IMF-能源補貼能幫助抑制通貨膨脹嗎？（英）

CanEnergySubsidiesHelpSlayInflation?

ChristopherErceg,MarcinKolasa,JesperLinde,

andAndreaPescatori

WP/24/81

IMFWorkingPapersdescriberesearchin

progressbytheauthor(s)andarepublishedto

elicitcommentsandtoencouragedebate.

TheviewsexpressedinIMFWorkingPapersare

thoseoftheauthor(s)anddonotnecessarily

representtheviewsoftheIMF,itsExecutiveBoard,

orIMFmanagement.

2024

APR

N

Ann

R

INTERNATIONALMONETARYFUNDii

?2024InternationalMonetaryFund

WP/24/81

IMFWorkingPaper

MonetaryandCapitalMarketsDepartment

ResearchDepartment

CanEnergySubsidiesHelpSlayInflation?*

ChristopherErceg,MarcinKolasa,JesperLinde,AndreaPescatori

AuthorizedfordistributionbyChristopherErceg

April2024

IMFWorkingPapersdescriberesearchinprogressbytheauthor(s)andarepublishedtoelicit

commentsandtoencouragedebate.TheviewsexpressedinIMFWorkingPapersarethoseofthe

author(s)anddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoard,orIMFmanagement.

ABSTRACT:Manycountrieshaveusedenergysubsidiestocushiontheeffectsofhighenergypricesonhouseholdsandfirms.AfterdocumentingthetransmissionofoilsupplyshocksempiricallyintheUnitedStatesandtheEuroArea,weuseaNewKeynesianmodelingframeworktostudytheconditionsunderwhichthesepoliciescancurbinflation.Wefirstconsideraclosedeconomymodeltoshowthataconsumersubsidymaybecounterproductive,especiallyasaninflation-fightingtool,whenappliedgloballyorinasegmentedmarket,atleastunderempiricallyplausibleconditionsaboutwage-setting.Wefindmorescopeforenergysubsidiestoreducecoreinflationandstimulatedemandifintroducedbyasmallgroupofcountrieswhichcollectivelydonothavemuchinfluenceonglobalenergyprices.However,theconditionsunderwhichconsumerenergysubsidiesreduceinflationarestillquiterestrictive,andthistypeofpolicymaywellbecounterproductiveiftheresultingincreaseinexternaldebtishighenoughtotriggersizeableexchangeratedepreciation.Sucheffectsaremorelikelyinemergingmarketswithshallowforeignexchangemarkets.Iftheprimarygoalofusingfiscalmeasuresinresponsetospikesinenergypricesistoshieldvulnerablehouseholds,thentargetedtransfersaremuchmoreefficientastheyachievetheirgoalsatlowerfiscalcostandtransmitlesstocoreinflation.

JELClassificationNumbers:E32,E52,E64,H23,Q43

Keywords:

EnergyPrices;EnergySubsidies;MonetaryPolicy;InternationalSpillovers

Author’sE-MailAddress:

CErceg@IMF.org,MKolasa@IMF.org,JLinde@IMF.org,andAPescatori@IMF.org

*WethankTianchuQiandKailiChenforexcellentresearchassistanceandTobiasAdrian,JaebinAhn,GiovanniDell'Ariccia,Pierre-OlivierGourinchas,ZoltanJakab,andMathiasTrabandtforusefulcommentsanddiscussions.ThepaperalsobenefitedfromcommentsfromtheparticipantstotheCEPRParisSymposium,ESSIMinTarragona,NBU-NBPARCinCracow,CEFinNice,SummerWorkshoponEconomicsandFinanceinWarsaw,andseminarsattheEuropeanCentralBank,Bancad'Italia,andUniversityofOttawa.TheviewsexpressedhereareoursanddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoard,orIMFmanagement.

