Philippe Andrade Banque de France CREM

Philippe Andrade

Description

Niveau: Supérieur, Doctorat, Bac+8
Inattentive professional forecasters? Philippe Andrade Banque de France & CREM Herve Le Bihan Banque de France October 2010 Abstract We use the ECB Survey of Professional Forecasters to characterize the dynamics of expec- tations at the micro level. We find that forecasters (i) have predictable forecast errors; (ii) disagree; (iii) fail to systematically update their forecasts in the wake of new information and disagree even when updating; (iv) differ in their frequency of updating and forecast perfor- mances. We argue that these micro data facts are qualitatively in line with recent models in which expectations are formed by inattentive agents. However building and estimating an expec- tation model that features two types of inattention, namely sticky information a la Mankiw-Reis and noisy information a la Sims, we cannot quantitatively generate the error and disagreement that are observed in the SPF data. The rejection is mainly due to the fact that professionals' forecasts are too sluggish compared to the ones our inattention model generates. Keywords: Expectations, imperfect information, disagreement, inattention, business cycle ?We thank our discussant Ernesto Pasten as well as Carlos Carvalho, Olivier Coibion, Anil Kashyap, Noburo Kiyotaki, Bartosz Mackowiak, Juan Pablo Nicolini, Giorgio Primiceri, Sergio Rebelo, Jonathan Willis, Alexander Wolman, Michael Woodford, Tao Zha and seminar participants at the Banque de France, ECB, New-York Fed, Philadelphia Fed, San-Francisco Fed, University Paris 1 and the ESEM 2009, AFSE 2010 and SED 2010 conferences for useful comments.

• lower than


• aggregate forecasting


• disagreement among


• experts disagree


• forecasting macroeconomic variables


• inattention


• professional forecasters

Sujets

• Nicolini
• Banque de France
• Le Bihan
• Andrade
• Woodford
• Carvalho
• Rubene (floorball club)
• Rebelo

Informations

• Publié par: profil-zyak-2012
• Nombre de lectures: 36
• Langue: English

Extrait

Inattentiveprofessionalforecasters ∗ PhilippeAndradeHerve´LeBihan BanquedeFrance&CREMBanquedeFrance October2010 Abstract WeusetheECBSurveyofProfessionalForecasterstocharacterizethedynamicsofexpec- tationsatthemicrolevel.Wefindthatforecasters(i)havepredictableforecasterrors;(ii) disagree;(iii)failtosystematicallyupdatetheirforecastsinthewakeofnewinformationand disagreeevenwhenupdating;(iv)differintheirfrequencyofupdatingandforecastperfor- mances.Wearguethatthesemicrodatafactsarequalitativelyinlinewithrecentmodelsin whichexpectationsareformedbyinattentiveagents.Howeverbuildingandestimatinganexpec- tationmodelthatfeaturestwotypesofinattention,namelystickyinformationa`laMankiw-Reis andnoisyinformationa`laSims,wecannotquantitativelygeneratetheerroranddisagreement thatareobservedintheSPFdata.Therejectionismainlyduetothefactthatprofessionals’ forecastsaretoosluggishcomparedtotheonesourinattentionmodelgenerates. Keywords:Expectations,imperfectinformation,disagreement,inattention,businesscycle ∗ WethankourdiscussantErnestoPaste´naswellasCarlosCarvalho,OlivierCoibion,AnilKashyap,Noburo Kiyotaki,BartoszMac´kowiak,JuanPabloNicolini,GiorgioPrimiceri,SergioRebelo,JonathanWillis,Alexander Wolman,MichaelWoodford,TaoZhaandseminarparticipantsattheBanquedeFrance,ECB,New-YorkFed, PhiladelphiaFed,San-FranciscoFed,UniversityParis1andtheESEM2009,AFSE2010andSED2010conferences forusefulcomments.Allremainingerrorsareours.WearealsogratefultoSylvieTarrieuforsuperbresearchassistance astoClaudiaMarchiniandIevaRubenefortheirhelpwiththeSPFdata.Thispaperdoesnotreflectnecessarilythe viewsoftheBanquedeFrance.e-mails: [e-mail] , [e-mail] 1

