Indeed one view is that inadequate access to credit prevents women from under taking highreturn business activities However one recent empirical study 64257nds low returns to capital in femalerun microenterprises Suresh De Mel David McKenzie and Chr ID: 43422
Download Pdf The PPT/PDF document "Do Traditional Institutions Constrain Fe..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
2PAPERSANDPROCEEDINGSMAY2010I.GendernormsinIndiaTheHinducastesystemdevelopedasanex-tremelyhierarchicalsocialsystem.Itsdeningprinciplesincludetheideasofpurityandpollu-tion.UCsmaintainpuritybyavoidingsexualrelations,marriageand,inextremecases,con-tactwithlowercastes.Premisedonmenbe-ingasourceofpollution,restrictionsareplacedonwomentolimitcontactwithmenotherthantheirhusbands.Requirementsincludethatamarriedwomanremainveiled,notremarryifwidowed,notinteractwitholdermen,andhaverestrictedmobilityoutsideofherhouse.Thesenorms{particularlythelattertwo{signi-cantlyrestrictfemalelaborforceparticipation.MaintainingpuritybyminimizingcontactwithlowercastesislessrelevantforSCs,whoranklowinthehierarchy.Inaddition,greaterpovertyimpliesincreasedrelianceonfemalewage-earning(KarinKapadia1997).Asaresult,SCwomenfacefewersocialrestrictionsand,byvirtueofbeingindependentearners,enjoygreaternancialautonomyandincreasedcontroloverhouseholdnancialdecisionsrelativetoUCwomen(JoanP.Mencher1988).Notably,theserestrictionsonfemaleautonomyamongUCsarenotlimitedtothewealthy(MukeshEswaran,BharatRamaswamiandWilimaWadhwa2009).RelativetoHindus,MuslimsinIndiaplacemorerestrictionsonwomen'scontactwithpeo-pleoutsidebutnotwithinthesphereofkinship.BecauseMuslimwomenareentitledtoashareinfamilyrealestate,controllingtheirrelationshipswithmalesoutsidethefamilycanbecrucialtothemaintenanceoffamilypropertyandprestige.II.InterventionandstudydesignWeconductedabusinesstraininginterven-tioninconjunctionwithSEWABank,whichisbasedinthecityofAhmedabadinwesternIndia.Its170,000member-clientsareprimar-ilypoorwomenwhoworkintheinformalsec-tor(forexampleasincense-stickmakers,tailors,andvegetablevendors).SEWABanko ersthesewomenawidearrayofnancialproducts.Allclientsarerequiredtohaveasavingsac-count,androughlyaquarterofclientshaveevertakenoutaloanfromSEWABank.Forseveralyears,SEWABankhasrunave-daynancialliteracytraining.Thecurricu-lum,developedbyFreedomfromHungerandusedwidelyaroundtheworld,coversbasicac-countingskills,interestratesandlife-cycleplan-ning.Itemphasizesnancialprudenceanden-courageswomentoavoidexcessdebt,savemoreandreduce\frivolous"spending.Morerecently,SEWABankstartedasecondve-daycoursethatteachesbusinessskillssuchascostreduc-tion,investment,andcustomerservice.IncollaborationwithSEWABank,wede-signedastreamlinedtwo-daytrainingmodulethatcombinedelementsoftheirnancialliter-acyandbusinessskillscurricula,andaddednewmaterialfocusedonaspirations.Theaspirationscomponentincludedashortlmshowcasingsuc-cessfulSEWAmemberswhousedgoodnancialpracticestobringthemselvesoutofpoverty.Ashomeworkaftertherstdayoftraining,partic-ipantslledoutaworksheetidentifyinganan-cialgoaltheywantedtoachieveoverthenextsixmonths,andontheseconddaybrokeitdownintosmallershort-runsteps.Fortheexperiment,636womenwereran-domlydrawn(intwophases)fromthepoolofSEWABankcustomersages18to50whowerebothactivesaverswithinthepasttwoyearsandemployed.Twothirdsofthesewomenwereran-domlyassignedtothetreatmentgroup.2Womenassignedtothetreatmentgroupwereapproachedintheirhomesandrecruitedtoat-tendaparticulartrainingsessionwithsevenotherparticipantsattheSEWAbranchnearesttothem.Fordatacollectionandanalysispur-poses,womeninthecontrolgroupwerealsoas-signedbutnotrecruitedtoaparticulartrainingsessionattheirnearestSEWAbranch,allowingustoclusterstandarderrorsbysession.Ouranalysissamplecomprisesthe597womenwhoweresuccessfullysurveyedatfollow-upandcouldbecategorizedintosub-castesbasedonsurname.3Wecategorizedwomenintothreebroadsocialgroups:Muslims,HinduSCs,andHinduUCs(non-scheduledcastesinclud-ingotherbackwardcastes).Inaddition,wescoredhowrestrictiveeachHinduandMuslim2TherandomizationwasstratiedbysamplingphaseandSEWAbranch.3Thesurveyattritionrate(5.3%)issimilaracrossexperimentalgroups.Wewereunabletoassigncastetosevenwomen. VOL.100NO.2TRADITIONALINSTITUTIONSANDFEMALEENTREPRENEURSHIP3sub-castewasinregardtovenormsgoverningwomen'sbehavior:abilitytosocializealone,re-quirementstocoverthefaceorwearaveil,abil-itytospeakdirectlytoelders,abilitytoleavethehouseorneighborhoodalone,andabilitytoremarry.Wecreatedanindexrangingfrom0to5equaltothenumberofnormsforwhichthesub-castewashighlyrestrictive.Figure1showsthevalueofthisindexacrossthethreesocialgroupsinoursample.SCwomenfacethefewestrestrictions,followedbyUCandMuslimwomen.UCsare50%morelikelytofaceseveresocialrestrictionsasSCs,andMus-lims,inturn,haveovertwicetherateofsevererestrictionsasUCs. Figure1.NumberofsocialrestrictionsTheinterventiontrained289womenin57two-daytrainingsessionsconductedfromSeptember2006toApril2007.Programtake-upwashigh,withover70%ofthoseinvitedchoos-ingtoattend.SCs,theleastrestrictedgroup,werenearlyonethirdmorelikelytoattendthetrainingthanMuslimsandUCs.Theresultsontake-up,whicharealsotherststageofourtreatment-on-the-treated(TOT)results,arere-portedintheonlineappendix.Baselinecharacteristicsarebalancedacrossthecontrolandtreatmentgroups(seeonlineap-pendix).Table1comparesbaselinecharacteris-ticsacrosssocialgroups.Womeninoursamplearestrikinglyhomogenousacrosssocialgroups:AverageeducationandfamilysizearealmostidenticalforMuslims,UCsandSCs,andSCwomenhavehigherhouseholdincomeandareslightlymorelikelytoownabusiness,thoughthedi erencesarestatisticallyinsignicant.Table1|BaselinecharacteristicsUCSCMuslimEducation6.336.626.36Income4852.65694.65189.5Householdsize5.255.255.42Ownsbusiness0.270.310.26N34670181Notes:Asterisksdenotesignicantdi erencesinmeansbetweenthegroupandUCs.Inalltables*denotessignicantat5%level**signicantat1%level***signicantat0.1%level.III.E ectsofbusinesstrainingOurestimationstrategyexploitstherandomassignmenttotreatment,i.e.,beinginvitedtoatrainingsession.Weexaminetheimpactofattendingthetrainingoneconomicoutcomes,instrumentingforattendancewithwhethertheparticipantwasinthetreatmentgroup.ThisIVspecicationprovidesTOTestimates.Weseparateoutthedi erentiale ectoftrainingbysocialgroupbyinteractingthetrainingdummywithindicatorsforbeingaSCandaMus-lim.Outcomevariablescomefromasurveyconductedonarollingbasisfourmonthspost-training.TheresultsinTable2revealthattrainingledtoasignicantincrease(13percentagepoints)inthelikelihoodoftakingoutaloanwithinfourmonthsoftraining(Borrowed)amongUCs,whoaretheomittedcategory.UCswhoat-tendedtrainingtookoutloansatnearlytwicetherateasUCsinthecontrolgroup.Mean-while,wecannotrejectthattherewasnoe ectonborrowingamongSCorMuslimwomen.Whenweestimatethetraininge ectonthelikelihoodthatawomanreportedproblemsmanagingherdebt,wendnooveralle ectandnodi erentiale