1

1Introduction

Centralbankshavetightenedmonetarypolicyaggressivelytocombathighinflation,whichsurgedtomulti-decadepeaksinmanyeconomies.Whiletheinflationrunuphasbeendrivenbyanumberoffactors,includingsupplydisruptionsandmassivefiscalandmonetarystimu-lus,sharpincreasesinenergypriceshavebeenamajorcatalyst.Manycountriesrespondedbyusingvariousformsofenergysubsidies,oftentargetedmainlytohouseholds,tohelpcurb

inflationaswellastocushiontheeffectsonhouseholdpurchasingpower(Arreguietal.

,

2022

;

Sgaravattietal.

,

2023).Energysubsidiesareviewedbymanypolicymakersashelpful

inthefightagainstinflationonthegroundsthattheyarelikelytolimitbothfirstandsecondroundeffects,includingbyreducingpassthroughtowages.

Buttherearemanyopenquestionsabouttheconditionsunderwhichenergysubsidiesareactuallylikelytobeeffectiveinloweringinflation,or,conversely,whentheymaybecoun-terproductive.Whilewagestickinesshaslongbeenrecognizedaspivotalinthetransmissionofenergyshockstooverallpriceinflation(e.g.,

BlanchardandGali

,

2007),therehasbeen

littleanalysisofhowtheeffectsofenergysubsidiesoninflationdependonthemodalitiesofwage-setting,includingthedegreeofindexation.Similarly,itisunclearhowtheeffectsofsubsidiesdependonthestructureoftheenergymarket:couldinsulatingconsumersfromtheeffectsofhigherpricesinahighlysegmentedmarket–suchastheonefornaturalgasinEurope–hurtfirmsenoughtocausecoreinflationtorise?Andhowdotheeffectsofsubsidiesdependonwhethertheyareusedbyasingle“smalleconomy”orabroadsetofcountries?

Inthispaper,weuseaNewKeynesianmodelingframeworktohelpanalyzetheseques-tions,butfirstseekempiricalgroundingbyestimatinghowoilshockstransmittotheeconomyintheUnitedStates(US)andeuroarea(EA).Specifically,weuseaproxySVARapproachwheretheinstrumentisgivenbysurprisechangesinoilpricesfollowingtheannouncementofOPECmeetingdecisions.Theseriesofsurprisesisadoptedfrom

Ka¨nzig

(2021)andaggre

-gatedtoquarterlyfrequencyoverthe1983to2019sampleperiod.Fortheseidentifiedshocks,wefindthatanegativeoilsupplyshockwhichdrivesupthepriceofoilleadstohighercoreandheadlineinflation,lowerGDP,andhigherpolicyratesinitiallyinboththeUSandtheeuroarea.Importantly,wefindsignificantdifferencesinthetransmissiontowagesbetweenthesetwoeconomies,withoilpricehikeshavingsmalleffectsonnominalwagesintheUS–sothatrealwagesfallnotably–whereastheytendtoboostnominalwagesenoughintheeuroareathatthereal(product)wageactuallyrises.Thesedifferencesinwagetransmissionturnouttohavecriticalbearingonhowenergysubsidiesaffectinflation.

Turningtoourtheoreticalframework,webeginbyconsideringasimpleworkhorseclosedeconomymodelwithnominalwageandpricestickinessthatbuildson

Bodensteinetal.

(2011)

.Asinthatmodel,weassumetheeconomyhasanendowmentofenergythatisutilizedbothinproductionandconsumption,butenrichthesupplyblockbyallowingfor

2

alternativeformsofdynamicwageindexationwherewagesmayadjusttoeithercoreorheadlineinflation.Wealsoincorporateenergysubsidiesthatareproportionaltoenergyuse,thusresemblingthedominanttypeofsubsidiesintroducedinmanycountries,includingtheUSandEurope.

Usingthismodel,wedemonstrateformallythatprovidingenergysubsidiestoconsumersisessentiallyequivalenttotaxingenergyusebyfirms:bothpoliciesdecreasetheeffectiveenergycosttohouseholdswhileincreasingthatfacedbyfirms,thusresultinginthesameallocations

.1

Theimplicationisthatwhileconsumerenergysubsidiescanbedesignedtoshieldhouseholdsfromariseintherelativepriceofenergy(sayduetoasupplyshock),suchapolicyisessentiallyataxonfirmsthatdrivesuptheirmarginalcostsofproduction.

