1154535 (1)

Embed Size (px)

Citation preview

  • 7/27/2019 1154535 (1)

    1/23

    Female-Headed Households and Female-Maintained Families: Are They Worth Targeting toReduce Poverty in Developing Countries?Author(s): Mayra Buvini and Geeta Rao GuptaSource: Economic Development and Cultural Change, Vol. 45, No. 2 (Jan., 1997), pp. 259-280Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/1154535 .

    Accessed: 16/10/2013 14:13

    Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

    .JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

    content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

    of scholarship. For more information about JSTOR, please contact [email protected].

    .

    The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to

    Economic Development and Cultural Change.

    http://www.jstor.org

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/action/showPublisher?publisherCode=ucpresshttp://www.jstor.org/stable/1154535?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/1154535?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=ucpress
  • 7/27/2019 1154535 (1)

    2/23

    Female-Headed Households andFemale-MaintainedFamilies: Are TheyWorthTargetingto Reduce Povertyin Developing Countries?*

    MayraBuvinid and Geeta Rao GuptaInternational Center for Research on WomenThe subjectof targetingpublic-andprivate-sectorprograms o female-headedhouseholds in developingcountries in orderto attackpovertyand social disadvantage s controversial and lackingin rigorousevi-dence. One arguments thatwomen who headhouseholds areworthyof specialattentionbecausetheyaretriplydisadvantaged: hey experi-ence the burdens of poverty, gender discrimination,and absence ofsupportas heads of household.' The opposite arguments thatfemaleheadshipshould not be used as the main targetingcriterion becausefemaleheadship s notalwayscorrelatedwithpoverty,therearepracti-cal difficulties n identifyingde facto headship,and there may be per-verse incentive effects as a result of targetingbenefits or services tosinglemothers-that is, it maypromoteratherthandiscouragesinglemotherhood.2

    Unfortunately,there are few documented experiences of pro-gramsthattargetfemaleheadshipand theirconsequences.Most of theprojectexperience has takenplace in the nongovernmental ector, issmall in terms of projectsize and coverage, and has not been evalu-ated. In this article we address the above controversy by first dis-cussingthe issues related to the definitionand measurementof femaleheadshipand the importanceof the concept for developmentpolicy,and then undertakinga systematicreview of the empiricalevidenceon the relationbetween female headshipandpoverty. If, on average,female-headedhouseholds are in fact poorer than other households,headshipshould seriouslybe consideredas a potentiallyuseful crite-rion for targetingantipovertyinterventions,especially in developingcountries where means testing is not feasible. Second, we examinepotentialcosts and benefits of targeting emale headshipand review? 1997by The Universityof Chicago.All rightsreserved.0013-0079/97/4502-0001$01.00

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    3/23

    260 EconomicDevelopmentand CulturalChangethe experience of Chile, one of the few countries that has targetedfemale headship throughgovernment ntervention the others are Co-lombia, Honduras,and India), and the only one that has evaluationdata available. We use the analysis of the project experience alongwith a review of the empiricalevidence to answer the question ofthe desirabilityand efficiencyof targeting emale headshipto reducepovertyin developingcountries.TheNatureof FemaleHeadshipDefinitionsThe problemswith definingandmeasuring emale-headedhouseholdsarewell known. First, countriesuse differentandthereforeoften non-comparabledefinitionsof both the terms "household" and "head ofhousehold" in their census instruments.Second, there is ambiguityinherentin the term "head of household" when the assignmentofheadship s left to thejudgmentof household members.Membersmayuse differentcriteria to make this assignment, rendering n-countrycomparisonsinvalid. Reliable identificationof female-headedhouse-holds is furthercompoundedby the fact that femaleheadship may bea transitoryphenomenonin the life cycle of families. The thirdandperhapsmost seriouslimitation s that the term "head of household"is not neutral.It is loadedwithadditionalmeanings hat reflecta tradi-tionalemphasison householdsas undifferentiated nitswith a patriar-chal system of governanceand no internalconflicts in the allocationof resources.3Inresponseto these definitional ndmeasurementproblems,someexpertsarguefor usingmore specific language hanthe term "female-headedhousehold"-language that would be more effective in captur-ingthe wide rangeof familystructures hat areeconomicallydependenton women. Additional ermssuggestedinclude "female-maintained,""female-led,""mother-centered,""single-parent,"or "male-absent"rather than "female-headed"to more accurately describe differenthouseholdstructuresandfamilysituations.It is also helpfulto distin-guish between female-headed households (residential units) and fe-male-maintainedamilies(kinshipunits)becausea female-headed am-ily mayreside as a subfamilyn a larger,often male-headedhousehold.Disaggregatingheconceptof femaleheadship s undoubtedlyuse-ful for the purposesof research and for advancementof knowledgeinthis field. However, for purposesof policy and program mplementa-tion, the term "female-headedhousehold" and theconditionof female-maintained amilieswithin male-headedhouseholdsare practical,al-beit imperfect, proxies for the whole range of family structures andhouseholdsin which women are the primaryprovidersfor their fam-ilies.

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    4/23

    MayraBuvinidand Geeta Rao Gupta 261PrevalenceThe importanceof the concept of female-headedhouseholdsarises, inpart, from the sheer numbers of such households in both developingand industrialsocieties. As table 1 shows, in the 1980s, 14%of allhouseholdsin Indonesia,up to 45%of all householdsin Botswana andBarbados,and 31%of households in the United States were headedby women. In developingcountries,femaleheadshipis morefrequentin Latin Americaand the Caribbeanand in sub-SaharanAfrica and isless frequentin Asia and the Near East. However, its prevalence isincreasingin the differentregions, as table 2 reveals, and thereforethis rise can no longer be attributedto isolated circumstances or toany specificculturalor ethnic heritage.Which factorsfoster the rise of female-headedhouseholds? Thereare demographicand social antecedents that occur with increasingfrequency in today's societies. There is sex-specific migration,re-sulting n "left-behind"female heads in the place of origin,as in ruralareasin sub-SaharanAfrica, or the creation of households headedbymigrantwomen in the place of destination,as often happensin largecities in Latin America;maritaldisruptionand increases in unpart-nered adolescent fertility (evident in both Latin America and sub-SaharanAfrica);erosion of extended family systems and traditionalsupportnetworks, which leaves single mothers and widowed women

    TABLE 1PERCENTAGE OF HOUSEHOLDSHEADED BY WOMEN IN 1980s

    IN SELECTED COUNTRIESCountry %Botswana(1981) 45Barbados 1980) 45Malawi* 29Cuba(1981) 28Ghana* 27Venezuela(1981) 22Honduras* 22Chile(1982) 20Bangladesh 1981) 17S. Korea(1980) 15Indonesia 1980) 14United States (1985) 31

    SOURCE.-UnitedNations,Demographic Yearbook(NewYork: UnitedNations, 1989).* Late 1970s data.

