世界銀行-小島嶼發(fā)展中國家的氣候變化脆弱性、適應(yīng)和公共債務(wù)可持續(xù)性

PublicDisclosureAuthorizedPublicDisclosureAuthorized

PolicyResearchWorkingPaper10787

ClimateChangeVulnerability,Adaptation

andPublicDebtSustainabilityinSmall

IslandDevelopingStates

MartinBrownbridgeSudharshanCanagarajah

WORLDBANKGROUP

ClimateChangeGroupJune2024

PolicyResearchWorkingPaper10787

Abstract

SmallIslandDevelopingStates(SIDS)areagroupof39UnitedNations(UN)memberstatesand18dependentterritoriesmostlylocatedintheCaribbeanandOceania.Theyareamongthemostvulnerablecountriestolossesanddamagesfromclimatechangeglobally,andtheyareespe-ciallyexposedtotropicalcyclonesandsealevelrise.Thispaperevaluatestheclimatechange-relatedvulnerabilitiesoftheSIDS.ItusesdatafromtheInternationalDisasterDatabase(EM-DAT),toexaminethemagnitudeofdam-agesincurredoverthelastthreedecadesandconductsaneventstudyanalysistoexaminethefiscalimpactsoflargetropicalcyclonedisastersinthelastdecade.SIDSneedtoinvestsubstantialresourcesoverthelongtermforclimate

changeadaption.Thepaperreviewsthelimitedstudiesthathavebeenconductedsofar,onthepotentiallong-termcostsofeffectiveadaptationinvestmentprograms.UsingthemostrecentDebtSustainabilityAnalysesofSIDS,thepaperdiscussesthechallengescountrieswithhighpublicdebtlevelsfaceandwhetherageneralprogramofdebtreliefmightbeafeasiblewaytofundtheadaptationinvestmentrequirementsofSIDS.Finally,thepaperdiscussesthetypeofassistancethatSIDSneedtostrengthentheirresiliencetoclimatechangeinacost-effectivemanner,andtheroledevelopmentpartnerscanplayinsupportingclimate-re-silientdevelopment.

ThispaperisaproductoftheClimateChangeGroup.ItispartofalargereffortbytheWorldBanktoprovideopenaccesstoitsresearchandmakeacontributiontodevelopmentpolicydiscussionsaroundtheworld.PolicyResearchWorkingPapersarealsopostedontheWebat

/prwp.Theauthorsmaybecontactedatscanagarajah@

.

ThePolicyResearchWorkingPaperSeriesdisseminatesthefindingsofworkinprogresstoencouragetheexchangeofideasaboutdevelopmentissues.Anobjectiveoftheseriesistogetthefindingsoutquickly,evenifthepresentationsarelessthanfullypolished.Thepaperscarrythenamesoftheauthorsandshouldbecitedaccordingly.Thefindings,interpretations,andconclusionsexpressedinthispaperareentirelythoseoftheauthors.TheydonotnecessarilyrepresenttheviewsoftheInternationalBankforReconstructionandDevelopment/WorldBankanditsaffiliatedorganizations,orthoseoftheExecutiveDirectorsoftheWorldBankorthegovernmentstheyrepresent.

ProducedbytheResearchSupportTeam

ClimateChangeVulnerability,AdaptationandPublicDebtSustainabilityinSmallIslandDevelopingStates

MartinBrownbridgeandSudharshanCanagarajah

Keywords:SmallIslandDevelopingStates,ClimateChange

JELClassificationQ54,H63

2

1.Introduction

Smallislanddevelopingstates(SIDS)areagroupof39memberstatesoftheUnitedNationsand18otherdependentterritories.Althoughtheyareaheterogenousgroup,mostofthemfacesimilarchallenges.Mostarerelativelysmalleconomieswhicharenotwelldiversifiedandsoarevulnerabletomacroeconomicshocks.ThiswasstarklyillustratedbytheeconomicimpactoftheCOVID-19pandemic;therealGDPofSIDScontractedonaverageby7.4percentin2020,andonlypartiallyreboundedby4.1percentinthefollowingyear.Thiswasmainlydrivenbythecollapseoftourism,onwhichtheeconomiesofmanySIDSdepend.

