Enabling High-frequency and Real-time Poverty Monitoring in the Developing World with SWIFT

PublicDisclosureAuthorized POVERTY PublicDisclosureAuthorized POVERTYANDEQUITY EQUITABLEGROWTH,FINANCE&INSTITUTIONSINSIGHT PublicDisclosureAuthorized EnablingHigh-frequencyandReal-timePovertyMonitoringinthe DevelopingWorldwith SWIFT SurveyofWellbeingvia InstantandFrequentTracking PublicDisclosureAuthorized NobuoYoshidaandDanielleVictoriaAron ©2024InternationalBankforReconstructionandDevelopment/TheWorldBank1818HStreetNW WashingtonDC20433 Telephone:202-473-1000Internet:www.worldbank.org ThisworkisaproductofthestaffofTheWorldBankwithexternalcontributions.Thefindings,interpretations,andconclusionsexpressedinthisworkdonotnecessarilyreflecttheviewsofTheWorldBank,itsBoardofExecutiveDirectors,orthegovernmentstheyrepresent. TheWorldBankdoesnotguaranteetheaccuracy,completeness,orcurrencyofthedataincludedinthisworkanddoesnotassumeresponsibilityforanyerrors,omissions,ordiscrepanciesintheinformation,orliabilitywithrespecttotheuseoforfailuretousetheinformation,methods,processes,orconclusionssetforth.Theboundaries,colors,denominations,andotherinformationshownonanymapinthisworkdonotimplyanyjudgmentonthepartofTheWorldBankconcerningthelegalstatusofanyterritoryortheendorsementoracceptanceofsuchboundaries. Nothinghereinshallconstituteorbeconstruedorconsideredtobealimitationuponorwaiveroftheprivileges andimmunitiesofTheWorldBank,allofwhicharespecificallyreserved. RightsandPermissions Thematerialinthisworkissubjecttocopyright.BecauseTheWorldBankencouragesdisseminationofitsknowledge,thisworkmaybereproduced,inwholeorinpart,fornoncommercialpurposesaslongasfullattributiontothisworkisgiven. Anyqueriesonrightsandlicenses,includingsubsidiaryrights,shouldbeaddressedtoWorldBankPublications,TheWorldBankGroup,1818HStreetNW,Washington,DC20433,USA;fax:202-522-2625;e-mail:pubrights@worldbank.org. ACKNOWLEDGMENT OurheartfeltthankstothecontributorsfortheiressentialinsightsonthecountrycasestudiesinSection6ofthisreport.SpecialmentionsincludeXueqiLiandElianaCarolinaRubianoMatulevichforParaguay;CarolinaDiaz-BonillaforBotswana;ParipoornaBaxi,withassistancefromShinyaTakamatsufortheDRC;AnjaliKini,withinputsfromDhirajSharma,ShinyaTakamatsu,andRobSwinkelsforZimbabwe;JeremySchneider,withinsightsfromAzizAtamanovforUganda;AnjaliKiniandXueqiLi,withLydiaKim'scontributionsforMongolia;MariaGabrielaFarfanBetranandSebastianPatrickAlexandreSilvaLeanderforZambia;andParipoornaBaxi,withMartaSchoch'sinputsforNigeria.Theirexpertisesignificantlyenrichedthisreport,highlightingthediverseapplicationsandimpactsoftheSWIFTmethodologyworldwide. SpecialappreciationgoestoJeremyEvanSchneiderforhisdedicatedassistanceindraftingthisreport.We'realsogratefulfortheinsightfulandconstructivefeedbackprovidedbyTaraVishwanath,GabrielLaraIbarra,MariaEugeniaGenoni,CarolinaDiaz-Bonilla,NandiniKrishnan,MariaFernandaGonzalezIcaza,LauraLilianaMorenoHerrera,RoyVanderWeide,Hai-Anh H.Dang,DavidNewhouse,DeanMitchellJolliffe,TalipKilic,KseniaAbanokova,KimberlyBlairBolch,HenryStemmler,andallparticipantsinthereviewmeetingattheWorldBank.OurappreciationalsoextendstoCarlosSabatinoforhiseditorialreview,andtoLuis-FelipeLopez-CalvaandBenuBidanifortheirleadershipandsupportinthepreparationofthisreport. TableofContents 1.Introduction6 2.FrameworkforSWIFTPovertyEstimaton8 A.Modeldevelopment9 B.Collectionorharmonizationofthetargetdata10 C.Imputation11 D.ImprovedmodelstabilitywithSWIFTPlusandSWIFT2.011 E.ReliabilityofSWIFT-basedpovertyestimates12 F.AdvantagesoftheSWIFTframeworkanditscaveats13 3.PovertyEstimationwithSWIFT:AFlexibleFrameworkforDiverseUseCases14 A.Increasingthefrequencyofpovertystatisticsusingavailablefrequentsurveys14 B.Producingpovertystatisticswhenanappropriatetrainingdatasetisnotalreadyavailable(SWIFT2.0)15 C.Rapidpovertymonitoringwithnon-traditionaldatacollection16 D.Restoringcomparabilitytoreestablishapovertytrend16 4.CostImplicationsoftheSWIFTFramework18 (i)FrequentHouseholdSurveyAlreadyAvailable18 (ii)NoAppropriateDataforTraining19 (iii)NoFrequentHouseholdSurveyAvailable,butPlansforaFrequentPhoneSurvey19 (iv)NoFrequentHouseholdSurveyAvailable,butPlansforCommunity-BasedDataCollection20 5.BriefSummaryoftheSWIFTFramework22 6.CountryExamples26 (i)Paraguay:utilizingquarterlysurveystoprovidepovertystatistics26 (ii)Botswana:utilizingquarterlylaborforcesurveysforpovertystatistics28 (iii)DemocraticRepublicofCongo:providingapovertytrenddespitealackofcomparablehouseholdsurveys31 (iv)Zimbabwe:usingSWIFTtoprovidepovertystatisticsafterunexpectedeconomicchanges32 (v)Malawi:providingtimelypovertydatainaclimatecrisiscontext34 (vi)Uganda:providingtimelypovertydatainarefugeecrisiscontext36 (vii)Mongolia:restoringpovertytrendsafterimprovementstothehouseholdsurvey38 (viii)Zambia:restoringpovertytrendsduetodifferencesinhouseholdsurveys40 (ix)Nigeria:usingavailablepastsurveystouncoverapovertytrend43 7.Conclusion46 Annex1.Metadataforcountrycases50 Annex2.Simulationprocess52 Figures Figure1StepsoftheSWIFTframework23 Figure2Paraguayquarterlypovertyprojectionresults28 Figure3Botswanapovertyprojectionresults30 Fig