conﬂicts over credit: re-evaluating the empowerment...

Con¯icts Over Credit: Re-Evaluating the

Empowerment Potential of Loans to

Women in Rural Bangladesh

NAILA KABEER *

Institute of Development Studies, Brighton, Sussex, UK

Summary. Ð This paper explores the reasons why recent evaluations of the empowerment potentialof credit programs for rural women in Bangladesh have arrived at very con¯icting conclusions.Although these evaluations use somewhat di�erent methodologies and have been carried out atdi�erent points of time, the paper argues that the primary source of the con¯ict lies in the verydi�erent understandings of intrahousehold power relations which these studies draw on. It supportsthis argument through a comparative analysis with the ®ndings of a participatory evaluation of arather di�erent credit program in Bangladesh in which the impact of loans was evaluated by womenloanees themselves. Ó 2000 Elsevier Science Ltd. All rights reserved.

Key words Ð microcredit, gender, empowerment, evaluation, Bangladesh, Asia

1. INTRODUCTION: CONFLICTINGEVALUATIONS OF CREDIT

Microcredit programs for the poor have cometo occupy a central place in poverty-orientedstrategies in Bangladesh. Such programs have anumber of features in common. They are largelytargeted at women from the poorest sections ofthe population; they lend small sums of moneyto individuals as members of groups and rely ongroup liability to ensure loan repayment; theysubsidize administrative costs rather thaninterest rates; and loans are repaid in weeklyinstallments. Debates as to the actual e�ective-ness of these programs in reducing povertycontinue. More recently, these debates havebeen extended to the possible implications ofsuch programs for women's empowerment, withsome evaluations claiming extremely positiveresults while others suggesting that microcreditleaves women worse o� than before.

In this paper I want to focus on a number ofattempts to evaluate the empowerment poten-tial of loans to women in order to ®nd out whysuch diametrically opposed claims can be madeabout the same, or very similar, programs. Iwill be exploring examples of both ``negative''and ``positive'' evaluations, interrogating themfor the methodologies they used, the questionsthey asked, the ®ndings they reported and theinterpretations they gave to their ®ndings. In

addition, I will be drawing on the ®ndings ofmy own evaluation of a rather di�erent creditprogram in Bangladesh in order to explore thequestion of empowerment when it is assessedon the basis of women loanee's own testimoniesrather than deduced from selected aspects oftheir behavior.

(a) Does access to credit ``empower'' women?the negative verdict

My ®rst example of a negative evaluation isby Goetz and Sen Gupta (1994). They use a ®ve-point index of ``managerial control'' over loansas their indicator of empowerment. At one endof their index are women who are described ashaving ``no control'' over their loans: these arewomen who either had no knowledge of howtheir loans were used or else had not providedany labor into the activities funded by the loan.At the other end are those who were consideredto have exercised ``full'' control, having partic-ipated in all stages of the activity funded bythe loan, including the marketing of produce.The study found that the majority of women,

World Development Vol. 29, No. 1, pp. 63±84, 2001Ó 2000 Elsevier Science Ltd. All rights reserved

Printed in Great Britain0305-750X/00/$ - see front matter

PII: S0305-750X(00)00081-4www.elsevier.com/locate/worlddev

* I would like to acknowledge the support provided by

NORAD in carrying out this evaluation and to the sta�

of SEDP for their co-operation in the ®eld. Final

revision accepted: 8 July 2000.

63

particularly married women, exercised little orno control over their loans by this criteria.Interpreting this as evidence of widespreadloss of control by women over their loans tomen, Goetz and Sen Gupta go on to suggestthree possible repayment scenarios, all withnegative implications for women.

In the ®rst, the male family member using theloan takes responsibility for its repayment, asatisfactory outcome from the woman's pointof view but one which the authors believenegates the developmental objectives of lendingto women. In the second, men are unable tosupply the requisite repayment funds andwomen loanees have to substitute funds fromother sources, drawing on their savings, cuttingback on consumption, selling o� utensils andother assets. They have responsibility withoutcontrol. In the third, men are unwilling to repaythe loans, leading to an intensi®cation oftensions within the household, often spillingover into violence. In addition, violence againstwomen was also exacerbated by the frustrationof husbands at the wives' delay or failure inaccessing credit. Facilities to enhance women'saccess to the market is put forward by theauthors `às the single most e�ective way ofenhancing their control over loans, as well astheir public presence and their self-con®dence''(p. 59). The provision of transportationrecommended to take women to the marketplace along with security measures to protectthem against the possibility of male resistanceto their presence in the market place arerecommended as supportive measures.

In her study, Ackerly (1995) noted thatunderpinning most credit interventions inBangladesh was an implicit model of theempowered woman:

Empowered, the borrower wisely invests money in asuccessful enterprise, her husband stops beating her,she sends her children to school, she improves thehealth and nutrition of her family, and she partici-pates in major family decisions (p. 56).

Rather than seeking to measure theseoutcomes directly, she takes `àccountingknowledge'' as her indicator of the likelihoodof these and other transformative outcomesoccurring. Women who were able to report onthe input costs for loan-funded enterprise, itsproduct yield and its pro®tability, were countedas empowered. She found that membership ofsome credit organizations was more likely thanothers to contribute to the likelihood of

women's empowerment by this criteria. Womenwho provided labor to loan-assisted enterprise,sold their own products, or kept their ownaccounts were also likely to be empowered. Shetoo concluded that women's access to themarket was the primary route for theirempowermentÐ``knowledge and empower-ment come through market access'' (p. 64)Ðand warned against the likelihood of overwork,fatigue and malnutrition were loans used topromote women's labor involvement withoutalso promoting their market access.

Our third example of a negative evaluation ofthe impact of credit for women's empowermentis by Montgomery, Bhattacharya, and Hulme(1996). Although the evaluation of theempowerment impact occupies only a smallsection of their more general evaluation ofcredit programs for the poor, I have included ithere because it exempli®es a particular under-standing of households and gender relationswithin the literature on Bangladesh. Accordingto their ®ndings, only 9% of ®rst-time femaleborrowers were primary managers of loan-funded activities while 87% described their rolein terms of ``family partnerships.'' By contrast,33% of ®rst-time male borrowers had soleauthority over the loan-assisted activity while56% described it as a family partnership. Theyalso found that access to loans did little tochange the management of cash within thehousehold for either female or male loanees.Interpreting reports of ``joint'' management asdisguised male dominance in decision-making,the authors concluded that access to loans haddone little to empower women. Its main e�ecthad been to increase the social status of loan-receiving women vis-�a-vis less well-o� womenrather than vis-�a-vis men within their householdor the wider community.

(b) Does access to credit `èmpower'' women?the positive verdict

In contrast to this set of evaluations areothers which paint a far more positive pictureof the impact of these same credit programs onwomen's lives. Rahman (1986) found that thatloanee households in general, regardless of thegender of the loanee, had higher income andconsumption standards than equivalent non-loanee households. Although loans to womenwere more likely to bene®t male consumptionstandards than male loans were to bene®tfemale consumption standards, women loaneesnevertheless did bene®t from their direct access

WORLD DEVELOPMENT64

to loans. Women who made active use of atleast some of their loans had higher consump-tion standards and were more likely to have arole in household decision-making, either ontheir own or jointly with their husbands, than``passive'' female loanees, and both in turn hadconsiderably higher consumption standardsand were more likely to participate in house-hold decision-making than women from maleloanee households or from households who hadnot received any credit.

A study by Pitt and Khandker (1995)explored the impact of male and female mem-bership of credit programs on a number ofdecision-making outcomes in order to establishthe extent to which they were di�erentiated bythe gender of the loanee. The outcomes inclu-ded the value of women's nonland assets, thetotal hours worked per month for cash incomeby men and women within the household, fer-tility levels, the education of children as well astotal consumption expenditure. These authorsalso concluded that households receiving loanswere largely better-o� than those not receivingloans. In addition, the ®ndings that the genderof the loanee did in¯uence the pattern ofhousehold decision-making outcomes was inter-preted as evidence that women's preferencescarried greater weight in determining decision-making outcomes in households where theyhad received a loan compared to householdswhere either men received the loans or inhouseholds where no loans had been received.

A third example of a ``positive'' verdict is byHashemi, Schuler, and Riley (1996). Theyexplored the impact of credit on a number ofindicators of empowerment: (i) the reportedmagnitude of women's economic contribution;(ii) their mobility in the public domain; (iii)their ability to make large and small purchases;(iv) their ownership of productive assets,including house or homestead land and cashsavings; (v) involvement in major decision-making, such as purchasing land, rickshaw orlivestock for income earning purposes; (vi)freedom from family domination, including theability to make choices concerning how theirmoney was used, the ability to visit their natalhome when desired and a say in decisionsrelating to the sale of their jewellery or land orto taking up outside work; (vii) politicalawareness such as knowledge of key nationaland political ®gures and the law on inheritanceand participation in political action of variouskinds; and ®nally, (viii) a composite of all theseindicators.

They found that women's access to creditwas a signi®cant determinant of the magnitudeof economic contributions reported by women;of the likelihood of an increase in asset holdingsin their own names; of an increase in theirexercise of purchasing power; of their politicaland legal awareness as well as of the value ofthe composite empowerment index. In addi-tion, BRAC loanees tended to report signi®-cantly higher levels of mobility and higherlevels of political participation while Grameenmembers reported higher involvement in``major decision-making.'' When women'seconomic contribution was used as an inde-pendent variable, the e�ect of access to crediton the empowerment indicators was reducedbut remained signi®cant, suggesting that oneimportant route through which women's accessto credit translated into ``empowerment'' wasvia their enhanced contribution to familyincome.

The study also found that access to creditappeared to be associated with an overallreduction of the incidence of violence againstwomen. Regression analysis suggested thatolder women, women who had sons andwomen with education were less likely to havebeen beaten in the past year (Schuler, Hashemi,Riley, & Akhter, 1996). These ®ndings areconsistent with the lower status of young wiveswho are relatively new in the husband's home,with the prevailing culture of son preferenceand with the greater agency attributed towomen with education (see, for instance, Dreze& Sen, 1995). In addition, they found thatmembership of a credit program was associatedwith a statistically signi®cant reduction in vio-lence, but that the magnitude of women'seconomic contribution did not have anysigni®cant e�ect. They concluded that it waswomen's participation in the expanded set ofsocial relationships embodied in membership ofcredit organizations rather than increases intheir productivity per se which explainedreductions in domestic violence.