1

Moreover,becauseoftheirgeographicalcharacteristics,manyofthemaremuchmorevulnerabletoclimate-shocks—especiallydamagesfromtropicalcyclones—thanaremostotherEmergingMarketsandDevelopingEconomies(EMDEs).Climatechangeisexacerbatingtheirvulnerability,makingextremeclimate-relatedshocksmorefrequentanddamagingoverthelongterm.

Alongsideothersocio-economicchallenges,suchasraisingrealincomelevels,reducingpovertyandenhancinghumandevelopment,mostSIDSalsofacetheurgentneedtoimplementadaptationtoclimatechange.Smallislandstatesmustadapttotropicalcyclones,floods,droughts,risingsealevelsandotherextremeweatherevents.Investmentinclimate-resilientinfrastructurecouldproducelongtermbenefitsintermsofhigherGDP(Cevik,2022).Climatechangeadaptationinvestmentsare,inlargepart,publicgoodsandthusrequirepublicinvestment.

2

Moreover,becauseofthecomparativelysmallersizesofsmallislands(bothintermsofpopulationandeconomy),itisverydifficulttorealizeeconomiesofscaleinpublicinvestments,makingsuchinvestmentsrelativelymoreexpensivethaninlargercountries.Themagnitudeofinvestmentsrequiredforeffectiveclimatechangeadaptation,combinedwiththelackofeconomiesofscaleformanyoftheseinvestments,unavoidablypresentsfiscalchallenges,withmostSIDShavinglimitedfiscalspace(ifany)forexpandingpublicinvestments.

FiscalspaceinmanySIDSisconstrainedbyhighpublicdebtratios.MorethanathirdoftheSIDSwhichareindependentUNmemberstateshaverelativelyhighpublicdebtratios(above60percentofGDP).IntheirmostrecentDebtSustainabilityAnalysis(DSA),14SIDSwereassessedasbeingathighriskofdebtdistress,mainlybecausetheirdebtratiosbreachsustainabilitythresholdsinDSAstresstests(seesection4formoredetails).Consequently,someresearchersandpolicyadvocatesarguethatdebtreliefmeasures,alongsideotherreformstotheglobalfinancialarchitecture,areessentialtorestoredebtsustainabilityandtoenableSIDStoundertaketheinvestmentsneededtostrengthentheirresilience(FresnillioandCrotti,2022;Rustomjee,2017).ThisisthemotivationfortheSIDSStrengtheningResiliencebyAlleviatingDebt(SRAD)program,whichadvocatesforpolicymeasurestoimproveclimateresilienceandincreasefiscalspaceinSIDS.Itfocusesondebtalleviationmeasuressuchas:i)

1Forcomparison,theaverageunweightedrealGDPofallEMDEswhicharenotSIDScontractedby1.8percentin2020butthenincreasedby6.9percentin2021.ThesourceofthedataistheWEOdatabase.ThedatafortheSIDSincludesthe36stateswhichareUNmembersandwhichareclassifiedasEMDEs(allSIDSexceptforSingapore),exceptCubaandtheCookIslandsforwhichtherearenodataintheWEOdatabase.

2Therearesometypesofadaptationinvestmentswhichareprivategoods,suchasstrengtheningprivatelyownedbuildingstomakethemmoreresilienttostorms,buteventhesemayrequiresomeformofpublicsubsidy,eitherbecauseprivatepropertyownersdonothavesufficientresourcesforthenecessaryinvestments(asmaybethecaseforhousesownedbylowincomehouseholdswhichoftenbearthebruntofdamagefromtropicalcyclones)orbecauseprivatepropertyownersdonotfullyunderstandthemagnitudeoftherisksthattheyfaceoverthelongtermandthereforerequireasubsidytoincentivisethemtoinvestoptimallyinadaptation.

3

debtswaps(forclimate/nature/lossanddamage);ii)insurance-basedapproachestocoverdebtrepaymentswhilehandlingshocks;andiii)disasterclausesinborrowingtermstoallowSIDStoprioritizerecoveryanddeferdebtservicing(principalandinterest)toalaterdate(alsocalledpauseclauses).Therearealsoproposalsforparametricriskinsurancetocoversovereigndebtservicepaymentsintheeventofanaturaldisaster,facilitatedbyaGlobalFundtoenableSIDStopoolrisksaswellasfinancinginvestmentstostrengthentheirresilience(Bharadwaj,MitchellandKarthikeyan,2023).

