世界銀行-印尼投資改革的早期影響：初步分析（英）

PublicDisclosureAuthorizedPublicDisclosureAuthorizedPublicDisclosureAuthorizedPublicDisclosureAuthorized

PolicyResearchWorkingPaper10478

EarlyImpactsofIndonesia’sInvestmentReforms

APreliminaryAnalysis

AngellaFaithMontfaucon

VictorKidakeSenelwa

AufaDoarest

Macroeconomics,TradeandInvestmentGlobalPractice&

Finance,CompetitivenessandInnovationGlobalPractice

June2023

ProducedbytheResearchSupportTeam

PolicyResearchWorkingPaper10478

Abstract

directinvestmentsuggestthattheincreaseinfullyliber-alizedsectorsislikelytocontinue.Theresultsondirectdomesticinvestmentsuggestapossiblecrowding-ineffect.Amongthefullyliberalizedsectors,thebasemetalindustrywasakeydriverofgrowth,andsectorslinkedtoSustainableDevelopmentGoalssectorshadmixedresults.Thefindingsprovidesuggestiveevidenceofthecomplementaryeffectoftradereforms,althoughtheanalysisdoesnotspecifywhichoftheseveralreformsmayhaveledtotheincrease.TheseresultsarerobustwhenthepossibleeffectsofCovid-19recoveryandothermacroeconomicfactorsarecontrolledfor.Theseresultsarealsorobusttoalternativeeventstudymodels.Furtheranalysiswouldbeneededtoobservethetrajectoryinboththequalityandquantityofinvestmentgoingforward,thedistributionaleffectsandtheneededcomplementaryreformstoensuresustainablegainsbeyondtheshortrun.

TheIndonesiangovernmentimplementedcomprehen-

siveinvestmentreformsin2021toencourageinvestment

inflowsandrelatedpositiveimpacts.Comparedtothe

previousinvestmentregulationin2016,thenewdecree

removedforeigndirectinvestmentrestrictionsinover500

businessactivities.Byestimatingthedifferenceindiffer-

enceindifferenceeventstudymodel,thispaperempirically

assessestheresponsein(realizedandplanned)foreigndirect

investmentandrealizeddomesticdirectinvestmenttothe

reforms.Thepaperalsoassesseswhethertherewasgrowthin

investmentsinfullyliberalizedsectorslinkedtoSustainable

DevelopmentGoalsasaproxyforthequalityofinvestment.

Theresultssuggestthattheinvestmentreformswereasso-

ciatedwithincreasesinrealizedforeigndirectinvestment,

realizeddirectdomesticinvestment,andplannedforeign

directinvestment,especiallyinfullyliberalizedsectors,

whiletherewasadeclineinallthreetypesofinvestmentsin

thenon-liberalizedsectors.Theresultsforplannedforeign

ThispaperisaproductoftheMacroeconomics,TradeandInvestmentGlobalPracticeandtheFinance,CompetitivenessandInnovationGlobalPractice.ItispartofalargereffortbytheWorldBanktoprovideopenaccesstoitsresearchandmakeacontributiontodevelopmentpolicydiscussionsaroundtheworld.PolicyResearchWorkingPapersarealsopostedontheWebat/prwp.Theauthorsmaybecontactedatamontfaucon@.

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

EarlyImpactsofIndonesia’sInvestmentReforms:APreliminaryAnalysis*

AngellaFaithMontfaucon?VictorKidakeSenelwa?AufaDoarest§

JELclassification:F13,F21,F23,D25,L51

Keywords:Investmentliberalization,foreigndirectinvestment,domesticdirectinvestment,tradepolicy

*TheauthorswouldliketothankBayuAgnimarutoforhelpinassemblingtherealizedinvestmentdataandPeterKusekforhelpinaccessingtheplannedinvestmentdata.WethankCsillaLakatosandYuCaofortheircomprehensivecomments.WethankGerlinMayU.CatanguiandFredericoGilSander,HabibRab,WaelMansour,Aristomene

VaroudakisandMassimilianoCaliforhelpfuldiscussions.TheviewshereindonotnecessarilyreflecttheviewsoftheWorldBank.Allerrorsareourown.

?WorldBank,amontfaucon@.

