trade openness and gender wage gap: evidence from indonesia
TRANSCRIPT
25
Journal of International Cooperation Studies, Vol.26, No.1(2018.7)
Trade Openness and Gender Wage Gap: Evidence from Indonesia
Maryam JAMIELAA*and
Koji KAWABATA**
Ⅰ.Introduction
Manydevelopingcountrieshaveengaged intotheglobalmarketsinceearly1980s.As
acountrybecamemore integrated to themarket, tradevolumeandattractedmany foreign
direct investments inthatcountryalso increasesubstantially.Asaconsequence, thiseconomic
condition impactsmanysectors including itsdomestic labormarket.Accordingto literatures,
tradeliberalizationwillcreatedomesticcompetitionwhichtendstoreducediscriminationinlabor
market,includinggenderwagedisparity(Becker,1957).
Inpast twodecades, Indonesiahasexperiencedan increase in total tradeandFDI.
Moreover,ascanbeseeninFigure1,theincreasingofFDIvolumeisalsofollowedbyawidening
trendon femaleandmaleearningsgap fromperiod2008-2014.Basedonthe facts, thisstudy
investigatestheimpactoftradeopennessongenderwagegapandhowitsimpactdiffersacross
wagedistribution.
Tradeliberalizationcanaffectgenderwagedifferentials,bothinnegativeorinpositive
ways.Firstly,accordingtoBecker(1957), inordertobecompetitivewith foreigncompanies,
localfirmsarepushedtobemoreefficient inoperatingtheirproduction.Tobemoreefficient,
localcompanieswilleliminate their “tasteofdiscrimination”andpayworkers’wagebasedon
theirproductivity.Throughthismechanism,female-maleearninggapisbelievedtobedecreased
(ArteconaandCunningham,2002;BlackandBrainerd,2004).Secondly, indevelopingcountries,
internationaltradewill inducetheirproductionshift into labor intensiveproductionwhichused
low-skilledlaborratherthanhigh-skilledlabor.Sincefemalestereotypeisattachedtolow-skilled
* DataandInformationstaff,CenterforStrategicStudies,SecretariatGeneral,MinistryofPublicWorks.** AssociateProfessor,GraduateSchoolofInternationalCooperationStudiesatKobeUniversity.
26 国 際 協 力 論 集 第 26 巻 第 1 号
labor,genderwagegapwillbereducedthroughthismechanism(Chenetal., 2013).Another
literaturesuggeststhattradeopennessalsocreatesskilled-biasedtechnicalchange(STBC)which
leadstotheincreaseofdemandforhigh-skilledlabors(Bermanetal.,1994;Juhnetal.,2014;Lee
andWie,2015).Since female labors fall into low-skilled labor inmostdevelopingcountries,as
aconsequence, thismechanismwilldecreaserelativewage for femaleworkersandwidenthe
earningsdisparity.
Genderinequalityhasbeenanimportantissuethatshouldbesolvedinordertogaina
sustainableeconomicdevelopment,andoneofdimensionsofgender inequality isgenderwage
disparity.Improvementingenderwageequalityishighlycorrelatedwithinclusivegrowth(ILO,
2015).Byencouragingwomentocontributeineconomicdevelopment,inclusiveeconomicgrowth
canbegainedthroughthreechannelswhichareviaaccumulatedphysicalandhumancapital,
participation in labormarket,and increasedsavings(Arora,2012).Inadditionto that,women
withhigheraccessto income labortendtouse largeproportionoftheir incomeforhealthand
educationwhichcanincreaseaccumulatedhumancapitalinsociety(Thomas,1997).