ThispaperevaluatesthechallengesSIDSface,focusingonthenexusbetweenvulnerabilitytoclimatechange,requirementsforinvestmentsinadaptationandpublicdebt,andtheimplicationsforpublicpolicy.Vulnerabilitytoclimatechangeandpublicdebtareclearlylinked;oneofthemainreasonswhymanySIDSareassessedasbeingathighriskofdebtdistressintheirmostrecentDSAistheirvulnerabilitytolargeclimaterelatednegativeshocks.However,thatdoesnotaxiomaticallymeanthatdebtreliefmeasuresareoptimalpolicyinstrumentstogeneratethefiscalspacerequiredtostrengthentheresilienceoftheSIDS.

Theorganizationoftherestofthispaperisasfollows:Sectiontwoexaminesthemainclimate-relatedvulnerabilitiesofSIDS,usingdatafromtheinternationalDisastersDatabase(EM-DAT).Italsoprovidesaneventstudyanalysisofthemacro-fiscalimpactofsomerecentlargecyclonedisasterswhichstruckSIDSintheCaribbeanandOceania.SectionthreediscussestheinvestmentrequirementsoftheSIDSforadaptationtoclimatechange,drawingonexistingliterature.SectionfourexaminescurrentlevelsofpublicdebtintheSIDSandtheirrisksofpublicdebtdistress,drawingontheassessmentsanddataintheirmostrecentDSAs.Italsobrieflyexaminesthecompositionofpublicdebt,differentiatingbetweenSIDSwithhigherandlowerdebt-to-GDPratios.SectionfivediscusseswhattypeofexternalassistancemostSIDSneedtohelpthemstrengthentheirresiliencetoclimatechange,andwhetherdebtreliefmeasuresofferfeasibleandoptimalinstrumentstodeliverthisobjective.Sectionsixprovidesaconclusionandmainmessages.

2.Vulnerabilitytoclimaterelatednaturaldisasters

ThissectionexaminesthevulnerabilityofSIDStoclimatechange-relatedhazards,whichislikelytoincreaseoverthelongterm,astheplanet’sclimatebecomeswarmer.SIDSareamongthemosthighlyexposedcountriesintheworldtohazardsfromclimatechange(Hallegatteetal,2018).ThisisillustratedbytheirrankingintheNotreDameGlobalAdaptationIndex(ND-GAIN).Oneofthesub-indicesoftheoverallND-GAINIndexisexposuretoclimatechange,whichcapturesthephysicalfactorsexternaltothesystemthatcontributetovulnerability—itisacomponentofvulnerabilityindependentofthesocio-economiccontext.Ofthe192countriesrankedinthissub-index,10ofthe19countriesinthemostexposeddecileareSIDS.Figure1showstherankingofallof36EMDESIDSwhicharerankedintheND-GAINsub-indexofexposuretoclimatechange.

4

Figure1Rankingof36SIDSintheND-GAINsub-indexofExposuretoClimateChange

250

200

150

100

50

0

185186187188189190192

180182

169173

158

152

143

139

136

129131

125126

120

105107

94

90

85

7979

73

6666

60

50

5354

46

Dominica

Grenada

CapeVerde

SaintLucia

Barbados

Guyana

Trinidad&Tobago

StVincent&Grenadines

Bahrain

Suriname

Bahamas

SaintKittsandNevis

Haiti

DominicanRepublic

Jamaica

Fiji

Comoros

AntiguaandBarbuda

SaoTome&Principe

Samoa

Mauritius

Vanuatu

PapuaNewGuinea

Cuba

Guinea-Bissau

Timor-Leste

Palau

Tonga

SolomonIslands

Nauru

MarshallIslands

Seychelles

Micronesia

Kiribati

Tuvalu

Maldives

Note:Thereareatotalof192countriesrankedinthissub-index.Exposurecapturesthephysicalfactorsexternaltothesystemthatcontributetovulnerability.Itisacomponentofvulnerabilityindependentofthesocio-economiccontext.Ahigherrankingdenotesgreaterexposure.