(c) Explaining the con¯icts: methods, questions,interpretations and models

These con¯icting conclusions about the``empowerment'' potential of credit for womenare both apparent and real. What appear to becontradictory ®ndings concerning, for instance,the extent to which credit exacerbates or lessensviolence against women, enables or fails toenable them to acquire independent assets, is

CONFLICTS OVER CREDIT 65

associated with an increase or decrease in theirliving standards is partly a di�erence in meth-odology. It re¯ects the fact that some studiesrelied largely on statistical data and signi®cancetests for their ®ndings while others relied onmore qualitative, sometimes anecdotal, evi-dence. Consequently, some of the di�erences in®ndings relate to di�erences in the incidence ofempirical outcomes, some ®ndings referringto `àverage'' and others to ``nonaverage''outcomes. Thus Hashemi et al's ®nding thatwomen's access to credit was associated with anoverall reduction in the incidence of domesticviolence is perfectly compatible with the ®ndingthat it exacerbated violence in a number ofindividual households reported both by them(see Schuler et al., 1996) as well as by Goetzand Sen Gupta.

Con¯icting conclusions about the impact ofcredit also re¯ect di�erences in the questionsasked by di�erent evaluations. By and large, thenegative evaluations focused on processes ofloan use while the positive ones focused onoutcomes associated with, and attributed to,access to loans. The validity of both sets ofmeasures depends on their conceptual clarityand on the validity of their underlying premises.There are, for instance, reasons to questionwhether some of the process-based measures doindeed measure what they are intended tomeasure. In conceptualizing the process bywhich women's access to resources translatesinto impact, Pahl (1989) had made an analyti-cally important distinction between what shecalls ``control,'' the ability to make policydecisions concerning the allocation of resources,``management'' which relates to decisions to dowith the implementation of policy and ``bud-geting'' which merely involves keeping track ofincome and expenditure. Ackerly's measure ofempowerment is ambiguous because it does notdistinguish between women who acquired their`àccounting knowledge'' through an activeinvolvement in the control and management oftheir loans, in the way that she appears toassume, or merely through a budget-keepingrole of the kind pointed to by Pahl.

Goetz and Sen Gupta's index of managerialcontrol is similarly ambiguous. It essentiallycon¯ates ``control'' and ``management,''making no distinction between decisions aboutloan use and decisions related to implementa-tion. But policy decisions about how loans areto be used are separate from, and indeed priorto, decisions relating to the management of theenterprise to which the loan is assigned. Since

the authors o�er no information on the deci-sion-making processes by which the loans wereallocated, we have no way of knowing theextent to which the observed allocationsre¯ected a sound economic calculus on the partof women, the speci®c individual circumstancesof their household or the blatant exercise ofmale power. Indeed, it is in principle possible(though in practice unlikely) that, with theexception of the unknown number of the 22%of women in their ``no control'' category whodid not even know how their loans were used,the remainder (at least 78% of their sample)participated fully in decisions about loan use.

There is also a need to be cautious about thecausality implicit in process-based indicators.The possession of accounting knowledge orexercise of managerial control do not, on theirown, su�ce as evidence of empowerment. Tobe persuasive as such evidence, we would needto know more about their relationship to othervalued achievements, perhaps of the kindoutlined in Ackerly's description of the ideal-typical `èmpowered woman.'' Indeed, theassumption that managerial control over loanuse is a necessary condition for women to beempowered by their access to loans is explicitlyrejected by Hashemi et al. In their study, theyclassi®ed all the women loanees in their sampleaccording to the ``control'' categories devel-oped by Goetz and Sen Gupta and con®rmedthat large percentages of women loanees intheir sample had indeed ``lost control'' overtheir loans by this criteria. This did not,however, prevent a signi®cant proportion ofthem from achieving a range of other valuedimpacts, although, as we noted, the likelihoodof these positive impacts was strengthened ifwomen used at least part of their loans toincrease the value of their own economiccontributions.

As far as outcome indicators are concerned,their validity depends on how well they capturechanges in the structures of gender inequalitywithin the household and community, notmerely on how well they capture changes inhousehold living standards or even in children'swelfare. One of the strengths of the study byHashemi et al. is that their indicators meet thiscriteria. They can all be seen as valuedoutcomes in their own right as well as beinglinked to the structures of constraint which giverise to gender inequality in Bangladesh. Bycontrast, the study by Pitt and Khandker isundermined by the absence of any obviousrationale for the particular decision-making

WORLD DEVELOPMENT66

outcomes selected for their study. Their ®nd-ings are consequently not always easy tointerpret. The only outcomes with relativelyunambiguous theoretical links with women'sempowerment are (i) women's ownership ofnonland assets, an increase in which could beinterpreted as a strengthening of their fall backposition, and (ii) the gender gap in education, areduction in which could be seen as addressinga longstanding gender inequality in the valuegiven to children. Their own interpretations ofsome of their ®ndings tend to be somewhat adhoc and open to other equally plausible andvery di�ering interpretations.

This takes us to yet another factor behind thecon¯icting conclusions we have been discussingwhich is the proclivity to ``read'' empirical®ndings in the light of preconceived notionsabout loan impact so that the same ®ndings aregiven extremely contradictory interpretations.Thus, Pitt and Khandker take their ®nding thatloans to women led to an increase in theirmarket-oriented work to indicate an empow-erment e�ect. By contrast, all three negativeevaluations warn against the intensi®cation ofwomen's workloads and fatigue. Pitt andKhandker interpret the higher level of house-hold consumption expenditure associated withloans to women as evidence of the greaterweight given to women's preferences in house-hold decision-making; Montgomery et al.suggest that such ®ndings demonstrate thatloans to women are `heavily compromized bythe persisting responsibilities of women tocover the consumption needs of the family'(Montgomery et al., 1996, p. 168). Similarly,the increase in women's welfare levels as aresult of their access to credit is linked to theirenhanced role in household decision-making byRahman, but given a much more passiveinterpretation by Goetz and Sen Gupta whosuggest that women give up their loans to men``in exchange for the right to have greaterexpenditures on their own or their children'sclothing and health.''

In short, there are di�ering judgementsembodied in these evaluations as to whatkinds of changes constitute evidence ofempowerment, di�erences which in turnre¯ect the di�ering models of households, andthe power relations within them, which theseevaluations draw on. While both positive andnegative evaluations accept the premise ofgender inequality in intrahousehold relations,they vary considerably in the signi®cance andmeaning attached to cooperation and con¯ict

between men and women within the house-hold and consequently to autonomy, depen-dence and interdependence within thehousehold.

By and large, the negative evaluations tend tobe negative because they stress gender antago-nism within the household and discount thesigni®cance of cooperation. Thus, for Mont-gomery et al., reports of ``jointness'' in themanagement of household enterprise andincome are merely examples of disguised maledominance; only the exercise of autonomousfemale authority is counted as evidence ofempowerment. Goetz and Sen Gupta's discus-sion of the circumstances under which theinvestment of women's loans in the purchase ofa rickshaw would, and would not, constituteexercise of ``control'' also reveals this individ-ualized notion of empowerment. Rickshawpulling in Bangladesh is a purely male activityso that the purchase of a rickshaw, an extre-mely common use of loans to poor women,represents investment in an activity to whichwomen are unable to contribute any labor.While such women would automatically beclassi®ed in the ``little'' or ``no'' control cate-gory by Goetz and Sen Gupta's criteria, theysuggest that a woman could still be classi®ed asexercising ``signi®cant control'' if the rickshawwas licensed in her name and if she establisheda contractual rental relationship with the rick-shaw puller. In the context of rural Bangladesh,however, this would constitute extremelyanomalous behavior on the part of a womanwho had an unemployed son or husband in thefamily who was able and willing to pull therickshaw and to take responsibility for loanrepayment.

The more positive evaluations, by contrast,are positive partly because they do not privilegeindividualized over joint forms of behavior. Pittand Khandker attempt to capture possibleincreases in the weight given to women's pref-erences in a series of household decisionsfollowing their access to loans, but do not ruleout joint decision-making. Both Hashemi et al.,as well as Rahman explicitly incorporate some``jointness'' on interests within the householdinto their indicators of empowerment. In the®nal analysis, the plausibility of one or other setof conclusions about the transformatoryimpact of credit for women will rest on thecredence attached to the models of powerwhich inform the analysis.

Despite their di�erences, however, both setsof evaluations share in common an absence of


testimonies by women loanees themselves as tothe impact of credit on their lives. Obviously, inthe context of evaluation studies where valuedresources are at stake, personal testimonies onimpact have to be interpreted with caution,given that there may be a strong incentiveamong bene®ciaries to present impact in apositive light. At the same time, participatoryimpact assessments can help to enrich academictheorisations of gender subordination byproviding important insights into inequality asa ``lived experience.'' In the rest of the paper, Iwant to report on the ®ndings of my ownevaluation of a rather di�erent credit programin Bangladesh in which I sought out the testi-monies of 50 female and 20 male loanees as tothe impact of loans on their lives. In addition, Ialso carried a quantitative survey of 700households to provide basic descriptive statis-tics on the loanees, their households, theirpatterns of loan utilization as well as on someof the impacts identi®ed in the evaluationliterature. I will be drawing on the loanees'testimonies as a di�erent vantage point fromwhich to contextualize and assess the ®ndingsof the various evaluations discussed here as wellas to consider what the perspectives of womenloanees themselves can add to our under-standing of the transformatory potential ofcredit targeted at women. I will be concludingwith some general conceptual and method-ological comments on the evaluation ofwomen's empowerment.

2. THE SMALL ENTERPRISEDEVELOPMENT PROJECT: LENDING TO

THE NOT-SO-POOR

The SEDP was started in 1990 in Faridpurdistrict in Bangladesh and extended in 1992 toMymmensingh. It acts as an intermediarybetween eligible loanees and a special NORAD-funded credit line which is managed by one ofthe country's nationalized banks. Its primarygoal is the promotion of small-scale labor-intensive enterprises, including women's enter-prises, in order to enhance income and expandemployment. To be eligible, potential borrow-ers must have at least half an acre of land andsome prior entrepreneurial competence. Theyare identi®ed by project sta� and attend a three-day training in basic entrepreneurial skills andsocial development issues during which they areassessed for their entrepreneurial potential.Loans range from 5000/- takas to 500,000/-.

