亞開行-新興經(jīng)濟(jì)體的債務(wù)沖擊與產(chǎn)出和通脹動(dòng)態(tài)（英）

DEBTSHOCKS

ANDTHEDYNAMICS

OFOUTPUTANDINFLATIONINEMERGINGECONOMIES

JohnBeirneandNuobuRenzhi

NO.739

ADBECONOMICS

August2024

WORKINGPAPERSERIES

ASIANDEVELOPMENTBANK

ADBEconomicsWorkingPaperSeries

DebtShocksandtheDynamicsofOutputandInflationinEmergingEconomies

JohnBeirneandNuobuRenzhiNo.739|August2024

TheADBEconomicsWorkingPaperSeries

presentsresearchinprogresstoelicitcommentsandencouragedebateondevelopmentissuesinAsiaandthePacific.Theviewsexpressed

arethoseoftheauthorsanddonotnecessarilyreflecttheviewsandpoliciesofADBor

itsBoardofGovernorsorthegovernmentstheyrepresent.

JohnBeirne(jbeirne@)isaprincipaleconomistattheEconomicResearchandDevelopmentImpact

Department,AsianDevelopmentBank.NuobuRenzhi(renzhinuobu@)isanassistantprofessor

attheSchoolofEconomics,CapitalUniversityofEconomicsandBusiness.

ASIANDEVELOPMENTBANK

CreativeCommonsAttribution3.0IGOlicense(CCBY3.0IGO)

?2024AsianDevelopmentBank

6ADBAvenue,MandaluyongCity,1550MetroManila,PhilippinesTel+63286324444;Fax+63286362444

Somerightsreserved.Publishedin2024.

ISSN2313-6537(print),2313-6545(PDF)PublicationStockNo.WPS240396-2

DOI:

/10.22617/WPS240396-2

TheviewsexpressedinthispublicationarethoseoftheauthorsanddonotnecessarilyreflecttheviewsandpoliciesoftheAsianDevelopmentBank(ADB)oritsBoardofGovernorsorthegovernmentstheyrepresent.

ADBdoesnotguaranteetheaccuracyofthedataincludedinthispublicationandacceptsnoresponsibilityforany

consequenceoftheiruse.ThementionofspecificcompaniesorproductsofmanufacturersdoesnotimplythattheyareendorsedorrecommendedbyADBinpreferencetoothersofasimilarnaturethatarenotmentioned.

Bymakinganydesignationoforreferencetoaparticularterritoryorgeographicareainthisdocument,ADBdoesnotintendtomakeanyjudgmentsastothelegalorotherstatusofanyterritoryorarea.

ThispublicationisavailableundertheCreativeCommonsAttribution3.0IGOlicense(CCBY3.0IGO)

/licenses/by/3.0/igo/

.Byusingthecontentofthispublication,youagreetobeboundbythetermsofthislicense.Forattribution,translations,adaptations,andpermissions,pleasereadtheprovisionsandtermsofuseat

/terms-use#openaccess

.

ThisCClicensedoesnotapplytonon-ADBcopyrightmaterialsinthispublication.Ifthematerialisattributed

toanothersource,pleasecontactthecopyrightownerorpublisherofthatsourceforpermissiontoreproduceit.ADBcannotbeheldliableforanyclaimsthatariseasaresultofyouruseofthematerial.

Pleasecontactpubsmarketing@ifyouhavequestionsorcommentswithrespecttocontent,orifyouwishtoobtaincopyrightpermissionforyourintendedusethatdoesnotfallwithintheseterms,orforpermissiontousetheADBlogo.

CorrigendatoADBpublicationsmaybefoundat

/publications/corrigenda

.

