speeches · September 3, 2003

Speech

William Poole · President
Prospects and Risks in the Bond Market St.LouisFinancialAnalystsSociety St.Louis,Missouri September4,2003 I’m pleased to be here today to discuss an theactualfedfundsrateclosetotheintended issue dear to all of our hearts: Long-term rate.Thisinstitutionalfacthastheunfortunate interest rates and their recent behavior. sideeffectofleadingsometobelievethatthe The gyrations of the bond market since FederalReservehasmorecontroloverinterest the beginning of May have created considerable ratesthanitinfactdoes. discussion.Someinthefinancialpressandsome Let’sstartthestoryatthebeginning.TheFed haspricestabilityandeconomicgrowthobjec- tradersaskwhethertheFedsentconfusingsignals tives.Pricestabilityrequiresmaintenanceoflow aboutprospectsformonetarypolicy;otherswon- andstableinflation.Pricestabilitycontributes derinsteadwhetherthebondmarketsmisunder- directlytoeconomicgrowth;inaddition,timely stood and overreacted to Fed announcements. policyadjustmentscanhelptoreducefluctua- Ratherthandiscusstheseissues—whichhave tionsinemploymentandoutput. reacheda“he-said,she-said”standoff—Iwould PutasidefornowthefactthattheFOMC liketostepawaytoreviewthebasicsofthebond implementspolicybysettingtheintendedfederal market.I’veoftenfeltthatlonger-runfundamen- fundsrate,andconcentrateinsteadontheobjec- talstendtogetlostinawelterofshort-runcon- tivesofpolicy.Inrecentyears,successisalmost siderationsthatfadeintooblivionquitequickly absolutewithrespecttoinflationandexpected asanewsetofshort-runconcernsdominatethe inflation,bothofwhicharelowandstable.Suc- news.I’lldiscussthefundamentaldeterminants cessisconsiderable,butincomplete,intermsof oflong-terminterestrates,whichI’llindexbythe outputandemployment.Idonotmeantoimply benchmarkyieldon10-yearTreasurybonds. thattheFedcouldhavedonemore,buttheecon- Beforeproceeding,Iwanttoemphasizethat omy’sperformanceoverthelastfewyearshasnot theviewsIexpressherearemineanddonot beenasrobustaseveryonewants.Unemployment necessarilyreflectofficialpositionsoftheFederal increasedasaconsequenceofthemild2001 ReserveSystem.Ithankmycolleaguesatthe recessionandhascontinuedtorisesincethe FederalReserveBankofSt.Louisfortheircom- recessionendedinNovember2001becausethe ments;ChristopherJ.Neely,Researchofficerat paceofrecoveryhasbeenmodest. theBank,providedspecialassistance.However, I’dliketoconvinceyouthatinpracticethe Iretainfullresponsibilityforerrors. Fedhasnolatitudeastohowtosettheintended federalfundsrate,exceptformattersofshort-term timing,ifitwantstoachieveitsobjectives.Low FUNDAMENTAL DETERMINANTS andstableinflation,andoutputgrowthalongits long-rungrowthpath,implyacertain,though OF INTEREST RATES notconstant,long-terminterestrate.TheFed TheFederalOpenMarketCommittee(FOMC) mustdeliverapathfortheintendedfederalfunds implementsmonetarypolicybysettingan rateconsistentwiththedesiredoutcomesforthe intended,ortarget,federalfundsrate,andthen economyinordertorealizethoseinflationand engaginginopen-markettransactionstokeep outputgoals. 1 FINANCIALMARKETS Economiststhinkofthebenchmark10-year italspendingbyissuingbonds.Iftherealinterest Treasuryrateashavingthreecomponents:the rateistoolowrelativetotheexpectedreturnon realrateofinterest,expectedinflation,andarisk newinvestmentinplantandequipment,then premiumforunexpectedinflation.I’lldiscussthe therewillbeexcessdemandforfundstobuild determinantsoftherealrateofinterestlast. physicalcapitalandtherealinterestratewillbe Theexpectedinflationcomponentisnotdif- bidup.Conversely,iftherealinterestrateistoo ficulttounderstand.Inflationeatsawaythepur- highrelativetoexpectedreturns,thentherewill chasingpowerofabond;borrowerscanpayoff