?WorldBank.vsenelwa@

§WorldBank.adoarest@

1

1Introduction

Regulationstowardtheliberalizationofglobalcapitalflowshavebeendebated,whichreflectstheconcernsaboutthelossofnationalsovereignrightsandotherpotentiallynegativeconsequencesoffreecapitalmovements.Foreigndirectinvestment(FDI),morethanothertypesofcapitalflows,hashistoricallysparkedsuchconcernsbecauseitofteninvolvesacontrollingstakebylargemulti-nationalcorporationsoverwhichdomesticauthorities,itisfeared,havelittlecontrol.Forthesereasons,governmentshavesometimesimposedrestrictionsoninwardFDI.Ontheotherhand,re-formsthatfacilitateforeigndirectinvestmententryandoperationsappeartohaveasignificantandsizableimpactonthelevelofinwardFDI,whereasanti-competitiveproductmarketregulationsarenegativelyassociatedwithinvestment(

MisturaandRoulet

,

2019

).

Beforethereformsanalyzedinthispaper,IndonesiahadthethirdhighestrestrictionsonFDIamongthehigh-andmiddle-incomecountriessurveyedbytheOrganisationforEconomicCo-operationandDevelopment(OECD)andwasthesecondmostrestrictivebehindallcountriesintheEastAsiaregionnexttothePhilippines(

WorldBank

,

2018

).Indonesia’snegativeinvestmentlist(DaftarNegatifInvestasi-DNI)includedwidespreadforeignequitylimits,sectoralreservationsforMSMEs,andspeciallicensingregimes.Untilearly2021,theDNIappliedatleastoneinvestmentrestrictioninalmostone-thirdofalleconomicsectors,andin20percentofthemiteitherlimitedforeignequityparticipationorprohibitedforeigninvestmentaltogether.Theserestrictionssignifi-cantlyreducedbothforeignanddomesticinvestments,reducedentryandperformance,increasedpricesinthesectorstowhichtheywereapplied(

WorldBank

,

2018

),andwereassociatedwithproductivitylosses(

GenthnerandKis-Katos

,

2022

).

In2021,theGovernmentofIndonesia(GoI)implementeditsmostcomprehensiveinvest-mentreformprogramindecades.MostdirectrestrictionsoninvestmentswereremovedthroughaFebruary2021presidentialregulation(Perpres)10/2021,lateramendedbyPerpres49/2021onInvestmentinMay2021(

WorldBank

,

2022

).Thenewpresidentialregulationoninvestmentsig-nificantlyreducedbusinessactivitiessubjecttotwomainrestrictions:(i)sectorsthatarereservedforMSMEsandCooperatives,oropenwithrequirementstopartner/cooperatewithMSMEsand

2

cooperativesand(ii)sectorsthatareopenwithcertainconditions–i.e.,maxforeignownershiporcertainrecommendations

1

Thereformeliminatedforeignequitylimitsacrossawiderangeofsec-torsandturnedmanysectorswhereFDIwasnotallowedorrestrictedbyminorityshareholdingintosectorsfullyopentoFDI.ThereformswereintendedtoboostFDIinflowsandattract/encouragetheentranceofmultinationalcorporations(MNCs).Inthispaper,theterm“fullyliberalized”referstotheremovalofalltheserestrictionsinallthebusinessactivitiesorsub-sectorwithinabroadersector,i.ethenumberofbusinessactivitieswithinasectorthathasbeenliberalized(morebelow).

Thisstudyassessestheimpactofthisreform,specificallymakingthefollowingcontribu-tions:First,weexaminetheearlyeffectsoftheregulatoryreformonbothplannedandrealizedFDIaswellasrealizeddirectdomesticinvestments(DDI),bycomparingthepre-andpost-policyperiodsandcomparingacrosssectorsinresponsetothemostrecentversionoftheregulation(Per-pres49/2021);byanalyzingDDI,wetestwhetherforeignfirmssubstituteorcrowd-indomesticinvestments(

Barriosetal.

,

2018

).Wealsoassesswhethertherewasgrowthamongfullyliberal-izedsectorsinsustainableinvestment(i.e.investmentslinkedtoSustainableDevelopmentGoals(SDGs)asdefinedby

UNCTAD

(

2022

)).