ABSTRACT

Thispaperempiricallyexaminestheimpactofpublicdebtshocksonoutputandinflationin34emergingmarketeconomies(EMEs)usingpanellocalprojectionsovertheperiod2000to2022.Theestimatedresultsshowthatrealgrossdomesticproduct(GDP)fallssignificantlyafteranunanticipatedincreaseinpublicdebtwhileinflationrises.WealsoexaminewhetherfundamentalcharacteristicsacrossEMEscouldaffecttheimpactofpublicdebtshocks.Theresultssuggestthathigherinitialdebtlevels,tighterdomesticfinancialconditions,andlowerincomelevelsamplifythenegativeresponsesofrealGDP,whiletighterglobalfinancialconditionsdampenthenegativeimpactsofdebtshocks.Forinflation,theresponsesvarydependingoneconomic-specificcharacteristics.Wealso

findothernonlinearitiesinthedynamics,withEMEsfacingmoresevereeffectsduring

recessionaryperiods.

Keywords:publicdebt,GDP,inflation,emergingmarketeconomiesJELcodes:E62,F40

1.Introduction

Thelinkbetweendebtandtheoutlookforeconomicgrowthandinflationisakeypolicyconsideration.Debtsustainabilityrisksinemergingmarketeconomies(EMEs)wereamplifiedintheaftermathofthecoronavirusdisease(COVID-19)pandemic.Fiscalsupportpackagesimplementedduetothepandemicledtohigherstocksofpublicdebt,whiletighterglobalmonetaryconditionsincreasedthecostoffinancingpublicdebt,wideningdeficits.Managingtheinflationaryeffectsoffiscalexpansionshascreatedsignificantchallengesforpolicymakers,alsoaffectingthetrajectoryofoutputgrowth.ThispaperconsiderstheramificationsofdebtshocksinEMEsonthetrajectoryofoutputandinflationovershortandlongerhorizons.

Thispapercontributestotheliteratureonthemacroeconomiceffectsofdebt.Thereissignificantliteratureontheroleofdebtinaffectingoutput.ReinhartandRogoff(2010)notethatforlevelsofdebttogrossdomesticproduct(GDP)above90%,anegativeimpactongrowthmaterializes.Asregardstheroleofdebtoninflation,thiscanbetracedbacktothefiscaldominanceworkofSargentandWallace(1981),wherebypersistentfiscaldeficitsandexcessivedebtconstrainthecentralbankfromtighteningmonetarypolicy,leadingtoamplifiedinflationexpectations.CevikandMiryugin(2023)findthattheextentoffiscalpolicyshocksoninflationiscloselydrivenbytheinitialleveloffiscalspace.Buildingontheliterature,thispapersimultaneouslyexaminestheimpactofexogenousdebtshocksonoutputandinflation.ThispaperaimstofillthisgapintheliteraturebyemployingpanellocalprojectionsasinJordà(2005)toestimatetheimpactofpublicdebtonoutputandinflationin34EMEsovertheperiod2000–2022.

2

Thepaperfocusesontheeffectofunexpecteddebtaccumulationsoneconomicgrowthandinflationinemergingeconomies.Researchontheimplicationsofdebtonoutputaremoreprevalentintheliteraturecomparedtothoseoninflation,withsomestudiesnotingnonlinearityintherelationship,wherebydebtoveracertainthresholdcanreducegrowth(Baum,Checherita-Westphal,andRother2013;Cecchetti,Mohanty,andZampolli2011;ReinhartandRogoff2010).However,thereisalackofconsensus,withotherworkfindinganegativerelationshipandnothresholds(EberhardtandPresbitero2015,PanizzaandPresbitero2014,WooandKumar2015).Morerecentworkhasfoundthatthetrajectoryofdebtmatterfortheimpactongrowth(deSoyres,Kawai,andWang2022,Chudiketal.2017).Otherworkhasfocusedonprivateandcorporatedebt,systemicrisk,andrecessions(Jordàetal.2013,BernardiniandForni2017).Asregardstheinflationaryrepercussionsofdebtshocks,thiscanbetracedbacktoSargentandWallace(1981),whofoundthatpersistentfiscaldeficitsandexcessivedebtconstrainthecentralbankfromtighteningmonetarypolicy,leadingtoamplifiedinflationexpectations.Studiespertainingtomonetarypolicycoordinationwithfiscalpolicyhavealsoshedlightontheroleoffiscalshocksoninflation,withsomepreviousworkfindingthatfiscaltighteningininflation-targetingcentralbankeconomieshelpstoreduceinflation(Cat?oandTerrones2005;deMendon?aandMachado2013;Coibion,Gorodnichenko,andWeber2021;Brandao-Marquesetal.2023).CevikandMiryugin(2023)alsofindtheextentoffiscalpolicyshocksoninflationiscloselydrivenbytheinitialleveloffiscalspace.ForEMEsinparticular,theriskofencounteringadebt-inflationtrapwasfoundforhighly-indebtedeconomies(Kwon,McFarlane,andRobinson2009).