1Introduction Modelsinwhichimperfectinformationandtheformationofexpectationsactasatransmission mechanismofeconomicfluctuations—inthespiritofFriedman(1968),Phelps(1968)andLucas (1972)—haverecentlyregainedinterestinthemacroeconomicliterature. 1 Severalauthors 2 relate imperfectinformationtotheinattentionofagentstonewinformation,abehaviorthatcanbe rationalizedbycostlyaccesstoinformationandlimitedprocessingcapacities.Oneappealofthese modelsistoprovideanalternativechanneltostickypricestoexplainthepersistenteffectsof transitoryshocks,andinparticularmonetaryshocks,ontheeconomy.Moreover,thisapproach canparsimoniouslyaccountforpatternsofindividualexpectationsobservedinsurveydatathat areatoddswiththestandardperfectinformationrationalexpectationframework,namelythat forecasterrorsarepredictableandforecastsdifferacrossforecasters. 3 Inthispaper,weexploitthepaneldimensionofsuchasurveyofforecasts,namelytheECB SurveyofProfessionalForecasters(SPF),toproducemicrofactscharacterizingtheformationof expectations.Wethenelaborateonthosecharacteristicstoassesswhethermodelsofinattention accuratelydescribethebehaviorofforecastersandthusmaycontributetoabetterunderstanding ofbusinesscyclefluctuations.Tobeconsistentwiththerecentliteratureweconsidertwotypes ofinattentionmodels.Ontheonehand,stickyinformationmodelsdevelopedbyMankiw&Reis (2002)andReis(2006)whereagentsupdatetheirinformationsetinfrequentlybutgetperfect informationoncetheydo.OntheotherhandnoisyinformationmodelsproposedbyWoodford (2002),Sims(2003)andMac´kowiak&Wiederholt(2009)inwhichagentscontinuouslyupdatetheir informationbuthaveanimperfectaccesstoitateachperiod. TheECB-SPFisaquarterlypanelstartingin1999surveyingaround90forecastingunitsineither publicorprivateinstitutions.Professionalforecastersmaynotberepresentativeoflesssophisti- catedagents,sinceprofessionalsobviouslyallocatesubstantiallymoretime,human,collectingand computingresourcestothetaskofforecastingmacroeconomicvariables.Howeverprofessionals’ opinionhasbeenshowntospreadtoothertypesofagentsandthereforeinfluenceexpectations 1 See,amongothers,Woodford(2002),Carroll(2003),Hellwig&Veldkamp(2008)andLorenzoni(2009).Veldkamp (2009)andMankiw&Reis(2010)providesurveys. 2 3 Mankiw&Reis(2002),Sims(2003),Reis(2006)andMac´kowiak&Wiederholt(2009). SeeforexampleMankiw,Reis&Wolfers(2003). 2

