speeches · October 15, 2006

Speech

William Poole · President

Data, Data, and Yet More Data AssociationforUniversityBusinessandEconomicResearch(AUBER)AnnualMeeting UniversityofMemphis Memphis,Tennessee October16,2006 PublishedintheFederalReserveBankofSt.LouisReview,March/April2007,89(2),pp.85-89 I’velonghadaninterestindata,andIthink Beforeproceeding,Iwanttoemphasizethat that this topic is a good one for this con- theviewsIexpressherearemineanddonot ference. The topic is also one I’ve not necessarilyreflectofficialpositionsoftheFederal addressed in a speech. ReserveSystem.Ithankmycolleaguesatthe Apersonalrecollectionmightbeagoodplace FederalReserveBankofSt.Louisfortheircom- tobegin.Intheearly1960s,inmyPh.D.studies ments.RobertH.Rasche,seniorvicepresidentand attheUniversityofChicago,Iwasfortunatetobe directorofResearch,providedspecialassistance. amemberofMiltonFriedman’sMoneyWorkshop. Friedmanstokedmyinterestinflexibleexchange rates,inanerawhenmainstreamthinkingwas ORIGINS focusedontheadvantagesoffixedexchangerates Thedistributionofeconomicdatabythe andcentralbankseverywherewerecommitted ResearchDivisionoftheFederalReserveBank tomaintainingthegoldstandard.Well,Ishould ofSt.LouiscanbetracedbackatleasttoMay saycentralbanksalmosteverywhere,giventhat 1961.Atthattime,HomerJones,thendirector Canadahadafloating-ratesystemfrom1950to ofresearch,sentoutamemowiththreetables 1962.Friedmangotmeinterestedindoingmy attachedshowingratesofchangeofthemoney Ph.D.dissertationontheCanadianexperience supply(M1),moneysupplyplustimedeposits, withafloatingexchangerate,andlaterIdida andmoneysupplyplustimedepositsplusshort- paperonnineotherfloatingrateregimesinthe termgovernmentsecurities.Hismemoindicated 1920s.ForthispaperIcollecteddailydataon exchangeratesfrommustypaperrecordsatthe thathe“wouldbegladtohearfromanyonewho BoardofGovernorsinWashington. thinkssuchtimeserieshavevalue,concerning Whatwasstrikingaboutthedebatesover promisingapplicationsorinterpretations.” floatingratesinthe1950sisthateconomists Recollectionsofdepartmentemployeesfrom weresowillingtospeculateabouthowcurrency thattimewerethatthemailinglistwasabout speculatorswoulddestabilizeforeignexchange 100addressees. marketswithoutpresentinganyevidencetosup- ApparentlyHomerreceivedsignificantposi- portthoseviews.Inthisandmanyotherareas, tivefeedback,sincevariousstatisticalreleases carefulempiricalresearchhasresolvedmany emergedfromthisinitialeffort.Amongthese disputes.Ourprofessionhascomealongwayin wereWeeklyFinancialData,subsequentlyU.S. institutionalizingempiricalapproachestoresolv- FinancialData;BankReservesandMoney,sub- ingempiricaldisputes.Theenterpriserequires sequentlyMonetaryTrends;NationalEconomic data,andwhatIwilldiscussissomeofthehis- Trends(1967)andInternationalEconomicTrends toryoftheroleoftheFederalReserveBankof (1978),allofwhichcontinuetothisdate.InApril St.Louisinprovidingthedata. 