3

Inthispaper,examiningtheeffectofdebtshocksonoutputandinflationacrossEMEs,weovercomepotentialendogeneityconcernsbycomputingaseriesofidentifiedpublicdebtshocksforeachofthe34EMEs.FollowingtheapproachbydeSoyres,Kawai,andWang(2022),weusetheforecasterrorsfromtheWorldEconomicOutlook(WEO)vintagestoidentifytheexogenousshocksinpublicdebt,whichisolatetheimpactfromthedebtvariableonly.Theestimatedpublicdebtshocksthereforemitigatetheendogeneityissuebetweenpublicdebtandothermacroeconomicvariables,includingrealGDPandinflationrate.WethenestimatetheresponsesofrealGDPandinflationratetotheidentifiedpublicdebtshocksandfindthatrealGDPsignificantlydecreasesafteranunanticipatedincreaseinpublicdebtwhiletheinflationraterespondsintheoppositedirection.Theseresultsarerobusttoasetofsensitivitychecks,includingaddinglagsofinitialdebtlevelascontrolsanddroppinglagsofdependentvariables.

WealsostudywhetherfundamentalcharacteristicsacrossEMEscouldaffecttheimpactofpublicdebtshocks.Theresultssuggestthatahigherinitialdebt,atighterdomesticfinancialcondition,andalowerincomelevelcouldstrengthenthenegativeresponsesofrealGDPwhiletighterglobalfinancialconditionsdampenthenegativeimpactsofdebtshocks.Fortheinflationrate,theresponsesvarydependingoneconomic-specificcharacteristics.Wealsofindsignificantstatedependency,withEMEsfacingmoresevereeffectsduringrecessionaryperiods.Theremainderofthispaperisstructuredasfollows.Section2describesthedataandoutlinestheempiricalmethodology.Section3presentstheempiricalresultswithrobustnesschecks.Section4presentstheresultsoftheextendedanalysis.Section5concludes.

4

2.DataandEmpiricalMethodology

Inthissection,wefirstdescribethedatasourceofvariablesweuseintheempiricalanalysis.Wethendiscusstheidentificationofthepublicdebtshocks.Finally,wepresent

oureconometricframeworkusedtoproducetheempiricalresults.

2.1Data

Weuseavailableyearlydatawithanunbalancedpanelfor34EMEsspanningfrom2000to2022.1WecollectdataontherealGDPasthedomesticoutputmeasure,theyear-on-yearchangeoftheconsumerpriceindexasameasureoftheinflationrate,andtheactualandprojectedpublicdebtrelativetoitsGDP.ThedataareallfromtheWEOoftheInternationalMonetaryFund(IMF).

Astheimpactofpublicdebtonthemacroeconomycanbeaffectedbydomesticandglobalfinancialconditions,wealsoincludethefollowingtwovariablesinourempiricalanalysis.First,tomeasurethedomesticfinancialconditions,weconstructafinancialconditionindexbyaveragingthecentralbankpolicyrateand10-yeargovernmentbondyield,obtainedfromtheIMF’sInternationalFinancialStatistics.Second,weusetheVIXindexwhichstandsfortheChicagoBoardOptionsExchange(CBOE)VolatilityIndex,asameasureofglobalfinancialconditions.

1ThesampleofEMEsincludesBrazil;Bulgaria;Chile;thePeople’sRepublicofChina;Colombia;CzechRepublic;Egypt;Greece;HongKong,China;Hungary;Indonesia;Israel;theRepublicofKorea;Kuwait;Mauritius;Mexico;Malaysia;Morocco;Nigeria;Pakistan;Peru;thePhilippines;Poland;Qatar;Romania;theRussianFederation;SaudiArabia;Singapore;SouthAfrica;Taipei,China;Thailand;Ukraine;VietNam;andtheUnitedArabEmirates.Thedatastartingyearforaspecificvariablevariesacrosseconomiesduetothedataavailabilityandreliability.