Weshowhowalog-linearizedversionofthemodelwithenergycanbeexpressedinaformisomorphictothatof

Ercegetal.

(2000),andusethisformulationtoderiveastriking

analyticalresult.Inparticular,weshowthatenergysubsidiestoconsumersnecessarilyboostcorepriceinflationifmonetarypolicykeepsemploymentatpotentialandwagesareindexedtocoreinflation(eitherfullyorpartially,includingthespecialcaseofnoindexation).Thisresultholdsforanyconfigurationofparametersassumingthatveryweakconditionsaresat-isfied(i.e.,implyingthatadverseenergyshocksreducethepotentialrealwage).Intuitively,whileconsumersubsidiesdamptheinitialriseinheadlineinflation,thesubsidiesdriveuptheenergypricesfacedbyproducers,whichtheypassalonggraduallythroughraisingproductprices.Hence,thesesubsidiesmakeinflation–bothcoreandheadline–morepersistent.

Theclosedeconomymodelisrelevantforconsideringthelikelyeffectsofconsumerenergysubsidiesintheglobaleconomywhenallcountriespursuesimilarpolicies,orinahighlysegmentedenergymarket(suchasfornaturalgasinEurope).Inshort,wefindusingquantitativesimulationsofourmodelthatevenahighdegreeofindexationtoheadlineconsumerpriceinflation–wellabovewhatwefindempiricallyfortheUS,oreventheeuroarea–isnotenoughforconsumerenergysubsidiestoreducecoreinflation.Thesesubsidiesindeedreducefirst-roundeffectsonheadlineinflationinresponsetoanenergypricespikeandthisreducesupwardpressureonwages,buttheyalsocausethepre-subsidyenergypricefacedbyfirmstoriseenoughthatthe“all-in”effectisariseincoreinflation.Welfareanalysisusingtheusualutility-basedmetricconfirmsthatconsumerenergysubsidieslowerwelfare.

Whileourworkhorsemodelcastsstrongdoubtontheabilityofconsumersubsidiestolowerinflationwhenappliedglobally,itstillmightbeexpectedthatasmallopeneconomy–orgroupofcountriestoosmalltoaffectglobalenergyprices–mightbeabletousesubsidiestolowerinflation.Toexplorethis,weuseatwo-countryextensionwithincompleteassetmarkets.Thisvariantallowsforthetransmissionofenergysupplyshockstodiffersubstantiallyacrossenergy-importersandexporters,andalsodependonthedepthofthe

1Thesealternativeapproachesdohavedifferentimplicationsforthepre-subsidyenergyprice,withtheconsumersubsidydrivinguptheenergyprice,andthetaxonfirmsdrivingitdown(hence,allocationswouldbeaffectedinasettingwithendogenousoilproduction).

3

foreignexchangemarketwhichheavilyinfluencestheexchangerateresponse.

Focusingonthesmallopeneconomycase,wefindmorescopeforconsumerenergysubsidiestoreducecoreinflationthanintheclosedeconomymodel.Thisseemsintuitive:suchsubsidiesshould“chokeoff”firstroundeffectsandthuslimitthescopeforsecondroundeffectstomaterialize,includingthroughwageindexation.Andfirmsshouldnotbe“bitten”byhigherglobalenergypricesgiventhehomecountry’ssmallshareintheglobalenergymarket.

Evenso,theconditionsunderwhichsubsidiesreducecoreinflationstillturnouttobesurprisinglyrestrictive.Althoughsubsidieshaveessentiallynoeffectonglobalenergyprices,theycausethetradebalancetodeteriorateandexchangeratetodepreciate.Thisexchangeratechannelcanbebigenoughthatsubsidiescancausecoreinflationtorise,unlesswageindexationtoconsumerpricesissufficientlyhigh.