Indonesiangovernmenthasbeenalreadyawaretheimportanceofwomenparticipation
ineconomicdevelopment.Toassureanequalparticipationofwomenandmen ineconomic
development, Indonesiangovernmentputgenderequalityprogramsasmainagendawhich
Figure 1 FDI and Wages (Full-time workers)
Source:StatisticsIndonesia(BPS)
27Trade Openness and Gender Wage Gap: Evidence from Indonesia
written inNationalMedium-TermDevelopmentPlan(RPJMN2015-2019).However,according
toHumanDevelopmentReport2016,IndonesiaisstilloneofASEANmemberswithhighgender
inequality(0.47).Intermofwages, femaleworkersstillearnlessthanmale labor,witharatio
ofabout0.80(StatisticsIndonesia,2015).Intheotherwords,femaleworkersearn20%lessthan
theirmalecounterparts fordoingthesame jobs.Thisdisparity,at theend, leads intoanother
majorproblemwhich is thedecreaseofsharetonational income.Ascanbeenseen inFigure
2,inperiod2008-2014,sharewomentonationalincomestillbelowmen’sandthegapisgetting
wider,which iscausedbyfemale-maleearningsdifferentialsand imbalanceproportionbetween
maleand femaleworkers in labor forceparticipation(IndonesianHumanDevelopmentReport
Gender-Based, 2014).Asimulationconductedby InternationalPovertyCenter showed that
Indonesia lostpotentialrevenueofUS$2.4billionperyearduetothedeclineofwomen labor
participationrateanddisparityinearnings(UnitedNations,2007).Lookingatmajorimpactsof
female-maleearningsdifferentialsoneconomydevelopment,conductingresearchregardingthis
issueisimportant,especiallyinconditionwhereIndonesiawillfaceagreatforeigncompetitionin
globalizationera.
Figure 2 Female and Male Share to National Income Growth
Source:IndonesiaHumanDevelopmentReport
28 国 際 協 力 論 集 第 26 巻 第 1 号
Furthermore,previousstudiesrelatedtogenderwagedisparity isstillnotabundant.
Mostofstudies tried to identifygenderwagegap in Indonesiawithout take intoaccount the
effectoftradeliberalization.ByconductingOaxaca-BlinderDecomposition,Feridhanusetyawan,
Aswicahyono, andPerdana(2001),Pirmana(2006)andTaniguchi andTuwo(2014),used
differentdataset, investigate female-maleearningsdifferentials inruralandurban. Inaddition,
applyingdifferentdecompositionmethod,Sohn(2015)analyzed thegenderwagegapusing
IndonesiaFamilyLifeSurvey(IFLS)dataacrossquantilewagedistributioninIndonesia.Astudy
conductedbyFitrania(2013)isthefirststudytoattempttodrawthelinkbetweentheimpact
ofglobalizationonthegendersalarygap,usinganalternativemeasurement(wagelevelofwage
nets), inprovincialanalysis.Therefore, tocontributetothe lackof literatureonthis issue, this
studywill focusonthe impactof tradeopennessongenderwagedifferentialsand its impact
acrossquantilewagedistribution.
Toinvestigatetheeffectoftradeopennessandgenderedwages,thisstudyfollowsthe
samemethodologybyBraunsteinandBrenner(2009),whichconductedanalysisbycombining
micro-data levelandmacro-data level. In thispaper,micro-data level isdrawn fromNational
LaborSurvey(SAKERNAS)fromStatisticsIndonesia,while,macro-levelwhichconsistsofFDI,
tradevolumeandGRDPperprovincelevelisalsogainfromStatisticsIndonesia.Thehypothesis
inthisresearchisthattradeopennessexposurewillaffectwagesformenandwomenworkers
differentlyandthelargerimpactwillbegainedbymaleworkers.Thedatausedinthisresearch
ispoolingcross-sectionaldatacovers34provincesinIndonesiaintheperiod2008-2014.
Therestof thispapercomposes foursections.Section II isa literaturereviewabout
tradeopennessandgenderwagegap. InsectionIII,methodology includingmodelspecification
anddatadescription ispresented.Section IV isdiscussionof theregressionresults.Finally,
sectionVwillpresenttheconclusionandlimitationofstudy.
Ⅱ.Literature Reviews
1.Empirical Evidence on Trade Openness and Female-Male Earnings Differentials SincethediscriminationtheorywasproposedbyGaryBeckerin1957,manyresearchers
attempted toexamine tradeandgenderwage inequality.Thereare several studies showed
thattradeliberalizationexposurebringsanarrowingimpactonthegenderwagegap.Artecona
andCunningham(2002)andBlackandBrainerd(2004)tried to testBecker’sDiscrimination
29Trade Openness and Gender Wage Gap: Evidence from Indonesia
Theorembycomparingcompetitiveandnon-competitive(concentrated)firms.Arteconaand
Cunningham(2002)useddifference in female-malemean loghourlywagesasameasurement
forthegenderwagegapandfoundthat,inBrazil,womenworkersexperiencedwageincreases
especially therewhoworked inconcentrated industries.Similarly,BlackandBrainerd(2004)
alsodiscoveredthatasimportshareis increased,thegenderwagegaptendstodeclineinU.S.