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    5/23

    262 Economic Development and Cultural ChangeTABLE 2

    TRENDSIN PERCENTAGE F HOUSEHOLDSHEADED BY WOMEN(de Jure)(All AvailableData)Earlier LaterCountry Date % Date %

    Demographic urveydata:Asia:Indonesia 1976 15.5 1987 13.6Sri Lanka 1975 15.7 1987 17.8Thailand 1975 12.5 1987 20.8Latin Americaand Caribbean:Colombia 1976 17.5 1986 18.4DominicanRepublic 1975 20.7 1986 25.7Ecuador 1979 15.0 1987 14.6Mexico 1976 13.5 1987 13.3Peru* 1977/78 14.7 1986 19.5TrinidadandTobago 1977 22.6 1987 28.6MiddleEast and North Africa:Morocco 1979/80 11.5 1987 17.3Sub-SaharanAfrica:Ghana 1960 22.0 1987 29.0Sudan 1978/79 16.7 1989/90 12.6Census data:Asia:Hong Kong 1971 23.5 1991 25.7Indonesia 1971 16.3 1980 14.2Japan 1980 15.2 1990 17.0Korea 1980 14.7 1990 15.7

    Philippines 1970 10.8 1990 11.3Latin AmericaandCaribbean:Brazil 1980 14.4 1989 20.1Costa Rica 1984 17.5 1992 20.0Panama 1980 21.5 1990 22.3Peru 1981 22.1 1991 17.3Uruguay 1975 21.0 1985 23.0Venezuela 1981 21.8 1990 21.3Sub-Saharan Africa:Burkina Faso 1975 5.1 1985 9.7Cameroon 1976 13.8 1987 18.5Mali 1976 15.1 1987 14.0SOURCE.-Judithruce,CynthiaB. Lloyd,AnnLeonard,et al., Fami-lies in Focus: New Perspectives on Mothers, Fathers, and Children (NewYork:PopulationCouncil,1995).NOTE.-De jure = "recognized" household headship.* De facto = headship on day of interview.

    on their own (in Bangladesh,Egypt, and India, among others); andsex ratio imbalancescaused by war deaths and civil conflicts, whichresultin a surplusof females in native or refugeepopulations.There are two trendsassociated with economic changethat maycontributefurther to the increasingprevalence of female headship.The first trend is the disruptionof traditionalsystems of patriarchal

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    6/23

    MayraBuvinid and Geeta Rao Gupta 263governance,which weakensexplicitandimplicitcontractsthatenforceincome transfersfrom fathers to mothers and their children. NancyFolbre states that this detachment of childrenfrom fathers' earningsis a converging ssue for women in manydevelopingandindustrializedcountries, and is more often than not indicative of a forced indepen-dence frommalewages ratherthanwomen's choice not to dependonmen's earnings.4As evidence of this trend, the great majority ofwomen and mothersin the United States were dependenton men inthe 1940s,while only a minority(under25%)were exclusively relyingon male earningsin the 1980s.5The second trend is declining realhouseholdincome and the increasingpoverty usually associated witheconomiccrises, "forcing"men to relinquishresponsibility or familymaintenance.In the Latin Americanand Caribbeanregion, for exam-ple, feminist researchershave hypothesizedthat the economic crisisof the eighties and the loss of gainful employmentamong men haveincreasedthe numbersof householdsthatdependonly or primarilyonwomen's income.ThePovertyof Female-HeadedHouseholdsThe relation between female headship and poverty and the conse-quences of female headshipfor child welfare have been reasonablywell studied. We reviewed information rom 65 studies carried out inthe past decade. Sixteen were done in Africa, 17 in Asia, and 32 inLatin Americaandthe Caribbean.Self-reportand the physicalabsenceof men as a result of migration,death, divorce, or abandonmentwerethe most commonlyused definitionsof femaleheadship n the studies.Some studiesdistinguishedbetween de facto and dejure female head-ship, and a few examined the situation of functional families headedby women residingin largerhouseholds.6Relationshipto PovertyOf the 65 studies reviewed, 61 examined the relation of female head-shipto poverty. Thirty-eightof the 61 studiesfound,by usinga varietyof poverty indicators(totalor per capitahouseholdincome, mean in-come per adultequivalence,total or per capitaconsumptionexpendi-tures, and access to services andownershipof land andassets, amongothers),that female-headedhouseholdsareoverrepresentedamongthepoor. Fifteen other studies found that poverty was associated withcertain types of female heads, or that the association emerged forcertainpoverty indicators.Only eight of the 61 reports(13%)showedno empiricalevidence for the hypothesis of the greater poverty offemale-headedhouseholds.Whyare they poor? The review of the evidence that shows apositive association between female headshipand poverty points tothreesets of factorsthatdetermine he greaterpovertyof these house-

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    7/23

    264 EconomicDevelopmentand CulturalChangeholds. They emerge, respectively, from characteristicsof householdcomposition,the genderof the main earner,and the unique circum-stances of beinga female-headedhousehold.First, female-headedhouseholds, despite their smaller size incomparisonwithothertypesof households,oftencarrya higherdepen-dencyburden.Thatis, they tend to contain a higherratio of nonwork-ers to workersthan do other households, as supportedby data fromruralBotswana,7Malawi,8Brazil,9Mexico, and Peru.'oThis type ofhouseholdcompositionwould not necessarily lead to poverty, how-ever, if the household received child-supportpayments from absentfathers,as is the case with some of the left-behindruralfemale headsin India who receive adequateremittances." The poverty of female-headed households thus reflects a disruptionof traditionalsystemsof family governancethat enforced income transfers from fathers tochildren.'2Second, the main earnersof female-headed amiliesareby defini-tionwomen,who have loweraverageearnings hanmen, fewerassets,and less access to remunerativeobs and productiveresources suchas land, capital, and technology. This gender-relatedeconomic gapcontributesto the economic vulnerabilityof female-headedfamilies.InBrazil,female-headedhouseholdshave a 30% o 50%greaterchanceof beingin povertythan do male-headedones, not because they havemore childrenor fewer adultsbut because the female head earns less.She earnsmore than other women do but less than men.13The lowerearningpowerof womenwho head households was a function of theirlower educationin Perul4 and of their restrictedaccess to land andcredit in El Salvador15and in villages in India.16This second set offactors, then, emergesfromgenderdifferencesin access to economicopportunities.It follows that the implementationof policies that ex-pandeconomicopportunitiesor all womenshould reducethe vulnera-bility to povertyof female-headedhouseholds.Third,reasons for the greaterpovertyof female-maintainedami-lies cannot be attributedto household structure factors per se, or,strictly speaking,to gender-relateddifferences in economic opportu-nity, but to the combinationof both. That is, there is an independenteffect of female headshipon household economic vulnerabilitythatcannot be reducedto the characteristicsof women or the household.This effect, in turn, can operate throughthree differentmechanisms.First, women who areheads of households and have no other(female)adulthelpalso have to fulfillhomeproductionor domestic roles. Theytherefore face greatertime and mobilityconstraintsthan male headsand other women do, which can result in an apparent"preference"for working fewer hours for pay, for "choosing" lower-paying jobsthat are nevertheless more compatible with child care, and for spend-ing more for certain services, such as water and housing, because