2

Thedepreciationeffectislargerthemorepersistentthesubsidy(whichtypicallyislinkedtothepersistenceoftheunderlyingshocktoenergysupply)andislargerineconomieswithshallowforeignexchangemarkets,asisthecaseinmanyemergingmarketeconomies.

Thepolicyimplicationisthatconsumerenergysubsidiesmayhelpreducecoreinflationifimplementedbycountriesthatarearelativelysmallshareoftheglobalenergymarket;ifthesubsidiesaretemporary;ifwageindexationtoconsumerpricesishigh(sothatsecondroundeffectswouldbelargeintheabsenceofsubsidies);andiftheactionsaretakenbyeconomieswithwell-developedfinancialmarketsandmodestdebtlevels.Theseconditionsalignwellwiththosethat

Daoetal.

(2023)foundinthecaseoftheeuroarea,wheretheyarguethat

consumerenergysubsidieshelpedmitigatebothheadlineandcoreinflationpressures

.3

Butasforeshadowedbytheclosedeconomyresults,ourtwocountrymodelhighlightshowconsumersubsidiesarelesslikelytobeeffectiveinreducinginflationastheirusebecomesmorewidespread.Thepre-subsidyenergypricerisesbymore,makingitmorelikelythatsubsidiesboostinflation.Moreover,ourmodelhighlightstheadversespilloverstootheren-ergyimporters–thelatterexperiencebiggertradedeficits,currencydepreciation,andupsideinflationpressure.Thespilloversareparticularlyacuteforvulnerableemergingeconomieswiththinforeignexchangemarkets.

Whileourfocusislargelyonaggregatesupply,consumerenergysubsidiesalsohavedemandsideeffectsthatmayvarysignificantlyacrosshouseholds,andwhicharenotwell-capturedbyourframework.Indeed,ourwelfareresultsinparticularmaybeviewedasundulybiasedagainstsuchsubsidiesinsofarthemodeldoesnotallowforhowsubsidiesmayshieldpoorerandliquidity-constrainedhouseholdsfrommajorhitstotheirpurchasingpowerandconsumptionarisingfromenergyshocks–whichwasclearlyakeyaimofpolicymakers

2Energysubsidiestendtobehelpfulundertheseconditionsbychokingofffirst-roundeffectsthatwouldotherwiseexertlargepassthroughtocore.

3Wealsoshowthatapackageofenergysubsidiestofirmsaswellashouseholdsmaybeattractiveforasmall“energy-pricetaking”economy–supportingaggregatedemandwhileloweringinflation.

inadditiontorestrainingoverallinflation.

Toexplorethis,weincorporatehand-to-mouthhouseholdsintoourclosedeconomymodelasinthetwo-agentNewKeynesian(TANK)setupdevelopedby

Galietal.

(2007)and

Bilbiie

(2008),andshowthatourresultsarelargelyrobustevenwhencapturingthedemand-sideef

-fectsofenergyshocks.

4

Interestingly,weshowthatconsumerenergysubsidiescouldimprovewelfareinourTANKframeworkbyreducingconsumptioninequalityacrosshouseholds(i.e.,permanentincomeversushand-to-mouth).However,targetedtransferstohand-to-mouthhouseholdsaremoreefficientastheycanachievethisgoalatlowerfiscalcostandwithabetterinflation-outputtradeoff.Themainadvantageofsuchtransfersoverenergysubsidiesisthattheydonotdistortrelativeenergypricesfacedbyhouseholds,andhencepreservetheirincentivetoconserveenergy.Thislimitstheincreaseinthepre-subsidyenergypricefacedbyfirms,andhencelowersinflationpressure.

Ourpaperiscloselyrelatedtoaninsightfulrecentpaperby

Auclertetal.

