging tematicallyinfluencedbytheintroductionof eventsconnectedwithhurricanesKatrinaand “forward-looking”languageinthepressrelease. Ritawillbeinterestingtoexaminecarefullyin Tocompletetheanalysis,itisnecessaryto thefuture.However,theseeventsaretoorecent examinethenatureofmarketreactionstoemploy- tobegoodcandidatesforcarefulanalysisnow, mentsurprisesonthosedateswhentherelease andI’llforgoaneffortatinstantanalysis. oftheemploymentstatisticswasnotaccompanied bylargechangesintheyieldonthe4-month-ahead fundsfuturescontract.Thesedataareshownin CONCLUSION Figure5.Asbefore,thepointsinlightblueare Thefederalfundsfuturesmarket,andother fortheperiodsincetheintroductionof“forward- marketsIhavenotdiscussedhere,providearich looking”languageinAugust2003.Thepoints sourceofinformationtobetterunderstandthe indarkbluearefromJanuary1999throughmid- effectivenessoftheFed’schangesindisclosure 2003.IncontrasttoFigure3,thedifference policiesovertheGreenspanera.Itisquiteclear betweenthedatafromthepasttwoyearsandthe thatthemarketsunderstandFedpolicytoamuch earlierperiodisstartling.Intherecentperiod,all greaterextentthanbefore.Myownviewisthat ofthesmallchangesinthefuturesyieldoccurred themarket’simprovedunderstanding,andthe whentheemploymentforecastwashighlyaccu- Fed’seffortstoimproveclarityofmonetarypolicy rate.Hence,duringthisperiodthereisaclear decisionsanddecisionprocesses,havemuchto distinctionintheresponseofthefuturesyieldto dowiththeeconomy’simprovedstability.Reces- theemploymentdata:Whentherearelarge sionshavebecomemilder,andcoreinflationmore employmentsurprisesthefuturesrateadjusts; stable.Maintainingthesegainsisimportantto whentherearenosurprises(orverysmallsur- economicwelfare.Iwouldnotclaimthatwehave prises)thereislittlemovementinthefuturesyield. enoughevidencetosaythatthegainsareperma- However,intheperiodfrom1999throughmid- nent,butwedohaveenoughtosaythattheeffort 2003thereweremanyoccasionswhenemploy- hasbeenveryproductive. 10

Cite this document

APA

William Poole (2005, October 3). Speech. Speeches, Federal Reserve. https://whenthefedspeaks.com/doc/speech_20051004_poole

BibTeX

@misc{wtfs_speech_20051004_poole,
  author = {William Poole},
  title = {Speech},
  year = {2005},
  month = {Oct},
  howpublished = {Speeches, Federal Reserve},
  url = {https://whenthefedspeaks.com/doc/speech_20051004_poole},
  note = {Retrieved via When the Fed Speaks corpus}
}