beanexcesssupplyoffundsandtherealinterest theirdebtswithdollarsofdepreciatedpurchasing ratewillbebiddown.Throughtheseopportuni- power.Ifexpectedinflationrisesby1percentage tiesforsubstitutingcapitalforbonds,orviceversa, point,thennominalinterestrateswillriseby1 therealrateofreturnonbondsislinkedtothe percentagepointtoo,allelseequal.Byfarthe realrateofreturnoncapital. mostimportantreasoninterestratestodayare Arobusteconomyandhighproductivity farbelowtheirlevelsin1981isthatactualand meansthatbusinesseswillseektoborrowto expectedinflationaremuchlowernow. financefutureproduction,whichwillbidup Nooneknowsexactlywhatinflationwillbe, interestrates.Higherlevelsofeconomicactivity solendersmustbecompensatedforthefactthat meanhigherinterestrates,allelseequal.The couponandprincipalpaymentswillhaveuncer- principalreasonthe10-yearbondrateislower tainpurchasingpower.Inflationuncertaintyisthe todaythanitwasatitsmonthlypeakinJanuary sourceofanothercomponentofnominalinterest 2000isthatthedemandforfundstofinancecapi- rates.Themorevolatileandunpredictableisinfla- talinvestmentismuchlowertodaythanitwas tion,thehigheristhiscomponentofthenominal attheheightofthepreviouseconomicboom. rate—theinflationriskpremium.Today,inflation Weseethissameeconomicforceatwork isexpectedtobefairlystable,sotheinflationrisk whencomparinginterestratesintheUnited premiumisprobablysmall,certainlyconsider- StateswiththoseinJapan.Japanhasbeengrow- ablylessthaninthelate1970s.Itisverydifficult ingveryslowlyforadecade,goinginandoutof tosortoutinflationexpectationsfrominflation recession.Thedemandforfundsislowbecause uncertainty;I’lllumpthetwotogetherandsimply profitableinvestmentopportunitiesareviewed talkofinflationexpectations. aslimited.Incontrast,theU.S.situationtodayis I’vementionedthatthelargedeclineininter- viewedmuchmoreoptimistically,eventhough estratesafter1981wasduetoalargedeclinein notasoptimisticallyasitwasfouryearsago. inflationexpectations.Ifweexamineinterest Allofthesecomponentsofinterestratesare ratesoveraperiodwheninflationchangedalot, forwardlooking;theydependonexpectations oracrosscountrieswithverydifferentinflation aboutthefuture.Wecan’tdirectlyobservethese rates,we’llfindthatmostofthevariationofrates componentsoflong-terminterestrates,butwe isduetovariationofinflation.Butwhenweexam- canestimatethem. ineratesoveraperiodcharacterizedbyinflation Measuringexpectationsisinherentlydifficult, stability,oracrosscountrieswithsimilarinflation butourtoolsforestimatingexpectedinflation rates,theninterestratevariationcannotbeattrib- haveimprovedsignificantlyinrecentyears. utedtoinflation.Theobservedvariationover Primarily,weusesurveyexpectationsofinfla- time,ordifferencesacrosscountries,reflects tionorthedifferencebetweenyieldsonregular forcesdeterminingtherealrateofinterest. TreasurybondsandTreasuryinflation-indexed Therealratedependsmostcentrallyonthe securities(TIIS),whichI’llcallsimply“indexed expectedproductivityofphysicalcapital.Investors bonds”forshort. canchoosewhethertoholdbondsorphysical Indexedbonds,firstissuedbytheTreasury capital,orequityclaimstophysicalcapital,in in1997,containaprovisionthatincreasestheir theirportfolios.Firmscanchoosetofinancecap- principalandeverysemiannualinterestpayment 2 ProspectsandRisksintheBondMarket bytheincreaseintheconsumerpriceindexfrom risesandfallsofdefaultriskarequitenaturally thedateofissue.Thesebonds,therefore,com- linkedtothestateoftheeconomy. pletelyprotecttheinvestorfromtheeffectsof Onefinalcomponentofinterestratesarises inflation;theyieldonthebondsisbydefinition fromtheliquidityorilliquidityofthebond.Bonds arealyield.Ifweassumethattheinflationrisk thatareactivelytradedinlargevolumearehighly premiumisnegligible,thenthedifferencebetween liquid.Aninvestorcanbuyorsellsubstantial theyieldonaconventionalbondandaTIISbond amountswithlittleornoimpactonyield.Even measuresthemarket’sexpectationoffutureinfla- withintheTreasurymarket,weobservesignifi- tion.Forexample,examiningbondsmaturingin cantdifferencesinliquidityofvariousissues. approximately10years,yesterdaytheindexed Tradingtendstobeconcentratedinbenchmark bondhadayieldof2.38percentwhileaconven- issues,suchasthe10-yearmaturity.Bondswith tionalTreasurybondhadayieldof4.53percent. 