Interest rates are subsidized and vary between10% and 14% according to loan size, whilerepayment is generally on a monthly basis.Repayment rates are high at over 90%.

The household survey of male and femaleloanees found that while male loanees in bothdistricts tended to be better o� in terms of landowned and cultivated and education level ofloanee, female loanees in Faridpur district weremuch better o� than those in Mymmensingh.Female loanees in Mymmensingh came fromthe poorest households in the survey sample;they were also more likely to be female-headedthan women loanees in Faridpur (14%compared to 7%). They were also given muchsmaller loans than those in Faridpur andreported correspondingly small rates of returnto loan investments. There was nothing in theSEDP rule book to explain this pattern. Itappeared to re¯ect di�erences in managementorientation in the two districts.

The SEDP thus di�ered from the main pov-erty-oriented programs in a number of signi®-cant ways which are summarized in Table 1.While these programmatic di�erences meanthat the ®ndings from my study cannot not bedirectly compared to those discussed earlier,they nevertheless provide a useful basis fordistinguishing between impacts which appearedto be associated with women's access to creditper se, regardless of delivery characteristics, andthose impacts which were clearly associatedwith particular kinds of program delivery. Inaddition, a tentative degree of direct compari-son was possible because of the pervasivenessof the poverty-oriented credit organizationsand inevitable contact with their operations inthe course of the ®eld work. Where SEDPloanees in my qualitative sample or a memberof their family had themselves borrowed fromone of these organizations, interviews wereextended to cover this experience.

(a) Female mobility and social status: thecontradictions of class and gender

In terms of impact, the central overall ques-tion framing the qualitative component of theevaluation was ``What di�erence did the loansmake to the women's lives?'' What changes didit bring about and how did women assess thesechanges? Let me start out by noting that, as inmost of the evaluations discussed earlier, therewas little evidence of any radical change in thegender division of labor as a result of women'saccess to loans. Access had increased their levels

WORLD DEVELOPMENT68

of economic activity, but not the range. Thehousehold survey showed that womenremained con®ned to a small number of``female'' occupations, with livestock rearingpredominating in Faridpur and paddy huskingin Mymmensingh. Production of cane andbamboo goods was second in importance inboth districts. In addition, a few women inves-ted in poultry raising, home-based tailoring anditinerant hawking. Male occupations were moreevenly distributed and over a wider range ofactivities: ``shops'' of various kinds, engineeringand other workshops of various kinds, seednurseries, managing power tillers and rice millsas well as farming and livestock rearing.

This gender patterning of the occupationalstructure suggests that adherence to purdahnorms continues to constrain women's publicmobility, limiting their choice of enterprise andtheir ability to carry out transactions in themarket place. Given that the resilience ofpurdah featured so centrally in some of theevaluations, it is worth analyzing what womenloanees themselves had to say on this question.Two key insights emerged out of their testi-monies. First of all, notions of purdah wereclosely interwoven with local understandings ofclass, social status and gender propriety so thatbehavior expressing gender norms were oftensimultaneously expressive of class hierarchyand social standing within the community.Conformity with purdah often featured inwomen's testimonies in terms of a voluntaryadherence to status norms rather than as adirect manifestation of male control, as is evi-dent from the following testimonies:

I do all my work within the house, it is not a matter offear, it is a matter of izzat (honour). Women who caneat by staying within the home are given greater value.Everyone gives value to women who work within thehome, people outside, as well as those in the family.Men work outside, and women inside. Otherwisewhy have men been made, you could have had only

women. I will go without food, but I will not go with-out izzat (F23).

It is alright moving around within the neighbourhood,but I have no time to go to the bazaar. Anyway, I ama woman, it is not possible for me to go to the bazaar.Some women go, those without husbands. But I havea husband and a son, I don't go. It is a matter of man-shonman (honour). People in the neighbourhood willsay, she has a husband, she has a son, how can she goto the bazaar? (F20)

The second point to come out of the women'stestimonies was that the distinction between``public'' and ``private'' space was not repre-sented as a simple dichotomy but rather as acontinuum of locations in the public domain,ranging from acceptable to unacceptable placesfor women to be seen. Many of the women inmy sample moved around freely within theirneighborhoods, were prepared to go into thedistrict headquarters to attend the initialtraining and subsequently to the local SEDPo�ce to deposit their monthly repayments.Rural markets, on the other hand, the weeklyhaat and the permanent bazaar, were located atthe other, unacceptable end of the spectrum.

Because the need to adhere to purdah was notequally subscribed to by all women, or by allclass groups, and because the decision to adhereto purdah did not impose the same costs, therelationship between women's presence in publicactivities and their empowerment was not astraightforward one. In this connection, we candistinguish between a number of di�erent cate-gories of women. There were those from better-o� households for whom there was a conver-gence between the economic logic of earning alivelihood and the social logic of maintainingtheir honor. They owned homestead land andother facilitating assets so that returns tohome-based work exceeded the returns to mostforms of waged employment available to them.The majority of these women had never sought,

Table 1. Di�erences in goals and organizational practice between SEDP and typical poverty-oriented lending inBangladesh

Characteristics SEDP Poverty-oriented lending

Goal Small enterprise development Poverty alleviationRole Intermediary Direct lendingLoan size (takas) 5000±500,000 1000±5000Class eligibility Own more than 50 decimals of land Landless or less than 50 decimals of landOther criteria Prior entrepreneurial experience No experience requiredGender Men and women Predominantly womenInterest rates Subsidized (10±14% annually) Nonsubsidized (18%)Repayments Monthly WeeklyEmphasis on Individual lending Group-based lending


or been forced to seek, outside employment. F23cited above was one such woman.

It was a di�erent story for women frompoorer households who, nevertheless, had somesocial standing within the community. Purdahnorms also constrained their mobility even if itcarried a high economic cost. Prior to accessingloans, these women had either starved invisiblyat home, or opted for badly paid anddemeaning domestic labor within the shelter ofother people's homes ``where nobody would seeus.'' Access to credit was a godsend for thisgroup because it allowed them to feed them-selves and their families without the humilia-tion associated with menial domestic labor inother people's homes.

The poorest women in the sample were leastlikely to have paid attention to notions ofpropriety. There were a number of them whohad been working the public domain prior toaccessing loans, some as agricultural wagelaborers in the ®elds, others as petty traders inlocal market or by the roadside. Even amongthis group, however, the prosperity which wentwith their loans often led to their withdrawalfrom public forms of activities. In some cases,the decision was on normative grounds. F25'stestimony points to the role played bycommunity opinion in her decision to withdrawfrom public transactions:

So many people say to my husband, ``Your wife goesoutside the house, she goes to the ®eld, she has gonebad''.. . . I survive by my own e�ort, I do not borrowfrom my neighbours any more, nor do I lend . . . Thereis no dishonour in work. But I don't sell milk in themarket anymore. My value has gone up from before,I feel ashamed, people say, she has improved so much,how can she still go and do this work?

More often, the decision to withdraw re¯ec-ted the conditions which prevailed in the femalesegments of the labor market. For instance,F33's testimony illustrates why agriculturalwaged labor in the public domain was unlikelyto be experienced as particularly empoweringby most women:

Before the loans, women used to work on other peo-ple's ®elds, cutting lentils, rice, wheat. They got 20/-to 30/- takas a day. That is happening less nowbecause so many women are getting loans, they areraising cows, goats, they can work for themselves sowhy should they work for someone else. If you canwork for yourself, well, look, I am sitting here withyou, could I do that if I worked for someone else?They would pay me less. I would pull up lentils, theywould give me 20/- a day, this was four years ago.

Before women used to clear the irri blocks, they wouldstand in the water and get leeches on them. Now theydon't. Now, with the loans, they have some peace.

By and large, women who remained inoutside forms of employment were femalehousehold heads, who often had little choice inthe matter, and a number of poorer womenwho had been itinerant traders before theiraccess to loans.

In contrast to this general picture, however, itshould be noted that there were a number ofwomen in the sample who gave a positive valueto their increased ability to move more freely inthe public domain, associating it with theacquisition of ``courage'' rather than as asource of shame. These women attributed theirnewly found self-con®dence in dealing withlocal elites, with the police and with others whohad previously intimidated them to theirinteractions with SEDP sta� rather than tosimply their access to credit per se. F29, whocame from the poorer end of the economicspectrum, valued the fact that her access toloans had allowed her to move from selling afew vegetables under a tree outside the villagebazaar to establishing her own permanent shopwithin it. She was the only woman in thatbazaar, but was now such a familiar sight thatshe no longer aroused any comment. F50, whocame from a poor, but status-conscious,household and had previously gone hungry athome rather than compromise her family'ssocial standing, also valued her new mobility:

By joining these samities, many women have got thecourage of men. Women now have the same rightsas men. If a man can go and cut earth, go to thehaat-bazaar, to the towns, why can't women? I cango everywhere now, even to the haat. If my husbandis not at home, if he has gone to the market, I willgo to the laborers' house to fetch them. If I neededto go to the bazaar and my husband was not at home,I would go.

Our analysis thus highlights the ambiguitiesassociated with the use of increased physicalmobility, particularly in relation to the marketplace, as an indicator of empowerment in thecontext of rural Bangladesh. On one hand, aslong as women adhere to norms of purdah anddo not participate signi®cantly in markettransactions, they will remain dependent onmale household members to undertake suchactivities on their behalf and to that extent theireconomic agency will be restricted. On theother hand, if empowerment entails the

WORLD DEVELOPMENT70

expanded capacity for making choices, fortaking actions which express own values andpriorities, then it has to be recognized that thesevalues and priorities are likely to be shaped bythe values and priorities of the wider commu-nity (Kabeer, 1999). The paradox is that inmany cases, this leads women to opt for someform of purdah if they can a�ord to, both tosignal their social standing within the commu-nity and to di�erentiate themselves from thosewomen who do not have this choice.

(b) Enhancing self-worth and perceivedeconomic contribution

If there had been no radical change in thegender division of labor of the kind consideredby many to be a necessary precondition towomen's empowerment, what kinds of changesdid occur as a result of women's access toloans? One important change that featured inmany of the women's testimonies related totheir sense of self-worth, of bringing somethingof value to their households. The signi®cance ofthis has to be understood in the context of theincreasing monetization of the Bangladesheconomy and the gap that it has opened upbetween women and men in terms of accessingnew opportunities. Men have been privilegedby their gender, class and education in gainingsuch access while the resilience of purdahnorms have kept women largely con®ned to theprecincts of their homesteads, dependent onmale members of their family for economicprovision and social protection.