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    8/23

    MayraBuvinidand Geeta Rao Gupta 265they cannot contributetime to offset transaction costs. C. Chipandedescribes how female farmersin Malawi were inclined to limit theirlabortime in farm activities due to a heavy commitmentto domesticchores." CatherineBerheideand MarciaSegal, andSherrieKossoudjiand Eva Mueller,reportsimilar indings.1" heyfound thatresponsibil-ity for childrenandhousekeepingmade it difficult or female heads toopt for regularor off-farm aboractivities to increasetheir earnings.Second, women who headhouseholdsmayencounter discrimina-tion in access to jobs or resourcesbeyond that which they encounterbecause of theirgender,or maythemselves, because of social or eco-nomic pressures, make inappropriate hoices that affect the house-hold's economic welfare. In Chile, for instance, MarianaSchkolnikfound that female heads had significantly ess access to governmentsubsidiesthan did other heads." Third,female heads may have a his-tory of prematureparenthoodandfamily instabilitythat tends to per-petuate poverty to succeedinggenerations.There are many social and economic situations that predisposecertainfamilytypes to poverty,andthese maybe highlyculturebound.For instance, out-of-wedlockteenage childbearing s a majorpredis-posingfactorin many LatinAmericanand in some Africancountries,while in India early widowhood is far more important.Researchersneed to investigatethe relativecontributionof the differentfactors ingenerating emale-headed amilies and in determining heirpoverty inorder to design appropriateand effective interventions.Theexceptions. Not all householdsthatreporta femalehead arepoor, and some of the measures used underestimatethe poverty offemale headship. Among the exceptions to a destiny of poverty arewomen with economicmeanswho choose to headfamilies,andwherethereis a traditionof womenlivingapartfromhusbandsandthe oldergeneration.Examplesof the latter include wives of polygamous menwho set up independenthouseholdsin West Africansocieties, such asin the Ivory Coast,20and women who inherit and and the rightto setup a householdthroughmatrilinealdescent, such as the Minangkabauof west Sumatra, ndonesia.21Theoppositeis also true,and in societieswhere the norm is to have men headingfamilies, female-headed ami-lies carrya particularlyhighrisk of poverty. These includehouseholdsheaded by single mothersin Brazil22 nd Kenya,23or by abandonedwomen in Bangladesh.24There is heterogeneityin the situation of female heads who havebeen left behindby the economic migrationof theirpartner,and thisheterogeneitydependsboth on the generosityandregularityof remit-tances as well as on the situation of the left-behindhousehold. Inimpoverishedrural areas-such as southernBotswana, where the re-turns from agricultureare uncertain-men's remittances, if any, donot begin to offset the costs of labor requiredto maintainadequate

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    9/23

    266 EconomicDevelopmentand CulturalChangeproductivity,andfemale-headed armhouseholds tend to be the poor-est.25However, in other, more promisingrural situations, female-headedfamilies with access to resources or remittancescan sometimesbe betteroff than male-headedones. Examplesincludefemale-headedcash-croppingarmhouseholds n Kenya,26eft-behind emale headsofTrivandrumDistrict27ndUttarPradesh28n India,andfemale-headedhouseholdsof the Teba tribein Malawi.29Methodologicalconsiderations. It is likely that some of the ex-ceptions are a result of the fact that the more sophisticatedconsump-tion-expendituremeasuresandadult-equivalence cales underestimatethe povertyof female headship, especially when poverty is measuredintergenerationally.The representationof female-headed householdsamongthe poordiminisheswhenpercapitarather hantotalhouseholdincome or expendituremeasures are used, because female-headedhouseholds are, on average, smaller than other households.30 Thequestion is whetherthe per capita adjustmentwipes out the povertyeffect associatedwith femaleheadship,and the answeris yes when thestudiesuse per capitaconsumptionexpendituremeasuresof householdpoverty.31Theanswer s no, however,whenper capitaincomeor assetmeasuresare used. Across regions, 13 out of 15studies that used thesemeasures found that there was a disproportionaterepresentationoffemale-headedhouseholdsamongthe poor.32Because of a number of properties (they are less distorted byincome underreporting nd seasonal variability n earningsand betterreflectthe use of savings), consumptionexpenditure s vastly preferredover income in the development iteratureas a measureof well-beingand poverty.33 Most poverty experts would thereforegive more cre-dence to the findingsof studies thatused the per capita consumption-expenditure indicator (that showed no poverty effect for female-headed households) rather than the per capita income indicator.However, unlikeincome, consumption ndicatorsare likely to be sub-ject to genderbiases that resultin underratinghe poverty of female-headed households because of differences in respondents'access toinformationandbecausethese indicatorsdo not pick up differencesinboth leisure availableto householdsand the intergenerational imen-sion of poverty.First, female heads of smallerhouseholds, since they are oftenbothchiefearnersandhousewives,arelikelyto reportmoreaccuratelyhouseholdconsumptionexpendituresthan are wives in largerhouse-holds who maynot havecompleteinformation n household,andespe-cially male, expenditures.This tends to inflatethe consumptionof thefemale-maintainedmaller household and understatethat of male- orjoint-headedhouseholds.34This potentialbias will be less relevant insocieties, some in East and SoutheastAsia, where wives have tradi-tionally controlled household expenditures, and continue to do so.

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    10/23

    MayraBuvinidand Geeta Rao Gupta 267Second, andperhapsmoresignificantlyntermsof capturinghelegacyof poverty, consumptionmeasuresmaybe augmentedsubstantiallybythe time women and childrenspendin the productionof home goods,especially in farmhouseholds, at the expense, however, of time de-voted to othercriticalactivities,including hild care and incomegener-ation for womenandschoolingfor children.35 he substitutionof workfor leisureto achieve a certain evel of consumption n female-headedhouseholdsmay signifythe perpetuationof povertyinto the next gen-eration.More than half of the studies we reviewed that examined theconsequencesof femaleheadshipfor childrenshowed thatchildren nfemale-headed households worked more often than did children inotherhouseholds,with potentiallynegative long-term mplications ortheirwell-being.36In supportof the hypothesisthat consumption-expendituremea-suresmaybe subjectto genderbiases, C. Lloyd andA. Brandondoc-ument the economic disadvantageof female-headedhouseholds inGhanathroughother indicators.These householdshave less access toland, credit, and education;have higher dependencyratios; and de-pend more on outsidesupport;and memberswork significantlyongerhours.37C. JohnsonandB. Rogers similarly ound thatfemale-headedhouseholdsin the DominicanRepublic,althoughnot overrepresentedamongthe poor,were moreeconomicallyvulnerable hanmale-headedhouseholds because of a greaterdependenceon income transfersinthe form of gifts and remittancesfrom others.38And SandraRosen-house reportsthat in Peru,in the poorer multiple-earneramilies,bothmale- and female-headed households are equally disadvantagedinterms of per capita consumption,but the workburdenof female headsis significantlyhigherthan that of maleheads, potentiallyresulting nlong-termcosts for women and children.39Many of the studies show that female-headedhouseholds have ahigher dependencyburdenthan other householdsdo, which is corre-lated with poverty. A problemwithpercapitaindicatorsof householdeconomic status is thatthey fail to capturedifferentdependencyratiosacross headshiptypes. Adult-equivalencycales, which are oftenusedto adjusthouseholdconsumptionby size andcompositionof the house-hold, furthermask dependencyburdensby assigningan adult maleequivalenceof less than one to females and children,on the assump-tion that theirconsumptionneeds are less thanthose of men.40Usingper capita indicatorsand, in particular,adult-equivalency cales willdistort poverty assessments if high dependency burdens increasehouseholds' currentand future poverty by increasingmothers' andchildren's work efforts to achieve a certain level of consumption.These per capitaindicatorswill yield artificiallyow povertyrisks forhouseholds with high proportionsof dependentchildrenrelative toadults, such as those headedby women.