1989,beforeasubscriptionpricewasimposed,the 1 MISCELLANEOUS circulationofU.S.FinancialDatahadreached RATIONAL EXPECTATIONS almost45,000.ABusinessWeekarticlepublished MACROECONOMIC EQUILIBRIUM in1967commentedaboutHomerthat“whilemost leadingmonetaryeconomistsdon’tbuyhistheo- Thecaseformakingdatareadilyavailableis ries,theyeagerlysubscribetohisnumbers.”As simple.Mostmacroeconomiststodayadheretoa anaside,asaChicagoPh.D.,Ibothboughtthe modelbasedontheideaofarationalexpectations theoriesandsubscribedtothedatapublications. equilibrium.Policymakersareassumedtohave asetofgoals,aconceptionofhowtheeconomy Bythelate1980s,accordingtoBerylSprinkel works,andinformationaboutthecurrentstate (1987,p.6),aprominentbusinesseconomistof andhistoryoftheeconomy.Theprivatesector thetime,“weeklyandmonthlypublicationsof understands,totheextentpossible,policymakers’ theResearchDivision,whichhavenowbecome viewsandhasaccesstothesameinformation standardreferencesforeveryonefromunder- aboutthestateandhistoryoftheeconomyas graduatestoWhiteHouseofficials,wereinitially policymakershave. Homer’sproducts.” Anequilibriumrequiresasituationinwhich Whyshouldacentralbankdistributedataas (i)theprivatesectorhasaclearunderstandingof apublicservice?LegendhasitthatHomerJones policygoalsandthepolicymakers’modelofthe viewedasanimportantpartofhismissionprovid- economyand(ii)thepolicymodeloftheeconomy ingthegeneralpublicwithtimelyinformation isasaccurateaseconomicsciencepermits.Based aboutthestanceofmonetarypolicy.Inthissense onthisunderstanding,marketbehaviordepends hewasanearlyproponent,perhapstheearliest centrallyonexpectationsconcerningmonetary proponent,ofcentralbankaccountabilityand policyandtheeffectsofmonetarypolicyonthe transparency.WhileHomerwasadedicated economy,includingeffectsoninflation,employ- monetarist,anddataonmonetaryaggregates ment,andfinancialstability.Ifthepolicymakers havealwaysfiguredprominentlyinSt.Louis andprivatemarketparticipantsdonothaveviews Feddatapublications,dataonothervariables thatconverge,nostableequilibriumispossible prominentinthemonetarypolicydebatesatthe becauseexpectationsastothebehaviorofothers time,includingshort-terminterestrates,excess willbeconstantlychanging. reserves,andborrowings,wereincludedinthe Theeconomyevolvesinresponsetostochas- datareleases. ticdisturbancesofallsorts.Thecontinuousflow Earlyon,thevariousSt.LouisFeddatapub- ofnewinformationincludeseverythingthat licationsincorporated“growthtriangles,”which happens—weatherdisturbances,technological trackedgrowthratesofmonetaryaggregatesover developments,routineeconomicdatareports, varyinghorizons.Accompanyinggraphsofthe andthelike.Thecoreofmypolicymodelisthat aggregatesincludedbrokentrendlinesthatillus- marketresponsesandpolicyresponsestonew tratedrisesandfallsingrowthrates.Thisinforma- informationarebothmaximizing—households tionfeaturedprominentlyinmonetaristcritiques maximizeutility,firmsmaximizeprofits,and of“stop-go”andprocyclicalcharacteristicsof policymakersmaximizetheirpolicywelfare monetarypolicyduringtheGreatInflationperiod. function. Doesthetraditionofdatadistributioninitiated Acriticalassumptioninthismodelisthesym- byHomerJonesremainavaluablepublicservice? metryoftheinformationthatisavailabletoboth Icertainlybelieveso.ButIwillalsonotethatthe policymakersandprivatemarketparticipants.In St.LouisFed’sdataresourcesarewidelyused caseswherethepolicymakershaveaninforma- withintheFederalReserveSystem.Thisinfor- tionaladvantageovermarketparticipants,policy mationisrequiredforFedresearchandpolicy likelywillnotunfoldinthewaythatmarkets analysis;theextracostofmakingtheinformation expect,andtheequilibriumthatIhavecharac- availablealsotothegeneralpublicismodest. terizedherewillnotemerge.Hence,publicaccess 2 Data,Data,andYetMoreData