Most studies on gender relations in Bangla-desh have pointed to women's status asdependants, but few have explored what suchdependency might mean as a ``lived'' experi-ence. The testimonies of women loanees madeit clear that many found the position ofsupplicant within the family galling andhumiliating, particularly those who were forcedto literally plead for money to meet theireveryday needs. This was vividly illustrated inF15's comments:

If I had not gone to that SEDP meeting, had not takena loan, had not learnt the work, I would not get thevalue I have, I would have to continue to ask my hus-band for every taka I needed. Once I had a headache,I wanted one taka for a bandage to tie around myhead, I wept for eight days, he still would not giveme the money. Just one taka.

Testimonies such as this help us to appreciatethe importance that women like F15 attached

to their new identities as bearers of valuedeconomic resources. Nor was it a case of purelypassive access to such resources. According tothe household survey, the majority of thewomen in the sample used at least part of theirloans to enhance their own productivity. Thosewho had not been economically active previ-ously were able to start up new activities.Others were able to put pre-existing enterpriseson a more secure basis and yet others were ableto move into their own home-based enterprisesrather than working for others in forms ofwork they considered demeaning.

Consequently, while most women experi-enced an increase in their workloads, they didnot give it the negative interpretations sugges-ted by some of the evaluations discussed earlier.The distinction that they made between paidand unpaid work helps to explain why. It wasnot that these women were idle prior to theiraccess to credit. Most were involved indomestic chores as well as in expenditure-sav-ing work, but such activities, as we well know,were generally unremunerated and receivedlittle recognition within society or within thehome. It was evident from the women'saccounts that they too shared the low socialvalue given to these activities. The new uses oftheir time made possible by their loans broughtabout an enhanced sense of self-worth as wellas giving a new meaning to ``work.'' As F43 putit: Ideas of the mind is everything. If you havemoney in your hand, you feel joy. If you haveno money, you feel pain. My labour hasincreased, but I don't feel it because the moneyis also coming in. It doesn't feel like hard work.

Nor was it only in relation to their ownactivities that women reported a sense ofachievement. Their testimonies also highlightedthe value they attached to the well-being anddignity of the work engaged in by otherhousehold members. There is little spaceallowed for such impacts, and their possibleimplications for gender relations, by models ofthe household which conceptualize it incon¯ictual terms and fail to recognize thepotential for solidarity between householdmembers. Yet many of the women I inter-viewed pointed to the release of male householdmembers from demeaning economic relation-ships as one of the valued achievements whichthey associated with their loans. F27 used herloan to mortgage in land for her husband tocultivate because, as she said, ``How long washe going to give labour on other people'sland?'' For other women, their husband's


dependence on moneylenders or wealthy land-lords for credit, usually at extortionate rates,had been the most humiliating aspect of theirpre-loan experience: My husband now worksalongside me. He no longer has to hear harshwords if he does not pay his debt on time (F43).

Greater social inclusion was another impactwhich was highlighted in the testimonies ofpoorer loanees, male as well as female. Theyspoke bitterly of how it had felt to be outsidethe orbit of community life, to be excludedfrom its social events and from everyday formsof hospitality. As a result of their loans-relatedprosperity, they had acquired a new respect inthe eyes of those who had previously despisedthem and position of strength from which todeal with them:

Before I had to sit under a tree and sell my goods, peo-ple would make comments about me, I could saynothing. Now since the loan, they don't know whatto say, they are nervous to say anything. After all, Ihaven't brought a loan just once, I have broughttwice, thrice, four times. Now even if people want tosay anything, they don't have the courage. Thosewho never acknowledged me now invite me, I havemoney, they might need to borrow. Before theylooked down on me, never came to my house, I waspoor, I could not feed them. And now even in houseswhere I do not expect to be invited, I am asked. (F29)

Have things improved for us? Listen, when you haveno money, there is nobody, but when you havemoney, you suddenly have so many friends andacquaintances. Money is all. All that time, when wehad no food, nothing to eat, no one wanted to giveus anything. And now, day and night, from house tohouse, it is ``have a betel leaf, tobacco leaf, cigarettes,chair, chowki . . ..'' (F37)

While women's own sense of self-worth wasenhanced by these various achievements, soapparently was their worth in the eyes of otherfamily members. This was evident in themarked improvement in the quality of familyrelationships that many reported, particularlyin the context of marriage. As primary, oftensole, breadwinners for their families, men inpoorer households experienced many di�cul-ties in making ends meet. The women I inter-viewed were well aware of the stress andfrustrations involved in this responsibility andthe extent to which their own dependencycontributed to it. Access to loans helped toreduce the burden for men since women werenow able to share some of the responsibility ofproviding for the family. The result was areduction in levels of tension and con¯ict andgreater a�ection from their husbands:

Before we had scarcity. Suppose we needed 5 seers ofrice, and he brought home 4 seers, we would be shortof food,. the children's stomach would not be ®lled,they would cry and he would know why they were cry-ing. I would keep it quiet, but the children wouldsometimes let it out. Now we sit down to eat together,those tears are gone. We eat properly, systematicallynow and there is no worry about food. He no longerhas to worry about whether we have eaten or not.When he couldn't give the money, there would bewords, I would say angry things to him, he would re-spond angrily: ``I don't have it, how can I give it.''Now we don't have those words. (F24)

The e�ects of women's enhanced economicvalue were particularly marked in householdswhere con¯ict between husband and wife haddeteriorated into violence. The question ofdomestic violence cropped up su�cientlyfrequently in the interviews to suggest both thatit had been a signi®cant problem in the pastand that at least those forms of violence whichstemmed from scarcity-related frustrations hadbeen reduced in the wake of women's access toSEDP loans. The link between credit andreduced violence was made directly by anumber of women, including F25:

My husband did not have clean clothes before, now hehas, and they know it is because of me. My husbandacknowledges this. He does not raise his hand to meany more. Before he used to hit me. What could onedo if one's husband hits one . . . ? In a house of scar-city, there is more kalankini. If he brought home fourannas, and I could not buy enough rice, he abused me.The house where there is no scarcity, there is no abuse.Because of this scarcity, this poverty, the lives of thepoor are so troubled.

(c) Gender, voice and decision-making

Another impact which featured widely inwomen's testimonies, and one which derived tosome extent from the impacts discussed in thepreceding section, related to their increasedvoice in household decision-making. The studyprovided both quantitative as well as qualita-tive evidence on this. The household survey hadincluded separate questions on loanees' roles indecisions related to loan use and loan-fundedactivities, in order to capture the distinctionbetween management and control noted earlier.As a result of the qualitative component of the®eld work in Mymmensingh, an additionalquestion relating to decisions relating to theallocation of loan-related pro®ts was includedin the Faridpur questionnaire. The results,

WORLD DEVELOPMENT72

reported in Table 2, make it clear that access tocredit had not obliterated gender asymmetriesin decision-making. Male loanees were muchmore likely to be primary decision-makers inrelation to loan use, enterprise management aswell as the allocation of pro®ts than werefemale loanees who were more likely to reportjoint decision-making. On the other hand,women's access to credit does appear to havemitigated some of the gender asymmetries indecision-making. In male loanee households,the percentage of women having some sort ofsay in decision-making did not exceed 20%whereas in female loanee households, the ®gurevaried between 40%, if we look at the exerciseof primary decision-making role, and 90%, ifwe also include joint decision-making. Thequalitative interviews with the women loaneessuggested some of the reasons for theirmarginal, joint and primary decision-makingroles.

The 10% of female loanees who played verylittle role in decision-making were made up ofthree subgroups. In some cases, their lack ofvoice re¯ected the straightforward appropria-tion of their loans by husbands. These womenhad not usually played a particularly active rolein the decision to seek out loans and maleappropriation was merely a further manifesta-tion of a pre-existing marginalization within thehousehold which the access to loans had donelittle to alter. Women who were either ill or hadsome disability also did not play much of a rolein decision-making. Finally there were women

who had conceded control over their loans tomale household heads in recognition of theirresponsibility for the collective welfare of thehousehold. As F2 pointed out, ``What need do Ihave to take decisions? Even if I die, myhusband will continue to take responsibility formy children. . .. I keep the money, but it is hisresponsibility to spend it so it does not stay toolong with me.''

Around 40±50% female loanees in the samplereported joint decision-making. In some cases,this re¯ected a taken-for-granted ``jointness'' ofhousehold interests and they saw it as irrelevantthat the loan had been granted in their name:As F40 put it, ``I may have brought in the loan,but I did it with my husband.'' For others,``jointness'' re¯ected their awareness of theirreliance on male family members to carry outcertain stages of production and hence their toneed to ensure male cooperation. F40 o�eredthe following explanation of how ``separate''and ``joint'' areas of decision-making weredetermined within her household:

We had cows and calves before, but they were myhusband's. . . I have bought cocks and ducks and goatswith my second loan and with the third, I bought acow and also gave my husband some money for hiswood business . . .. My husband takes decisions todo with looking after the cow, but with the goatsand poultry, I decide. You see, the cow has to be takenout in the morning and brought back in the evening,and if some man comes to buy the milk, well, I ama woman, I can't go in front of him, my husbandhas to do the talking and running around. He has a

Table 2. Decision-making in loan use, enterprise management and loan income by gender (percentages of loanees)

Use of loan Running of business Use of pro®t

Male Female Male Female Male Female

Faridpur: ®rst loansSelf 81 47 81 51 79 32Others 11 8 11 12 11 5Joint 8 45 8 37 10 63

Faridpur: second loansSelf 84 41 85 45 83 31Others 14 9 13 15 11 4Joint 3 50 2 40 6 65

Mymmensingh: ®rst loansSelf 88 36 85 34Others 6 18 7 25Joint 5 45 7 40

Mymmensingh: second loansSelf 86 36 83 35Others 5 20 8 27Joint 8 43 8 38


role in it. I may get my husband to take my goats tothe bazaar for sale, but I make all the decisions aboutit.