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    11/23

    268 EconomicDevelopmentand CulturalChangeBecause of these problemswith poverty indicators,a variety ofindicators,rather hanonly one measureof well-beingandpoverty, is

    advisable when assessing the condition of female-headedhouseholds.Equallyas revealingof well-beingas incomeandexpenditureare socialincome measures, such as access to social services andsubsidies, andinformationon the determinantsof income, such as assets andaccessto creditmarketsand agricultural echnologies.41The Transmissionof DisadvantageA most persuasiverationale or targeting emaleheadship s to reducethe transmissionof povertyinto the next generation.In the 29 studiesreviewed, when the indicatorsof disadvantageare children's nutri-tional status and school performance,female headship sometimesseems to protectagainst poverty, and at other times to reproduceit.There is a slightbias towardfindingmoreprotectiveeffects in Africa,but recent studiesreportthis phenomenonalso in LatinAmerica andthe Caribbean.42Children'snutrition. There are varyingresults from the studiesthat have examined nutritionaleffects on children. Of the 18 studiesthat examined these effects, roughlyhalf reported positive and halfreportednegativeeffects of femaleheadship. S. Kumar'sstudy fromZambiaand M. Garcia'sstudyfromthe Philippines ound that a greaterpercentage of children in female-headed households were malnour-ished comparedwith children rom otherhouseholds.43CharlesWoodfound that the survival probabilitiesof children in female-headedhouseholdsin Brazil were significantlyower thanthose of children nmale-headedhouseholds.44This differencein child mortalitywas notthe result of female headship per se, but rather was the outcome ofdifferences n race, region,education,housingquality, monthlyhouse-holdincome, andotherstandard-of-livingndicators.Similarly,Brazil-ian childrenof female-headedhouseholdswere more likely to work,butthis was accountedforby characteristicsassociatedwiththe lowerincome andliving standardsof households withfemale heads.45The studies found that when controllingfor variablesassociatedwithfemaleheadship,the negativeeffect of headshipon child welfaredisappeared.Thisoutcomeimpliesthat,at least in Brazil,female head-ship per se does not add extra burdensto being a woman or beingblack, and wouldthereforearguefor less targetedmeasuresto attackthe intergenerational overtytransmittedby female headship.Studies that reporta positive effect of female headshipon childnutrition indthis effect to be moresignificantn poorerthanin better-off households.46The more credible explanationfor the positive ef-fect is that there are gender differences in expenditure preferences(whether hispreferenceresultsfromnatureor nurture s immaterial ntermsof policy).47 Thisexplanationrests on the notion thata woman's

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    12/23

    MayraBuvinidand Geeta Rao Gupta 269greaterpreferenceto invest in children is more easily realized in ahouseholdshe heads, wherethereareno conflicts or negotiationswitha male partnerover the use of household resources. This preferenceappears in poorer rather than in better-offfamilies, either becauseinvestmentsin childrenyield greaterreturnsat lower levels of incomeor becausethere are fewercompetingalternativenvestments n lower-thanin higher-incomehouseholds.48The gender-preference explanation receives support from re-search that finds similar effects when analyzingunearnedincome ofmothers who are not female heads.49 An alternativeexplanationisthat more competentmothershave more success at earninga higherproportionof familyincome, living on theirown, and caringfor theirchildren.Children's education. Of the studies that examined the impactof female headshipon children'seducation,four reporteda positiveeffect: that children's educationis more likely to receive priorityinfemale-headedhouseholds than in male-headedhouseholds."5Six re-porteda negativeeffect: because of the lack of additionaladultlaborand because of low income levels in households headed by women,childrenare often forced to dropout of school to assist in houseworkand child care." These results are not necessarilycontradictory.It ismore likely that they representevidence of the conflict that womenheads of household mustface, that is, the need to use every availableresource to survive versus a desire to invest in their children. Thesame canbe argued or the observationof bothprotectiveandhigh-riskeffects of female headshipon child nutrition. That is, the protectiveeffects fromgender-relatedpreferencesare likely to breakdown withincreasing mpoverishment.

    Overall, the evidence suggests that poor female-headedhouse-holds prefer to invest scarce resources in children,which translatesinto increased child welfare relative to income. When women haveaccess to insufficientincome, they cannot act on their preferencessufficiently o makeup the difference.In this lattercase, the economicdeprivation hatthey suffer s transmittedo thenextgeneration.Whilethe primarymediation for the reproductionof poverty appearsto beeconomic, some evidence from Latin America suggests that the ab-sence of fathersmaytransmit ocialas well as economicdisadvantagesto the next generation,eitherdirectly by the fathers'absence or indi-rectly by affectingthe caretakingbehaviorof mothers andotherchild-rearingagents.52Targeting Female-Headed HouseholdsAnticipating Costs and BenefitsUnderpressureto reduceboth povertyandpublicexpenditures, gov-ernmentstarget nterventions o increase cost effectivenessandinsure

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    13/23

    270 EconomicDevelopmentand CulturalChangethat assistance reaches the neediest. Using female headshipas a tar-geting criterionis, in principle,attractivebecause of the associationbetween female headship and poverty. Targetingthese householdsmay be especially useful when there are no other reliable ways toidentify poor households and when their prevalence is not too high.There is logic to the argument hat the morepervasivea problem,thegreater s the case for resolvingit with a universalprogram.53But all targetinghas costs andbenefits,for the poor as well as forthe agency that finances targetingprograms.For the governmentorthe agency that financestargetingprograms,the costs of targeting n-clude the cost of the goods and services delivered to the poor, theamountof leakageto the nonpooras a result of errorsin the screeningprocess, the administrativecosts of the program, political economycosts, and second-roundcosts or perverse incentive effects that arebuilt into the nature of targeting nterventions o some but not to all.Errors in screening female-maintainedhouseholds and familiescan arise frommisidentifying he genderof the head or the povertyofthe household. The former errors shouldbe higherin terms of falsenegatives(i.e., failingto capturehouseholdsthatareclassified as maleheaded when in reality they are female headed) than false positives.This is a result of the increasingnumberof women who in actualitymaintainhouseholdsin developingcountriesbutare not recognizedasdoing so because of culturalprescriptionsthat identify the man asthe main breadwinnerand householdauthority.54 hiftingthe unit oftargetingfrom female-headedhouseholdsto female-maintained ami-lies (i.e., kinshipunits residingon their own or in largerhouseholds)should reduce the number of false negatives. The number of falsepositives (or leakages to male-headedhouseholds)may neverthelessincreasewith time. This is because over time womenmay shift in andout of the status of familyhead(e.g., by changing heir maritalstatus),andprograms hat do not monitorbenefitsclosely canendupprovidingbenefits to some male-headedhouseholdsthat were formerlyfemale-headed. Programdesignersshould examine the empiricalquestion ofhow easily the status of female headshipmay be altered. This easeshouldvarywith the originsof femaleheadship,women'sphysicalandsocialmobility,and the life-cyclestageof thefamily.De factoheadshipshouldin theorybe more subjectto changethande jure headship.A second type of misclassificationerror occurs when female-headed households are used as a proxy for poverty households andthere are leakagesto nonpoorfemale-headedhouseholds.In this case,the numberof falsepositives (identifyingas poorfemale-headedhouse-holds that are not so) shoulddecrease with the increasingproportionof female heads that are poor in the population. The likelihood ofproducing false negatives (failing to identify poor female-headed

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    14/23

    MayraBuvinidandGeeta Rao Gupta 271households) should increase as the proportionof female heads whoarepoor declines.