Source:ND-GAIN

AlthoughSIDSareaheterogeneousgroupofcountries,manyofthemsharecommonfactorswhichexposethemtothehazardsofclimatechange.Mostarelocatedgeographicallyintropicalcycloneregions.Manyalsohaveextensivelow-lyingcoastlines.Mostalsohavesmallundiversifiedeconomiesheavilydependentonasingleexportsector,suchastourism.Smallislandsareuniqueinthatalmostalloftheirterritoryisexposedtoanaturalhazardandtheirsmalllandareameansthatasingledisastercanhaveasystemicimpact.

Themaintypeofclimate-relatednaturaldisasters

3

whichaffectmostSIDS—bothintermsofincidenceandespeciallyintermsoftheeconomicdamagecaused—aretropicalcyclones,whichaccountedfor95percentofthetotaldamagesfromclimate-relatednaturaldisastersthatSIDSintheCaribbeanandOceaniasufferedfrom1995to2022asrecordedintheEM-DAT.SIDS

arealsoaffectedbyriverinefloodsanddroughts,althoughtheeconomicdamagestheycausearegenerallymuchlessthanthosecausedbytropicalcyclones.

4

3Theseareclassifiedasmeteorological(e.g.tropicalcyclone,invectivestorm),climatological(e.g.drought,wildfire)andhydrological(e.g.landslide,riverineflood,coastalflood,flashflood)intheEM-DAT.

4SomeSIDSarealsohighlyexposedtogeologicalhazardssuchasearthquakesandvolcaniceruptions.Theserisksalsoneedtobeincorporatedintothedisasterriskplanningandmanagementofgovernmentsandtheprivatesector.

5

ManySIDSarehighlyexposedtotropicalcyclones,becauseoftheirgeographicalposition,andarehighlyvulnerablebecauseofthesmallsizeoftheirlandmassandlackofeconomicdiversification.TheEM-DATrecords185incidencesoftropicalcyclonedisastersinthe16CaribbeanSIDSduringthe28-yearperiodbetween1995and2022.Italsorecorded74incidencesinthe12SIDSinOceaniaduringthisperiod.

5

TheEM-DATrecordstotalestimateddamages,comprisingalldamagesandeconomiclossesrelatedtothedisaster,butthesedatapertaintolessthanhalfoftherecordedtropicalcyclonedisasters.Table1showsthetotaldamagescausedbytropicalcyclonesandrecordedinEM-DATintheCaribbeanandOceanianSIDSfrom1995to2022,bycountryandregion,estimatedin2022USdollarprices.Thesedatamustbeunderestimatesoftheactualdamagecausedbytropicalcyclones,giventheabsenceofdataontotaldamagesformorethanhalfofthetropicalcyclonedisastersrecordedbyEM-DAT.ItislikelythatsomeSIDSprovidemorecomprehensivedataondamagesthanothers;forexample,datafromCubaappearstobemorecomprehensivethanthatfromHaiti.

6

Thedataindisasterdatabasesisunavoidablyimperfect,andaresubjecttoproblemssuchasoverorunderreportingofcertainhazardtypesandotherformsofbias(Gall,2015).

Table1TotalRecordedDamagescausedbyTropicalCyclonesinSIDSinthe

CaribbeanandOceania;1995-2022(USdollarmillions,2022prices)

Caribbean

DamagesUSDM.

Oceania

DamagesUSDM.

Cuba

18,374

Fiji

1,080

TheBahamas

8,478

Vanuatu

557

DominicanRepublic

4,586

Tonga

253

Haiti

3,241

Samoa

170

Dominica

2,709

Jamaica

2,273

Grenada

1,387

StKittsandNevis

1,193

AntiguaandBarbuda

1,167

Belize

985

AllCaribbeanSIDS

44,606

AllOceaniaSIDS

2,076

Source:EM-DAT

TotalrecordeddamagesfromtropicalcyclonessufferedbySIDSintheCaribbeanduring1995-

2022wasUSD44billion,whileSIDSinOceaniasuffereddamagesofUSD2billion.in2022

5SometropicalcycloneshitmultipleSIDS.