Theanalysisisdoneusingadifferenceindifferenceindifference(DDD)eventstudymodelandcontrollingforthesectorandtime-fixedeffects.Asthedataisatamoreaggregatedlevelthanthelistedbusinessactivitiesinthereformregulation,theterm“fullyliberalized”referstoallthesub-sectorsorbusinessactivitieswithinthebroadsectorthathavehadalltherestrictionsdiscussedremoved(moreinthedatasection).Wefocusonliberalizednon-commoditysectorstohighlightinvestmentslesslikelytohavepressureonthenaturalenvironmentfromtheexpectedincrease.Therefore,inlinewiththeWorldBank’ssupporttothereform.

2

Asrobustness,wefurtheraccountfortwokeyeventsthatweretakingplacearoundthereformtime:therecoveryfromCovid-19lockdownsaroundtheglobeandrelatedtradereforms,especiallyonnon-tariffmeasuresthegovernmentundertookinNovember2021(

Montfauconetal.

,

2023

).Thetradepoliciesalso

1Perpres49/2021deletedthespecificappendixcontainingtheclosedbusinessactivities.TheclosedsectorswerelistedinJobCreationOmnibusLawandthemaintextofthenewinvestmentregulation.

2ThisisinlinewiththeWorldBanksupporttothereformasdetailed

https://projects.worldbank.

org/en/projects-operations/project-detail/P172439

3

servesasaproxyfortheinstitutionallandscapewhichiskeytoretainingbenefitstoFDI(

Obuobi

etal.

,

2022

).

ResultssuggestthattheinvestmentreformswereassociatedwithincreasesinrealizedFDI,realizedDDI,andplannedFDI.Investmentsincreasedmostinthefullyliberalizednon-commoditysectors(23percentforFDI,25percentforDDI,and19percentforplannedFDIrelativetothegrowthofthepre-policyperiod).Ontheotherhand,therewasadeclineinallthreetypesofinvestmentsinthenon-liberalizedsectors.PlannedFDIresultssuggestthattheincreaseinfullyliberalizedsectorsislikelytocontinueandtheresultsonDDIsuggestacrowding-ineffect.Amongthefullyliberalizedsectors,thebasemetalindustrywasakeydriverofgrowth.WhiletherewasincreasedplannedFDIinwatermanagement,andinsomeinfrastructure-relatedsectorsforrealizedDDIandFDI,otherSDG-relatedsectorssawmixedresults.Toaccountfortheinstitutionalcontext,weassesstheeffectofsignificantchangesintraderegulationsthattookplace6monthsaftertheinvestmentreforms.WefindthatFDIinfullyliberalizednon-commoditysectorsincreasedfollowingthetradepolicychanges,buttheanalysisdoesnotspecifywhichofthechangesmayhaveledtothisincrease.

3

Thissuggeststhecomplementarityoftradeandinvestmentreforms.WefindthattheseresultsarerobustevenwhenpossibleeffectsofCovid-19recoveryarecontrolledfor,whenweincludeGDPandnominalexchangeratevariablesascovariatesandtheresultsarerobusttoalternativeeventstudymodels.

EarlyresearchonFDItraceditsoriginstointernationaltradetheory,whichidentifiedthecomparativeadvantageofhostcountriesasthemostimportantdeterminantofFDI.Whenforeigncompaniesinvestinacountryinordertogetaccesstoitsdomesticmarketanduseitasaplatformforexportingtoothercountries,itmighthavea”homemarketeffect,”accordingto

Dunning

(

1993

).ThiscouldleadtoincreasedFDIinflowsandimproveddomesticeconomiccompetitiveness.Ourmainresultsareinlinewiththisliterature.Recently,studieshavehighlightedtheimportanceofenhancingandsustainingthehost-countryeffectwiththegovernment’srole,thusshiftingthefocustogovernmentpoliciestoincreaseFDIinflows.Asidefromtheoutstandingroleofreformsinthe

3

Montfauconetal.

(

2023

)providedetailofthesetradereformsandlinktothedataonwhichmeasureswerechanged.

4

hostcountry,institutionsinthehostcountryplayanimportantrole,asnotedby

AitkenandHarrison

(

2017

).CountrieswithstronginstitutionsandeffectivegovernanceattractmoreFDI.MNCsprefercountrieswiththebestinstitutionalarrangements,whicharecharacterizedbyfriendlyglobaltradepolicies(

Cantahetal.

,

2018

;

Obuobietal.

,

2022

).Ourresultsontradepolicyareinlinewiththislineofliterature.