9yearstomaturityand11yearstomaturitytrade Thesetwobondswillturnouttohaveidentical atayieldafewbasispointshigherthanthe10- yieldsiftheinflationratebetweennowand2013 yearbond. averages2.15percent,thedifferencebetween thetwoquotedyields.Asafirstapproximation, wecansaythatthesetwoassetmarketsarepro- A REVIEW OF MAJOR INTEREST vidinguswiththeinformationthatinvestors RATE DEVELOPMENTS expectthattheinflationratewillaverage2.15 Thesefundamentalconceptsarewhatwe percentoverthenext10years. needtounderstandinterestratedevelopments Theexpectedrateofinflationdependsboth inrecentyears.Usingannualdata,thepeakyear onwhattheFedsaysandwhatithasdone.Mar- forthe10-yearTreasurybondratewas13.91per- ketspayattentiontowhattheFedsaysitwants cent,in1981.It’shardtoknowexactlywhatthe todobecauseithasbuiltupareservoirofcredi- realratewasatthattimebecauseunambiguous bilitybyprovidinglowandstableinflationfor dataarenotavailable.However,thepremiumfor sometime.Butwordsaren’tenough.Ultimately, anycentralbank’spolicyactionsmustbeconsis- inflationexpectationsandinflationuncertainty tentwithitswordsoritwilllosethepublic’strust. wassurelyquitelarge. IftheFeddoesn’tmaintainpoliciesconsistent After1982,inflationfelltothe4percentrange withlowinflation,inflationexpectationswill butinterestratesremainedquitehigh.The10- begintorise.Andifinflationexpectationsrise, yearratewasabove8percenteveryyearinthe long-terminterestrateswillalsorise—holding 1980s,exceptfor1986,whenitaveraged7.68 allelseconstant. percent.Astheeconomygrewafterthemild So,yieldsonlong-termTreasurybondshave recessionof1990-91,thebondrateactuallyfell anexpectedinflationcomponentandarealcom- slowlyonaveragethroughtheendof1998.Along- ponent.Theyieldsonotherbonds,suchascor- termdeclineinbondratesduringaneconomic poratebonds,containanothercomponent:credit expansionisadecidedlyrareevent,asratestypi- risk.Theinterestdifferentialdependsonthepos- callyriseduringanexpansion.Decliningexpected sibilitythattheborrowerwilldefault.Creditrisk inflationanddeclininginflationriskexplain dependsontheindividualbondissuerandoften thisoutcome.Basedonsurveydata,expected changesoverthecourseofthebusinesscycle, CPIinflationdeclinedfromroughly5percentat increasingasbusinessconditionsdeteriorate.The thebeginningof1990toabitover2percentat mostdramaticexampleofthisphenomenonis theendofthedecade.Isuspectthatdeclining thatthespreadbetweenyieldsonjunkbonds inflationriskalsocontributedtothedeclinein andgovernmentbondstendstowidenduring bondrates,butnosatisfactoryseriesoninflation recessionsandnarrowduringexpansions.The riskexists. 3 FINANCIALMARKETS Thedeclineinratesinthesecondhalfof duringthisperiod,butmostofthedeclineinthe 1998reflectedconcerns,inthemarketsandin bondratereflectedadeclineintherealrateof theFOMC,abouttheeffectsoftheRussiandefault interest.Bythemiddleofthisyear,the10-year inAugustandthesevereproblemsfacedbyLong indexedbondyield,whichhadbeenabout4.4 TermCapitalManagement,whichbrokeintothe percentinlate1999,wasdowntoabout1.5per- openinSeptember.Bytheendoftheyear,the cent.Thatyield,bytheway,hasnowrebounded economyshookofftheseconcernsandtheeco- toabout2.4percent. nomicboomresumedinfullforcein1999.The Ithinkthattherightinterpretationofthe