    Targetingfemale-headed households and families may producesignificantbenefits, both direct and indirect, to participating amiliesand to society. Targetingpoor female heads will producedirectbene-fits for women if the gross amount of the benefit or transferexceedsparticipation osts. The time costs of participatingn a targeted nter-vention should be greaterfor these women than for men who headhouseholdsor for womenin male-headedhouseholds. In addition,par-ticipationcosts arisingfrom social mores may be higher for femaleheads than for other women (since in many cases a social stigma isattachedto the conditionof female headship). Targeting emale head-ship may have largerbenefits for children than targetingpoor male-headed households with equal amounts of benefit, as the evidencereviewed on consequences of female headship for child well-beingpoints out. As an indirector second-roundbenefit, targetingfemaleheadshipcan increase the fairness of developmentinterventions.An-otherindirectbenefit s makingvisible theunrecognizedeconomiccon-tributionof poor women.The costs to society of targetingfemale-maintained amilies canbe as largeas the benefits,andhighanticipatedcosts are likely one ofthe main reasons why female headshiphas so seldom been targetedby governmentprograms.Any additionalscreeningcriterion mposesadministrativecosts. Further,since targetingfemale-headedfamiliescan imply excludingmale-headed amilies, the result could be signifi-cantpoliticalcosts, both realandperceived.These costs may be sub-stantially ower, however, if female-headed amilies are targetedwithinterventions hatare specificto women, such as nutritionsupplemen-tationsfor pregnantwomen, maternaland child health interventions,or training n female-specificoccupations.Politicalconsiderationssug-gest avoidingthe risk of antagonizinghe morepowerfulmaleconstitu-ency by targeting emale-headed amilies with resources that are notperceived as female specific, such as housing subsidies, agriculturalloans, food coupons, and cash transfers.In addition,it suggests thattargetingwomen who head familiesmaybe less politicallyviable thantargetingall women. These politicalconsiderationsdo not bode wellfor the long-termsurvival of initiatives that target female headshipwith significantdevelopmentresources.Perhapsthe majorperceived cost of targeting emale-maintainedfamilies is the potential of changingthe way families behave in re-sponse to programincentives. Possible perverse incentive effectscould includeincreasingthe prevalenceof female-headedhouseholds.A dilemmamay arise for programs hat targetfemale-headed amilieson the rationaleof theirpoverty:do the short-termbenefits of poverty

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    15/23

    272 EconomicDevelopmentand CulturalChangereliefcompensatefor the possiblelong-term ncreasein the prevalenceof female-headedhouseholds?TargetingFemale Headshipin ChileIn 1991,the governmentof Chile, through ts newly createdWomen'sNational Service (Servicio Nacional de la Mujer-SERNAM), choseas a prioritya program argeting emale headshipin order to increaseincomes, improvewelfare,andfightgenderdiscrimination."552-yearpilot project launchedin five poor municipalitiesaroundthe countryprovided2,500female heads withjob training,housingimprovement,child care (includingeducationalprogramsfor the children), healthservices, and legal aid. It included some universalcomponents, suchas buildinginfrastructureor child care, as well as extending to allusers the hours of operation of health clinics to accommodate thetime constraintsof workingfemale heads. Female household headsparticipatingn the pilot projecthadpriorityaccess to these universalbenefits.Theproject ncluded n the beneficiarypopulationhouseholdsheadedby women with childrenyoungerthan 14 as well as unwed orpartnerlessmothers and their children, even if the latter resided assubfamiliesin largerhouseholds. The project included only a smallsubsidyfor transportation nd no cash transfers. It emphasizeda pro-ductiveratherthan a welfareorientation,offeringno free handoutsordirect cash transfers and requiringwomen to devote significanttimeto the project'strainingcomponent.It was financedby state and inter-nationaldonors and it was implemented hroughthe municipalities.56The pilot projecthas now become a nationalprogram,and a priorityprogramwithin the government'snew strategyto combatpoverty.The projecthas been successful in targetingthe poor as well asfemaleheads, with very few leakagesto the nonpoorandwomen whoare not de facto heads. In 1992we estimated the overallprojectcostsin one of the municipalities(Conchali, in Santiago)to be $450 perwomanper year, whichis roughlysimilar o the cost of targetednutri-tionalinterventionsset at roughly$420perchildper yearfor a comple-mentaryfeeding program-the ProgramaNacional de Alimentaci6nComplementaria.57 survey of halfof the beneficiaries,carriedout in1992,showed thata full 91%of the project participantshadper capitafamily incomes below the poverty line (poorest 30%),and 57% weredestitute (poorest 10%).Only 0.2%of the householdsfell amongthenonpoor.58Table 3 breaks down these results by municipality.Thepovertyline used in the studywas below the governmentpoverty lineset at the cost of two minimumfood baskets for the poor and oneminimumfood basket for the destitute. Only one other governmentprogram outof 11)rankshigherthan this projectin terms of targetingbenefitsto the poor with very few leakagesto the nonpoor, as table 4shows. For instance, at a similarcost, 69%of the participants n the

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    16/23

    MayraBuvinidandGeeta Rao Gupta 273TABLE3

    POVERTYARGETINGFFEMALEEADSHIPILOTROJECTYMUNICIPALITY:HILE(June1992;%)Arica Conchalf Santiago S. Antonio Coronel Total

    Destitute(1 decile) 55.4 62.1 43.4 61.0 63.1 57.0Poor(1-3 deciles) 93.2 91.3 79.2 94.1 95.4 90.7Nonpoor 6.8 8.7 20.8 5.9 4.6 9.3SOuRCE.-Juan abloValenzuela,"Caracteristfcase Pobrezade las Participantesen el Plan Piloto" (Povertyfeaturesof participantsn the Pilot Plan for Female Headsof Households), nDe MujerSolaa JefadeHogar:Genero,Pobrezay PoliticasPTiblicas(Fromunpartneredemaleto headof household:Gender,poverty,andpublicpolicies),ed. M. E. Valenzuela,S. Venegas,and C. Andrade Santiago:Women'sNational Ser-vice, 1994),p. 257.

    PNACwere below the povertyline, andonly 38%were in the poorest20%of the per capitaincomedistribution,while a full 81.4%of benefi-ciaries in the female headshipprojectwere amongthe 20%poorest.59There were very few leakagesof projectbenefitsto women whowere not de facto heads. To avoid screeningerrors,the projectuseda detailed operationaldefinitionof female headship, required signifi-cant time commitments romwomen, andencouragedsomethingakinto groupsolidaritymechanisms,wherethe participantshemselves as-sumedresponsibility or identifyingotherfemale heads. The Conchaliproject, for instance, defined as female-headed households families

    TABLE4TARGETINGFSOCIALROGRAMSYPERCAPITANCOMEUINTILES:HILE1990)

    1 2 3 4 5Welfarepensions(PASIS) 45.5 22.2 19.4 10.0 2.9Family subsidy(SUF) 50.8 29.3 13.5 4.8 1.6Familybenefit 21.6 25.7 21.2 17.6 13.9Unemploymentbenefit 57.6 28.3 7.6 4.5 2.0Complementaryeeding(PNAC) 38.1 30.8 18.0 9.5 3.6Preschool 27.6 28.8 20.3 13.6 9.7Primary chool 35.7 26.9 18.1 12.6 6.7School lunches(PAE) 52.4 27.3 13.2 5.3 1.8Housing 29.4 28.4 23.5 13.5 5.2Sites and services 26.9 28.0 27.0 14.3 3.8SENAME* 100.0 0 0 0 0

    SOuRCE.-JuanabloValenzuela,"Caracteristicase Pobrezade las Participantesen el PlanPiloto"(Povertyfeaturesof participantsn the PilotPlanfor Female Headsof Households), n De MujerSola aJefadeHogar:Gdnero,Pobrezay Politicas Pablicas(Fromunpartneredemaleto head of household:Gender,poverty,andpublicpolicies),ed. M. E. Valenzuela,S. Venegas,and C. Andrade Santiago:Women'sNationalSer-vice, 1994),p. 258.* Data for 1987,Children'sNational Service.