Chen

(

2018

)establishedthatinvestmentliberalizationcanbenefitdomesticinvestmentwhenaccompaniedbycomplementarypoliciesthatpromotedomesticenterprisegrowth.Liberalizationmayharmdomesticinvestmentbyreplacingindigenousenterpriseswithforeigncompetitors,henceFDIliberalizationmeasuresshouldbeviewedasjustonecomponentofaliberalizationstrategyandmustbesupportedbyadjustmentstoliberalizelocalmarketsinordertohelpmaximizethebenefitsofFDI.OurresultssuggestthatthereformsIndonesiatookbroadlyhadapositiveimpactondomesticinvestmentsimilartoresultsby

Setiyanto

(

2023

).

Overall,thebodyofresearchillustratesthecomplexitiesandcontext-dependenceoftheef-fectsofinvestmentliberalizationonbothdomesticandinternationalinvestmentflows.Morere-searchisneededtofullyunderstandthevariablesthatgoverntheamountanddirectionoftheseim-pacts,whichthisstudycontributesto,withmorerecentmethodologyinthedifference-in-differenceliteratureandanattempttocloselylinkreformstothespecificsectorstargeted.Therefore,whiletheseresultssuggestthatthereformswereakeyroleinboostinginvestmentsinIndonesia,comple-mentarytradeandotherreformswouldensuregrowthissustained,especially,forexample,sincerestrictionsintheservicestradesectorremainthehighestinIndonesia(

LuandBatten

,

2023

).Fur-ther,thebroaderimpactsofthereformsontheeconomyandvariousgroupsarenotcapturedandarebeyondthescopeofthispaper.

Therestofthispaperisorganizedasfollows.In

section2

wedetailthepolicycontextanddata.In

section3

themethodologyisdiscussedandsubsequentlyresultsarepresentedin

section4

,includingfromrobustnesschecks.Finally,

section5

concludes.

5

2Context,DataandTrendsinFDI

2.1InvestmentRestrictioninIndonesia

TheGovernmentofIndonesia(GoI)implementeditsmostcomprehensiveinvestmentreformpro-gramindecadesthroughJobCreationOmnibusLaw.PromulgatedinNovember2020,theJobCreationOmnibusLawaimedtoimprovetheinvestmentclimatebyamendingdozensofindivid-uallaws.TheOmnibusBillanditsimplementingregulationonInvestment(presidentialDecree-Perpres10/2021-lateramendedbyPerpres49/2021onInvestment)

4

significantlyreducedbusinessactivitiessubjecttotwomainrestrictions:(i)sectorsthatarereservedforMSMEsandCoopera-tives,oropenwithrequirementstopartner/cooperatewithMSMEsandcooperativesand(ii)sectorsthatareopenwithcertainconditions–i.e.,maxforeignownershiporspecificrecommendations.Whilethesereformsdonotresultinfullinvestmentliberalization,theypavethewayforthesubse-quentreductionininvestmentrestrictions.

Perpres49/2021combinedpositiveandnegativeapproachestoFDIinflowtoIndonesia.Sim-ilartoInvestmentLawbefore2007(LawNo.1/1967),thePerpres49/2021introducedthelistofprioritysectorsforInvestmentinIndonesia.Investorsinindustrieslistedintheprioritylist(e.g.,basicchemicalandspecificfoodproductsindustries)willreceivefiscalandnon-fiscalincentivesfromtheGovernmentofIndonesia(GoI).Atthesametime,thenewregulationstillimplementedthenegativelistapproachasinInvestmentLawNo:25/2017.Inthisnegativelistapproach,allbusinessfieldsareopento100percentforeignequityownershipunlessotherwisestatedintheim-plementingregulation.Anydeviations(specificallyintheformofrestrictions)areoutlinedinaPresidentialDecree(Perpres)onNegativeInvestmentList-DaftarNegatifInvestasi(DNI).

IndonesiaremovedmostinvestmentrestrictionsthroughtherecentPresidentialRegulationonInvestmentbutupheldtheprotectionofselectsectors,suchasagricultureandtransport.ThePresidentialRegulationeliminatedforeignequitylimitsacrossawiderangeofsectors.Comparedtothepreviousinvestmentregulationin2016,thenewdecreeremovedFDIrestrictionsin553

4Fromtheauthors’estimations,thisamendmentre-addedrestrictionstosomebusinessactivitiesthatwereprevi-ouslyliberalizedinPerpres10/2021butoverallthereformremainedlargeandcomprehensive.

