Regional President Speech

TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR DISCUSSION OF ELLISON AND SARGENT: WHAT QUESTIONS ARE STAFF AND FOMC FORECASTS SUPPOSED TO ANSWER? JamesBullard* PresidentandCEO FederalReserveBankofSt. Louis 10thEABCNWorkshoponUncertaintyovertheBusinessCycle EuropeanCentralBank,Frankfurt March30,2009 AnyopinionsexpressedherearemineanddonotnecessarilyreflectthoseofotherFederalOpenMarketCommitteeemembers. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR SARGENT’S ADDRESS IN SEOUL MotivatedbyaspeechbyChairmanBernankeinOctober2007. Bernanke’sspeechwasabouthowpolicymakersmightthink aboutmodeluncertainty: Bayesianversusrobustcontrol methods. Policymakersthatcanplaceapriorovercompetingmodelsare Bayesian. Policymakersthatcannotinsteadusemultiplepriorsandrobust control,usingaworst-caseanalysistoputboundson performance. Relevant,ortechno-babble? ParaphrasingGreenspan:Uncertaintyisthedefiningfeatureofthe economiclandscape. Re-analyzeacommentDavidandChristinaRomeronFOMC forecasting. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR MODEL UNCERTAINTY ThelevelofmodeluncertaintyfacedbytheFOMCisprofound. Manyofthe“bigideas”fromacademiaaresimplynotpartofthe analysis. Thisleavesonewonderinghowimportanttheseomissionsare. Examples:Outputgapmeasures,creditmarketfrictions, multiplicityofequilibria. Isitrighttobesofaroffoftheresearchfrontier? Idonotthinkwewouldallowitinthespaceprogramorinthe designofweaponssystems. Oneimportantdefense: forecasting. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR FORECASTING IS TRACKING Mostofthefocusinpolicydiscussionsconcernstoday’sstate vector. Furtheroutisnormallyslowmeanreversion. Throughexperience,forecasterslearnedthatthenearrandom walkmodelworksbest. Trackingrequirescarefulattentiontoincomingdataand revisionstorecentdata. Whatwewouldreallylikeistoaccuratelypredictthe consequencesofanunparalleledpolicyaction. Thisisthebigticketitem. Forthiswewouldneedmuchbettermodelsthanwehave. Result:FOMCmustmakeitsownjudgementsfrommodelswhich areplainlydeficientinmanyrespects. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR WHAT THE FOMC DOES TheFOMCmembersmakequarterlyforecastswhichare summarizedandpublished. Thestaff“Greenbook"forecastisreleasedtothepubliconlyafter fiveyears. OnepurposeoftheFOMCforecastsistocommunicatetothe public,especiallyconcerningmediumrunexpectations. Thisaloneisvaluable. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR MORE ON WHAT THE FOMC DOES TheFOMCmembersforecastsaremade"underappropriate monetarypolicy." TheGreenbookforecastnecessarilycontainsapolicy assumption. Formanyyearsthiswasaconstantornearlyconstantpolicyrate. Greenbookforecastisheavilyjudgementally-adjusted. Ifmembersfeelthatactualpolicywoulddiffermateriallyfrom theassumption,theyshouldturninadifferentforecast. Also,shouldasinglemembertrytopredictwhattheCommittee willdo? Ifso,onemightforecastpoormacroperformancebecauseofalack offaithintheCommittee. Ifnot,onemightforecastwhatmayachievableiftheCommittee followedthefullyoptimalpolicy. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR DOES THE FOMC EXERCISE MAKE SENSE? Notasapureforecastingexercise. Forthat,onewouldhavetoassume,likePrescott,thattheactual pathofmonetarypolicyhasaverylimitedimpacton macroeconomicoutcomes. ThisisprobablytheoppositeviewfromRomerandRomer. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR WHAT ROMER AND ROMER DO RomerandRomer(AERPapersandProceedings,2008)compare forecastsoftheCommitteeandthestaff. TheyfindthattheCommittee’sforecastsdonotimproveupon theGreenbook. TheyfindthattheCommitteesometimesactsondifferencesin theforecast. Thissuggeststhatthestaffforecastshouldbeworse,sincethe staffis"surprised"bytherenegadeactionsoftheCommittee. Thisdoesnotseemtohavehappened. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR DOES THE ROMERS’ EXERCISE MAKE SENSE? No. The"appropriatepolicyclause"arguablymakesthequestion: Whatcanbeachievedinthecurrentcircumstances? Sinceforecastingistrackingandcontainslittleinformation, perhapsnotsointeresting. TheequilibriumisthatthestaffandtheCommitteethinkexactly alike. Butthemodelsinuseareplainlymissingmanyelementsthat maybeveryimportantatcertainjunctures. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR WHAT ELLISON AND SARGENT DO TheCommitteemaybebetterinterpretedasrespondingto specificationdoubts. DoubtsareabundantontheCommittee. Thenmembersdoaworst-caseanalysisovermultiplepriors. Thisbiasestheforecaststowardworst-caseoutcomes. Question: Whydothis? Tosignalothermembersabouta potentialpitfall? TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR MORE ON WHAT ELLISON AND SARGENT DO Themodel: Primiceri(QJE2006),unobservedNAIRU. Comment: thelevelofmodeluncertaintyisfarmoreprofound thanthedetailsofthismodelwouldsuggest. Staffhastheapproximatingmodel. FOMChaspreferencesforrobustness. Question: ThestaffworksfortheFOMC,shouldtheyalsohave preferencesforrobustness? Question: Learningandrobustcontrol—afteralongenough timeperiod,policymakerswouldlearnthatpessimismdoesnot payoff. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR DOES THE ELLISON AND SARGENT EXERCISE MAKE SENSE? ItmakesmoresensethanRomerandRomer. Itallowspolicymakerstohaveaconcernforrobustnesswhich differsfromthestaffconcerns. Thisconcerncoulddriveasystematicdifferencebetweenthe forecastsoftheCommitteeandthestaff. Still,thestaffcouldbeaskedtocalculateforecastswithaconcern forrobustness. And,thisinterpretationblursthecommunicationroleoftimely publicationofFOMCforecastsbutnotstaffforecasts. Levelofmodeluncertaintyfargreaterthansuggestedhere. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR POLICYMAKING WITH A CONCERN FOR ROBUSTNESS Trackingwillremainimportant. Trackingincorporatesaveragepastpolicybehavior. Forecastingtheeffectsofimportantpolicyinterventionsis critical. Forthatweneedbettermodels. Currentforecasting/trackingdoesnottellusmuchandis unlikelytogetbetter. TRUTHANDPROBABILITY SOMEPROBLEMSWITHROMERANDROMER ASUGGESTEDDIVISIONOFLABOR SOME QUESTIONS I WOULD LIKE TO BE ABLE TO ANSWER Whatarethechancesthattheglobaleconomycouldbecome entrenchedinthedeflationarytrapequilibriumdescribedby Benhabib,Schmitt-GroheandUribe(2001)? Whatistheoptimalregulatoryresponseifweinterpretprivate informationtobeattheheartofthecurrentworldwidefinancial turmoil? Towhatextentcanglobalcoordinationimproveonworldwide equilibriumoutcomesinthecurrentenvironment? IfIputsomeweightonafiscaltheoryofthepricelevel,shouldI beconcernedaboutinflationaryprospectsgoingforward? Thesearethebigticketitems. Togetgoodanswers,weneed goodmodels.

James Bullard (2009, March 29). Regional President Speech. Speeches, Federal Reserve. https://whenthefedspeaks.com/doc/regional_speeche_20090330_james_bullard

@misc{wtfs_regional_speeche_20090330_james_bullard,
  author = {James Bullard},
  title = {Regional President Speech},
  year = {2009},
  month = {Mar},
  howpublished = {Speeches, Federal Reserve},
  url = {https://whenthefedspeaks.com/doc/regional_speeche_20090330_james_bullard},
  note = {Retrieved via When the Fed Speaks corpus}
}