BeenEffectiveAllovertheWorld?
Kurihara,Yutaka
Abstract
Inflationtargetinghasbeenadoptedallovertheworldsincethe1990s.Muchat‑
tentionhasbeenpaidtotheprosandconsofadoptingthispolicy.Countriesthat performedinflationtargetingreceivedgoodeconomicperformances.Sincethe 1990s,centralbankshavebecomeincreasinglyconcernedwithinflationstabiliza‑
tion.However,morefocusshouldbeputonoutputstabilization.Thispaperhas twopurposes:(1)Tocomparethetargetingtonontargetingcountriesininflation andoutput.(2)Toevaluatethecostsofinflationstabilization,inparticularre‑
ducedsacrificeratio.Thispaper・showsthatcountriesthattargetinflationhave successfullyintroducedthepolicyfromtheviewofreducinginflationratesandre‑
ducingmacroeconomicshocksempirically.Also,thesacrificeratiodeclinesfor countriesduringinflationtargetingperiod.
1.INTRODUCTION
About30countrieshaveadoptedinflationtargeting.Countriesthat haveadoptedinflationtargetinghavegenerallyshowngoodeconomic performance,includinglowinflationandstableeconomicgrowth.
Pricestabilityisoneofthemostimportantpointsforcentralbanksre一
*FinancialsupportfromGrant‑in‑AidforScientificResearch(B‑20330055)is
appreciated.
gardlessofwhetherornotinflationtargetingisadopted.Inflationtar°‑
getinghasbeenusedrecentlyinmanycountries.Sincethe1990s,policy makershavebecomeincreasinglyconcernedwithinflationstabilization.
Sectionlexplainsrecenttrendsandinnovationsininflationtargeting.
Section2analyzesonetheoreticalviewandanempiricalmethodforex‑
aminingwhyinflationtargetinghasbeenintroducedandhowithascon‑
tributedtoinflation.Section3performsempiricalanalyses.Section4 analyzeswhethertheintroductionofinflationtargetinghasreducedthe costsofinflationstabilization,theaccumulatedoutputlossduetoaper‑
manentreductionininflation.Finally,thispaperendswithabriefsum‑
mary.
2.ADOPTIONOFINFLATIONTARGETING
Inflationtargetingisapracticeinwhichcentralbankspubliclysetthe targetrateforinflation;monetarypolicyisthencarriedoutaccordingto thistarget(Fountas,Karanasos,andKim,2002).
TheUnitedStateshasnotadoptedityet,althoughtheUSCongress discussedit,especiallyin2002.TheBankofJapanalsodoesnothavea policyofinflationtargeting.TheECBdoesnotallowtheexistenceofin‑
flation:however,inflationtargetingmightbesubstantiallyadopted.
Ininflationtargeting,centralbanksareresponsibleforachievingan announcedobjectivefortheinflationrate.About30centralbanks adoptedthisframeworkfortheconductofmonetarypolicy,andithas showntobeeffectiveinmostcases.SvenssonandWoodford(2005)men‑
tinnedthatcountriesthatadoptinflationtargetingseemtohaveat‑
tainedasignificantreductioninboththerateofinflationandinflation expectations.
Manybenefitscanbeobtainedfromtheintroductionofinflationtar‑
geting.First,therealizationofthecentralbank'sgoalofpricestability
mightnotbejudgedaccuratelyintheabsenceofclearstandards.Second, showingcentralbank'sgoalsandmakingthemmoretransparentguar‑
anteesaccountabilityforthetargetandindependencefromthegovern‑
ment.Third,thisapproachgetsstabilityoftheexpectedinflationrate'.
Finally,inflationincountriesthathaveintroducedthepolicyhasbeen reduced.