5

2.2IdentificationofPublicDebtShocks

Thegoalofthispaperistoestimatethecausaleffectofpublicdebtonoutputandinflationinemergingeconomies.Assuch,wefacetheusualendogeneityissuebetweenpublicdebtandothermacroeconomicvariables,includingrealGDPandinflation.Toidentifyexogenousshocksinpublicdebt,followingdeSoyres,Kawai,andWang(2022),weusetheforecasterrorsfromvariousvintagesoftheIMFWEOpublications.Wecomputethedebtshocksindebt-to-GDPterms,withtheshockappliedonlytopublicdebtandnottoGDP,toisolatetheimpactfromthedebtvariableonly.Specifically,publicdebtshockscanbedefinedasfollows:

debtock=Δlndebt?Δlndebt,(1)

whereΔlndebtreferstotheloggrowthrateoftheactualdebt-to-GDPratioand

Δlndebt,istheloggrowthrateoftheforecasteddebt-to-actualGDPratio

projectedbyIMFanalystsinOctoberofthesameyear.Thechangebetweentheactualandprojectedpublicdebtlevelsisthusidentifiedasanexogenousshocktopublicdebtastheforecasterrorsabsentanyunanticipatedpolicychanges(e.g.,BlanchardandPerotti2002;AuerbachandGorodnichenko2012,2013;Abiad,Furceri,andTopalova2016;andFurcerietal.2018).

6

2.3EconometricMethodology

FollowingtheframeworkproposedbyJordà(2005),weusethepanellocalprojection(LP)toestimatethemodelandcalculateimpulseresponsestoexogenouspublicdebtshocks.Thebaselinemodelcanbegivenasfollows:

L

l=0

(2)

wherei=1,?Nreferstothespecificeconomyinthesample,yisthevariableof

interest(e.g.,realGDPorinflationrate),debtockistheseriesofidentifiedpublicdebt

shocks,zisavectorofcontrolvariablesincludinglaggedvaluesforyanddebtock

aswellasothercontrolvariables,andδlhisavectorofcoefficientsassociatedwiththelagsofz.Specifically,wesetL=2,thereforeweinclude2yearsoflaggedvaluesofz.Thecoefficientβhmeasurestheimpactofa1%publicdebtshocks,namelya1percentunanticipatedincreaseinthepublicdebttoGDPratio,onyathorizon?.Thus,oneconstructstheimpulseresponsesasasequenceoftheβhestimatedinaseriesofseparateregressionsforeachhorizon?.αidenoteseconomy-specificfixedeffects,

controllingforthetime-invariantcharacteristicsoftheeconomy.λrepresentsthetime

fixedeffects2.Finally,εi,t+hdenotesameanzeroerrortermcapturingothershocks.Weclusterourstandarderrorsattheeconomylevel.

WecanfurtheradapttheLPframeworktoallowfornonlinearitiesinthespecificationsthatareassociatedwithdomesticandglobalfinancialconditions.Wethereforemakethe

2ThetimefixedeffectsalsocontrolforstructuralbreaksduetotheCOVID-19pandemic.

7

responseofoutputorinflationtoapublicdebtshockalsodependentonthecontemporaneouschangeindomesticandglobalfinancialconditionsbyinteractingthepublicdebtshockwiththechangeinfinancialconditionvariables.Thespecificationisasfollows:

L

yi,t+h?yi,t-1=αZi,t-l+βhdebtock+θhFci,t×debtock+εi,t+h

l=0

(3)

whereFci,tisavariablerepresentingourfinancialconditionvariables,includingthedomesticfinancialconditionindex,andglobalfinancialconditions(measuredastheVIXindex).Therefore,βhmeasurestheresponseofoutputorinflationtothepublicdebtshockateachhorizon(year)?whenthefinancialconditionsareisolated,andβh+σθhrepresentsthetotaleffectsofpublicdebtshockswhenweconsidertheimpactoffinancialconditions.σmeasuresonepercentagechangeinthespecificfinancialconditionvariable.