manufacturingindustries.InIndia,ReillyandDutta(2005)alsodidnotfoundsupportingevidence
thattrade liberalization isassociatedwithhighergenderpaygaps.Moreover,astudybyChen
etal.(2013)whichusedindustry-leveldatahasconductedtoexaminetheimpactofglobalization
ontheChineselabormarketin2004.Theyfoundthatforeignandexportingfirmsemploymore
femaleworkersthandomesticnon-exporters,whichsignificantlyencouragedfemaleemployment
and reduced the genderwage gap in themanufacturing sector.Moreover,Oosterndorp
(2009)analyzedthegenderwagegapusing largecross-countrydatasetby introducinganew
measurementof thegenderwagegap,which isoccupationalgenderwagegap.Hediscovered
thattradeopennesslowersthegenderwagegapinuppermiddleincomecountries,whilethereis
alittleevidencefoundinlowermiddleincomecountries.
Ontheotherhand,inseveraldevelopingcountries,tradeliberalizationincreasesfemale-
maleearningsdifferentials.Seguino(1997)conductedastudy inSouthKoreaand foundthat
women’srelatively lowerwagesandwagediscriminationareassociatedwithweakerwomen’s
fallbackposition in labormarkets. In addition,Berik et al.(2004)compared the impact of
internationaltradeongenderwagegapinTaiwanandSouthKorea.Itappearsthat,incontrast
toneoclassical theory, rising import share ispositivelyassociatedwithwagediscrimination
againstwomeninconcentratedindustries.Moreover,MenonandMeulen-Rodgers(2009)applied
adifferentapproachbyreplacingcross-sectionallong-differencedobservationswithpaneldataset
of28industrieswitha3-digit level.Interestingly,theyfoundanoppositeresultfromReillyand
Dutta’sstudy,anincreaseintradeopennessevenworsenedthegenderwagedifferentialsinthe
industrialsector.
Mostofpreviousstudiesaboveweremore focusingon theeffectof tradeopenness
ongenderearningsdifferentialswithinmanufacturing industriesor firmsdirectlyaffectedby
tradereforms.Thesekindof studieswere typically ignoring thecontributionof individuals’
characteristics,whichusuallyembodied inmicrodata level, that canprovidemorescope for
discoverthemechanismsthroughwhichtradeliberalizationaffectsgenderinequality.Thereare
severalstudiesthatattemptedtocombinemacro-data levelandmicro-data level.Hazarikaand
30 国 際 協 力 論 集 第 26 巻 第 1 号
Otero(2004)conductedaresearchtocomparegenderdifferentialwagesbetweenmaquiladora
andnon-maquiladoraworkersinurbanMexicoi.UsingtheMincerianearningsfunctionandadded
interactiontermsbetweendummyfemaleanddummyformaquiladorasector, they foundthat
genderwagedifferentialdecreasedmoresignificant innon-maquiladora than inmaquiladora.
Inaddition,BraunsteinandBrenner(2007)examinedthe impactofFDIonearningsbetween
femaleandmaleworkersinChinabetweenyear1995and2002bycombiningindividual-datalevel
withFDIandtradeperprovinces.Theresultswerethatwomenexperiencedlargergainsfrom
FDIthanmenin1995,whilemenexperiencedlargerwagegainsfromFDIthanwomenbecause
of industrialupgradingandgender-basedemploymentsegregation in2002.Moreover,arecent
studyconductedbyHanetal.(2012)alsoappliedthesamemethodologyastwopreviousstudies
above tosee the impactof tradereformsonwage inequalityonChina.They found that the
WTOaccessionwassignificantlyassociatedwithrisingwage inequalityandmore importantly
alsorisingthereturnofeducationtocollegeamongChinaworkers.Bycombiningmicrodataand
macrodatalevel,itisexpectedthatinvestigatingtheimpactoftradeopennessongenderwage
gapcanbemorespecificallyandcomprehensive.
2. Indonesia-specific Research Studies on Trade Openness and Female-Male
Earnings Differentials PreviousstudiesinIndonesiarelatedtotradeliberalizationandthegenderwagegapare
notabundant.Feridhanusetyawanetal.(2001), investigatedthegenderwagegap inIndonesia
usingmicro-leveldatafrom1986-1997,theresultsshowedthatthegenderwagegapstillexists
inIndonesiaeventhoughthetrendisdeclining.Pirmana(2006)usesanewdataset1996-2004,
foundresultcontradictingwithpreviousstudies.Hereportedthatthereisanincreaseingender
earningsinequalitiesinIndonesia,whichisattributed41.6%toendowmentdifferencesand58.4%
tounobservedandunexplained factors. Inaddition,TaniguchiandTuwo(2014), foundnew
evidencethattheurbangenderwagegapislessthanthegapinruralareas,moreover,unlikein
ruralareas,theurbangenderwagegapiswideramongyoungeragecohorts.Anotherstudywas
conductedbySohn(2015)using2007IndonesiaFamilyLifeSurvey(IFLS)dataacrossquantile
wagedistributioninIndonesia.Byapplyingquantileregression,hepointedoutthattheexplained
gapremainssimilaracrosswagedistribution,while,unexplainedfactorsdecreaseinthetotalgap.