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    17/23

    274 EconomicDevelopmentand CulturalChangewith childrenunder14yearsof age thatare maintainedby women, aswell as unwed or partnerlessmothers and their children. The femalehead had to be underage 55. If a husbandor partnerwas present, hehad to be infirm,unemployed,or unableto work.If the pilot projectis successful in increasingthe income of poorwomen who are heads of households, it may increasethe prevalenceof femaleheadshipby changing heliving arrangements f unpartneredmothersand their children now living in someone else's household.One in every five householdsin Chile has a subfamilyresidingin thehousehold,something hat few wereawareof before 1990.60More thanhalf of these subfamiliesare headed by single motherswith childrenwho wouldhave the optionof movinginto an independenthousehold.This hasbeenone of the maineffects of welfareprogramsnthe UnitedStates:it enablessinglemotherswithchildren o establish ndependentresidence.61f the U.S. effects are repeatedin Chile, the project mayincreasethe prevalenceof female-headedhouseholds without or evenbefore alteringthe costs and benefitsof marriage or women. On thepossibilitythat the projectcould have a perverseincentiveeffect, proj-ect implementersbelieve that the opposite may occur; that is, theproject,by reducingwomen's feelings of stigmatizationand failureinpartnerrelationshipsandby raising heirself-esteem,mayincreasethelikelihoodthat female heads, especially the younger ones, will formstablepartnerrelationships,reducingthe prevalenceof female head-ship. Fortunately,because the project includedbaseline informationon participantsand a control group of nonparticipants,over time itwill be possible to answer the question of the project's impact onprevalence and disaggregate he contributionsof residentialchangesandperverseincentives to possible changesin prevalence.DiscussionThe Chileexperienceshows that targeting emaleheadshipcan be anefficient way of reachingthe poor and it reinforces the associationbetween female headshipand povertyfoundin the review of the evi-dence. Butfemaleheadship s heterogeneous,andnot all femaleheadsare poor; therefore,the decision to targetshouldbe precededby ananalysis of indicatorsrelated to the nature, rise, and vulnerability opovertyof femaleheads, includingmaritalchanges,fertilityrates andthe rise of unpartneredertility,migrationrendsby sex, andchangingfamilyforms. Circumstances hatmay ustifytargetingemaleheadshipinclude economic crises and civil dislocations.Improvedmeasures offemale family headshipand carefulanalyses of costs and benefits aswell as of the targetingenvironmentcan reduce screeningerrors andminimizeprojectcosts.Reportedfemale headshipis a crude indicator of actual femaleheadship.A more refined ndicatorof femalefamilymaintenance akes

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    18/23

    MayraBuvinidand Geeta Rao Gupta 275into account the age and maritalstatus of the head (to gauge origins)andthe presenceandages of dependents.In addition,attemptscan bemade to identifysubfamiliesmaintainedby women withinmale-headedhouseholds (as the governmentof Chile has done), and the targetedpopulationcan be expandedto cover families as well as households.The costs to the agency that finances such a projectare adminis-trativeandpolitical.Leakagesto male-headedhouseholds and to non-poor female-headedhouseholds can be reducedby usinga set of indi-cators beyond reported headshipto identify female family headship;devising and implementingpublic works and other interventions de-signed to attract female participants(by includingfeatures such aschild care facilities, flexible hours, transportation,and gender-awareimplementingagencies)withoutspecificallytargetingwomen; and de-livering emale-specificpublicgoods. But theevidence shows thatpoorfemale heads also need services,jobs, andtraining hat arenot female-specific. Because targetingthese benefits may antagonize the morepowerful male constituency, such programs should anticipate andguard against potentiallyhigh political costs. Antipoverty programsthattarget emale heads withresourcesor benefitscan, however,antici-pate increases in the prevalence of female-headedhouseholds as aresult of changes in the living arrangementsof female family headswho, as a result of the intervention,are able to afford independenthousing. These programscan quantifythese predictable ncreases tominimizepotentialpoliticalbacklash.There can be sizable direct and indirect benefits to participatingfamilies from antipovertyinterventions that are targetedto poor fe-male-headedfamilies. Such interventions can reduce the poverty ofwomenand offergreaterwelfarebenefitsthanwouldresultfromdirect-ing equalresources to male-headed amilies.In addition,when femaleheadshipadds additionalburdensto familypoverty-burdens that arenot entirely explained by household composition factors or genderdifferences n economicopportunities-governmentswould do well totarget female heads with antipovertyprograms.The benefits of in-vesting in poor women, for example,will not trickle down as well-ifat all-to female heads. Female heads of familiesrequire nterventionsthataredirectedspecifically o them, such as income-earning pportu-nitiesand child-caresupport,as well as affirmativepolicies to preventdiscriminationn access to markets and resources, aggressive healthand educationcampaigns(e.g., services for pregnant eenagers), andthe establishmentof effective social supportnetworksthrough ormalor informalorganizations.In answer to the question raised in the title of this article, wehold that targetingfemale headshipto reduce poverty in developingcountries s worthwhile n theory,and thatit can work in practice.Butthere areexceptionsand constraints acingthe designandimplementa-

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    19/23

    276 Economic Development and Cultural Changetion of interventions that are targeted to poor female-headed families.We have developed our arguments on the basis of increasingly substan-tial research evidence on the situation of these families, but there islittle documentation of experiences with targeting them. The next stepis to invest more in experimental antipoverty interventions that targetdiverse categories of female-headed or female-maintained households(including public initiatives to promote the economic responsibilityattached to fatherhood) and in sound evaluations of the experiencesof both governments and nongovernmental organizations. Policy-oriented research is far ahead of program and policy implementation.Still, significant research challenges remain, including the use of morediscriminating measures of the poverty of female-headed families andan exploration of the origins of family formation and female headship,of the stability of female-headed households and female-maintainedfamilies, and of the mechanisms that predict the likelihood that female-headed families will transmit poverty to the next generation.Notes* This articlewas prepared or the PopulationCouncilandInternationalCenterfor Research on WomenJointProgramon FemaleHeadshipand Pov-erty in Developing Countries,with supportfrom the Ford Foundation,thePopulationCouncil,and the United Nations Fundfor PopulationActivities.We wish to thankMariaElena Valenzuela SERNAM,Chile)for her substan-tive contributionso thisreport;JudithBruce(PopulationCouncil),LawrenceHaddad(IFPRI), and Joanne Leslie (UCLA) for their comments on earlierdrafts of this article;Susan Kalish (PRB)for her editorialsuggestions;andLinda Sturgeonfor her patience in preparing his version of the article andearlierdrafts.1. IdrissJazairy,MohiuddinAlamgir,and TheresaPanuccio, The Stateof WorldRuralPoverty(New York:New YorkUniversityPress [forthe Inter-nationalFundfor AgriculturalDevelopment],1992).2. Margaret E. Grosh, Administering Targeted Social Programs in LatinAmerica: From Platitudes to Practice, World Bank Regional and SectoralStudies(Washington,D.C.: WorldBank, 1994).3. Nancy Folbre, "Motherson Their Own:Policy Issues for DevelopingCountries"(paperprepared or thejoint ICRW/PopulationCouncil series onthe Determinantsand Consequencesof Female-HeadedHouseholds, 1990).4. Ibid.5. Sara McLanahan,"Family Structureand the Reproductionof Pov-erty," American Journal of Sociology 90, no. 4 (1985): 873-901.6. A table summarizinghese studies (samplesize, methodology,defini-tions, and studyfindings n termsof povertyconditions andconsequences)iscontained n MayraBuvinidandGeetaRao Gupta,"TargetingPoor Woman-Headed Households and Woman-Maintained amilies in Developing Coun-tries: Views on a PolicyDilemma" paperpreparedor thejointICRW/Popula-tion Council program on Female Headship and Poverty in DevelopingCountries, 1994).7. Sherrie Koussoudji and Eva Mueller, "The Economic and Demo-graphicStatus of Female-HeadedHouseholds n RuralBotswana,"EconomicDevelopment and Cultural Change 21 (July 1983): 831-59.