Ontheotherhand,manycriticalthinkinghavebeenpresented.First, becausecontrollinginflationusingmoneysupplygrowthorexchange ratemaybelesseffective,thetrustofacommitmenttoinflationtarget‑
ingmaybeunstable.Second,ifmarketparticipantsbelieveandcreditthe target,thereissomepossibilityofincreasesinlong‑terminterestrates, forexample.Third,targetingmaybeattainedatthesacrificeofother importanteconomicfactors.Finally,sincethe1990s,becausemarket authoritieshavenoeffectivemeasuresagainstdeflation,introducingin‑
flationtargetingmayimpactthesuccessofotherpolicies.Moreover, thereisnopreviouscaseinwhichacentralbankintroducedinflationtar‑
getingduringdeflation.
AsRotembergandWoodford(1997)andSvensson(1999)mentioned, thelossfunctionofcentralbankscanbeexpressedasshowninequation
(1):
L=(πt一 π 。。w*)2+λ(yt‑y*)2 (1)
πisaninflationrate,π 奉isatargetinginflationrate,yisanoutput (logofGDP),y'isapotentialoutput(logofitsvalue),and(ỳ‑ỳ)
meansGDPgap.Centralbankswouldliketominimizethislossfunction.
π 。,.零isdefinedasnewlyannouncedinflationtargetingandπdd率isthe oldinflationtarget.
π,canbeexpressedasfollows:
π,=α π 。,w冷+(1一 α)π 。ld率+β(y㌧y*) (2)
Thefirst‑orderconditionisgivenby
つ職π
㌧醐πく α
■
1 .
λβ+
2 β く
一 =
y
t 一 y
(3)
Theleftsideof(3)istheoutputlossthatresultsfromthepolicy.This lossisdeterminedbythedegreeofnominalinertia,theweightonthe outputgapinthelossfunctionandcredibilityinadditiontothesizeof thereductionintheinflationtarget.
3.SACRIFICERATIO
Thesacrificeratiohasbeenusedtoproxytherelativesuccessofpolicy regimes.Itsreductionfollowingaregimechangecouldbearesultof goodpolicycredibilityfortheinflationtargeting.Animportantquestion ofthispaperiswhethertheintroductionofinflationtargetinghasre‑
ducedthecostsofoutput.
Earlypapersofthesacrificeratioshowedthesacrificebyestimatinga Phillipscurve(Okun,1978;GordonandKing,1982).Insomelaterstud‑
ies,sacrificeratioshavebeenestimatedthroughofVARmode.Thispa‑
per'smethodologyhasmostincommonwithCecchetti(2000)'.
Recently,muchattentionhasbeenpaidtosacrificeratiosbecauseof therecentdeflationseeninmostdevelopedcountries.Cunado(2003)sug‑
gestedthatsacrificeratiotakeshighervaluesinperiodsoflowinflation rates.Zhang(2005)foundthatthereisanegativerelationshipbetween sacrificeratiosandinitialinflationrates,andthecostofreducinginfla‑
tionisgenerallylowerwhenthespeedofdisinflationisfaster.
ChristopherandChristopher(1997)showedthatthelostoutputfroma
disinflation‑inducedrecessiontypicallywillberecoupledin10to15years.
GoncalvesandCarvalho(2009)showedthatinflationtargeterssuffered smalleroutputlossesduringdeflationswhencomparedtonontargeters.
Chortareas,etal.(2003)foundthatahigherdegreeofcentralbank transparencyisassociatedwithlowersacrificeratios.Down(2004) showedthatcentralbankindependenceaffectsbothoutputandtheun‑
employmentcostsofdisinflation.Hetzel(2007)andBowdler(2009)in‑
vestigatedPhillipscurveslopeandfoundthatsacrificeratiosareweakly negativelyrelatedtoopenness.
Thispaper'smethodologyallowsustomeasurethesacrificeratioas thecostsofadeliberatedisinflationpolicyincontrasttotheoutputcosts ofaseriesofnegativecostpushshocks.Toderiveoursacrificeratioesti‑
mates,VAR(vectorautoregressive)modelisemployedforbothmen‑
tinnedtimeperiods.VARmodelscanidentifytheinterestratereaction functionthathasbeenpursuedovertheperiods.