A studybyFitrania(2013) is the only research that tried todrawa relationship
betweenglobalizationand thegenderwagegapacrossprovinces in Indonesia.Following the
31Trade Openness and Gender Wage Gap: Evidence from Indonesia
samemethodasOosterndrop(2009),sheusedtheoccupationalgenderwagegapandclassified
provinces into low-andhigh incomeprovinces and occupations into low-skill andhigh-skill
occupations.Shedrewtheconclusion that inadevelopingcountry like Indonesia,globalization
mainlyreducesthegenderwagegapinlow-skilloccupations.
3.Research Contribution Since the literature ongenderwage inequality and trade liberalization is still not
abundant in Indonesia, thisstudyattemptstocontributeto thisfield.Therearetwoapparent
differencesbetween this studyandotherprevious studies.First, followingBraunstein and
Brenner(2009)andHazarikaandOtero(2004), thisresearch focusesonthe impactof trade
openness at a household-level perspective by combiningmacro-data andmicro-data level.
AccordingtoBraunsteinandBrenner(2009),bymergingmacro-dataandmicro-datalevels,the
effectsofFDIonfemale-maleearningsdisparitycanbeexaminedasawholewithoutabandoning
the individual’s characteristicvariables,whichcan impairapurelymacroeconomicapproach.
Secondly,sincemanyotherpreviousstudiesonlyfocusonexamininggenderpaygaponaverages
without taking into account the inequality acrosswagedistribution, this papergobeyond
averages toseeacompleteunderstandingabout thedifferenteffectof trade liberalizationon
genderearningsdifferentialsacrosswagedistributionusingthesamemethodasHanetal.(2012).
Ⅲ.Methodology and Data
1.Econometric Model and Estimation Methods FollowingHazarikaandOtero(2004)andHanetal.(2012),thisresearchusesextended
Micherianwagemodelbyaddinginteractiontermbetweendummyvariablegenderandvariable
tradeopenness.Thebaselinespecificationasexpressedasfollows:
Where:
=individual; =province; =timeperiod(year).
isanaturallogarithmofrealhourlywageforeachworker.
isavariabletomeasuretradeopennessexposureineachprovince.
32 国 際 協 力 論 集 第 26 巻 第 1 号
isafemaledummy.
iscontrolvariablesthatcaptureindividualcharacteristicsandtheeconomicenvironment.
Followingprevious studiesbyHazarikaandOtero(2004),BraunsteinandBrenner
(2007)andHanetal.(2012),themodelspecificationabove isestimatedusingOrdinaryLeast
Square(OLS)estimationmethod.Furthermore,quantileregressionisalsoappliedinthisstudy
toinvestigatetheimpactoftradeopennessacrosswagedistribution.
2.Data Description Thisstudyusessecondarydata fromNationalLaborSurvey(SAKERNAS)published
byStatisticsIndonesia.SAKERNAScovers200,000householdsandonlyincludesfamilymembers
over10yearsold.Thisstudyusesthedataofindividualswhoarebetween15and65yearsold
andwhoearnedwage in2008-2014.Furthermore, this studyalsousesmacroeconomicdata,
whichareprovincial tradeopennessand the relatedvariables.Thesedataaredrawn from
InvestmentCoordinatingBoard(BKPM)andStatistics Indonesia. Finally,weconductall the
dataintopoolingcross-sectionaldata.
This studyuses logarithm formof real hourlywageas adependentvariable. It is
becausehourlywagegivesanappropriatecomparableunitofanalysis forwagesince female
workersusuallyworkshorterthanmaleworkerinIndonesia.SAKERNASprovidesonlymonthly
wagedata,soweconvertedmonthlynominalwageintohourlyrealwage.Moreover,thisstudy
usestwotypesofmeasurementfortradeopenness:theratiooftradetoGRDPandtheratioof
FDItoGRDP,foreachprovince.Itisexpectedthattradeliberalizationdoesstructurallyimpact
labormarkets inwaysbeyond individualcharacteristicsofworkers(BraunsteinandBrenner,
2007).