tocurrentinformationontheeconomyatlow multipledataseriescanbedownloadedintoa costisaprerequisitetogoodpolicyoutcomes. singleExcelworkbook,anddatalistscanbestored forrepeateddownloadsofupdatedinformation. InthefirstsixmonthsafterthisversionofFRED THE EVOLUTION OF ST. LOUIS wasreleased,3.8millionhitswererecordedto FED DATA SERVICES thewebsite.Inarecentsix-monthperiod,FRED received21millionhitsfromover109countries DataservicesprovidedbytheFederalReserve aroundtheworld.FREDcurrentlycontains1,175 BankofSt.Louishaveevolvedsignificantlyfrom nationaltimeseriesand1,881regionalseries. thepaperpublicationsinitiatedbyHomerJones. FREDdataareupdatedonareal-timebasisas TheinitialphaseofthisevolutionbeganinApril informationisreleasedfromvariousstatistical 1991whenFRED®,FederalReserveEconomic agencies. Data,wasintroducedasadial-upelectronicbul- After45years,HomerJones’smodestinitia- letinboard.Thisservicewasnotnecessarilylow tivetodistributedataonthreevariableshasdevel- cost.ForusersintheSt.Louisarea,accesswas opedintoabroad-baseddataresourceonthe availablethroughalocalphonecall.Foreveryone U.S.economythatisavailablearoundtheglobe else,long-distancephonechargeswereincurred. attheclickofamouse.Throughthisresource, Nevertheless,withinthefirstmonthofservice, researchers,students,marketparticipants,and usagewasrecordedfromplacesaswideranging thegeneralpubliccanreachinformeddecisions asTaipei,London,andVancouver.1FREDwas basedoninformationthatiscomparabletothe relativelysmallscale.Theinitialimplementation informationpolicymakershave. includedonlythedatapublishedinU.S.Financial Inthepastyear,wehaveintroducedanumber Dataandafewothertimeseries.Subsequently, ofadditionaldataservices.Oneofthese,ALFRED® itwasexpandedtoincludethedatapublishedin (ArchivalFederalReserveEconomicData),adds MonetaryTrends,NationalEconomicTrends,and avintage(orreal-time)dimensiontoFRED.The InternationalEconomicTrends.Attheendof ALFREDdatabasestoresrevisionhistoriesofthe 1995,theprintversionsofthesefourstatistical FREDdataseries.Since1996,wehavemaintained publicationscontainedshorthistoriesonapproxi- mately200nationalandinternationalvariables; monthlyorweeklyarchivesoftheFREDdatabase. initiallyFREDwasofcomparablescope. Alltheinformationinthesearchiveshasbeen Thenextstepoccurredin1996whenFRED populatedtotheALFREDdatabase,andtheuser migratedtotheWorldWideWeb.Atthatpoint, canaccesspoint-in-timerevisionsofthesedata.2 403nationaltimeseriesbecameavailableinstan- Wehavealsoextendedtherevisionhistoriesof taneouslytoanyonewhohadapersonalcomputer manyseriesbackintimeusingdatathatwere withaWebbrowser.Anadditional70seriesfor recordedinU.S.FinancialData,MonetaryTrends, theEighthFederalReserveDistrictwerealsoavail- andNationalEconomicTrends.Forselected able.Thedataserieswereintextformatandhad quarterlynationalincomeandproductdata,we tobecopiedandpastedintotheuser’sPC.InJuly havecompleterevisionhistoriesbackto1959 2002,FREDbecameatruedatabaseandtheuser forrealdataand1947fornominaldata.Revision wasofferedawiderrangeofoptions.Datacanbe historiesareavailableonhouseholdandpayroll downloadedineithertextorExcelformat.Shortly employmentdatabackto1960.Asimilarhistory thereafter,useraccountswereintroducedsothat forindustrialproductionisavailablebackto1927. 1 EighthNote(1991,p.1). 