6FormostofthetropicalcyclonedisastersinCubaforwhichdeathswerereportedorwheretherewerelargenumbersofpeopleaffected,totaldamagesarealsorecorded.ButHaitiwasstruckbyseveraltropicalcyclonedisasterswhichkilledmanypeoplebutforwhichnofigurefordamageswererecorded;e.g.HurricaneNoelin

2007whichkilled90andaffectedmorethan100,000people,HurricaneGustavin2008whichkilled85andaffected73,000people,HurricaneHannain2008whichkilled529andaffected48,000peopleandHurricaneLaurain2020whichkilled39andaffected44,000people.Jonesetal(2023)discusstheproblemofmissingdataindisasterdatabasesandhowitishandledinempiricalresearch.

6

USdollarprices.ThelargedifferencebetweenthetworegionsreflectsnotonlythatCaribbeanSIDSsufferedfrommoretropicalcyclonedisastersthanSIDSinOceania,butalsothattheformerhaveeconomicassetsexposedtotropicalcyclonedamagethataremuchgreaterinvaluethanthoseofthelatter,becausethecombinedGDPofCaribbeanSIDSiseighttimeslargerthanthatofSIDSinOceania.

Table2showsalltheclimate-relatednaturaldisastersaffectingSIDSintheCaribbeanandOceaniafrom1995to2022,forwhichrecordedtotaldamagesexceeded10percentofGDP.AlthoughthelargereconomiesintheCaribbeansufferedthegreatesttotaldamages,someofthesmallerislandssufferedthelargestlossesrelativetotheirGDP,notablyAntiguaandBarbuda,Dominica,Grenada,StLucia,St.KittsandNevisandVanuatu,allofwhichincurredlossesgreaterthan50percentofGDPfromasingletropicalcyclonedisaster.

Table2Climaterelateddisasterscausingdamagesexceeding10percentofGDPin

theCaribbeanandOceania,1995-2022

Year

Country

TypeofDisasterEvent&Name

Damagesas%ofGDP

TheCaribbean

1995

AntiguaandBarbuda

HurricaneLuis

61

1995

StLucia

HurricaneLuis

67

1995

Dominica

HurricaneMarylin

64

1998

AntiguaandBarbuda

HurricaneGeorges

14

1998

St.KittsandNevis

HurricaneGeorges

109

1999

St.KittsandNevis

HurricaneLenny

11

2000

Belize

HurricaneKeith

25

2001

Belize

HurricaneIris

22

2004

TheBahamas

HurricaneFrances

11

2004

Grenada

HurricaneIvan

148

2005

Guyana

Riverineflood

27

2013

StVincentandGrenadines

Riverineflood

14

2015

Dominica

HurricaneErika

89

2016

Haiti

HurricaneMathew

14

2017

AntiguaandBarbuda

HurricaneIrma

17

2017

Dominica

HurricaneMaria

279

2019

TheBahamas

HurricaneDorian

26

Oceania

2001

Tonga

CycloneWaka

28

2012

Samoa

CycloneEvan

17

2014

Vanuatu

CyclonePam

60

2016

Fiji

CycloneWinston

12

2020

Tonga

CycloneHarold

23

Source:Em-Dat

7

WorldBank(2021)reportsestimatesoftheannualaveragelosses(AALs)inCaribbeancountriesfromfourdifferenttypesofnaturaldisaster,madebytheUnitedNationsOfficeforDisasterRiskReduction,usingestimatesofexposuretohazardsandassumptionsofassetvulnerability.HurricanesarethelargestsourceofriskformostCaribbeancountries,withtheBahamasfacingAALsfromhurricanesof4.8percentofGDP,AntiguaandBarbuda3.8percentofGDPandDominica2.7percentofGDP.DetailedcountryriskprofilespreparedbytheWorldBankforfourCaribbeancountriesprovideestimatesofboththeAALandProbableMaximumLoss(PML)

7

foreachcountry.Singlefamilybuildingsarethecategoryofassetwhichsuffersthelargestlossesineachcountry—between20percentand30percentofAALs(WorldBank,2021,p46).