ectsacrosssocialgroups(seeappendix).Administrativebankdataonloandefaultconrmthisresult.Thus,thetraining 1.70.480.81.21.6 MuslimsUppercasteScheduledHindus 4PAPERSANDPROCEEDINGSMAY2010doesnotseemtohaveinducedUCwomentoborrowbeyondtheirmeans.Table2|TreatmenteffectsonfinancesBorrowedSavingsTrained0.13*-315.32(0.07)(492.83)TrainedSC-0.16444.71(0.14)(540.02)TrainedMuslim-0.14317.51(0.11)(630.07)SC0.04-298.12(0.10)(301.87)Muslim0.04-46.50(0.06)(333.99)Meanofdep.var.0.17277.59Notes:Standarderrorsclusteredbytrainingsession.RegressionsincludeSEWAbranch,treatmentmonth,andsamplingphasexede ects.MeanofdependentvariableisforUCsinthecontrolgroup.N=597.Seetheappendixforfurtherdetails.ThesenotesalsoapplytoTable3.Savingsduringthepastmonth(Savings,measuredinrupees)shownosignicantdi er-encesacrosstreatmentandcontrol,thoughthepointestimatesagaingoinoppositedirectionsforUCscomparedtoSCsandMuslims,withUCssubstitutingawayfromsaving,consistentwiththeirincreasedborrowing.Tomoredirectlymeasurethee ectoftrainingonbusinessactivity,inTable3weexaminethewomen'spersonalbusinessactivity,includinganindicatorofwhethershereportsanypersonallaborincomeoverthepastweek(AnyIncome)andanindicatorofwhethershereportstalkingtofamilyaboutbusinessplans.Althoughwesampledonbeingemployed,manyofthewomendonotreportregularearnings:inthecontrolgroup,only80%ofwomenreportanyearninginthepastweek.Weobserveapositiveandsignicante ectofthetrainingonbusinessincomeamongUCsonly,suggestingthatthenewloanswereputto-wardbusinessinvestments.Theestimatedef-fectoftrainingonearninganybusinessincomeishighlysignicantamongUCs,indicatinga25%increaseinthelikelihoodthatawomanen-gagesinlabormarketactivity.Whenwelookatamountofincomeearnedoverthepastweek,thepointestimatessuggestanincreaseofaround30%,buttheresultsaretooimprecisetodrawconclusions.Togetherwiththeloanresults,thispatternsuggeststhatthetrainingencour-agedUCstostartorexpandtheirmicroenter-prises.Furthersupportingtheseresults,wendthattrainingledUCsbutnototherwomentotalkmorefrequentlywithfamilymembersaboutbusinessplans(TalkBusiness).Table3|TreatmenteffectsonbusinessTalkAnyIncomeBusinessTrained0.19**0.17*(0.09)(0.10)TrainedSC-0.37***-0.51***(0.14)(0.15)TrainedMuslim-0.22*-0.35***(0.12)(0.12)SC0.25***0.35***(0.08)(0.09)Muslim0.090.18**(0.07)(0.08)Meanofdep.var.0.790.69IV.DiscussionGiventhesimilarityineducation,householdwealthandtypesofbusinessesacrosssocialgroups,thedi erenceacrossgroupsintheirre-sponsetotrainingisstark.Itismadeevenmorestrikingbythefactthatdatacollectedduringtrainingrevealidenticalpatternsofbusinessandnancialgoalsacrosssocialgroups.Onepossibilityisthatimbalancesintreat-mentassignmentwithinsocialgroupsarere-sponsiblefortheobservedpatterns.However,whilethereishigherbaselinebusinessownershipamongUCtreatedrelativetocontrolwomenandtheoppositeistrueforMuslimsandSCs,treat-mentdi erencespersistevenwhenwecontrolforthis(and/orother)observables(seeappendix).Anotherpossibilityisthatdi erentialtreat-mente ectsre ecthigherprogramtake-upamongSCs.ThetrainingmayhaveattractedaselectedsampleofUCswhowereespeciallyre-sponsivetotraining.However,observablechar- VOL.100NO.2TRADITIONALINSTITUTIONSANDFEMALEENTREPRENEURSHIP5acteristicsappeartobeequallybalancedacrosscasteamongthesubsetofwomenwhoweretreatedrelativetotheITTsample,indicatingthatheterogeneityintake-updoesnotseemtoexplaintheheterogeneoustreatmente ects(thoughwecannotruleoutunobservabledi er-encesacrossgroups).Furthermore,take-upcan-notexplainobserveddi