demandforcapitalwashigh,andhandsome reboundintherealrateofinterestinrecentweeks returnsinthestockmarketledinvestorsinthat isthatthemarketexpectsaresumptionofeco- directionawayfrombonds.Tokeepinflation nomicgrowth,withaccompanyingresumption undercontrolandtempertheboom,theFed ofstrongercreditdemands.Ibelievethatthe raisedtheintendedfederalfundsrate. evidenceforthisinterpretationissubstantial, Usingdataonindexedbonds,firstissuedin becauseweknowfrompreviousbusinesscycles 1997,wehaveaprettygoodfixonrealinterest andfromcross-countryevidencethatthereal rates.Astheeconomicboomintensifiedoncethe rateofinteresttendstobehigherineconomies economygotpastthedisruptioninthefallof withhighergrowthrates. 1998,realratesrose.The10-yearindexedbond yieldpeakedatabout4.4percentinJanuary 2000,butinflationexpectationsremainedinthe THE IMPORTANCE OF neighborhoodof2percent.Accordingly,the10- TRANSPARENCY yearnominalTreasurybondyieldalsopeakedin I’veconcentratedonlonger-runfundamentals January2000,atabout6¾percent. becauseIthinkthatistherightplacetolookfor Astheeconomysoftenedoverthecourseof explanationsofsignificantchangesininterest 2000andreachedabusinesscyclepeakinMarch rates.Themarketisconstantlyseekingtounder- 2001,bondratesbegantofall.TheFedbeganto standthelonger-rundirectionoftheeconomy; easepolicyinJanuary2001.Interestingly,bond however,trendsarealwayseasytoidentifyafter ratesinitiallyrose,asthemarketapparently thefactanddifficulttoreadinrealtime. believedthatFedeasingwouldstimulatethe Ontopofthelonger-runtrendsininterest economyfairlyquicklyandleadtoaresumption ratesisanoverlayofshort-runnoise.By“noise” ingrowthofcreditdemands.Thatwasnottobe; Imeansmallday-to-dayandweek-to-weekfluc- asweknownow,theeconomycontinuedtodrift, tuationsthatlaterturnouttoreflectmispercep- firstinthemildrecessionthatlastedfromMarch tions,verytemporaryliquiditychanges,andsuch toNovemberof2001andthenonlymodestlyup things.Ifyoulookatagraphofquarterlyaverage in2002. data,muchofthenoisedisappears.Ifyoulookat Onseveraloccasionsoverthelastthreeyears, agraphofdailydataandgobacktothedaily forecastsofmorebuoyantgrowthweredisap- financialpress,youwillseethenews,rumors, pointed.Thecombinationofcapitaloverhang andspeculationsthatliebehindmuchofthe fromtheinvestmentboomofthelate1990s, noise.Ifyougobacktothedailynews,whichI shocksfrommajorbankruptcies,thetragedyof didatgreatlengthasanacademicbeforeIcame 9/11,andcorporategovernancescandalsheld totheFed,youwilllikelybeboredbymostof theeconomyback.Whenexpectedincreasesin whatyouread.Mostoftheaccountsdescribe creditdemandsdidnotmaterializeandwiththe noisethathasnovalueinunderstandingthe disappointingperformanceofthestockmarket, fundamentalsdrivinglonger-rundevelopments. investorsbidbondratesdown.Long-terminfla- Thestanceofmonetarypolicyisoneofthe tionexpectationsmayhavedrifteddownalittle fundamentalsinthelargerpicture,becauseifthe 4 ProspectsandRisksintheBondMarket Fedgetsitwrong,thentheeconomywillnotgrow WhatoughttobepredictableistheFed’scommit- alongitpotentialgrowthpathwithlowandstable menttoitspolicygoalsanditsresponsetoevents inflationandinterestrateswillnotsettleat,or astheyoccur.Fedresponsesarenotperfectly fluctuatearound,thelevelappropriatetoan predictabletodaybecausenooneinsideoroutside economygrowingatitsfullpotential. theFedknowshowtowriteanexplicitrecipe Becauseofthemarket’sintenseinterestin forconductingpolicy.However,thereisample monetarypolicy,theFOMChasamajorcommu- evidencethatpolicyissubstantiallypredictable, nicationschallenge.Ispokeinsomedetailonthat asthemarketandtheFedmostoftendoreadthe topicinaspeechtwoweeksagoinPhiladelphia, implicationsofarrivinginformationthesameway. anddonotwanttorepeatthatentirediscussion