6

businessactivities.Asaresult,theshareofsectorsthatwereexposedtoforeignequitylimit(FEL)fellfrom34.1%in2016to7.6%in2021(figure

1

a).ThusthereformleftseveralsectorsfullyopentoFDI,suchas:(i)relativelysmallactivitiesfullyorpartiallyreservedforMicro,Small,andMediumEnterprises(MSMEs)and(ii)sectorsconsideredtobeofpublicinterestthatrequireminorityornoforeignownership.Withtheremainingrestrictions,theshareofbusinessactivitiessubjecttoatleastoneinvestmentrestrictiondecreasedfrom38.5%to20.4%ofalleconomicsectors(figure

1

b).

Figure1:EvolutionofInvestmentRestrictioninIndonesia

a.ShareofsectorwithFELrestrictionb.ShareofsectorwithanyrestrictionSource:AuthorcalculationbasedonseriesofPresidentialDecreeonInvestment

2.2Data

Twosetsofdataareusedinthisanalysis:realizedFDIandDDIvalues,sourcedfromIndonesia’sMinistryofInvestment(henceforthBKPM);andplannedFDIsourcedfromfDimarketsdatawhichalsoproxiesinvestorsentiments/expectations.TheBKPMdataincludesalltypesofFDIwhilethefDimarketsdataincludedinthisanalysisdoesnotincludemergersandacquisitions.BKPMdata

isataquarterlyleveluptoSeptember2022whilefDimarketdataisatamonthlyleveluptoMarch2022.

KlasifikasiBakuLapanganUsahaIndonesiaorKBLIistheofficialbusinessclassificationinIndonesia,publishedbytheIndonesiaCentreofStatisticswithreferencetotheInternationalStan-dardClassificationofAllEconomicActivities(ISIC).Theregulationwasatahighlydisaggregatedlevelatbusinessactivitiesof5-digitKBLI(seelistintheAnnex

TableA1

).However,bothdata

7

areonlyavailableatamoreaggregatedlevelatthe2-digitKBLIlevelofaggregation.Assuch,mappingsectors1:1tothelistofbusinessactivitiesintheregulationisnotpossible.Instead,weareabletomapthesectorsintheregulationto5-digitKBLIandthencalculatetheshareof5-digitKBLIthatareliberalizedwithina2-digitsector.

Thereareatotalof68sectorsat2-digitKBLI.Amongthese,6ofthesectorsbelongtotheprimarysector,22tothesecondarysectors,and40inthetertiarysector,whichrepresents9percent,32percentand59percentrespectively(see

TableA1

whichlists62sectorsexcludingtheprimarysector).Between2019andQ3of2022,theprimarysectorrepresentedabout7percentofFDIvalues,thesecondarysector49percent,andthetertiarysector43percent.TheFDIvaluesinthesecondarysectorincreasedinthistimeperiod(35percentsharein2019to58percentshareinQ3of2022)whiletheshareofthetertiarysectordecreased(58percentin2019to43percentinQ3of2022)andprimarysectorsstayedaboutthesame.Ouranalysisfocusesonnon-commoditysectors,thatis,weexcludetheprimarysector,andtrendsinFDIbeforeandaftertheperiodofthereform.

Atotalof24ofthe2-digitKBLInon-commoditysectors(secondaryandtertiary),repre-senting39percent,had100percentofthe5-digitsub-sectorswithinthemliberalizedbyPerpres49/2021(i.efullyliberalized).Thesharebetweenprimary,secondary,andtertiarysectorsintermsofFDIvalueswithinthesefullyliberalizedsectorsisbroadlysimilartoallofFDI.

ThisanalysiswillbebasedonPerpres49/2021whichcameintoeffectinMay2021asanamendmentto10of2021onInvestmentBusinessSectors.Thenewregulationrevokesthe2016NegativeList(promulgatedunderPerpres44of2016ontheListofBusinessActivitiesClosedandOpentoInvestment)andprovidesanew,significantlyliberalized,investmentlist.Therefore,intheremainderofthissection,thebefore-policyperiodisdefinedastheperiodfromthefirstquarterof2019tothefirstquarterof2021andtheafter-policyperiodisfromthethirdquarterof2021tothethirdquarterof2023forrealizedFDI.ForplannedFDI,whichisatamonthlylevel,thepre-policyisfromJanuary2019toMay2021,andthepost-policyisfromJune2021toMarch2022.