Finally there were those who had beenpreviously been disenfranchised in householddecision-making processes. They associated thetransformation of their marginal role to one ofjoint decision-making to their access to creditand the resulting improvements in their earningcapacity: ``My say has increased now that Iknow how to earn. I did not used to say muchbefore but now I am malik (mistress) of myown shongshar (household economy).''

The third group of women, those whodescribed themselves as primary decision-mak-ers, were analytically the most interesting fromour point of view since they appeared to goagainst the cultural norm. Here again, anumber of di�erent factors were at play. The®rst and most predictable category in thisgroup were women heads of households whoseprimary decision-making role occurred bydefault. A second and less expected categorywere those who explained their key role inhousehold decision-making in terms of theirsuperior entrepreneurial competence, an opin-ion that was usually shared by other familymembers. F11 was an example of this category.She pointed out:

The money from our business stays with me. Whenmy husband needs it, he asks for it. He is not so goodwith accounts, so it all stays with me...I memorise theaccounts, I can't read or write. 7000/- worth of busi-ness is not so much that you need to write it down.. . .My husband knows whatever I do. He will never stopme from doing anything, whatever I say, he goes alongwith. I take all the business decisions. I keep all the hi-sab. If I tell him not to go to the bazaar today, he willnot go.

An important point to make about thewomen in this category is that while access toloans may have expanded their sphere of deci-sion-making, many of them were already exer-cising considerable voice within their ownhouseholds on the basis of recognized mana-gerial skills. This should not negate theimportance of credit in their lives since itallowed them to realize their hitherto suppres-sed entrepreneurial potential, but it does meanthat the extent of voice they exercized cannotbe attributed solely to their access to credit.

A ®nal category of women in the primarydecision-making category were those who hadextremely con¯ictual relationships with their

husbands. While violence within marriageappeared to be a fairly widespread phenome-non, a certain degree of empathy characterizedwomen's accounts of such violence when it wasseen as an outcome of household poverty, andof the struggles of the male breadwinner tomake ends meet. Exceptional violence, on theother hand, di�ered in that it was not explainedin terms of the shared su�ering of the poor, butin terms of the husband's character (abusiveand foul-tempered) and habits (alcohol anddrugs). There were three women out of the 50in our qualitative sample who reported beingmarried to such men. They had not left theirhusbands, but had e�ected a form of ``divorcewithin marriage,'' using their loans to create aparallel economy for themselves which gavethem considerable ®nancial independence oftheir husbands.

F48 had used her loans to set up her ownlivestock business and then to purchase arickshaw which she registered in her own namebut which her son pulled. She was on her fourthloan when we interviewed her. By this time, shewas managing the household budget, her rela-tionship with her husband had improved andshe had used her current loan to set him up inhis own transport business. Here was how shedescribed the changes in her relationship withher husband:

My husband is working well now, he gives his earn-ings to me. Before he did not used to give it to meregularly. Now he doesn't drink any more and hehas even reduced his biri smoking. I have cut it down,I have said I don't want to see you smoking, But hesteals a little money and still smokes a bit, he can'tdo without it. And he used to drink in the beginningbut not now . . .. And he has not raised his hand tome for the past 3 or 4 years. Not since the loan. Hisanger has subsided. Scarcity creates a lot of problems. . .. If womankind has no money in her hand, man-kind tolerates her less. When I had no money in myhand, he gave me no regard. Now he sees the womanhas money in her hands, so that now if anything hap-pens to me, it is his head that hurts. That is how itseems.

In F29's case, her husband's violence towardher had diminished when he had lost his legsbut not his abusiveness. While she used some ofher loan money to improve his business, herrelationship toward him remained antagonisticand it was in her son that she invested most ofher a�ection and e�orts. She related withsatisfaction her ability to ignore her husband'sabuse and go her own way, now that she waseconomically self-reliant:

WORLD DEVELOPMENT74

Now I am eating out of my own e�ort, I don't have togo to him for a single paisa. . . If before I said, we needmoney, he would get angry, now he can't. If he givesme money, then he gives it and if he doesn't give it,then he doesn't. It is all the same to me. My son isnow doing his training in Dhaka so I don't have tocook for him at present. If my husband gives memoney for the bazaar, I will cook for him, if hedoesn't, I won't.

As a footnote to this discussion, I shouldpoint out that the in-depth interviews with the20 male loanees con®rmed that women were farmore marginal to household decision-makingin their households and also helped to illumi-nate why. They reminded us that men did notrely on female family labor input in theirenterprises in the critical way that women reliedon men. In any case, men could take women'sassistance in certain activities associated withtheir enterprises for granted on the basis oftheir authority as household heads: cooking forextra workers; adding the ®nishing touches to agarment; assistance with pottery or weaving.Consequently, they had neither practical nornormative imperative to take steps to ensurewomen's cooperation. Male loanees saw them-selves as the primary and usually sole bread-winners of their families. Many were not onlyagainst their wives taking up income-earningactivities but had explicitly forbidden them totake out their own loans. By and large, thisgroup gave fairly unequivocal descriptions ofmale dominance within their households:

I take the decisions about the business, she does whatshe understands, she doesn't get involved in extra jha-mela (hassle). She has neither hisab or kitab (literacyor numeracy).. I take all the decisions around thehouse.

Once again, however, there were exceptionsin that a number of male loanees did reportmaking decisions jointly with their wives. Acommon factor appeared to be women's abilityin these cases to make a contribution beyondthat prescribed by the traditional gender divi-sion of labor. In some cases, women in maleloanee households had taken out loans of theirown, from organizations such as GrameenBank or BRAC so that we were observing thee�ects of their access to credit on householddecision-making. In others, it was their educa-tion which allowed them to assist their husbandin keeping household accounts, particularly incases where the husband was himself poorlyeducated.

(d) Transformatory investments: assets andeducation

The increase in women's voice in householddecision-making processes was important in itsown right for those who had been previouslydisenfranchised. It had an added signi®cance ina context where access to loans had led to anincrease in household income. While it proveddi�cult to calculate precise returns to loan-funded activities, 1 what can be said withcon®dence is that most loanees reported satis-factory levels of pro®tÐless than 5% reported alossÐand repayment was not generally regar-ded as a problem. The ®nding that access toloans generally increased levels of householdincome, and that women's access to creditenhanced their voice in household decision-making, also supports similar ®ndings reportedby some of the evaluations cited earlier, givingcredence to the idea that expenditure patternswithin the household were di�erentiated by thegender of the loanee. By and large, we foundthat male loanees were more likely than femaleto reinvest part of their pro®ts in their busi-nesses, that better-o� women loanees (inFaridpur) were more likely to invest in someform of savings and that poorer ones (fromMymmensingh) were most likely to spend theirpro®ts on purely consumption needs.

Here I would like to focus not so much ongender di�erences in allocational priorities perse but on those which had the potential foraddressing some of the inequalities whichunderpin women's subordinate status. My®ndings rea�rmed the ®nding, reported byboth Pitt and Khandker (1995) and Hashemi etal. (1996), that women's access to credit hadallowed a number of them to accumulate assetsof their own. Table 3 reports on savingspatterns of male and female loanees whileTable 4 looks at patterns of owned andpurchased homestead land. It will be seen that``secret'' savings, a longstanding practice bywhich rural women in Bangladesh ensuredsome degree of economic autonomy for them-selves, persisted among women loanees. It willalso be seen, however, that women loanees inboth districts, but particularly the better-o�ones from Faridpur, were also engaging in theless traditional practice of opening bankaccounts in their own names.

Table 4 o�ers both direct evidence thatwomen's exclusion, at least from homesteadland, was beginning to break down, and indi-rect evidence that women were using their loans


to purchase homestead land in their ownnames. Homestead land has a particularsigni®cance for women given the home-basednature of their enterprises. 2 Although womenare entitled to half their brothers share ofparental property under Islamic law, most havenot claimed land, waiving it voluntarily (as wellas involuntarily) in favor of their brothers so asnot to alienate their support should theirmarriages break down. Given the greater like-lihood of male inheritance of land, the greaterpercentage of male loanees in the sample whoreported that some or all of their homesteadland was registered in their names is to beexpected. But women loanees were as likely asmale loanees in each district to report thepurchase of some or all of their homestead landand, furthermore, 19% of those in Faridpur and29% of those in Mymmensingh reported thatsome or all of it was registered in their names.

There are understandable reasons whywomen have not generally sought to assertindividual property rights within the family inthe past. There is also a broad cultural associ-ation between men and property. Morever,given men's collective responsibility for house-hold welfare, including the welfare of women,investment in male assets can be seen as aninvestment in the most widespread form ofsocial capital available to women, their familialnetworks. In addition, certain assets such asrickshaws, irrigation pumps and tractors tendto be considered ``male'' because they arealmost invariably used by men. For a woman

to register such an asset in her name, particu-larly when she had a husband, meant goingagainst the grain of local notions of masculin-ity. As F13, who had purchased a rickshawwith her loan money, explained, He told me toput it in my name, but the thing is that therickshaw has a signboard with the owner'sname and if people see it has a woman's name,they will say, ``look at that fellow, he pulls arickshaw but it is in a woman's name.'' I foundthat shameful.

Whatever the rationale, it remains the casethat the investment of women's loans in assetsregistered in men's names did little to altercustomary gender asymmetries in the distribu-tion of resources. For the purposes of ouranalysis, therefore, the more interesting caseswere those where women's investment strate-gies did represent a departure from past prac-tice and consequently evidence of the exerciseof new kinds of choices. It was most often incon¯ictual marriages that women sought toseparate out their asset holdings. In such cases,the practice of clandestine savings became acovert declaration of independence rather thanan attempt to retain some control overpurchasing power. These savings were nolonger the traditional minuscule amounts, theresidual income after basic needs had been met,or the ®stful of rice accumulated painstakinglyon a daily basis, but substantial sums of cash.F29, for instance, kept her ®nances separatefrom her husband's and chose not to let himknow about her various savings:

Table 3. Saving patterns by gender (%)a

Faridpur Mymmensingh

Male Female Male Female

Cash at home 18 16 17 23Bank account 25 20 20 9Saving society 13 18 3 4Lent on ) 1 ) 6Secret saving ) 12 ) 6No saving reported 44 45 61 54

Table 4. Acquisition and registration of homestead land (%)


Acquisition Male Female Male Female

Inherited 67 61 54 56Purchased 29 34 34 31Both 2 2 9 10Registration in own name 74 19 64 29

WORLD DEVELOPMENT76

I have two DPS accounts in the bank and I save withour market samity (society). I also have saved 60,000/-, my husband knows nothing about that. I lend it toother business men and I get 50/- for 1000/-monthly.We are ignorant people, our money does not earn inthe bank, but if I lend outside, I earn 250/- in 5 monthsfor every 1000/- I lend.