Thesacrificeratiomeasuresthecumulativeoutputlossunderdisinfla‑
tionforeachpercentagepointreductionintheinflationrate,formally:
X。 バy零!o+sSacrificeRatio
=Σ δs't°E、,,/P
y㍉ 。ゆs
wheretoisthestartpointofthedisinflationperiod,Etistheexpecta‑
tionsoperatorconditionaloninformationattimet,8isthediscount actor,Yistheactualproduction,Y'isthepotentialproduction,andpis thechangeintheinflationrateovertheperiod.
Keynesiantheoryingeneralsuggeststhatpricesaresetinastaggered mannerandfixed.Pricessettersareconcernedwithhowpricesareex‑
pectedtorise/fall,andthuswiththemonetarypolicy.Ifrationalexpec‑
tation'sformationplaysacrucialroleinpricesetting,thedegreeof policycredibilitymayinfluencethesacrificeratio.Underrationalexpec‑
tationsandcostlypriceadjustments,imperfectcredibilityisanessential
ingredientinaccountingforapositivesacrificeratio.
Supposethatmonetarypolicycredibilityisbroughtunderinflation targeting.Whatimplicationsmaythishavefortheconductofmonetary policy?Accordingtothestandardcredibilityargumentspresentedby KydlandandPrescott(1977)andBarroandGordon(1983),commitment toadesiredpolicyrulewillbewelfare‑enhancingifthecommitmentis viewedascredibletothepublic.Non‑crediblecommitmenttoadesired rulecouldreducewelfareevenwhentheactualpolicyturnsofcredibility bythepublic.
Theendogenousvariablesinthispaper'smodelsaregeneraleconomic activity,thelogarithmoftheseasonallyadjustedconsumerpriceinfla‑
tion,theshort‑termnominalinterestrate,andthechangeinthenominal effectiveexchangerate.Thelengthoftheendogenousvariableswascho‑
senwiththefewobservationsavailableunderinflationtargetingin mind.Fortheoutputgap,wehavecomputedtheHodrick‑Prescott detrendedseriesofthelogarithmofrealGDP.Thelengthoftheendoge‑
nousvariableswaschosenwiththefewobservationsavailableunderin‑
flationtargeting.Akaikeinformationcriteria(AIC)wascomputedfor lengthuptosixquartersforthetwoquarters.Aconstanttermisin‑
eludedinallequations.ThedataseriesarefromIFS(IMF).
VAR(vectorautoregressive)modelsareemployedastheseimpose onlyaminimumofrestrictionsforanalyzingmonetarypolicy.Fifteen countriesinthesamplehaveadoptedinflationtargeting(dateofadop‑
tionshowinparentheses):Australia(1994:Q4),Canada(1992:Q1), Finland(1994:Q1),Korea(1998:Q1),Norway(2001:Q1),Spain(1994:Q1), Sweden(1993:Q1),Switzerland(2000:Q1),andtheUnitedKingdom (1993:Q1).Foreachcountry,wedefinethebeginningoftargetingasthe firstfullquarterafterthetargethadbeenpubliclyannouncedinwhich aspecificinflationtargetortargetrangewasineffectandcompareinfla‑
tion"targeters"totheother"nontargeters."Thesampleperiodendswith
therecent2007:Q4.Finally,allofthedatesareI(0)'.
Thesacrificeratiosareestimatedfortheindividualcountriesofthe samplethroughMonteCarlosimulationofestimatedVARswith1000
replications,drawingeachtimeanewsetofcoefficientsinaccordance withtheirestimateddistribution.
4.RESULTS
Estimatesforboththepreinflationtargetingandtheinflationtarget‑
ingperiodarereportedwithstandarderrorsinparentheses.Theresults areshownintable1.