anddecisionsoffirmsandhouseholds(Carroll,2003).Furthermorewemayexpectprofessional forecasterstobetheagentsinthebestpositiontopayattentiontotherelevantmacroeconomic information.Asaresult,theextentofinattentiontonewsamongprofessionalforecasterscanbe seenasalowerboundforotheragents’inattentiontoaggregateconditions. WehighlightfourmaincategoriesofmicrofactsfromtheseSPFdatathatweargueareconsistent withprofessionalforecastersbehavingasiftheywereinattentive.First,theforecastsofexperts exhibitpredictableerrorsandsystematicbias.Second,expertsdisagreeastheyreportdifferent predictionsforthesamevariableatthesameforecastperiod.Moreover,thedisagreementbetween forecastersevolvesovertime.Third,agentsdonotsystematicallyupdatetheirforecastseven whennewinformationisreleased.Furthermore,theforecasterswhoupdatealsodisagreeontheir forecast.Fourth,thefrequencyofupdatingaforecast,theaverageforecasterrorandtherevision offorecastvaryacrossindividuals.However,thereisnoclearlyinterpretablecorrelationbetween theseindividualcharacteristics. Toourknowledge,thepresentstudyisthefirsttodocumentinfrequentupdatinginsurveyforecasts. TheoriginalityofourapproachistoexploitthefactthattheEuropeanSPFprovidessequences ofindividualforecastsforthesameevent(variableanddate).Wecanthereforeconstructadirect micro-dataestimateofthefrequencyofnon-updatingaforecast.Theresultsshowthat,onaverage, eachquarter25%professionalforecastersfailtoupdatetheir1-yearor2-yearforecastsdespitethe macroeconomicenvironmentevolves.Thisfrequencyhasastructuralinterpretation:itcorresponds tothedegreeofinattention,akeyparameterinasticky-informationtypeofmodel.Furthermore, identifyingforecastersthatupdatetheirforecasts,wealsouncoverthattheytoodisagree.Thus thelackofinformationupdatingisnotthesoleresponsiblefordisagreementamongexperts.Even whenupdatingtheirforecasts,theydonothaveaccesstothesameinformation.Thisresultisin linewiththepredictionsofanoisy-informationmodel.Lastly,theindividualdimensionofthedata allowsustoanalyzethecrosssectiondistributionofthedegreeofinattentionandrevealsthatour resultsarenotdrivenbyaspecificgroupofprofessionalforecasters. Wethenturntoaformalempiricalassessmentofinattentionmodels.Moreprecisely,weargue thatthepreviousresultsqualitativelysupportamodelfeaturingtwotypesofinattentionnamely sticky-informationa`laMankiw-Reisandnoisy-informationa`laSims.Wethereforedevelopsuch 3

anexpectationmodelandthenuseitalongwiththeSPFdatatocarryoutaminimumdistance estimation.Wefindthatthisinattentionmodelfailstoquantitativelyreproducetheobserved persistenceoftheaverageforecastingerrorstogetherwiththerelativelysmalldisagreementbetween forecasters.Moreover,thesmoothnessobservedintheaverageSPFforecastswouldrequireamuch higherinattentiondegreethanourmicrodataestimates.Suchahighinattentionwouldinturnlead tomuchmoredisagreementthanobservedintheSPFdata.Therefore,elementsothersthanthe mereinattentionincludedinourexpectationmodelareneededtoreconcileboththerelativelylow disagreementamongprofessionalsandtherelativelyhighpersistenceoftheaggregateforecasting error. Ourpaperrelatestothevastliterature,mostlyrelyingonUSdata,thatstudiesthebehaviorof surveyforecastsandcomparesitwiththeimplicationsoftheoreticalexpectationmodels.Numer- ousstudies(seeThomas,1999andPesaran&Weale,2006forsurveys)foundsystematicaggregate forecasterrorsanddisagreementinthesedata,atoddswiththeperfectinformationrationalexpec- tationframework.WeconfirmtheseresultsforarecentsampleperiodandforEuropeanSPFdata. Wealsocomplementtheseresultswithnewempiricalmicroevidenceonindividualexpectations. OurworkisalsorelatedtoMankiwetal.(2003),Branch(2007),Coibion&Gorodnichenko (2008)orPatton&Timmermann(2009)whorelyonthecharacteristicsofsurveyexpectationsto assessinattentionand,moregenerally,imperfectinformationtheories.Mankiwetal.(2003)and Branch(2007)focusonthecross-sectiondistributionofforecaststocalibratethesticky-information inattentionparameterdiscussedabove.Bycomparison,weunderlinetheimportanceofinvestigat- ingtheconsistencyofthisparametercalibratedvalueswithboththecross-sectionofforecasts andtheaggregateforecasterrors.Furthermore,weimproveontheirapproachbyconsideringa modelwhichcanexplaindisagreementamongforecastersthatupdatetheirinformation.Coibion &Gorodnichenko(2008)lookattheconditionalresponsetovariousstructuralshocksoftheag- gregateerroranddisagreementimpliedbysurveystodisentanglethesticky-informationandthe noisy-informationmodelsofinattention.Theyfindmixsupportinfavorofthetwo,aswedo.We goastepfurthersinceweestimatedirectlyamodelthatfeaturesthetwotypesofinattention. Patton&Timmermann(2009)relyontheevolutionofforecastsoverdifferentforecasthorizonsto stressthatdifferentperceptionsofinformationratherthandifferentinformationsetsarethemain 4