It was also in situations of con¯ict thatwomen were registered ``male'' assets in theirown names. Both F15 and F48, who hadsu�ered extreme violence in the past at herhusband's hands, registered the rickshaws theypurchased with their loans in their own names.F15 hired out her rickshaw in return for a dailyrent while F48 registered it in her own name``so that no one could sell it o�,'' but gave it toher son to pull and took a daily contributionfrom his earnings. F4, whose husband hadmoved out to live with his second wife, regis-tered her rickshaw in her own nameÐ``I bringin the loan, I will be the one to make therepayments''Ðbut handed it over to her son torun as his business and collected her repaymentmoney from him.

In other cases, it was a sense of generalizedinsecurity which led women to invest in someassets of their own. This motivation tended notto be openly discussed because it appeared tocast doubts on the reliability of familynetworks. F39 was one the few female loaneeswho referred explicitly to the insecurities whichunderpinned her desire to save in her ownname. She had adopted a two-pronged strategyof investing in ``joint'' assets (life insurancepolicy and mortgaged-in land) in her husband'sname but also in a second life insuranceaccount in her own name, both safeguardingfamily loyalty but also creating an independentresource for herself:

I have saved what I could and made a life insurancepolicy, it is in his name and I am the nominee. Thatwas for 50,000/-. I also took some mortgaged landwith the loan money for 15,000/-. That is in my hus-band's name. Now I have raised 30,000/-, they havegiven 20,000/-. I put some in my own life insurance.It is for 30,000/-. Women have to look after them-selves, can a husband and son do everything for them?These days, the left and right hand must work sepa-rately, they can't work together. Suppose somethinghappens in future, where will I go? I don't want tohave to su�er. Understanding this has determinedmy actions. He knows about the second account,but it is in my name. I didn't take the mortgage inmy name, I have a husband, I have children, won'tthey be upset if I put it in my name. They will say,look we work to feed and clothe our mother, andshe puts the land in her name.

It is also worth noting that while the regis-tration of land in women's names can be seenas an important strengthening of their fall-backposition, most women tended to explain it as adecision taken on their behalf by theirhusbands as an act of love and gratitude. But,its other signi®cance was explicitly acknowl-edged by one of the male loanees we inter-viewed, a man who had registered one acre ofthe four acres he had purchased in his wife'sname, both as recognition of her labourcontributions to the household economy, butalso to strengthen her bargaining position inthe future, when he was no longer around:

We have both worked hard. That is why I have putsome of the land in her name, she has struggled alongwith me. If I die, my sons may not look after theirmother or when they marry, their wives might misbe-have with her. Now my sons will know she has prop-erty, their wives will know that their mother-in-lawhas property, they will give her importance. They willsay, come and eat with us . . ..

Along with material assets, the other form ofinvestment reported by women loanees whichhad the potential for transforming genderrelations in the long-run related to girls' edu-cation. In many cases, children, particularlythose with educated parents, were alreadyattending school prior to loanees' access tocredit. It was also clear, however, that access toloans, and the enhanced income levels which itgenerated, made education a�ordable for manyhouseholds who could not previously havea�orded it. This sometimes introduced a birth-order factor in educational di�erentials: edu-cation levels tended to be lower among olderchildren whose school-going years coincidedwith the pre-loan phase of the household life-cycle and higher among younger children whoreaped the bene®ts of credit access.

Of greater signi®cance from the point of viewof this paper is the fact that loan access alsointroduced a gender dimension to the decisionto invest in children's schooling. Table 5reports on mean ``gross enrollment rates,''measured as boys and girls aged 6±18 within ahousehold currently attending school as apercentage of boys and girls aged 6±18 presentin that household. In both districts, grossenrollment rates for boys were higher on aver-age than for girls among male loanee house-holds than female (although the di�erence wasnegligible in Mymmensingh) while in bothdistricts, gross enrollment rates for girls wereconsistently higher than for boys in female


loanee households. It is worth noting that asimilar pattern was reported by Pitt andKhandker (cited in World Bank, 1995, p. 36).Not only should such results be welcomed inthe light of the longstanding gender gap ineducation in Bangladesh but also on thegrounds of the various transformatory e�ectsattributed to female education by a wide-ranging body of academic ®ndings, and also bythe female loanees themselves.

Many of the rationales given by the womenloanees for wanting to educate their daughtersre¯ected a change in attitudes that appearedwidespread and were often also expressed bymale loanees: the idea that education enabledgirls to marry more educated, and hence better-behaved, husbands, that less dowry would beasked of an educated bride, that husbands wouldrespect a working woman; that it was no longeracceptable for women to be uneducated. Therewas also evidence, however, of a gender-speci®crationale in the particular stress that womenloanees put on the need for women to ``standon their own two feet,'' both within marriage orin case the marriage failed. This was often basedon their own bitter experiences of what it meantto be totally dependent on husbands for theirevery need, particularly at a time when marriagewas no longer a very secure option. 3 Somewomen made a very explicit equation betweenfemale education, greater self-reliance withinthe marriage and reduced likelihood of abuseand violence: I will educate my daughter as faras is within my means. The reason is that thesedays if you don't educate girls, you marrythem o� to some no-good boy who will beatthem. Why should I get my daughter beaten?This belief does of course receive some statisti-cal backing from Schuler et al. ®nding citedearlier that women with some education wereless likely to report having been beaten. 4

(e) Program-speci®c impacts

It will be seen that there were many conver-gences between the impacts reported for theSEDP and those described in relation topoverty-oriented micro-credit interventions,suggesting that certain impacts can be attrib-

uted to access to credit per se rather than tospeci®c models of credit delivery. At the sametime, it is important to highlight two importantdivergences in ®ndings which appeared tore¯ect speci®cities of organizational practice.First of all, there was general agreement amongSEDP loanees, including those who had previ-ously borrowed from BRAC and Grameen,that there were greater stresses and strainsassociated with repayment of loans frompoverty-oriented programs. These often spiltover into con¯ict, sometimes between husbandand wife, as noted by Goetz and Sen Gupta,sometimes between ``irresponsible'' loanees andother group members worried about theirfuture creditworthiness (this was also noted byMontgomery et al.) but most often betweenloanees' families and program o�cers seekingto recover repayments. The testimonies of theloanees pointed to some of the programmaticdi�erences between SEDP and poverty-orientedorganizations which they believed accountedfor the di�erence.

First, SEDP tended to target women and menwith some prior entrepreneurial experience.Many of the tensions reported in connectionwith poverty-oriented lending occurred betweenprogram sta� and loanees who were havingtrouble meeting their weekly repaymentsbecause of the failure to use loans pro®tably.Second, the larger size of SEDP loans also madea di�erence. As one woman said in relation toGrameen lending: ``They only give you 1000takas, what can you do with that?'' SEDP loanswere large enough for women to invest in theirown enterprises, hence enhancing the value oftheir own contributions to the household. andstill be able to share them with male householdmembers, thereby reducing potential resent-ment and ensuring joint bene®ts. 5 Finally, andperhaps most important, SEDP loans weregiven on easier terms: subsidized interest rates,monthly repayment and possibility for post-ponement of repayments in times of trouble.

The discipline built into poverty-relatedlending, which gave rise to the stresses remarkedon by the loanees, re¯ected a concern with loanrecovery and with long-term sustainability onthe part of these programs. SEDP could a�ord

Table 5. `Gross enrollment rates' for children aged 6±18 (in percentages)


Boys Girls Boys Girls

Male loanee households 77 69 79 78Female loanee households 71 77 55 62

WORLD DEVELOPMENT78

to run a more relaxed lending regime because aconcern with sustainability had not been builtinto program design while its loan recoverye�orts were backed up by the perceivedauthority of a government bank. It was one ofthe constant ironies thrown up by the ®eldworkthat relatively well-o� households could accessloans at subsidized interest rates with greater¯exibility built into their repayment scheduleswhile all around us, poverty-focused creditorganizations were lending far smaller sums ofmoney to much poorer sections of the popula-tion at much higher interest rates with far morein¯exible weekly repayment schedules. Indeed,the pressures of meeting weekly repayments wasmentioned as the single most important source ofthe tensions generated by poverty-orientedlending. As F18 said bitterly, having experi-enced the repayment discipline imposed byGrameen Bank:

If you take say, 1000/- from Grameen, you have to re-pay 10/- takas a month or the members of your samitywill have to make it up for you. The cashier refuses toget up and says to you, Until you have given yourrepayment, I will not leave. With SEDP, they allowyou to give it 2 months late. In Grameen, your samitymembers will come and sell whatever is in your houseto repay your loan. Grameen says, even if your hus-band or your son has died, even then you will haveto make sure that you have made your repayment.

What was missing from the testimonies of theSEDP loanees was the kind of political aware-ness and mobilization documented by Hashemiet al. in the context of BRAC. This is notsurprising since, aside from a brief initialtraining which covered both social andeconomic issues, SEDP did not set out explic-itly to ``empower'' women in the way that someof the other credit programs did. Most of itspractices, including its training, were geared toenterprise development. The di�erence betweenthe lessons o�ered by SEDP training and thoseo�ered by a local, explicitly feminist develop-ment organization was spelled out by F46 whohad experience of both:

Training is good for women... Before I joined Sapta-gram, you could say I was stupid. . . I was like a child.Saptagram taught me to think for myself. With SEDPtraining,I also learnt something new, how to do busi-ness. . . Which is better? Both are important.

But there was little evidence among SEDPloanees of a concern with wider political issuesor with challenging the larger structures of

gender subordination. Indeed, the practice ofsome of the loanees of lending out the pro®tsfrom their loans to other sections of thecommunity at the kind of usurious interestrates that the SEDP had rescued them from,while a sound use of their money from themicro-perspective, raises questions about thepossible widening of the gap between those whowere able to borrow from these organizationsand those who could not.