erencesbetweenUCsandMuslims,whohadsimilarattendancerates.IftheexplanationforUCwomenbeingespe-ciallyresponsivetotrainingisthatsocialrestric-tionscausedthemtohaveknowledgedecitsorthetrainingallowedthemtochallengeso-cialnormsthatweredistortingtheirbusinesspractices,thenanimportantquestioniswhyMuslims,whofacethehighestdegreeofre-strictions,didnotrespondmoretothetrain-ingthanSCsdid.Onepossibilityisthat,al-thoughrestrictionsaregreaterfortheaverageMuslimwomanthantheaverageSCwomaninAhmedabad,thereislittledi erenceinrestric-tionsacrossMuslimandSCmembersofSEWABank.Unfortunately,withoutindividual-leveldataonrestrictedness,wecannottestthisstory.AnotherpossibilityisthatMuslimsinAhmedabad,whichhasahistoryofreligioustension,faceconsiderablediscriminationinthemarketplace,whichbusinesstrainingcouldnotundo.Alternatively,religiousrestrictionsoninterest-bearingloansmightexplainwhyde-mandforcreditdidnotincreaseamongMuslims.However,thisexplanationisunlikelygiventhattherateofborrowingissimilaracrossMuslimsandHindusinthecontrolgroup.Analpossibilityworthmentioningisthat,althoughMuslimwomeninthissettingfaceahighdegreeofsocialrestrictions,forthenormsthatmostdirectlya ectbusinessactivity{abil-itytoleavethehomealoneandtalktostrangers{wendthattherateofbeinghighlyrestrictedisinfactlowerforMuslimsthanUCs.However,forthesetwonorms,onlyabout5%ofwomeninoursamplewerecodedashighlyrestricted,andaveragerestrictednessisstillconsiderablyhigheramongMuslimsthanUCorSCHindus.Bearinginmindthesecaveats,aprimafacieexplanationforourresultsisnon-monotonicityinthee ectofsocialrestrictions:Thetrain-inghelpedwomenwhosebusinesseshadbeenhelddownbysocialrestrictions,butwomensub-jecttoextremerestrictionshadtoolittleagencytoeasilychangetheiraspirationsoractivities.Evenwithmoreknowledgeorhigheraspira-tions,themostrestrictedwomenmightfacetoomanysocialstricturestoavailthemselvesofen-trepreneurialopportunities.Ourbusinesscounselingprogramsignicantlyreducedthebusinessincomegapbetweensocialgroups.Thus,anotherreadingofourresultsisthatmodernization{intheabsenceofinterven-tionsthatcounteracttraditionalnorms{mayyieldgreaterbenetsforwomenlowerinthecastehierarchy,apointalsomadebyKaivanMunshiandMarkRRosenzweig(2006).Thisview,however,assumesthatgendernormsforlowercasteswillcontinuetobelessrestrictive.If,instead,modernizationheightenssanskritiza-tion{thedesireoflowercastestoemulateuppercastes{andSCsincreasinglyadoptthegendernormsofUCs,theneconomicgrowthmayfailtoemancipatewomentothesameextent.REFERENCESDeMel,Suresh,DavidMcKenzie,andChristopherWoodru .2008.\ReturnstoCapitalinMicroenterprises:EvidencefromaFieldExperiment."QuarterlyJournalofEco-nomics,123(4):1329{1372.Eswaran,Mukesh,BharatRamaswami,andWilimaWadhwa.2009.\Status,Caste,andtheTimeAllocationofWomeninRuralIn-dia."mimeo,UniversityofBritishColumbia.Kapadia,Karin.1997.\MediatingtheMean-ingofMarketOpportunities:Gender,CasteandClassinRuralSouthIndia."EconomicandPoliticalWeekly,35(52):3329{3335.Karlan,DeanS.,andMartinValdivia.2008.\TeachingEntrepreneurship:ImpactOfBusinessTrainingOnMicronanceClientsandInstitutions,."mimeo,YaleUniversity.Mencher,JoanP.1988.\Women'sWorkandPoverty:Women'sContributiontoHouseholdMaintenanceinSouthIndia."InAHomeDi-vided.,ed.DaisyDwyerandJudithBruce.Stanford,CA:StanfordUniversityPress.Munshi,Kaivan,andMarkRRosen-zweig.2006.\TraditionalInstitutionsMeettheModernWorld:Caste,GenderandSchool-ingChoiceinaGlobalizingEconomy."Ameri-canEconomicReview,96(4):1225{1252.