WealsodivideEMEsintogroupsaccordingtotheirlevelsofinitialpublicdebts,incomeclassifications,andregionalcategorizationsandestimateseparateimpulseresponsesforeachgroup.Regardingtheestimationoflocalprojections,weincorporateadummyvariableIthattakesavalueof1forEMEswherebytheyfallwithinacertaingroupm∈M.FollowingCloyneetal.(2023),weextendthelocalprojectionasfollows:

yi,t+h?yi,t-1=α+λ+∈MI[Σl0δ,lZi,t-l+βdebt,ock+θFci,t×debt,ock]+εi,t+h(4)

wherethenotationisasinequation3.

8

3.EmpiricalResults

3.1BaselineResults

Figure1showstheestimatedimpulseresponsesofrealGDPandinflationratetopublicdebtshocks,basedonthelinearmodelofequation2.Thesolidlineineachgraphrepresentstheestimatedimpulseresponsesinpercentageoverthefollowing5yearstoa1%positivepublicdebtshock,namelya1%unanticipatedincreaseinthepublicdebttoGDPratio.Theshadedarearepresents68%(dark)and90%(light)confidencebands.

Figure1:ResponseofRealGDPandInflationtoPublicDebtshocks,Baseline

GDP=grossdomesticproduct.

Notes:ThefigureplotstheimpulseresponsesofrealGDPandinflationratetoa1%positivepublicdebt

shock.Shadedareasrepresent68%and90%confidencebands.Source:Authors’estimates.

TheimpulseresponsesindicatethatrealGDPdecreasesafteranunanticipatedincreaseinpublicdebtwhiletheinflationraterespondspositivelytodebtshocks.3Followingapositivepublicdebtshock,thelevelofrealGDPdecreaseswithamaximumimpactof–0.015%percentaround2yearsaftertheshock.TheestimatedresultisinlinewiththeliteraturethatalsoshowsanegativeimpactofpublicdebtshocksonrealGDP

3Ourresultsalsoshowthata1%debtshockalsotriggersasignificantpositiveresponseindebtitselfbyaround0.03%to0.07%withinthefirstyear.

9

(e.g.,ReinhartandRogoff2010;WooandKumar2015;deSoyres,Kawai,andWang

2022).Onthecontrary,publicdebtshockshaveapositiveimpactontheinflationrate.4Anunexpectedincreaseinpublicdebtisassociatedwithapersistentincreaseintheinflationrate,whichisbroadlyconsistentwithCat?oandTerrones(2005)whofindastrongpositiveassociationbetweendeficitsandinflationamonghigh-inflationand

developingeconomies.

3.2Robustness

3.2.1AddingLagsofInitialDebtLevelasControls

Tochecktherobustnessofourbaselineresults,weaddtwolagsoftheinitialdebt-to-GDPratioasadditionalcontrols.Figure2reportstheestimatedimpulseresponsesofrealGDPandinflationratetoa1%positivepublicdebtshock,addinglagsofinitialdebtlevelasadditionalcontrols.TheresultsshowthatrealGDPdecreasespersistentlyandsignificantlyaftertheshockwhiletheinflationraterespondspositively,broadlyconsistentwithourbaselineestimates.

4Whilethereisastrongunderlyingeconomicrationaleforincludingtimedummiesinthebaseline(i.e.,tocontrolforanytrendbehaviorintheforecasterrors),theresultsexcludingthetimedummiesarenotmateriallydifferent.Globaldebtshocksindeedwillbereflectedwithvariationacrossdomesticdebtdynamics.Thebaselineapproachinidentifyingthedebtshocksusingforecasterrorsremainsconsistentattheeconomyleveltherefore,eveninthepresenceofglobaldebtshocks.Theinclusionoftimedummiesfurthercontrolsforanyresidualtrendbehaviorthatcouldbedrivenbyglobalfactors.Nonetheless,theresultsthatexcludethetimedummiesdonotsignificantlychangethebaselineresults.Heterogeneitiesintimedummyestimatesacrossalternativecategoriesofeconomies,suchashighversuslowincome,alsodonotsignificantlychangethebaselineresults.