Table1showseachvariabledescriptionandTable2shows thesummarystatistics,
respectivelyii.
Table 1 Variable Description
Variable Descriptions
wage RealwageperhourinIDR
trade RatiooftradetoGRDPinprovince
FDI RatioofFDItoGRDPinprovince
female =1iffemale,0otherwise
33Trade Openness and Gender Wage Gap: Evidence from Indonesia
age Worker’sage
tenure Workingyearsincurrentjob
workhours(perweek) Workinghoursperweek
urban =1ifurban,0otherwise
married =1ifmarried,0otherwise
educ_elementary =1iffinishedelementaryschool,0otherwise
educ_junior =1iffinishedjunior-highschool,0otherwise
educ_senior =1iffinishedsenior-highschool,0otherwise
educ_college =1iffinishedcollegeandmore,0otherwise
sector_manufacture =1ifworkingatmanufacturing,0otherwise
sector_service =1ifworkingatservice,0otherwise
GRDP ProvincialGDP(trillionIDR)
Table 2 Summary Statistics
Variable Mean S.D. Min Max
wage(hourly) 7,440 11,627 36 2,010,417
trade 0.331 0.374 0.001 3.060
FDI 0.025 0.035 0.000 0.414
female 0.344 0.475 0 1
age 36.641 11.408 15 65
tenure 8.122 8.632 0 60
workhours(perweek) 43.826 15.538 1 98
urban 0.589 0.492 0 1
married 0.712 0.453 0 1
educ_elementary 0.195 0.396 0 1
educ_junior 0.170 0.376 0 1
educ_senior 0.340 0.474 0 1
educ_college 0.189 0.391 0 1
sector_manufacture 0.139 0.346 0 1
sector_nonmanufacture 0.701 0.458 0 1
GRDP(trillionIDR) 345 478 4 5,920
n=971,044
34 国 際 協 力 論 集 第 26 巻 第 1 号
Ⅳ.Results and Discussion
Theregressionresultsarepresented inTable3.Mostof theestimatedcoefficientsof
keyvariables,suchas trade, FDI, femaleandthe interactionterms,havesimilarandconsistent
signswiththepreviousstudies.
Theestimatedcoefficientof trade fromOLSresult ispositiveandsignificant,while
thevalueoftheestimatedcoefficientoftradeisbiggerinhighquantilewagegroupthaninlow
quantilewagegroup.Assuming thatworkers inhighquantilewagegroupareworkerswith
high-skilledoccupations, it indicates that tradeshareofGRDP increaseswages forhigh-skilled
labormorethanlow-skilled labor.Thisfindingis in linewithapreviousstudybyLeeandWie
(2015).TheyfoundthatinIndonesia,tradecauseddemandtoshifttowardmoreskilledlaborand
increasedtheirwages.
TheestimatedcoefficientsofFDI, fromOLSresult andquantile regression results,
arepositiveandsignificant.Moreover, thevaluesarequite largerthanthevaluesof tradeand
arealmostthesameamongallquantilewagegroups,unliketheestimatedcoefficientsoftrade.
TheresultofFDIisnotinlinewithpreviousstudiesbyLeeandWie(2015)andFeenstraand
Hanson(1997)focusingonMexicoduringthe1980s.
Fromoverall results,dummyvariable femalearenegativeandsignificant. Incolumn
OLS, thecoefficient isnegativeandsignificantwithvalue-0.3094. Inotherwords,onaverage,
afemaleworker’shourlywageis26.6%belowacomparablemaleworker’shourlywageiii.This
finding is in linewith theresultofHazarikaandOtero(2004) in thecaseofMexico.From
quantileregressionresult, itappears that thevalueofestimatedcoefficientof female in10th
quantilewagedistributionislowerthanin90thquantilewagedistribution.Itimpliesthatfemale-
malewagedifferentialsarelargerinlowerwagegroupthaninhigherwagegroup.