3. CONCEPTUALIZING ANDEVALUATING EMPOWERMENT: SOME

LESSONS

Evaluations are attempts to document, assessand weigh up the social and economic signi®-cance of changes attributed to a particularintervention. Which particular changes aregiven signi®cance in an evaluation will dependon the intervention in question but also onwhose understanding of reality is given priority.My own evaluation of the SEDP prioritized theunderstandings of the loanees themselves andconsequently used their testimonies as the basisof the analysis. This stress on personal testi-monies should not be taken as a negation ofother more objective forms of data. Thequanti®cation of ®ndings plays a valuable rolein providing some idea of their incidence andmagnitude, helping to distinguish betweenthose which are widespread and those whichare relevant only to a minority. Qualitativemethodologies can be used to place personaltestimonies in their larger context. I chose torely on personal testimonies because empow-erment contains an irreducibly subjective ele-ment, but I sought to interpret my ®ndings onthe basis of my understanding of this largercontext and to support with quantitative evi-dence, either from the household survey orfrom the secondary literature.

The ®rst part of this paper discussed the verycontradictory conclusions arrived at by anumber of evaluations which set out to explorethe empowerment impact of credit to women.Although these various evaluations, includingmy own, were conducted at di�erent points intime (from the mid-1980s to the mid-1990s) andrepresent somewhat di�erent models of creditdelivery, I would suggest that the di�erences intheir conclusions do not re¯ect either di�erencesin timing of evaluation or in speci®cs ofprogram delivery. Indeed, con¯icting conclu-sions were evident for evaluations of the same


set of programs carried out within a year or twoof each other. Conversely, my own evaluation,while relating to a di�erent model of creditdelivery than the rest, nevertheless convergedwith some of their ®ndings. Instead, I suggestthat the main reasons for these con¯ictingevaluations lie in the questions asked, and theinterpretations given to the answers, both ofwhich re¯ect the underlying model of intra-household relationships which underpin theseevaluations.

Some of the evaluations prioritized structuralaspects of intrahousehold relations, the normsof female seclusion, and the gender division oflabor which they legitimized. Others focused onmore individual aspects, such as welfareoutcomes and decision-making roles. Someevaluations conceptualized households as sitesof gender con¯ict while others tended to stresstheir cooperative aspects. The particular modelof gender relations which emerged out of thetestimonies of the loanees interviewed in mystudy drew attention to the relations of unequalinterdependence which underpined the speci®ccon®guration of ``cooperative-con¯ict'' 6 whichcharacterized intrahousehold relations in theBangladesh context. Interdependence within thehousehold was partly emotional. Familymembers who have shared adversity and facedthe humiliations of poverty together, who wereworking toward common goals, are likely todevelop ties of a�ection and loyalty towardeach other. Interdependence also had a mate-rial basis, deriving from the division of rolesand responsibilities within the family. Cooper-ative endeavor was a logical outcome of suchinterdependence. Interdependence within thefamily was also highly unequal. Gender asym-metries in relation to resources and opportu-nities made women far more dependent on menthan men were on women. It gave them a muchstronger stake in strengthening cooperation,and minimizing con¯ict within the family, thanmen and hence less able to bargain for theirown needs and priorities.

Unequal interdependence within the family,and women's greater vulnerability outside it,explain why the women loanees sought greaterequality within the family as a result of theiraccess to credit rather than greater indepen-dence from it. It explains, for instance, thesigni®cance they invested in their ability tobring a valued resource into the household andto contribute directly to household income. Italso explains the value they attached toimprovements in the quality of family rela-

tionships as a result of the increase in theirperceived economic contribution to the house-hold. Some experienced this change as anincrease in the a�ection, love and considerationthat they received, others as a reduction in thetensions and violence within the householdwhich arose out of men's frustrations at theirinability to ful®ll their obligations as primarybreadwinners. As women took on a greatershare of this responsibility, they also reportedgreater ``voice'' in household decision-making,sometimes in the context of joint, sometimesindividual, decision-making.

While these changes re¯ected cooperativesolutions to household inequality, women'sattempts to strengthen and democratizehousehold relationships, a di�erent pattern ofbehavior was reported those in exceptionallycon¯ictual marriages, con¯icts which wereoften associated with male irresponsibility asbreadwinners and hence the partial breakdownof interdependence. These women utilized theirloans, not so much to leave their husbandswhom they continued to rely on for socialprotection, but rather to e�ect a form of ``di-vorce within marriage.'' They separated outtheir ®nances, made decisions about loan useindependently of their husbands, althoughgenerally to the bene®t of their children, andthey retained control over loan-assisted activi-ties. They were more likely to report indepen-dent decision-making and to register assets,including traditionally male assets, in their ownnames. Thus individualized forms of behavioroften signalled greater con¯ict within thehousehold rather than greater ``empowerment''for women. Nevertheless it was women's inde-pendent access to loans that allowed new formsof both cooperative as well as con¯ictual solu-tions to emerge.

The ®rst key point that emerges out of thediscussion in this paper is therefore the need toground the conceptualization of empowermentin an understanding of the relationships ofdependence, interdependence and autonomywhich characterize gender relations in di�erentcultures, the structures of risks, incentives andopportunities which they generate and thereforethe particular trajectories which processes ofempowerment are likely to take. A secondimportant point is that, even within the samecontext, empowerment is a complex, rather thana simple, phenomenon. It has multiple dimen-sions and can occur through a multiplicity ofroutes. By way of illustration, I have summa-rized below the various impacts attributed to

WORLD DEVELOPMENT80

credit by the various evaluations discussed inthis paper, making a rough distinction betweenthose which relate to process and those whichfocus on outcomes (Table 6).

By and large, ``process''-related changes arethe hypothesized pathways through whichempowerment is believed to occur, while``outcome''-related impacts relate to thoseachievements associated with access to creditwhich have transformatory implications forgender relations. There is however, nostraightforward cause-and-e�ect relationshipbetween process and outcomes. Nor is it alwaysclear when a change is cause and when e�ect,when process and when outcome. Someimpacts are means to valued ends, others arevalued ends in themselves but also a necessaryprecondition for achieving yet other valuedends. Some outcomes may be conditional onhypothesized prior processes while others occurindependently of them. Thus women were stillable to achieve valued impacts in their lives as aresult of their access to credit, regardless of whomanaged their loan-funded activities or whokept the accounts. We also saw that wherewomen used at least part of their loans toenhance their own productivity, they were morelikely to experience such impacts. At the sametime, a growth of women's self-con®dence, intheir knowledge of their rights, their willingnessto participate in public action and even thereduction of domestic violence may haveoccurred as a result of women's participation inthe new forms of social relationships embodiedin credit organizations; they bore little rela-tionship to the productivity of their loans.

The third point, which is an extension of thesecond, is the importance of avoiding narrow,unidimensional conceptualizations of empow-erment which feed into dichotomous models of

change: women are judged to be eitherempowered or not empowered on the basis onhow closely they conform to a particular indi-cator. If instead we see empowerment as anexpansion in the range of potential choicesavailable to women so that actual outcomesre¯ect the particular set of choices which thewomen in question value, it becomes possibleto make sense of what appear at ®rst sight to berather contradictory ®ndings in my study. Itbecomes possible, for instance, to reconcile the®nding that many of the women who subscri-bed most strictly to notions of purdah as amatter of family honour and female proprietywere also some of the most successful entre-preneurs in my sample, women who not onlymanaged and made a ®nancial success of theirenterprises but also described themselves as themain decision-makers in their households. Italso allows us to make sense of women loaneeswho registered assets purchased with theirloans or with their hard-earned incomes in theirhusband's rather than their own names and yetdisplayed enormous physical courage andinitiative on occasions when their property wasunder threat. 7 It also explains why womenwho had previously ignored the norms ofgender propriety, working as agricultural laborin the ®elds or taking their own goods to themarket withdrew from these public forms ofactivity as soon as their increased prosperityallowed them to and opted instead for self-employment within the con®nes of their homes.

This suggests, as a fourth point, that we needto make a distinction between forms of changewhich have been prioritized in the feminist or inthe developmental literature and forms ofchange valued by those whose lives an inter-vention is seeking to transform. Many of theimpacts reported by the women and men in my

Table 6.

Process Outcomes

Decision to access loans Enhanced sense of self worthAccess to loans Increase in perceived economic contribution

Enhanced role in minor decisionsDecisions about loan use/repayment Enhanced role in major decisions

Exercise of purchasing powerMobility in the public domain

Decisions about loan-funded activities Political participationLabour contribution in loan-funded activities Reduction of domestic violenceMarketing of loan-funded products Increase in women's savings and assetsAccounting control Reduction of gender gap in well-being

Reduction in gender gap in educationTraining Greater social inclusionGroup participation Self-reliant livelihoods


study were also identi®ed by ``etic'' approachesto the evaluation of loan impact: the reductionin domestic violence, increased voice in deci-sion-making and enhancement of their assetbase. There were some aspects highlighted inthe women's testimonies which were notre¯ected in these other evaluations, whileothers were given a signi®cance not shared bythe women themselves. The stress that womenplaced on their own sense of enhanced self-worth as economic actors, of being able tomake a contribution to household livelihoodsand the value they attached to both their ownas well as husbands' ability to move out ofdemeaning forms of waged labor into their ownenterprises, all of these are forms of socialchange with implications for intrahouseholdinequality which had no place in the evalua-tions cited earlier. Indeed, the women appearedto give a very di�erent interpretation to theincrease in their work associated with theiraccess to credit to that given by some of theexternal analysts. The fact that this increasewas a product of their enhanced ability tocontribute to household livelihoods and theconsequent mitigation of their status asdependants led many to describe it as a valuedtransformation of the meaning of work ratherthan an intensi®cation of their work burdens.

On the other hand, they did not attach quitethe same degree of value to individualizedforms of control over resources that featured insome of the evaluations cited earlier. Althoughmost did seek to utilize some part of their loansthemselves, sharing their loans with husbandsand sons did not necessarily carry connotationsof loss of control. It was the ability to partici-pate in making decisions about how loans wereused and how the income from loans was to beused that mattered; this ability was valuedwhether exercised jointly or individually. As faras the ability to move around freely outside thehome was concerned, the picture was mixed.Most women did not see this as a particularlyvalued aspect of change in their lives. The valuethey gave to working in a self-employedcapacity on their own homesteads is not hardto understand when we consider the pittancethat they earned as agricultural wage laborersand the arduous and demeaning conditionsunder which they worked. There was also near-unanimous antipathy among women loanees tothe idea of marketing their own produce inrural bazaars or haats because of what suchaction signalled to the rest of the community.Consequently, the idea that women are exclu-

ded from the market place, and need transportand protection to overcome this exclusion,misses the point that many women excludethemselves from this arena and that they do soon the basis of what people might say ratherthan what men might do. Such self-imposedexclusion is likely to continue as long as thealternative is equated with poverty and with theabsence or failure of male protection.