(2023)which

exploresthetransmissionofoilshocksinanopeneconomyHANKmodel,andtheeffectsofconsumerenergysubsidies.Theseauthorsalsoshowhowsubsidiesmaybebeneficialinreducinginflationandboostingconsumptionforanindividual“small”country,butcanhavecounterproductiveeffectsifdeployedbyabroadsetofcountries.Ourpaperdiffersfromtheirworkinseveralimportantaspects.First,theirpaperfocusesheavilyonheterogeneityacrosshouseholdsinaHANKframeworkandassociateddistributionaleffects,whereasourstudydevotesmuchmoreattentiontothesupplyside(differentwageindexationschemesratherthanassumingfullindexationofwagestoconsumerprices).Second,weincorporatefeaturesaimedatcapturingfinancialvulnerabilities(suchasthinFXmarkets)thatinfluencehowsubsidiesoperatethroughopeneconomychannels.Third,weassumethecentralbankfollowsoptimalpolicyorusesaTaylor-stylereactionfunctionthatprovidesagoodapproximation.Thisturnsouttohaveimportantimplicationsforthepass-throughofenergypriceshocksandsubsidies.Alltold,ourpaperismorepessimisticintheabilitytouseconsumerenergysubsidiessuccessfullytoreduceinflation,eveninthecaseofasmallgroupofcountries(e.g.,showinghowsubsidiescanbecounterproductivethroughtradechannels).

Ourpaperalsodrawsontheextensiveearlierliteraturestudyingthetransmissionofen-ergypriceshocks.Keyempiricalcontributionsinclude

Hamilton

(1983),

Hamilton

(2003),

BarskyandKilian

(2004),

BlanchardandGali

(2007),

Kilian

(2009),

BaumeisterandHamil-

ton

(2019),and

Ka¨nzig

(2021)

.Weuseenergysupplyshocksidentifiedinthislastpapertoanalyzethereactionofwages.Anumberofpapersincorporateenergyintostructuralmacroeconomicmodels,butdonotstudytheeffectsofenergysubsidies,see

KimandLoun-

gani

(1992),

BackusandCrucini

(2000),

LeducandSill

(2004),

Bodensteinetal.

(2011),

BlanchardandRiggi

(2013),

Bodensteinetal.

(2018),

BaqaeeandFarhi

(2019),

Gagliardone

4DebortoliandGali(2018)arguethatasimpletwo-agentNewKeynesian(TANK)modelcanbea

goodapproximationofheterogeneousagentNewKeynesian(HANK)modelsregardingthetransmissionof

aggregateshockstomacroeconomicvariables.

4

5

andGertler

(2023),and

Pieroni

(2023),amongothers.Similarlyto

Auclertetal.

(2023),the

lastpaperusesaHANKsetupwhileweoptforamoreparsimoniousTANKmodelasanextensiontocapturehouseholdheterogeneityandanalyzethescopefortransfers.BylookingattheinteractionsbetweenpriceandnominalwagerigiditywithinNewKeynesianmodels,ourworkisrelatedtotheliteratureoriginatedby

Blanchard

(1986)anddrawsparticularly

on

Ercegetal.

(2000)and

Bodensteinetal.

(2008).

Arecentcontributioninthislineby

LorenzoniandWerning

(2023)considersboththetransmissionofsupplyanddemandshocks

andoptimalpolicyinacontinuoustimeframework(withinflationinterpretedasarisingfromconflictbetweenfirmsandworkersoverwages).

Thepaperisstructuredasfollows.Section2discussestheempiricalevidenceontheeffectsofenergypriceshocks,withaspecialfocusontheresponseofwages.Section3presentsasimpleclosedeconomymodelwithenergy.Section4offersanalyticalinsightsonthetransmissionofenergypriceshocksandsubsidies.ThesearecomplementedwithquantitativesimulationsdescribedinSection5.Section6presentstheversionofthemodelwithhand-to-mouthagentsandisfollowedbyanopeneconomyextensioninSection7.Section8concludes.

2EmpiricalEvidenceonOilShocksandWageAdjustment

Inthissection,weprovideempiricalevidenceontheresponseofwagesandconsumerpricestooilpriceshocksintheUnitedStatesandtheeuroarea.