Theestimatesofthesacrificeratioshowthattheydeclineformost countriesduringinflationtargetingperiod'.Thistendencyismost
Table1.Sacrificeratioestimates(mean,percent)
Startingperiod Preinflation targetingperiod
Inflation targetingperiod
Australia 1994:Q4 1.66
(1.89)
一〇.53 (2.99)
Canada 1992:Q1 2.43
05.78)
0.12 (4.96)
Finland 1994:Q1 一 〇.05
(0.33)
一〇.56 (3.28)
Korea 1998:Q1 1.05
(6.77)
一1.55 (3.53)
Norway 2001:Q1 0.36
(4.98)
0.04 (4。88)
Spain 1994:Q1 0.55
(5.44)
1 (3.65)
Sweden 1993:Q1 0.15
(7.08)
一〇.22 (10.12)
Switzerland 2000:Q1 一 〇.44
(9.54)
一〇.32 (12.92)
UnitedKingdom 1993:Q1 0.55
(2.88)
1・:
(5.02)
markedforCanadaandKorea,whichexperiencedrelativelyhighsaci・i‑
Ticeratiosinthefirstperiod.Somecountriessawariseinthesacrificera‑
tios.Thesecountriesexperiencedrelativelylowsacrificeratios.
Thepropertiesofthismodelcanbecheckedbytheresponseofavariable toashocktothemodel.Sincethispaperfocusesontheeffectofmonetary
policyinoutputandinflation,theshocktotheshort‑terminterestrate istheinterestrateforthepurposeofevaluatingandprovidingacompact accountofthepropertiesofthismodel.Theresponsestotheshockindi‑
Catetheextenttowhichthemodelsbringtheexpectedeffectstochanges inmonetarypolicy.TheeffectsonoutputgapsandCPIinflationofaone standarderrorinterestrateshockareshownforbothandwholeperiods infigure1‑9.
Itshouldbenotedthatthesacrificeratiosareimpreciselyestimated.
Theuncertaintyisinpartduetotheuseofonlyaminimumofexclusion restrictionsfromeconomictheoryinthemodels.Also,thesampleof countriesissmallandthetimeseriesareshort,weneverthelessinterpret ourfindingsassupportiveofacredibilityenhancingeffectofinflation targeting.However°,theresultsarenotfarfrompuzzlingthanwhatis commoninmostVARstudies.
5.CONCLUSIONS
The1990swereamazinginmanyways.NotonlytheInternetandcel‑
lularphonescameintouse,butalsoeconomicconditionsimproved greatlyallovertheworld(exceptinJapan).Growthwashigher,infla‑
tionwaslower,andbothwerestable.Thatsituationmayhaveinfluenced thefindingthatinflationratehasnotaffectedthedeterministicelements ofintroducinginflationtargeting.
Thispaperalsoshowedthattargetingcountrieshavesuccessfullyin‑
troducedinflationtargetingintermsofreducinginflationratesand
reducingeconomicshocks.Thispaperprovidessomeevidenceinfavorof anaffirmativeanswertothisquestion.However,whetherornottarget‑
ingimprovesacountry'seconomicperformanceasawholeisadifferent issue.Growth,forexample,isatypicalcase.Stockprices,somemacro‑
economicvariablesandtheirvariabilityaresometimesimportantfactors indeterminingeconomicperformanceandwelfare.Muchfurtherre‑
searchisneededtoinvestigatethedegreeofrobustnessofthesefindings.
Notes
1.King(2002)showedthatnotonlyinflationratebutalsoitsstandarddevia‑
tionhavebeenmorestableinrecentdecadesintheU.K.SeealsoBernanke etal.(1999).
2.Durandetal.(2008)usedstructuralVARanddiscoveredthenegativerela‑
tionshipbetweenthelevelofinflationandthecostofdisinflation.
3.ADFtestforunitroottestwasperformedforallthevariables(rates).They aresignificantatleast5%level.
4.Hofstetter(2008)suggestedlargesacrificeratioswerefoundinthe1970s and1980sinLatinAmericaandtheCaribbean.