10

Figure2:ResponseofRealGDPandInflationtoPublicDebtShocks,AddingLagsofInitialDebtLevels

GDP=grossdomesticproduct.

Notes:ThefigureplotstheimpulseresponsesofrealGDPandinflationratetoa1%positivepublicdebtshock.Shadedareasrepresent68%and90%confidencebands.

Source:Authors’estimates.

3.2.2DroppingLagsofDependentVariables

Thusfar,wehavefollowedtheliteratureonstandardlocalprojections,addinglaggeddependentvariablesascontrols.Ontheotherhand,onemaybeconcernedthatendogeneityissuesmayariseifthereisacorrelationbetweenthelagsofthedependentvariablesandtheexcludedlagssubsumedintheerrorterm(e.g.,Nickell-bias).Forthisreason,wenowdropthelaggeddependentvariablestofurthersubstantiateourbaselineestimates.Figure3reportstheestimatedimpulseresponsesofrealGDPandinflationratetoa1%positivepublicdebtshock,droppinglaggeddependentvariables.TheresultsshowthatrealGDPdecreasessignificantlyaftertheshockwhiletheinflationraterespondspositively,furtherindicatingtherobustnessofourbaselineestimates.

11

Figure3:ResponseofRealGDPandInflationtoPublicDebtShocks,DroppingLaggedDependentVariables

GDP=grossdomesticproduct.

Notes:ThefigureplotstheimpulseresponsesofrealGDPandinflationratetoa1%positivepublicdebtshock.Shadedareasrepresent68%and90%confidencebands.

Source:Authors’estimates.

3.2.3AlternativePublicDebtShockMeasurements

Tofurtherstrengthenthecasethattheforecasterrorscanberegardedasexogenousdebtshocks(evenpotentiallycapturinginflationorgrowthsurprises),wefollowthemethodologyintheOctober2017IMFWEOandMagudandPienknagura(2022)byorthogonalizingthedebtshocks.WeregressthedebtforecasterrorsonforecasterrorsofrealGDPgrowthandinflation,includingasetofeconomy-fixedeffects.Theresidualsfromtheseregressionsarethenusedasouralternativedebtshocks.Usingthealternativedebtshockmeasurement,theresults(Figure4)indicatethatourbenchmarkestimatesarequantitativelyandqualitativelyrobusttoalternativedebtshocks.

12

Figure4:ResponseofRealGDPandInflationtoPublicDebtShocks,AlternativePublicDebtShockMeasurements

GDP=grossdomesticproduct.

Notes:ThefigureplotstheimpulseresponsesofrealGDPandinflationratetoa1%positivepublicdebt

shock.Shadedareasrepresent68%and90%confidencebands.Source:Authors’estimates.

4.Extensions

4.1ImpulseResponsesConditioningonInitialDebtLevels

Sofar,wehaveexaminedtheimpactofpublicdebtshocksonoutputandinflationirrespectiveofinitialdebtlevels.Toexploretheextenttowhichtheex-antelevelofpublicdebtmattersinshapingtheresponsedynamicstothedebtshock,wefocusonthedifferenceintheimpactofthedebtshockonEMEswithdifferentinitialpublicdebtlevels.Specifically,weseparateoursampleEMEsintotwogroups:highandlowinitiallevelofpublicdebt,usingasathresholdthemedianofoursample(41.07%).Figure5illustratestheresponseofrealGDPandinflationforEMEswithlowandhighinitiallevelsofpublicdebt.Thefigureshowsthatforeconomiesinthehighdebtgroup,thenegativeimpactoftheunanticipatedpublicdebtincreaseonrealGDPismuchstrongerthanthebaselinecase,leadingtoasignificant–0.33%decreaseinrealGDParound1yearaftertheshock.