Damagescausedbytropicalcyclonesarelikelytoincreaseoverthecourseofthe21stcenturybecauseofglobalwarming.Walshetal(2016)reviewresearchontropicalcyclonesandclimatechange.Theyarguethat,whilewestilllackagenerallyacceptedtheoryofclimateandtropicalcycloneformation,thereisawell-establishedtheoryofhowclimateaffectsthemaximumintensityandpotentialintensityoftropicalcyclones.Ingeneral,modelspredictthat,withglobalwarmingoverthecourseofthe21stcentury,therewillbeareductionintheglobalnumbersoftropicalcyclonesby5-30percent,butanincreaseincategoryfourandfive(highintensity)tropicalcyclonesby0to25percentandincreasesinmaximumintensityandrainfalloftropicalcyclones,althoughthenumericalpredictionsaresubjecttoahighdegreeofuncertainty.Thereisalsoapredictedincreaseinstormsurgescausedbythecombinationofsealevelrisesandmorehighintensitytropicalcyclones.

Themostintensetropicalcyclonescausethemostdamage.Acevedo(2016)estimateswindspeeddamageelasticityforhurricanesintheCaribbean,whichisbetweentwoandthreedependinguponwhetherthestormmakeslandfall.Hisfindingssuggeststhatdamageshavehistoricallybeenunderestimated.AverageannualdamagesfromstormsintheCaribbeanareestimatedat1.6percentofGDPbut,giventheunderestimation,couldbeashighas5.7percentofGDP.Acevedoalsomakesestimatesofhowmuchstormdamageswillincreasewithwarmertemperatures.Thisinvolvesthreeparameters:i)howmuchariseinglobaltemperature(GT)translatesintoariseinseasurfacetemperature(SST);ii)howmuchariseonSSTtranslatesintohighermaximumwindspeedsoftropicalcyclones;andiii)thewindspeeddamageelasticity.IntheRCP6.0scenario,

8

stormdamagesincreasebybetween11and46percentoverthecourseofthiscentury,althoughthesepredictionsarehighlyuncertain.

SomeofSIDSarehighlyexposedtodamagefromsealevelrises(SLR)causedbyglobalwarming.Acrosstheglobe,sealevelsarealreadyrisingbyanaverageofaround3.4millimeters(mm)perannum,andSLRwillcontinueastheplanetgetswarmer.SIDShaveextensivecoastalareas,someofwhichliebelowthefuturehightidewaterlevel,andhencewillbeexposedtopermanentinundationfromSLRandtheassociatedhighertidesaswellasepisodicfloodingfromextremesealevel(ESL)events.Modellingindicatesthattheexpectedannualfloodedarea(EAFA)ofallSIDSwillmorethantripleby2050underallglobalwarmingscenariosandwill

7TheProbableMaximumLossisthelargestlossesthatmightbeexpectedtooccurforagivenreturnperiod.Thereturnperiodistheinverseoftheannualprobabilitythatadisasterwilloccur;i.e.thelengthoftimeoverwhichaspecificdisasterwouldbeexpectedtooccuronce.

8ThisisaRepresentativeConcentrationPathwaymodelledbytheIntergovernmentalPanelonClimateChange(IPCC).TheRCP6.0isanintermediatepathwayintermsoftheriseinglobaltemperaturesandregardedbytheIPCCasthebaselinescenario,inwhichglobaltemperaturesriseby2.50Cwitharangeof1.60Cto3.60C.

8

risemorethansix-foldby2100(Vousdoukasetal(2023).Kiribatiistheworstaffected,withaprojectedEAFAofbetween35and55percentofitstotallandareaby2100.Tuvalu,theMarshallIslands,theBahamas,MicronesiaandtheMaldiveswillalsobebadlyaffected.

9

Withpopulationgrowthincoastalareasandeconomicgrowth,by2100,theexpectednumberofpeopleexposedtoannualfloodingisprojectedtoincreasearound4.5toseventimesdependingontheglobalwarmingscenario.Expectedannualdamages(EAD)willalsomultiplyseveraltimesinrealterms.TheMaldivesisprojectedtoincurEADof12.5percentofGDPby2100.