First,weidentifyasourceofoilpricemovementsthatisnotcorrelatedwithotherfun-damentalshockstoavoidreversecausalityoraomittedvariablebias.Todoso,werelyon

Ka¨nzig

(2021)whoidentifiessurprisechangesinoilfuturespricesrelatedtoOPEC’s

meetingsdecisionsaboutproductiontargets.AsKnzigmeasuresthesechangesinatight

windowaroundtheannouncements,

5

itislikelythattheycanisolatetheimpactonoilpricesoftheunexpectedcomponentofOPEC’sdeliberations–analogouslytotheliteraturethatidentifiesmonetarypolicyshocks(e.g.,

Gurkaynaketal.,

2005;

GertlerandKaradi,

2015;

NakamuraandSteinsson,

2018,amongothers).SincemarketspaycloseattentiontoOPEC’s

decisions,areversecausalityoftheglobaleconomicoutlookshouldplausiblyberuledoutbecauseitshouldalreadybepricedinatthetimeoftheannouncement.OPEC’sdecisions

arenotrandomeventsbutastheKnzigshocksmeasureonlythecomponentnotanticipated

bymarkets,theirimmediateimpactonpricesshouldreflecttheelementofthedecisionthatsurprisedmarketparticipants

.6

Adistinctadvantageofusingtheseshocksrelativetothe

5Thewindowistakentobeoneday,usingWTIcrudefuturesstartingin1983tradedattheNewYorkMercantileExchange(NYMEX),whichisoneofthemostliquidandlargestmarketforcrudeoil.Forfurtherdetails,see

Ka¨nzig

(2021)

.

6NazerandPescatori

(2022)showthatitisthesurprisecomponentoftheOPEC(orOPEC+)decisions

thataffectsthemarketsincethesamedecisioncaninducepriceresponsesofdifferentsign.

6

standardapproachpioneeredby

Kilian

(2009)istobypassthedifficultiesofdisentangling

oildemandandoilproductionshocks.Evenso,wewillshowthatourresultsalsoholdwhenusingshocksmeasuredby

BaumeisterandHamilton

(2019)

.

Theseriesofoilpricesurprisesspanningfrom1983to2019areaggregatedatquarterlyfrequencyandusedasaninstrumentinanSVARmodel.Inadditiontothestandardoilmarketvariablessuchasrealoilprices(deflatedbyeithercoreorheadlineprices),oilproduction,andglobalindustrialproduction,ourproxySVARmodelalsoincludesrealGDP,coreorheadlineCPIinflation,theshort-termpolicyrate,aswellasnominalwageinflation.Allquantitiesandrelativepricesareinlogs,thepolicyratesinpercentagepoints,whilepriceandwageinflationaremeasuredaslogdifferences.WeuseidenticalSVARspecificationsfortheUSandtheeuroarea.Appendices

A.1

and

A.2

containadditionaldetailsontheproxySVARanddatadefinitions.

OurbaselineresultsarereportedinFigures

1

and

2

fortheUSandeuroarea,respectively.AccordingtotheproxySVARfortheUS,anincreaseinoilpricesdrivenbyanoilsupplynewsshockleadstoadeclineinnominalandrealwages.

7

Moreconcretely,anoilsupply-driven10percentincreaseinoilpriceswitha2-yearhalflifeincreasescoreinflationbyabout0.1percentagepointforafewquartersandreducesthenominalwagebyabout0.2percentonimpact.Consequently,theimpliedproducerrealwage–calculatedbyaccumulatingthenominalwagegrowthimpulsesandsubtractingtheaccumulatedcoreinflationimpulses–fallsbyroughly0.3percentoverthefirst2yearsandbynearly0.6percentafter5years.ThereductioninrealwagesisinlinewithanoilsupplyshocktransmittingsimilarlytoanadverseTFPshock,withrealGDPfallingbymorethat0.4percentafter2years.Cruciallyforourfurtheranalysis,theseresultssuggeststhattheabilityofUShouseholdsandlaborunionstoachievecompensationforhigherpricesthroughhigherwagesisverylimited.

TheimpactontheeuroareaisshowninFigure

2.