Furthermore,theinteractiontermbetweentradeandfemale(trade*female)isexpected
toexplaintheimpactoftradeliberalizationongenderwagegap.Fromquantileregressionresults,
theestimatedcoefficientoftrade*female ispositiveandsignificant,exceptinhighquantilewage
group. It implies that tradehelps toreducethewagegapbetween femaleandmaleworkers
in lowandmiddle-quantilewagegroup.Meanwhile, inhighquantilewagegroup, tradehasa
wideningeffectonthegenderwagegap.ThesefindingsaresimilarwiththeresultsofFitrania
(2013),which foundthat the impactofglobalizationongenderpaygapreduction is larger in
low-skilledoccupationthan inhigh-skilledoccupation.Thenarrowingeffectof tradeon female-
35Trade Openness and Gender Wage Gap: Evidence from Indonesia
maleearningsdifferentialsinlowquantilewagedistributionscanbeexplainedbytheHeckscher-
Ohlin(H-O)tradetheorythat thepresenceof international trade induceshigherdemandand
increasesrelativewagesforlow-skilledworkersinmostdevelopingcountries.Ontheotherhand,
awidening impactof tradeongenderearningsdifferentialmightbecausedby thepresence
of technologicalchangeswhichaffectsrelativewagesbyshiftingdemandforhigh-skilled labor
(LeeandWie,2015).Anupgradingtechnologyalsoinducestheneedtoemploybetter-qualified
workerswhich ismorebeneficial formaleworkers. It isbecause femaleworkers in thesame
occupationaresubjecttostereotypesthatharmtheirpositioninthelabormarket(Oostendorp,
2009).
Moreover, the estimated coefficient of interaction termbetweenFDI anddummy
variablefemale(FDI*female)showthesamepatternlikevariables trade*female,whichishigher
estimatedvalueoflowwagegroupthanthatofhighwagegroup.Theseresultsindicatethatboth
tradeandFDIcanhelptoreducegenderwagegap.Ontheotherhand,theapparentdifference
is that thevaluesofcoefficients forFDI*femalearehigher thanthevaluesof trade*female. It
indicatesthatthemagnitudeofFDI’seffectonfemale-maleearningsdifferentials is largerthan
trade’seffectonthegenderpaygap.
For thecontrolvariables,almostall thesignsof theestimatedcoefficientsare in line
with theory.Variables, suchasage, tenure,urban,married,andeducation,exceptworkhour,
significantlyincreaseindividual’swage.
Table 3 OLS and Quantile Regression Results
DependentVariable:lnwage OLS Quantile
IndependentVariables Overall 10th 50th 90thtrade 0.0491 *** 0.0047 0.0430 *** 0.1013 ***
(0.0025) (0.0050) (0.0026) (0.0036)
FDI 1.9604 *** 2.1612 *** 2.0096 *** 2.1892 ***(0.0268) (0.0540) (0.0281) (0.0385)
female -0.3094 *** -0.4169 *** -0.2773 *** -0.2479 ***(0.0023) (0.0047) (0.0025) (0.0034)
trade*female 0.0331 *** 0.1256 *** 0.0201 *** -0.0326 ***(0.0040) (0.0082) (0.0043) (0.0059)
FDI*female 0.5334 *** 0.6490 *** 0.5522 *** 0.5409 ***(0.0475) (0.0891) (0.0464) (0.0636)
age 0.0399 *** 0.0466 *** 0.0359 *** 0.0329 ***(0.0005) (0.0009) (0.0005) (0.0007)
36 国 際 協 力 論 集 第 26 巻 第 1 号
Ⅴ.Conclusion
Thepurposeofthisstudyistoinvestigatetheimpactoftradeopennessongenderwage
gapacrosswagedistributioninIndonesia.UsingtheIndonesianHouseholdSurveydatafor2008-
2014,weexaminetheimpactbyquantileregressionanalysiswithanextendedMincermodel.