here.ThebottomlineisthatoneoftheFed’sjobs istocommunicateasaccuratelyaspossibleso CONCLUDING COMMENTS themarketcandetermineinterestratesefficiently. NowI’llpulltheanalysistogether.Withregard ThatmeansthattheFedneedstodothebestit toinflationexpectationsandinflationrisk, cantoconveytheessentialelementsofpolicy prospectsgoingforwardareexcellent.Actual clearlyandnotitselfbeasourceofshort-run andexpectedinflationhavebeenquitestablein noiseinthemarket.Tome,anessentialingredi- recentyears,andthereiseveryreasontoantici- entofgoodFedcommunicationsistofocuson patethattheseattractiveconditionswillremain longer-runfundamentalsandonhowtheFOMC inforce.Themarketappearstohavegreatconfi- pursuesitsobjectivesbyadjustingthestanceof denceintheFederalReserve’scommitmentto policytothearrivalofnewinformation. pricestabilityanditspowerstomaintainthe EveniftheFedweretocommunicateits inflationratewithintherangeofexperienceof objectivesandmethodsperfectly,thefuturepath recentyears.Moreover,thestunningincreasein ofthefederalfundsratewouldneverbeperfectly productivityinthesecondquarterannounced predictablebecausetheFOMCmustchangethe intendedratefromtimetotimeasnewinforma- thismorningandthestrongcasethathandsome tionarrives.Nottodosowouldcreateproblems, productivityincreaseswillcontinue—evenifnot suchastheGreatInflationofthe1970s.Thatinfla- asstunningasthesecondquarterdata—makes tionwasaconsequenceofpolicyadjustments inflationcontrolconsiderablyeasierthanitother- thatweretoofrequentlytoolittleandtoolate.In wisewouldbe.Althoughwemustalwaysbealert theend,thosepolicymistakesledtomoreuncer- toinflationordeflationsurprisescomingoutof tainty,becauseoftheinflation,andlargerinterest thewoodwork,thereis,inmyview,theprospect ratechangesthanwouldhaveoccurrediftheFed goingforwardthatinflationwillbebenignand hadbeenwillingtoactearlieranddecisively. thattherisksinthisdirectionareaslowaswe GiventhattheFOMCmustrespondintimely haveseeninthelast40years. fashiontonewinformation,andcertainlyhasfor Withregardtoprospectsandrisksonthereal atleast20years,I’llmaketheclaimthatmis- rateofinterest,mymessageisthattherisksare communicationaccountsforonlyatrivialfraction tiedtoriskswithrespecttoeconomicgrowth.As ofinterestratechangesinrecentyears.Incident- Iexaminegrowthexpectationsofprofessional ally,“respondintimelyfashion”doesnotneces- forecasters,suchastheBlueChippanel,myread sarilyimplyfrequentpolicyadjustments.Inthe isthattheconsensusoutlookisforsolidandbal- morestableinflationenvironmentweenjoytoday, ancedeconomicgrowthgoingforward.However, theFedhasfarmorefreedomthanithadinthe asIalwaysemphasizewhendiscussingtheeco- 1970stowaitforinformationtoaccumulate. nomicoutlook,forecastschangeovertime,some- Becausetheeventsthatdrivechangesinthe timessignificantly,andatanygiventimethereis stanceofpolicyareunpredictable,theintended arangeofprofessionalopinionontheoutlook. federalfundsratecannotitselfbepredictable. Shouldweseeacontinuationofasluggishrecov- 5 FINANCIALMARKETS ery,thentheprospectsarethatbondrateswill fallsomewhatfromcurrentlevels.Shouldwe seeagangbustersrecovery,thentheprospects arethatrisingcreditmarketswilldrivebondrates abovecurrentlevels.Ineithercase,theaction willbeprimarilyintherealrateofinterestand not,Ibelieve,intheinflationpremiumcompo- nentofrates. 6
Cite this document
APA
William Poole (2003, September 3). Speech. Speeches, Federal Reserve. https://whenthefedspeaks.com/doc/speech_20030904_poole
BibTeX
@misc{wtfs_speech_20030904_poole,
  author = {William Poole},
  title = {Speech},
  year = {2003},
  month = {Sep},
  howpublished = {Speeches, Federal Reserve},
  url = {https://whenthefedspeaks.com/doc/speech_20030904_poole},
  note = {Retrieved via When the Fed Speaks corpus}
}