Doanetal(2023)developamethodologyforestimatingthenumbersofpeoplewhoareexposedtoextremeweathereventsandvulnerabletosufferingseverelossesiftheseeventsmaterialise.Theextremeweathereventsconsideredaredrought,flood,heatwaveandtropicalcyclone,andanextremeeventisdefinedintermsofagivenintensityandreturnperiod;forexample,anextremeweathereventforatropicalcycloneisonewithawindspeedofcategory2oraboveandareturnperiodof100years.Themethodologyenablesanestimatetobemadeofthetotalnumberofpeopleexposedtoeachextremeweathereventineachcountry,basedontheirgeographicallocationandtheprobabilityofsuchaneventoccurring.Vulnerabilityisdeterminedonthebasisofincome(peoplebeloworjustaboveapovertyline)andotherindicatorswhichareknowntobecorrelatedwithvulnerabilitytoshockssuchaseducationalattainment,accesstopipedwaterandelectricity,availabilityofasocialsafetynet,etc.Hencethepopulationwhichisvulnerabletoanextremeweathereventisasubsetofthetotalpopulationexposedtothatevent.ManyoftheSIDShaveaverylargeshareoftheirpopulationexposedtoanextremeweatherevent.Forexample,Belize,theDominicanRepublic,Fiji,Haiti,Jamaica,Mauritius,Samoa,St,Lucia,Tonga,TrinidadandTobagoandVanuatueachhavemostoralloftheirtotalpopulationexposed.However,thenumbersofpeoplewhoarecategorisedasvulnerabletoextremeweatherevetsismuchlowerinmostoftheSIDS,withtheexceptionofHaiti,becausetheyhaverelativelylowpovertyratesandgoodsocialindicators,comparedtootherEMDEs.Assuch,althoughalargeshareofthepopulationisexposedtoextremeweathereventsinmanySIDS,theirrelativelyhighincomelevels,educationalattainment,etcprovidethepopulationwithsomeresiliencetotheseshocks.

Empiricalresearchontheeconomiccostsofclimaterelatednaturaldisasters

ResearchershaveusedvariousmethodstoestimatetheimpactoftropicalcyclonesandothernaturaldisastersonGDPandmacro-fiscalvariables,includingvectorautoregressions(VARs)andeventstudies.Acevedo(2014)usedapanelVARmodeltostudytheimpactoffloodsandhurricanesonGDPandpublicdebtovera40-yearperiodintheCaribbean.BothfloodsandstormshaveanegativeimpactonGDP.Floodsleadtoariseinthedebt-to-GDPratio,butstorms(especiallysevereones)donot,possiblybecausetheyleadtoincreasedaidinflowsordebtreliefmeasures.Stormshavemorepronouncedeffectsonasub-sampleofCaribbeancountries—thememberstatesoftheEasternCaribbeanCurrencyUnion(ECCU),whicharegenerallysmallerandthereforemorevulnerabletosystemicdamagefromhurricanes.StormsreduceGDPintheECCUcountries,withseverestormshavingalargerimpact,andboth

9SeveraldependentterritoriesarealsoprojectedtosufferextensiveEAFAfromSLR,especiallytheTurksandCaicos,BermudaandtheCaymanIslands.

9

moderateandseverestormsincreasethedebt-to-GDPratio,althoughonlytheformerisstatisticallysignificant.

Cabezonetal(2016)estimatedthefiscalimpactofnaturaldisastersinfivePacificIslandcountriesusingaVARcoveringyearsbetween1970and2013.Disastersarescaledintermsofintensity(deathsandnumberofpeopleaffectedrelativetototalpopulation).NaturaldisasterswhichcausedamagesandlossesequivalenttoonepercentofGDPcauseanaveragefallinGDPof0.7percentintheyearofthedisasterandadeteriorationinthefiscalbalanceof0.5percentofGDP(alsointheyearofthedisaster),withtaxrevenuesfallingby0.2percentofGDPandspendingrisingby0.7percentofGDP(theimpactonthefiscalbalanceismitigatedbyariseingrants).

Inthefollowinganalysis,weuseaneventstudymethodologytoexaminetheimpactofhurricaneswhichcausedlargedamagesonislandsinthePacificandtheCaribbean,onGDPgrowth,generalgovernmentrevenuesandexpenditures,theoverallfiscalbalanceand(wh