TheregressionresultsindicatethatfemaleworkersinIndonesiastillearnlessthanmale
workers,onaverage,especiallyinlowerwageclasses.Moreover,itisfoundthattradeopenness
agesq -0.0004 *** -0.0005 *** -0.0004 *** -0.0003 ***(0.0000) (0.0000) (0.0000) (0.0000)
tenure 0.0177 *** 0.0183 *** 0.0174 *** 0.0131 ***(0.0001) (0.0002) (0.0001) (0.0002)
lnwork_hour -0.6084 *** -0.4777 *** -0.6336 *** -0.7581 ***(0.0020) (0.0036) (0.0019) (0.0025)
urban 0.0576 *** 0.0746 *** 0.0350 *** 0.0401 ***(0.0016) (0.0033) (0.0017) (0.0024)
married 0.1281 *** 0.1381 *** 0.1228 *** 0.1152 ***(0.0020) (0.0040) (0.0021) (0.0028)
educ_elementary 0.1307 *** 0.1657 *** 0.1222 *** 0.1006 ***(0.0027) (0.0056) (0.0029) (0.0040)
educ_junior 0.2988 *** 0.3289 *** 0.2804 *** 0.2707 ***(0.0028) (0.0060) (0.0031) (0.0043)
educ_senior 0.4858 *** 0.4940 *** 0.4760 *** 0.4769 ***(0.0028) (0.0058) (0.0030) (0.0041)
educ_college 0.9473 *** 0.9427 *** 0.9318 *** 0.8700 ***(0.0036) (0.0072) (0.0038) (0.0052)
sector_manufacture -0.0957 *** -0.1140 *** -0.0948 *** -0.0902 ***(0.0065) (0.0151) (0.0079) (0.0108)
sector_service -0.1259 *** -0.2038 *** -0.1133 *** -0.0652 ***(0.0063) (0.0146) (0.0076) (0.0104)
lnGRDP 0.0535 *** 0.0612 *** 0.0505 *** 0.0607 ***(0.0007) (0.0015) (0.0008) (0.0011)
_cons 9.1046 *** 7.7397 *** 9.4274 *** 10.2288 ***(0.0188) (0.0391) (0.0203) (0.0279)
regiondummies yes yes yes yes occupationdummies yes yes yes yesobs. 971,044 971,044 971,044 971,044 R-squared 0.3996PseudoR-sq. 0.1521 0.263 0.2889
Robuststandarderrorsinparentheses(Clusteredonprovince).
***1%significance
37Trade Openness and Gender Wage Gap: Evidence from Indonesia
contributestoreducinggenderwagegapasthelowerwagelevel,andthattheimpactofFDIon
wageislargerthanthatoftrade.
Fromthe findingsabove, it showed that tradeopennesscanchange thedynamicof
genderwage in local labormarket.Theresults indicate thepossibilityapolicy intervention
canreducegenderwagegap.Namely, if thegovernmentandpolicymakerscan formulatethe
appropriatesocialandlaborpoliciesandlinkthemwithpoliciesthatpromoteinternationaltrade
andinvestment,genderwagegapcanbereduced.Thisissuealsoindicatesthatthereisroomfor
policyintervention.Inotherwords,governmentsandpolicymakerscanencouragecompaniesto
provideequalvocationaltrainingandvocationaleducationprogramsinordertoreducethegap
betweentheskillsofwomenandmen’sworkers.Ultimately,thispolicycanincreasethechances
ofdevelopingwomen’sabilitytoacquirebroademploymentandraisingwomen’srelativewage.
Asalimitationofthestudy,theremaybeareversecausalitybetweenFDIandwagesin
spitethatFDIisusedasameasurementoftradeopennessinthisstudy.Oneofthemethodsfor
thisproblemisInstrumentalVariable(IV)approach.BraunsteinandBrenner(2007)suggested
infrastructureorgeography factorsasseveralpotential instrumentvariables.However,dueto
thelackofdataavailabilityofthosevariables,IVapproachdoesnotaddressinthisstudy.This
pointisforfurtherresearch.
Appendix: Table A1 Variable Description for Region and Occupation Dummy
Variable Description
regiondummies
reg_sumatera =1ifSumatera,0otherwise
reg_kalimantan =1ifKalimantan,0otherwise
reg_sulawesi =1ifSulawesi,0otherwise
reg_maluku =1ifMaluku,0otherwise
reg_java =1ifJava,0otherwise
occupationdummies
j_prof =1ifprofessional,1otherwise
j_manager =1ifmanager,0otherwise
j_admin =1ifadministrationstaff,0otherwise
j_sales =1ifsalesclerk,0otherwise
j_services =1ifservices,0otherwise
j_farmer =1iffarmer,0otherwise
j_prod =1ifproduction,0otherwise
38 国 際 協 力 論 集 第 26 巻 第 1 号
ReferencesArtecona,R.,&W.Cunningham(2002)Effects of Trade Liberalization on The Gender Wage Gap in Mexico,
WashingtonD.C.:TheWorldBank.Becker,G.(1957)The Economics of Discrimination,Chicago:TheUniversityofChicagoPress.Berik, G.,Y.V.MeulenRodgers,& J. Zveglich, Jr.(2004) “InternationalTrade andGenderWage
Discrimination:EvidencefromEastAsia,”Review of Development Economics,8(2):237-254.Berman,E.,J.Bound,&M.Stephen(1998)“ImplicationsofSkill-BiasedTechnologicalChange:International
Evidence,”The Quarterly Journal of Economics,113(4):1245-1279.Black,S.,&E.Brainerd(2004)“ImportingQuality?TheImpactofGlobalizationonGenderDiscrimination,”
Industrial and Labor Relations Review,57(4):540-559.Braunstein,E.,&M.Brenner(2007)“ForeignDirect InvestmentandGenderedWages inUrbanChina,”
Feminist Economics,13(3-4):213-237.Chen,Z.,Y.Ge,H.Lai,&C.Wan(2013)“Globalization andGenderWage Inequality inChina,”World
Development,44:256-266.Dollar,D.,&R.Gatti(1999)“GenderInequality, IncomeandGrowth:AreGoodTimesGood forWomen?”