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    20/23

    MayraBuvinidand Geeta Rao Gupta 2778. CatherineW. Berheideand MarciaT. Segal, "LocatingWomen n theDevelopmentProcess:FemaleSmall-Holdersn Malawi"(paperpresentedatthe meetingof the National Women'sStudiesAssociation,June 1989).9. ThomasW. Merrickand MarianneSchmink,"Households HeadedbyWomen and UrbanPoverty in Brazil," in Womenand Poverty in the ThirdWorld,ed. M. BuviniC,M. Lycette, and W. McGreevey(Baltimore:JohnsHopkinsUniversityPress, 1983).10. SandraRosenhouse,"IdentifyinghePoor:Is Headshipa UsefulCon-cept?" Living StandardsMeasurementStudy WorkingPaperno. 58 (WorldBank, Washington,D.C., 1989).11. Leela Gulati,"Impactsof Male Migrationo the Middle East on theFamily: Some Evidence from Kerala," WorkingPaper no. 76 (Center forDevelopment Studies, Trivandrum, Kerala, India, 1983); Devaki Jain,

    "Women and TheirHouseholds-the Importanceof Women in Macro Poli-cies" (paperpresentedatjoint ICRW/Population ouncilSeminar II, "Deter-minantsof HouseholdsHeadedor Maintainedby Women:Considerationsofthe Lifecycle," New York, April10-11, 1989).12. Folbre(n. 3 above).13. Ricardo Barros, Louise Fox, and Rosane Mendonqa, "Female-HeadedHouseholds,Poverty, and the Welfareof Children n Urban Brazil"(paperprepared or the joint ICRW/PopulationCouncilprogramon FemaleHeadshipandPovertyin DevelopingCountries,1993).14. MartaTiendaandSylviaOrtegaSalazar,Female-HeadedHouseholdsandExtendedFamilyFormations n Rural and UrbanPeru (Madison:Centerfor DemographyandEcology, Universityof Wisconsin, 1980).15. Susana Lastarria-Cornheil, Female Farmers and AgriculturalPro-duction in El Salvador,"Developmentand Change 19 (1988):585-615.16. Jain.17. G. H. R. Chipande,"InnovationAdoption among Female-HeadedHouseholds: The Case of Malawi," Development and Change 18 (1987):315-27.18. BerheideandSegal(n. 8 above);Koussoudjiand Mueller n. 7 above).19. MarianaSchkolnik,"Chile:Impactodel GastoSocialen los Hogarescon JefaturaFemenina" Chile:Impactof socialexpenditure nfemale-headedhouseholds),Draft PREALCWorkingPaper(Santiago,Chile, 1991).20. OdileFrank,"TheChildbearing amily n Sub-SaharanAfrica:Struc-ture,Fertility,and the Future"(paperpresentedatthejoint ICRW/PopulationCouncil SeminarI, "Conceptsand Classificationsof Female-HeadedHouse-holds:ImplicationsandApplications or NationalStatistics," New York, De-cember12-13, 1988).21. LaurelK. Schwede,"FamilyStrategiesof LaborAllocationand Deci-sionMaking na MatrilinealslamicSociety:TheMinangkabuf West Suman-tra, Indonesia"(Ph.D. diss., CornellUniversity, 1991).22. Barroset al. (n. 13above).23. CarolynBarnes,"Differentiation y Sex amongSmall-ScaleFarmingHouseholdsin Kenya," RuralAfricana15-16 (Winter/Spring 983):41-63.24. CarolSaldert,Female-HeadedHouseholdsinRuralBangladesh,Re-portfroma MinorResearchTaskForce (Department f Social Anthropology,Universityof Stockholm,January-February 984).25. Koussoudjiand Mueller(n. 7 above).26. EileenKennedy,"TheSignificanceof Female-HeadedHouseholdsinKenya" (paperpresentedat the joint ICRW/PopulationCouncilSeminarII,"Consequencesof FemaleHeadshipandFemaleMaintenance,"Washington,D.C., February27-28, 1989).

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    21/23

    278 Economic Development and Cultural Change27. Gulati(n. 11above).28. Jain(n. 11above).29. Eileen KennedyandPaulinePeters, "Household Food SecurityandChildNutrition:The Interactionof IncomeandGenderof HouseholdHead,"WorldDevelopment 20, no. 8 (1992): 1077-86.30. PravinVisaria,"PovertyandLivingStandardsn Asia," Populationand Development Review 6, no. 2 (1980): 189-223.31. F. Louat, M. E. Grosh,andJ. van der Gaag, "WelfareImplicationsof FemaleHeadship n JamaicanHouseholds"(paperpresentedat the Interna-tional Food Policy ResearchInstituteWorkshopon InterhouseholdResourceAllocation:PolicyIssues and ResearchMethods,Washington,D.C., February12-14, 1992);CynthiaB. Lloyd and AnastasiaJ. Brandon,"Women's Rolesin MaintainingHouseholds:PovertyandGenderInequality n Ghana"(paper

    preparedorthe ICRW/PopulationCouncil oint program n FamilyStructure,FemaleHeadship,and Povertyin DevelopingCountries,1991);CatherineF.JohnsonandBeatriceLorgeRogers,"Children'sNutritionalStatus in Female-Headed Households in the DominicanRepublic,"Social Science and Medi-cine 37, no. 11(1993): 1293-1301.32. Koussoudji and Mueller (n. 7 above); Farida E. Arif, "Self-Employment orRuralDistressed Female Headed Households-a Case Studyof a PilotProject n Bangladesh" paperpresentedat the InternationalLabourOrganisationNationalWorkshopon Female-HeadedHouseholdsandthe De-velopmentof Guidelines or TheirParticipationn Development,New Delhi,April26-28, 1988);ShahnazKazi and Bilquees Raza, "Households Headedby Women:Income, Employment,andHouseholdOrganization" paperpre-sentedat the fifth annualgeneralmeetingof the PakistanInstituteof Develop-mentEconomics, Islamabad,January4-6, 1989);Barros et al. (n. 13above).33. MartinRavallion,"PovertyComparisons:A Guide to ConceptsandMethods," Living Standards MeasurementStudy Working Paper no. 88(WorldBank, Washington,D.C., 1992).34. Susan Horton and BarbaraDiane Miller, "The Effect of GenderofHouseholdHead on Food Expenditure:Evidence fromLow-Income House-holds in Jamaica"(paperpresentedat the Conferenceon FamilyGender Dif-ference and Development,Economic GrowthCenter, Yale University, Sep-tember4-6, 1989).35. JudithBruceandCynthiaB. Lloyd, "BeyondFemaleHeadship:Fam-ily Research and Policy Issues for the 1990s," in IntrahouseholdResourceAllocation: Policy Issues and Research Methods, ed. L. Haddad et al. (Balti-more: JohnsHopkinsUniversityPress, in press).36. MerrickandSchmink n. 9 above);RanjanaKumari,Women-HeadedHouseholdsinRuralIndia(New Delhi:Radiant,1989);Schwede(n. 21above);Jeanne Frances I. Illo, Irrigation in the Philippines: Impact on Women andTheir Households: The Aslong Project Case, Women's Roles and Gender Dif-ferences in Development:Cases for Planners,Asia 2 (Bangkok: PopulationCouncil, 1988);Barrosetet al. (n. 13above).