REFERENCESReferences
Bernanke,B.S.,T.Laubach,F.S.Mishkin,andA.S.Posen(1999)Inflation Targeting:LessonsfromtheInternationalExperience,Princeton:Princeton UniversityPress.
Bowdler,C.(2009)"Openness,ExchangeRatesRegimesandthePhillipsCurve,"
JournalofInternationalMoneyandFinance28(1),148‑168.
Cecchetti,S.(2000)"MakingMonetaryPolicy:ObjectivesandRules,"Oxford ReviewcリfEconomicPoli(:ソ16(4),43‑59.
Chortareas,G.,D.Stasavage,andG.Steme(2003)"DoesMonetaryPolicy TransparencyReduceDisinflationCosts?"ThemanchesterSchool71(5), 521‑541.
Christopher,N.AndW.Chirstopher(1997)"ABenefit‑CostAnalysisofDisinf‑
lation,"ContemporaryEconomicPolicy15(1),50‑65.
Cunado,J.(2003)"SacrificeRatios:SomeLessonsfromEMUCountries,1960一
2001,"InternationalReviewofAppliedEconomics17(3),327‑347.
Down,1.(2004)"CentralBankIndependence,Disinflations,andtheSacrifice Ratio,"Cornpar・atiuePolitical.Studies37(4),399‑419.
Durand,J.J.,M.Huchet‑Boudon,andJ.Licheron(2008)"SacrificeRatioDis‑
persionwithintheEuroZone:WhatCanBeLearnedaboutImplementinga SingleMonetaryPolicy,"InternationalReview(ゾAppliedEconomics22(5), 601‑620.
Fountas,S.,M.Karanasos,andJ.Kim(2002)"lnflationandOutputGrowth UncertaintyandTheirRelationshipwithInflationandOutputGrowth,"
EconomicsLetters75,1607‑1617.
Gonsalves,C.AndA.Carvalho(2009)"lnflationTargetingMatters:Evidence fromOECDEconomies'SacrificeRatios,"JournalofMoneyandBanlzing41 (1),233‑253.
Hetzel,R.(2007)"TheCaseforInflationTargeting,"CatoJournal27(2),273‑
293.
Hofstetter,M.(2008)"DisinflationinLatinAmericaandtheCaribben:AFree Lunch,"JournalofMacroeconomics30(1),327‑347.
Kurihara,Y.(2003)"WhyDoEUCountriesUseInflationTargeting?,"Global BusinessandEconomicsReview7(1),74‑84.
Mishkin,F.(2001)'IInflationTargetinginEmergingMarketCountries,'りVBER WorkingPaper,7618.
Svensson,L.E.O.(1997)"lnflationForecastTargeting:Implementingand MonitoringInflationTargets,"EuropeanEconomicReひiew41,1111‑1146.
Svensson,L.E.O.(2002)"lnflationTargeting:ShouldItbeModeledasan InstrumentRuleoraTargetingRule,"EuropeanEconomicReview46,393‑
419.
Svensson,L.E.O.andM.Woodford(2005)"lmplementingOptimalPolicy throughInflation‑ForecastTargeting,"inS.BernankeandM.Woodford,
TheInflation‑TargetingDebate,Chicago:TheUniversityofChicagoPress.
Walsh,C.(1998)MonetaryTheoryandPolicy,Cambridge:MITPress.
Zhang,L.H.(2005)"SacrificeRatioswithLong‑LivedEffects∴Internatめnal
Finance51(2),231‑251.