2 WedonotmaintainhistoriesofdailydataseriesinALFRED.InterestratesandexchangeratesappearatdailyfrequenciesinFRED.Inprin- ciple,thesedataarenotrevised,thoughoccasionalrecordingerrorsdoslipintotheinitialdatareleases.Suchreportingerrorsarecorrectedin subsequentpublications,sothereissometimesavintagedimensiontooneoftheseseries. 3 MISCELLANEOUS Preservingsuchinformationiscrucialto includesBankingandMonetaryStatisticsandthe understandinghistoricalmonetarypolicy.For AnnualStatisticalDigestspublishedbytheBoard example,Orphanides(2001,p.964)shows“that ofGovernors,aswellastheBusinessStatistics real-timepolicyrecommendationsdifferconsider- supplementstotheSurveyofCurrentBusiness ablyfromthoseobtainedwithex-postrevised publishedbytheDepartmentofCommerce.We data.Further,estimatedpolicyreactionfunctions arecurrentlyworking,inajointprojectwiththe basedonex-postreviseddataprovidemisleading BoardofGovernors,tocreatedigitalimagesof descriptionsofhistoricalpolicyandobscurethe theentirehistoryoftheFederalReserveBulletin. behaviorsuggestedbyinformationavailabletothe Finally,wearepostingimagesofhistoricalstatisti- FederalReserveinrealtime.”Orphanidescon- calreleasesthatwehavecollectedintheprocess cludesthat“relianceontheinformationactually ofextendingthevintagehistoriesinALFRED availabletopolicymakersinrealtimeisessen- backintime.Theseimagesshouldallowscholars, tialfortheanalysisofmonetarypolicyrules.” analysts,andstudentsofeconomichistoryto Suchvintageinformationalsoisessentialfor reconstructvintagedataonmanyseriesinaddi- analysisofconditionsatsubnationallevels.For tiontothosewearemaintainingonALFRED. example,inJanuary2005theBureauofLabor Statisticsestimatedthatnonfarmemploymentin theSt.LouisMSAhadincreasedby38.8thousand TRANSPARENCY, betweenDecember2003andDecember2004. ACCOUNTABILITY, AND Thisincreasewaswidelycitedasevidencethat INFORMATION DISTRIBUTION theMSAhadreturnedtostrongemployment growthafterfouryearsofnegativejobgrowth. Asjustindicated,thescopeofthearchival However,thesedatafromtheCurrentEmployment informationinFRASERextendsbeyondnumeric Statisticswerenotbenchmarkedtomorecompre- data.Readyaccesstoawidevarietyofinformation hensivelabormarketinformationthatisavailable isessentialfortransparencyandaccountability onlywithalag.3Thecurrentestimateofnonfarm ofmonetaryauthoritiesandthepublic’sfull employmentgrowthintheSt.LouisMSAforthis understandingofpolicyactions.Since1994,the period,afterseveralrevisions,isonly11.6thou- FederalReserveSystemandtheFOMChave sand,lessthan30percentoftheincreaseoriginally improvedthescopeandtimelinessofinformation reported. releases.Ihavediscussedthisprogressinprevious Anotherdatainitiativethatwelaunchedsev- speeches.4Currently,theFOMCreleasesapress eralyearsagoisFRASER®—theFederalReserve statementattheconclusionofeachscheduled ArchivalSystemforEconomicResearch.The meetingandthreeweekslaterfollowsupwith objectiveofthisinitiativeistodigitizeanddis- thereleaseofminutesofthemeeting.Thepress tributethemonetaryandeconomicrecordofthe releaseandtheminutesofthemeetingsrecord U.S.economy.FRASERisarepositoryofimage thevoteonthepolicyaction.Thepolicystate- filesofimportanthistoricaldocumentsandserial mentandminutesgivethepublicaclearunder- publications.Atpresentwehavepostedtheentire standingoftheactiontakenandinsightintothe historyofTheEconomicReportofthePresident, rationalefortheaction. EconomicIndicators,andBusinessConditions Contrastthecurrentsituationwiththeonein Digest.Wehavealsopostedimagesofmostissues 1979.Atthattime,actionsbytheBoardof oftheSurveyofCurrentBusinessfrom1925 Governorsondiscountratechangeswerereported through1990andareworkingonfillinginimages promptly,buttherewasnopressreleasesubse- oftheremainingvolumes.Thecollectionalso quenttoanFOMCpolicyactionandFOMCmeet- 3 WallandWheeler(2005). 