13

Bycontrast,high-debteconomiesexperienceasignificantdecreaseininflationrateinresponsetoapublicdebtrise,indicatingthatdebtsustainabilityconcernsmayfurtherweakentheaggregatedemandforEMEswithhighdebtlevels.Forthelow-debteconomies,wefindthattheresponsesofrealGDPandinflationaresimilarwiththebaselinecasewhiletheimpactofpublicdebtincreaseisslightlyweakerintheshortterm.5

Inthecaseofhigh-debteconomies,debtshocksineconomieswheredebtisalreadyelevatedcantriggernegativeaggregatedemandeffects(Mian,Straub,andSufi2022).Thechannelworksthroughaweakeninginthecredibilityoffiscalpolicy,andlowerconsumptionduetoresultingnegativewealtheffects.Inparticular,alackofcredibilitybyagentsinfiscalpolicycanleadtoanticipatedrisesintaxationlateron,triggeringnegativedemandeffects,andlowerinflation(MontesanddeHollandaLima2022).

Figure5:ResponseofRealGDPandInflationtoPublicDebtShocks,byInitialDebtLevels

(a)RealGDP

Continuedonthenextpage

5Notethatforcomparativepurposesandforclarityonthedifferencesinimpactbymagnitude,aconsistentscaleisusedforlowversushighdebtgroups.

14

(b)Inflation

GDP=grossdomesticproduct.

Notes:ThefigureplotstheimpulseresponsesofrealGDPandinflationratetoa1%positivepublicdebtshock.Shadedareasrepresent68%and90%confidencebands.

Source:Authors’estimates.

4.2ImpulseResponsesConditioningonDomesticFinancialConditions

Wetestifaneconomy’sdomesticfinancialconditionsaffecttheimpactofpublicdebtshocksonoutputandinflation.Specifically,weinteractthepublicdebtshockwithourconstructedfinancialconditionindexasinequation3.Figure6showstheimpulseresponsesofrealGDPandinflationratetopositivepublicdebtshocksdependingondomesticfinancialconditions.Ourresultsshowthattighterdomesticfinancialconditionsfurtherstrengthentheimpactofanunanticipatedincreaseinpublicdebtontheoutput.ThelevelofrealGDPdecreasesbyabout–0.38%1yearaftertheshockanddropspersistentlyafter3years.Regardingtheinflationrate,theoverallresponsestoapositivepublicdebtshockaremoremutedandshowoppositesignscomparedtothebaseline.

15

Figure6:ResponseofRealGDPandInflationtoPublicDebtShocks,ConditioningonDomesticFinancialConditions

GDP=grossdomesticproduct.

Notes:ThefigureplotstheimpulseresponsesofrealGDPandinflationratetoa1%positivepublicdebt

shock.Shadedareasrepresent68%and90%confidencebands.Source:Authors’estimates.

4.3ImpulseResponsesConditioningonGlobalFinancialConditions

WefurtherexaminewhetherglobalfinancialconditionsmayimpacttheeffectofdebtincreasebyinteractingthepublicdebtshockwiththeVIXindex,basedonequation3.Figure7showstheimpulseresponsesofrealGDPandinflationratetopositivepublicdebtshocksconditioningonglobalfinancialconditions.Ourresultsshowthattheimpactofpublicdebtshocksonoutputandinflationissomewhatweakerwithtighterglobalfinancialconditions.AlthoughthelevelofrealGDPdecreasesbyabout–0.11%1yearaftertheshock,thenegativeresponsesdisappearandbecomestatisticallyinsignificantafter2years.Similarly,whiletheinflationrateincreasesbyabout0.1percentagepointsrightafterthepositivepublicdebtshock,theoverallresponsesaremoremutedcomparedtothebaseline.

16

Figure7:ResponseofRealGDPandInflationtoPublicDebtShocks,ConditioningonGlobalFinancialConditions

GDP=grossdomesticproduct.

Notes:ThefigureplotstheimpulseresponsesofrealGDPandinflationratetoa1%positivepublicdebtshock.Shadedareasrepresent68%and90%confidencebands.

Source:Authors’estimates.

4.4ImpulseResponsesConditioningonIncomeLevels

Toanalyzehowtheincomelevelaffectstheimpactofanunanticipatedincreaseinpublicdebt,wedivideou