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofAUINTtoAUGDPResponseofAUINTtoAUPRI .8.8
.6.6 .4a‑1.4 ,2.2 .0.0
‑.2ai‑.z
‑.41‑一.'一.4 51015205101520
FigurelaImpulse‑responsefunctions:Thewholeperiod:Australia
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofAUINTtoAUGDPResponseofAUINTtoAUPRI 1.51.5
1.01.0
0.50.5
0.OO,0
《).5‑0,5
‑tO‑7.0
1234567891012345678910
FigurelbImpulse‑responsefunctions:Thepreinflationtargetingperiod:Australia
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofAUINTtoAUGDPResponseofAUINTtoAUPRI .4
.3.3 .2 .1.1 .0.0
‑.1‑ .1
z‑.2
123456ア891012345678910
FigurelcImpulse‑responsefunctions:Theinflationtargetingperiod:Australia
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofCAINTtoCAGDPResponseofCAINTtoCAPRI .8.8
6
4
2 0
一.2
一.4 51015205101520
Figure2aImpulse‑responsefunctions:Thewholeperiod:Canada
一 .「㌧
「 隔
」し 1剛
脚 幽
一 駈
,
「1、 、
「 、、
、唱1
△ 、＼.層
、
、、 一 、
"1'1',,1ト1'},II'唱
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofCAINTtoCAGDPResponseofCAINTtoCAPRI
1.2
o.s ' o.a
o.o
イ).4
1.2
o.a
o.a
o.o
く》.4
'﹂﹂
' ﹂
r ︿
く」.8 ℃3
1234567891012345678910
Figure2bImpulse‑responsefunctions:Thepreinflationtargetingperiod:Canada
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofCAINTtoCAGDPResponseofCAINTtoCAPRI
.6
.4
.2
.o
一.2
6
.4
.2
0
一.2 ρ '
' ''
一.4‑.4
1234567891012345678910
Figure2cImpulse‑responsefunctions:Theinflationtargetingperiod:Canada
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofFIINTtoFIGDPResponseofFIINTtoFIPRI
・,4 51015205101520
Figure3aImpulse‑responsefunctions:Wholeperiod:Finland
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofFIINTtoFIGDPResponseofFIINTtoFIPRI
・0.8‑0.8
1234567891012345678910
Figure3bImpulse‑responsefunctions:preinflationtargetingperiod:Finland
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofFIINTtoFIGDPResponseofFIINTtoFIPRI .4.4
.3.3
.2.2
.1.7
.o.0
‑.1・.1
‑.2・.2
1234567891012345678910
Figure3cImpulse‑responsefunctions:inflationtargetingperiod:Finland
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofKOINTtoKOGDPResponseofKOINTtoKOPRI
一 .8 51015205101520
Figure4aImpulse‑responsefunctions:wholeperiod:Korea
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofKOINTtoKOGDPResponseofKOINTtoKOPRI
・,8‑.8
1234567891012345678910
Figure4bImpulse‑responsefunctions:preinflationtargetingperiod:Korea
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofKOINTtoKOGDPResponseofKOINTtoKOPRI
..4‑.4
1234567891012345B78910
Figure4cImpulse‑responsefunctions:inflationtargetingperiod:Korea
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofNOINTtoNOGDPResponseofNOINTtoNOPRI .2.2
1 0
一.1
一.2
印.3
、、
1
.o
・.1
一.2
一.3 51015205101520
Figure5aImpulse‑responsefunctions:wholeperiod:Norway
門
・ ρ隔臨 ' '
「 , '
一、覧
、、、、
1一、
門、99亀 騙.
一 L
、, 辱
、
、
、、
、讐,「r'♂「.一
一舶一、旬曹.‑一 1‑一一
. 脚 i曜 「", .唱 唱r 1' , 脚卜 1脚 .■
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofNOINTtoNOGDPResponseofNOINTtoNOPRI
3 2 ﹂ ゆ d 4 適
、
3 2 ︒ ‑ 刀 d 4 弓
12345B7891012345678910
Figure5bImpulse‑responsefunctions:preinflationtargetingperiod:Norway
ResponsetoCholeskyOneS.D.Innovations±2S.E.
ResponseofNOINTtoNOGDPResponseofNOINTtoNOPRI
3
2
1
0
一1‑1
1234567891012345678910
Figure5cImpulse‑responsefunctions:inflationtargetingperiod:Norway
■
一
",7醒'
1,舳
lllllIll