4 See,forexample,Poole(2005). 4 Data,Data,andYetMoreData ingminuteswerereleasedwitha90-daydelay. REFERENCES OnSeptember19,1979,theBoardofGovernors Berry,John.“FedListsDiscountRatetoPeakof11% votedbythenarrowmarginoffourtothreeto onCloseVote.”WashingtonPost,September19, approvea½-percentage-pointincreaseinthe 1979,p.A1. discountrate,withallthreedissentsagainstthe increase.Thisinformationgeneratedthepublic BusinessWeek.“MaverickintheFedSystem.” perceptionthatFedofficialsweresharplydivided November18,1967. and,therefore,thattheFedwasnotpreparedto actdecisivelyagainstinflation.JohnBerry(1979, EighthNote.“IntroducingFRED.”FederalReserve p.A1),aknowledgeablereporterattheWashington BankofSt.Louis,May/June1991,p.1. Post,observedthat“thesplitvote,withitsclear signalthatfromtheFed’sownpointofviewinter- Orphanides,Athanasios.“MonetaryPolicyRules estratesareatorclosetotheirpeakforthisbusi- BasedonReal-TimeData.”AmericanEconomic nesscycle,mightforestallanymoreincreasesin Review,September2001,91(4),pp.964. marketinterestrates.”However,theinterpretation ofthe“clearsignal”waserroneous.Onthatsame Poole,William.“FOMCTransparency.”Federal day,theFOMChadvotedeighttofourtoraisethe ReserveBankofSt.LouisReview,January/February rangefortheintendedfundsrateto11¼to11¾ 2005,87(1),pp.1-9. percent.Moreimportantly,threeofthefourdis- sentswereinfavorofamoreforcefulactionto Sprinkel,BerylW.“ConfrontingMonetaryPolicy restraininflation(seeLindsey,Orphanides,and Dilemmas:TheLegacyofHomerJones.”Federal ReserveBankofSt.LouisReview,March1987, Rasche,2005,pp.195-96).NeithertheFOMC’s 69(3),p.6. action,thedissents,northerationaleforthedis- sentswererevealedtothepublicunderthedis- Wall,HowardJ.andWheeler,ChristopherH. closurepoliciesthenineffect.Theresultwasto “St.LouisEmploymentin2004:ATaleofTwo destabilizemarkets,withcommoditymarkets,in Surveys.”CRE8OccasionalReportNo.2005-1, particular,exhibitingextremevolatility. FederalReserveBankofSt.Louis,February2005. CONCLUSION Thetraditionofdataserviceswaswellestab- lishedwhenIarrivedinSt.Louisin1998,andI mustsaythatIamproudthatleadershipinthe Bank’sResearchDivisionhasextendedthattra- dition.Dataarethelifebloodofempiricalresearch ineconomicsandofpolicyanalysis.Ourrational expectationsconceptionofhowthemacroecon- omyworksrequiresthatthemarketsandgeneral publicunderstandwhattheFedisdoingand why.Ofallthethingsonwhichwespendmoney intheFederalReserve,surelythereturnonour dataservicesisamongthehighest. 5

Cite this document

APA

William Poole (2006, October 15). Speech. Speeches, Federal Reserve. https://whenthefedspeaks.com/doc/speech_20061016_poole

BibTeX

@misc{wtfs_speech_20061016_poole,
  author = {William Poole},
  title = {Speech},
  year = {2006},
  month = {Oct},
  howpublished = {Speeches, Federal Reserve},
  url = {https://whenthefedspeaks.com/doc/speech_20061016_poole},
  note = {Retrieved via When the Fed Speaks corpus}
}