    37. Lloyd and Brandon.38. JohnsonandRogers.39. Rosenhouse(n. 10above).40. Ravallion.41. Nora Lustig, "MeasuringPoverty in Latin America: The EmperorHas No Clothes"(paperprepared or "Poverty:New Approaches o Analysisand Policy-a Symposium,"International nstitutefor LabourStudies, Ge-neva, Noyember 22-24, 1993).42. MayraBuvini" J. P. Valenzuela,T. Molina,and E. Gonzalez, "The

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    22/23

    MayraBuvinidand Geeta Rao Gupta 279Fortunesof Adolescent Mothersand Their Children:A Case Study on theTransmissionof Poverty in Santiago,Chile," Populationand DevelopmentReview 18, no. 2 (1992):269-97; PatriceEngle, "MaternalWorkand Child-Care Strategies in Peri-Urban Guatemala,"ChildDevelopment 62 (1991):954-65, and"Influencesof Mother'sandFather'sIncomeon Children'sNutri-tional Status in Guatemala,"Social Science and Medicine37, no. 11(1993):1303-12;JohnsonandRogers(n. 31 above).43. Shubh K. Kumar, "Income Sources of the MalnourishedPoor inRuralZambia,"InternationalFood Policy Research InstituteWorkingPaper(Washington,D.C., 1991);MaritoGarcia,"Income Sources of the Malnour-ished Rural Poor in the Provinces of Arba Antiqueand South Cotabatointhe Philippines,"International ood PolicyResearchInstituteWorkingPaper(Washington,D.C., 1991).

    44. CharlesH. Wood, "Women-HeadedHouseholdsand ChildMortalityin Brazil, 1960-1980"(draftpresentedat the joint ICRW/PopulationCouncilSeminarII, "Consequencesof FemaleHeadshipandFemaleMaintenance,"Washington,D.C., February27-28, 1989).45. DeborahLevison, "Family Compositionand Child Labor:SurvivalStrategiesof the BrazilianPoor" (paperpresentedat the annualmeetingofthe PopulationAssociationof America,March30-April 1, 1989).46. Eileen Kennedy, "Effects of Gender of Head of Household onWomen'sandChildren'sNutritionalStatus"(paperpresentedat theworkshopon the Effects of PoliciesandPrograms n Women,International ood PolicyResearchInstitute, Washington,D.C., January16, 1992);Engle, "MaternalWorkand Child-Care trategies";Buvinidet al., "TheFortunesof AdolescentMothersandTheir Children."47. See Victor R. Fuchs, "Women's Quest for Economic Equality,"Journalof EconomicPerspectives3, no. 1 (1989):25-41.48. Kennedy, "Effects of Genderof Headof Household."49. DuncanThomas,"Intra-HouseholdResourceAllocation:An Inferen-tial Approach,"Journalof HumanResources25, no. 4 (1990):635-64.50. Dov Chernichowskyand ChristineSmith, "PrimarySchool Enroll-ment and Attendance n Rural Botswana"(WorldBank, Washington,D.C.,1979,mimeographed); ouatet al. (n. 31above);SylviaChant,"Single-ParentFamilies: Choice or Constraint?The Formationof Female-HeadedHouse-holds in MexicanShantyTowns," Developmentand Change 16(1985):635-56; Gulati(n. 11above).51. Barroset al. (n. 13above);IsabelVial, EugeniaMuchnik,and JulianaKain, "Evaluationof Chile'sMain NutritionInterventionProgram" Univer-sity of Chile and CatholicUniversity, Santiago,May 1988, mimeographed);Kumari n. 36 above); Kazi and Raza (n. 32 above); Schwede (n. 21 above);DeborahS. DeGraffandRichardE. Bilsborrow,"Female-HeadedHouseholdsand FamilyWelfare n RuralEcuador,"Journalof PopulationEconomics6,no. 4, pp. 317-39.52. Engle, "Influences of Mother'sand Father's Income on Children'sNutritionalStatus nGuatemala";Buvinicet al., "TheFortunesof AdolescentMothersandTheirChildren."53. IrwinGarfinkeland Sara S. McLanahan,Single Mothers and TheirChildren Washington,D.C.: UrbanInstitute, 1986).54. MayraBuvinid,Nadia Youssef, and B. von Elm, "Women-HeadedHouseholds: The Ignored Factor in Development Planning"(Washington,D.C.: InternationalCenterfor Researchon Women, 1978).55. While the establishment f SERNAMencountered ignificantopposi-tion fromconservativesin the revived ChileanCongress,the programon fe-

    This content downloaded from 189.138.204.46 on Wed, 16 Oct 2013 14:13:55 PMAll use subject to JSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 7/27/2019 1154535 (1)

    23/23

    280 Economic Development and Cultural Changemale headshipraised no eyebrows, in part because giving priorityto poorfemale-headedhouseholdswas a safepoliticalalternativewhencomparedwithpromoting egislationon divorce or abortion.56. M. Elena Valenzuela,Sylvia Venegas, and CarmenAndrade, eds.,De MujerSola a Jefa de Hogar:G"nero,Pobrezay Politicas Pablicas (Fromunpartneredemale to head of household:Gender,poverty, and publicpoli-cies) (Santiago:Women'sNationalService, 1994).57. Vial.58. Juan Pablo Valenzuela,"Caracteristicasde Pobrezade las Partici-pantesen el Plan PilotoparaJefas de Hogar" (Poverty eaturesof participantsin the PilotPlanfor FemaleHeads of Households),in Valenzuelaet al., eds.59. Ibid.60. IgacioIrarrizabalandLuciaPardo,"JefaturaFamiliarFeminina,Es-tructuradel Hogar y Pobreza"(Femaleheadship,householdstructure,andpoverty), in Valenzuelaet al., eds.61. Garfinkel ndMcLanahann. 52 above).

    D d t e l tAm erican l?I?t Sor eighty years

    INTERNATIONALOURNAL O FAMERICAN LINGUISTICS

    FoundedbyFranzBoas in1917, IJAL xplores hedevelopmentndunderstandingfgeneral inguisticheoryandtherelations etweenlanguageand culture nNativeAmericananguages.Eachquarter,JALsubscribers eceiverigorousnvestigationsftexts andlinguisticatafrom nternationallyenownedcholars n thefield. Inaddition,heyreceivepresentationsfgrammars,rammaticalragments, ndotherdocumentsanddiscussionsrelevant oAmericanndiananguagesandtheir peakers.EditedyDavid . RoodPublisheduarterlyyTheUniversityfChicagoressRegular one-year subscription rates: $44.00 Individuals;33.00 Students (withcopy of validID);$120.00 Institutions.Outside USA, please add $6.00 forpostage; Canadianresidents, please add 7%GST plus postage. Visa and MasterCardaccepted. To order,send check, purchaseorder,or completecreditcard informationo The University f Chicago Press, Joumals Division,Dept.SF7SA, P.O. Box37005, Chicago, IL60637. Formore information nd onlineordering,visit our website athttp://www.joumals.uchicago.edu.