8

2.3TrendsinRealizedFDI

Overall,totalrealizedFDIgrewby64.7percentbetweenthepre-andpost-policyperiods.RealizedFDIfortheprimarysector,secondarysector,andtertiarysectorincreasedby73.9percent,92.5

percent,and21.5percentrespectively,fortheperiodafterthepolicyrelativetobeforethepolicy.FDIinfullyliberalizednon-commoditysectors(i.esecondaryandtertiarysectors)increased

by40percentbetweenthepre-andpost-policyperiodcomparedtojust6percentinnon-fullyreal-izedsectors(

Figure2

,a).AlthoughrealizedFDIinthefullyliberalizedsectorswasalreadyhigherthaninthenon-fullyliberalizedsectors,itincreasedmoresteeplyandconsistentlywhileothersectorsremainedflat,orgrowthwasmorevolatile(

Figure2

,b).Forthenon-commoditysectorsliberalizedat100percent,realizedFDIincreasedby49.8percent,whiletheaveragegrowthforthesectorsliberalizedby,between40percentto99percentwas24.4percentbetweenthepreandpost-policychangeperiods.Thissignalsthatliberalizationmayhavetargetedsectorsinwhichinvestorswerealreadyinterestedin,anddoublingdownonthereformwouldlikelyproducemoreresults.

Figure2:Pre-andPost-Policygrowthinnon-commoditysectors

a.PreandPostPolicyAveragesb.AveragesbyLevelofLiberalizationSource:AuthorestimationsfromBKPMdata

Growthwasdrivenmainlybythepaperindustry,basemetalsandtobaccoprocessingwhichhadthelargestvaluecontribution

Figure3

.Thelargestgrowthbetweenthepre-andpost-policychangeperiodwasalsoevidencedinactivitiesofserviceprovidersforbuildingsandparks;sup-portingactivitiesforfinancialservices,insurance,andpensionfunds.Duetothenatureoftheaggregateddata,weareunabletodissectFDIlinkedtosustainabilityinIndonesiaexhaustively.

9

However,fromthelistof2-digitsectorsthatarefullyliberalized,weseethattherewasa190per-centincreaseinFDIineducationinthepost-policyrelativetopre-policyanda212percentincreaseinwatermanagement.

Figure3:GrowthinRealizedFDIinselected100percentLiberalizedSectors

Source:Author’scompilationusingBKPMData

2.4TrendsinPlannedFDI

PlannedFDIgrowthgravitatedtowardsfullyliberalizednon-commoditysectors(

Figure4

,a).PlannedFDIforfullyliberalizednon-commoditysectorsgrewby30.6percentafterthepolicy,whilesectorsliberalizedat40-59percentdeclinedby16percent,60-79percentdeclinedby27.4percent,and80-99percentincreasedby0.1percent(

Figure4

,b).Thus,thisisaclearshiftinFDItowardsmoreliberalizedsectors.

Asthesetrendsarepurelydescriptive,amoreformalanalysiswouldenabledisentanglingtheeffectsofFDIgrowthfollowingthePerpres49/2021investmentreformpolicy,andtheeconomicrecoveryfromCovid-19thatmakesupmostofthepost-policyperiod.

10

Figure4:Pre-andPost-PolicygrowthinplannedFDIinthenon-commoditysector

Source:fDiMarkets

3EmpiricalMethodology

EmpiricalMethodologyToinvestigatethecausaleffectofthereformondirectinvestment,weuseapaneldatadifferenceindifferenceindifference(DDD)eventstudyapproach.Ourresearchsetupisbasedonthestandardcanonicaldifference-in-differenceapproach,withtwotimeperiods(beforeandafterthepolicy),andtwogroups:inthefirst,nosectoristreated/liberalized,andinthesecond,somesectorsaretreated(liberalized),butsomearenot(thecontrolgroup).

StandardDIDapproaches,however,asproposedby

CallawayandSant’Anna

(

2019

),facebackdropsbasedonpotentialoutcomesandfailtoaccountfortreatmenteffectheterogeneity,whichcanalsovarywithobservedcovariatesandtime.Toaddresssuchshortcomings,weusethe

Sun

andAbraham

(

2021

)approachbasedonastaggeredDIDsetup,whichstatesthatonceagroupistreated/liberalized,itremainstreatedforthenextperiod,andattheend,someunitsremainuntre