WhiletheoilpricepathisalmostidenticaltothatobtainedintheSVARfortheUS,thetransmissiontootherkeymacrovariablesdiffersnotably.Mostnotably,nominalwagegrowthpersistentlyincreases.ThisfindingispossiblydrivenbymorecentralizedwagebargainingandhigherunionizationintheeuroareathanintheUS,leadingtoahigherdegreeof(explicitorimplicit)indexation

(BabaandLee,

2022)

.IncreasingwagesputsomeupwardpressureoncoreinflationandsotheECBraisestheinterestratesignificantly.ThetighterpolicystanceimpliesthattheassociatedoutputcontractionisaboutthesameasintheUS,althoughtheeuroareaeconomyislessdependentonoil.Turningtotheimpliedrealproducerwage,thepointestimateofitsimpulseresponseismodestlypositive,butnotsignificantlyso.WeinterpretthisfindingasonlymildevidenceinfavorofsomeindexationmechanismsinEuropeanwagesettingthatcompensateshouseholdsforthelossofpurchasingpowerduetoenergypriceshocks.

7Theconfidencebandsarecomputedusingamovingblockbootstrap,asproposedby

JentschandLunsford

(2019).Thismethodproducesasymptoticallyvalidconfidencebandsunderfairlymildalpha-mixingcondi

-tions.Theblocksizeissetto24anddealswiththedifferenceintheestimationandidentificationsamples.

7

Figure1:USProxy-VAREvidence:OilPriceShocks

Notes:Allvariablesareexpressedinpercentdeviations,realoilpricesaredeflatedwithcoreCPI,impliedrealwageisdefinedasaccumulatednominalwageinflationminuscoreinflation.Theconfidencebandsarethe68and90percentlevelbandsbasedon10,000bootstrapreplications.

8

Figure2:EAProxy-VAREvidence:OilPriceShocks

Notes:Allvariablesareexpressedinpercentdeviations,realoilpricesaredeflatedwithcoreCPI,impliedrealwageisdefinedasaccumulatednominalwageinflationminuscoreinflation.Theconfidencebandsarethe68and90percentlevelbandsbasedon10,000bootstrapreplications.

AsshowninAppendix

A.3

,theimpactofanoilsupplyshockonheadlineinflationisnotablystrongeronimpactinboththeUSandtheeuroarea,upto1(0.5)percentinthe

9

samequarterbutmuchlesspersistent–consistentlywithoilpricesrevertingtotheirpre-shocklevels.Evenso,theimpliedconsumerrealwage,i.e.nominalwagedividedbyheadlineCPI,fallsnotablymoreinitiallythantheproducerrealwageinFigure

1

.Thesmallerpass-throughtoheadlineinflationrelativetocoreinflationintheeuroareathanintheUScanbeatleastpartiallyexplainedbyhigherexcisetaxesonliquidfuelsandadownstreamoil

industrythattendstohavemorerigidretailprices(Ahn,

2024).Importantlyforourfurther

analysis,regardlessifweworkwithcoreorheadlineinflationintheSVAR,wefindthattheimpliedrealwageintheeuroareafallslessthanintheUS.

WehaveexaminedtherobustnessofourresultsforanumberofalternativeSVARspec-ifications.Figure

3

presentstheimpulseresponsesfortheimpliedproducerrealwage(USintheleftcolumnandeuroareaintheright),whileAppendix

A.3

reportstheresultsforallvariablesincludedineachoftheconsideredalternatives.ThefirstrowinthefigurejustrestatesourbaselineresultsfromFigures

1

and

2.

Thesecondrowreportsanalternativewhereweimposemeanreversionoftherealwageimpulseresponsesbyintroducingthisvari-abledirectlyintheSVARinsteadofnominalwagegrowthandimposingstabilityconditions.Ascanbeseen,realproducerwagesintheUSstilldeclinesignificantlyinresponsetoanenergysupplyshockwhiletheyrisesomewhat(althoughnotsignificantlyatthe90percentconfidencelevel)inEurope.

Wealsotestedforpossiblenon-linearitiesintracingtheeffectsofoilpriceshoc