PolicyResearchReportonGenderandDevelopmentWorkingPaperSeries,No.1.Driffield,N.,&K.Taylor(2000)“FDIandTheLaborMarket:AReviewofTheEvidenceandPolicy
Implications,”Oxford Review of Economic Policy,16(3):90-103.Feenstra,R.C.,&G.H.Hanson(1997)“ForeignDirect InvestmentandRelativeWages:Evidence from
Mexico’sMaquiladoras,”Journal of International Economics,42(3-4):371-394.Feridhanusetyawan,T.,H.Aswicahyono,&A.A.Perdana(2001)“TheMale-FemaleWageDifferentials in
Indonesia,”CSISWorkingPaperSeries,No.59,CenterforStrategicandInternationalStudies(CSIS).Fitrania,S.(2013)“The ImpactofGlobalizationonGenderWageGap in Indonesia:ASub-NationalLevel
Analysis,”MasterThesis,TilburgUniversity.Fussell,E.(2000)“MakingLaborFlexible:TheRecompositionofTijuana’sMaquiladoraFemaleLaborForce,”
Feminist Economics,6(3):59-79.Han,J.,R.Liu,&J.Zhang(2012)“GlobalizationandWageInequality:EvidencefromUrbanChina,”Journal of
International Economics,87(2):288-297.Hazarika,G.,&R.Otero(2004)“ForeignTradeandtheGenderEarningsDifferential inUrbanMexico,”
Journal of Economic Integration,19(2):353-373.Juhn,C.,G.Ujhelyi,&C.Villegas-Sanchez(2014)“Men,Women,andMachines:HowTradeImpactsGender
Inequality,”Journal of Development Economics,106:179-193.Lee, J.-W.,&D.Wie,(2015)“TechnologicalChange,SkillDemand,andWage Inequality:Evidence from
Indonesia,”World Development,67:238-250.Menon,N.,&Y.V.Rodgers(2009)“InternationalTradeandtheGenderWageGap:NewEvidence from
India’sManufacturingSector,”World Development,37(5):965-981.Oostendorp,R.H.(2009)“GlobalizationandtheGenderWageGap,”The World Bank Economic Review,23(1):
141-161.Pirmana,V.(2006)“EarningsDifferentialbetweenMale-Female in Indonesia:Evidence fromSAKERNAS
Data,”WorkingPaperinEconomicsandDevelopmentStudies,No.200608,PadjajaranUniversity.Reilly,B.,&P.V.Dutta(2005)“TheGenderPayGapandTradeLiberalisation:EvidenceforIndia,”PRUS
WorkingPaper,No.32,UniversityofSussex.Seguino,S.(1997)“Genderwageinequalityandexport-ledgrowthinSouthKorea,”The Journal of Development
Studies,34:102-132.Sohn,K.(2015)“GenderDiscrimination inEarnings in Indonesia:FullerPicture,”Bulletin of Indonesian
Economic Studies,51(1):95-121.Taniguchi,K.,&A.Tuwo(2014)“NewEvidenceontheGenderWageGap inIndonesia,”ADBEconomics
WorkingPaperSeries,No.404.UnitedNationsDevelopmentProgramme(2016)Human Development Report 2016: Human Development for
Everyone,NewYork:UNDP.
39Trade Openness and Gender Wage Gap: Evidence from Indonesia
Notesi Maquiladora industry isan industry forexportspecializedmanufacturingunits inMexico
whichestablishedundertheagreementwithUnitedStatescalledBraceroProgram.ii SeealsovariabledescriptionforregionandoccupationdummyinTableA1inAppendix.iii (e-0.3094-1)× 100%=-26.6%