The ®nal point to make is a variation on oneoften made by feminist scholars. Women arenot a homogenous category. While this point isgenerally made to highlight the relevance ofclass, caste, race and culture in di�erentiatingwomen's needs and interests, I want to make ithere in relation to women as individuals. Thereis no reason to expect women, even those fromthe same class, to respond identically to newopportunities. Our understanding of theprocesses of empowerment needs to bear inmind the important distinction between womenas a socially subordinate category and womenas a highly diverse group of individuals. Wehave to allow for the fact that di�erent womenwill experience and act on new opportunities inways that re¯ect some combination of theirstructural positioning and their own uniqueindividual histories. On the one hand, thismeans that even the best planned interventionis unlikely to be automatically empowering forall women. At best, it can create the kind ofenvironment or provide the kind of resourceswhich are most likely to help as many womenas possible to empower themselves. But thereare always likely to be some women who willnot, or are not permitted to, take up thepossibilities on o�er. What we are likely toobserve at any point are distributions ofresponses to these di�erent possibilities.

On the other hand, however, the individual-ity of women also means that not all evidenceof empowered behavior on their part can beascribed to a particular intervention. Thetendency to do this rests on a version of thedichotomous model of empowerment notedearlier where it is assumed that prior to theintervention, the women in question werecowed, fearful and mute while after it, theybecame articulate, entrepreneurial and active.Interviews with both men and women in mysample made it abundantly clear not all womenhad been passive or silent actors within theirhouseholds prior to the arrival of SEDP.Indeed I found many examples of women whowere already exercising considerable entrepre-neurial initiative and playing key roles in

WORLD DEVELOPMENT82

managing their households Some of this can betraced to their recognized individual compe-tence relative to male household members. Inaddition, it can also be traced to some of themajor social changes in Bangladesh in the pastdecade or so which have e�ected what has beencalled a ``quiet revolution'' (Chen, 1983) in theideas and practices of gender relations. Thegreater availability of such loans for women,and women's willingness to take them up, canbe seen as both an e�ect of this revolution aswell as contributing to further changes. Here isF18's account of these larger changes:

We village girls, we understood less before, we neverwent into the town or city. Before this area was idle,there was very little education. It was jungle here,there was no decent road. But when the CNB roadcame, people became smarter. Before you could notsell a marrow here for 2 or 3 takas. Now since theroad, you won't get less than 20 or 25 takas. This ishow we have prospered . . .. I want both my son anddaughter to study till IA. I hope she can get a job infamily planning . . .. Many girls even in the villagesare working now, they become cashiers with samities(cooperatives), they get paid. Before women did notgo out of the house because people might say some-thing. Before we were idle, now if there is money tobe made, we are no longer idle . . .. Since independencewomen in towns got more opportunities, but since the®rst woman prime minister, women in the countrysideare also getting more opportunities.

4. CONCLUSION

Let me conclude by making a general pointabout micro®nance and women. While therecent questioning of the empowermentpotential of loans to women helps to counterthe earlier, somewhat single-minded preoccu-pation with ``repayment rates,'' the recom-mendations which come out of the morenegative evaluations cited in this paper carrythe danger of overloading micro®nance organ-isations with empowerment-related goals to theextent that their ability to deliver e�ective andsustainable ®nancial services is likely to beseriously undermined. This point is made more

generally by Rutherford (2000) who suggeststhat many NGOs promoting microcredit in theSouth Asian context have failed to develope�ective ®nancial services for the poor ``becausethey are not primarily interested in ®nancialservices but in much wider social issues'' (p. 9).

There are multiple rationales for lending towomen, apart from empowerment. The factthat women are much more likely to share theirloans with male household members than menare with women, in my view, merely strength-ens the argument for lending to women. Theentire family is much more likely to bene®teconomically, and women are much more likelyto bene®t personally and socially, when loansare directed at women rather than men. Loansto men do little to challenge the internal genderinequalities of households, and indeed appearto reinforce them by giving men an a�ordablebase from which to prevent their wives fromengaging in their own income-earning activities.

There are other arguments as well. It is oneof the injustices of the way that society isorganized in Bangladesh that extremely ablewomen, even those from better-o� households,are unable to realize their entrepreneurialpotential because their gender acts as a barrierto gaining access to the necessary resources.Men, even poor men, have always had morechoices in terms of accessing economic oppor-tunities than women from an equivalent class.Women's higher repayment records do notmerely re¯ect their socialized compliance in theface of the instrumentalist authority of NGO orgovernment o�cials, as the more negativeevaluations tend to suggest, but also thecompliance which comes with having fewchoices. If purposive interventions can help todirect resources to women, thereby overcomingpast barriers which have led to the suppressionof their entrepreneurial potential, then theymust be welcomed on grounds of e�ciency andequity. If greater e�ciency and equity help tolay the grounds for women to tackle otheraspects of injustice in their lives, then we willhave found a di�erent and perhaps moresustainable route to women's empowerment.

NOTES

1. Clearly, impact was likely at least partly to re¯ect

returns to loan-related investment, but these proved

extremely di�cult to calculate. Loanees were at di�erent

stages of their loan cycle and loans were often repaid

from sources other than the loan-funded enterprise so

that pro®t calculations required calibration between

costs and returns to more than one enterprise see Kabeer

(1998).


2. A nationwide participatory poverty assessment by

UNDP (1996) found ownership of homestead land the

second most important priority identi®ed by rural

women, second only to productive opportunities.

3. I found a similar value attached daughters' educa-

tion as a route to greater self-reliance by women workers

in the garment industry who I interviewed in the context

of a study exploring the impact of wages on women's

empowerment (see Kabeer, 2000).

4. It is also supported by studies from other parts of

the world, see Kabeer (1999).

5. Schuler et al. (1996) found that while access to credit

by itself appeared to have some e�ect in diminishing

violence against women, women's economic contribu-

tion appeared to only start to have an e�ect once it had

reached a certain level. The signi®cance of loan size also

crops up in the study Goetz and Sen Gupta: they found

that while very few women in their sample received loans

greater than 4000/-, those that did were much more

likely to make some use of their loans themselves.

6. See Sen (1990).

7. For instance, was one of the women who had

registered the rickshaw purchased with her loan in her

husband's name and refused to let him write his land

over into her name despite his desire to do so. Yet she,

together with another woman, had sat guard, with

machetes in their hands, over this land all night for

several nights while her husband was away because she

feared that their newly planted rice might be run over by

a tractor by another family which was disputing their

claim to the land. She told us: ``Afterwards my husband

wanted to register the land in my name but I said, it is

your father's property, what will people say. But he said,

you have struggled so hard for it. I said, it is enough for

me that you want to register it in my name but as long as

I have a husband, the land will be there and if I don't

have a husband, what use is the land to me?''

REFERENCES

Ackerly, B. A. (1995). Testing the tools of development:credit programmes, loan involvement and women'sempowerment. IDS Bulletin, 26(3), 56±68.

Chen, M. A. (1983). A quiet revolution. Women intransition in rural Bangladesh. Cambridge, MA:Schenkman.

Dreze, J., & Sen, A. (1995). India, economic developmentand social opportunity. Oxford: Oxford UniversityPress.

Goetz, A. M., & Sen Gupta, R. (1994). Who takes thecredit? Gender, power and control over loan use inrural credit programmes in Bangladesh. WorldDevelopment, 24(1), 45±63.

Hashemi, S. M., Schuler, S. R., & Riley, A. P. (1996).Rural credit programs and women's empowermentin Bangladesh. World Development, 24(4), 635±653.

Kabeer, N. (1998). Can't buy me love? Re-evaluatinggender, credit and empowerment in rural Bangladesh.IDS Discussion Paper No. 363. Institute of Devel-opment Studies, Sussex.

Kabeer, N. (1999). Resources, agency, achievement:re¯ections on the measurement of women's empow-erment. Development and Change, 30(3), 435±464.

Kabeer, N. (2000). The power to choose: Bangladeshiwomen and labour market decisions in London andDhaka. London: Verso.

Montgomery, R., Bhattacharya, D., & Hulme, D.(1996). Credit for the poor in Bangladesh. In D.Hulme, & P. Mosley (Eds.), Finance against poverty.London: Routledge.

Pahl, J. (1989). Money and marriage. London: Macmil-lan.

Pitt, M., & Khandker, S. (1995). Household andintrahousehold impacts of the Grameen Bank andsimilar targeted credit programs in Bangladesh.Paper presented at workshop on Credit programsfor the poor: household and intrahousehold im-pacts and program sustainability by the Educationand Social Policy Department, Washington, DCand Bangladesh Institute of Development Studies,Dhaka.

Rahman, R. I. (1986). Impact of Grameen Bank on thesituation of poor rural women. BIDS Working PaperNo. 1, Grameen Evaluation Project. Dhaka: Bangla-desh Institute of Development Studies.

Rutherford, S. (2000). Self-help groups as micro®nanceproviders: how good can they get? Paper presented atworkshop on `Best practices in group dynamics andmicro-credit. Organized by the Micro-credit Bureau,Department of Women and Children, Governmentof India. Delhi February 15±17.

Sen, A. K. (1990). Gender and co-operative con¯icts. InI. Tinker (Ed.), Persistent inequalities: Women andworld development (pp. 123±149). Oxford: OxfordUniversity Press.

Schuler, S. R., Hashemi, S. M., Riley, A. P., & Akhter,A. (1996). Credit programs, patriarchy and men'sviolence against women in rural Bangladesh. SocialScience and Medicine, 43(12), 1729±1742.

UNDP (1996). A pro-poor agenda: poor people's perspec-tives. Report on Human Development in Bangla-desh. Dhaka: UNDP.

World Bank (1995). Toward gender equality: the role ofpublic policy. Washington, DC: World Bank.

WORLD DEVELOPMENT84

conﬂicts over credit: re-evaluating the empowerment...

Documents