تعارضات بر سر اعتبار: ارزیابی مجدد پتانسیل توانمندسازی وام ها برای زنان روستایی در بنگلادش

This paper explores the reasons why recent evaluations of the empowerment potential of credit programs for rural women in Bangladesh have arrived at very conflicting conclusions. Although these evaluations use somewhat different methodologies and have been carried out at different points of time, the paper argues that the primary source of the conflict lies in the very different understandings of intrahousehold power relations which these studies draw on. It supports this argument through a comparative analysis with the findings of a participatory evaluation of a rather different credit program in Bangladesh in which the impact of loans was evaluated by women loanees themselves.

Microcredit programs for the poor have come to occupy a central place in poverty-oriented strategies in Bangladesh. Such programs have a number of features in common. They are largely targeted at women from the poorest sections of the population; they lend small sums of money to individuals as members of groups and rely on group liability to ensure loan repayment; they subsidize administrative costs rather than interest rates; and loans are repaid in weekly installments. Debates as to the actual effectiveness of these programs in reducing poverty continue. More recently, these debates have been extended to the possible implications of such programs for women's empowerment, with some evaluations claiming extremely positive results while others suggesting that microcredit leaves women worse off than before. In this paper I want to focus on a number of attempts to evaluate the empowerment potential of loans to women in order to find out why such diametrically opposed claims can be made about the same, or very similar, programs. I will be exploring examples of both “negative” and “positive” evaluations, interrogating them for the methodologies they used, the questions they asked, the findings they reported and the interpretations they gave to their findings. In addition, I will be drawing on the findings of my own evaluation of a rather different credit program in Bangladesh in order to explore the question of empowerment when it is assessed on the basis of women loanee's own testimonies rather than deduced from selected aspects of their behavior. (a). Does access to credit “empower” women? the negative verdict My first example of a negative evaluation is by Goetz and Sen Gupta (1994). They use a five-point index of “managerial control” over loans as their indicator of empowerment. At one end of their index are women who are described as having “no control” over their loans: these are women who either had no knowledge of how their loans were used or else had not provided any labor into the activities funded by the loan. At the other end are those who were considered to have exercised “full” control, having participated in all stages of the activity funded by the loan, including the marketing of produce. The study found that the majority of women, particularly married women, exercised little or no control over their loans by this criteria. Interpreting this as evidence of widespread loss of control by women over their loans to men, Goetz and Sen Gupta go on to suggest three possible repayment scenarios, all with negative implications for women. In the first, the male family member using the loan takes responsibility for its repayment, a satisfactory outcome from the woman's point of view but one which the authors believe negates the developmental objectives of lending to women. In the second, men are unable to supply the requisite repayment funds and women loanees have to substitute funds from other sources, drawing on their savings, cutting back on consumption, selling off utensils and other assets. They have responsibility without control. In the third, men are unwilling to repay the loans, leading to an intensification of tensions within the household, often spilling over into violence. In addition, violence against women was also exacerbated by the frustration of husbands at the wives' delay or failure in accessing credit. Facilities to enhance women's access to the market is put forward by the authors “as the single most effective way of enhancing their control over loans, as well as their public presence and their self-confidence” (p. 59). The provision of transportation recommended to take women to the market place along with security measures to protect them against the possibility of male resistance to their presence in the market place are recommended as supportive measures. In her study, Ackerly (1995) noted that underpinning most credit interventions in Bangladesh was an implicit model of the empowered woman: Empowered, the borrower wisely invests money in a successful enterprise, her husband stops beating her, she sends her children to school, she improves the health and nutrition of her family, and she participates in major family decisions (p. 56). Rather than seeking to measure these outcomes directly, she takes “accounting knowledge” as her indicator of the likelihood of these and other transformative outcomes occurring. Women who were able to report on the input costs for loan-funded enterprise, its product yield and its profitability, were counted as empowered. She found that membership of some credit organizations was more likely than others to contribute to the likelihood of women's empowerment by this criteria. Women who provided labor to loan-assisted enterprise, sold their own products, or kept their own accounts were also likely to be empowered. She too concluded that women's access to the market was the primary route for their empowerment—“knowledge and empowerment come through market access” (p. 64)—and warned against the likelihood of overwork, fatigue and malnutrition were loans used to promote women's labor involvement without also promoting their market access. Our third example of a negative evaluation of the impact of credit for women's empowerment is by Montgomery, Bhattacharya, and Hulme (1996). Although the evaluation of the empowerment impact occupies only a small section of their more general evaluation of credit programs for the poor, I have included it here because it exemplifies a particular understanding of households and gender relations within the literature on Bangladesh. According to their findings, only 9% of first-time female borrowers were primary managers of loan-funded activities while 87% described their role in terms of “family partnerships.” By contrast, 33% of first-time male borrowers had sole authority over the loan-assisted activity while 56% described it as a family partnership. They also found that access to loans did little to change the management of cash within the household for either female or male loanees. Interpreting reports of “joint” management as disguised male dominance in decision-making, the authors concluded that access to loans had done little to empower women. Its main effect had been to increase the social status of loan-receiving women vis-à-vis less well-off women rather than vis-à-vis men within their household or the wider community. (b). Does access to credit “empower” women? the positive verdict In contrast to this set of evaluations are others which paint a far more positive picture of the impact of these same credit programs on women's lives. Rahman (1986) found that that loanee households in general, regardless of the gender of the loanee, had higher income and consumption standards than equivalent nonloanee households. Although loans to women were more likely to benefit male consumption standards than male loans were to benefit female consumption standards, women loanees nevertheless did benefit from their direct access to loans. Women who made active use of at least some of their loans had higher consumption standards and were more likely to have a role in household decision-making, either on their own or jointly with their husbands, than “passive” female loanees, and both in turn had considerably higher consumption standards and were more likely to participate in household decision-making than women from male loanee households or from households who had not received any credit. A study by Pitt and Khandker (1995) explored the impact of male and female membership of credit programs on a number of decision-making outcomes in order to establish the extent to which they were differentiated by the gender of the loanee. The outcomes included the value of women's nonland assets, the total hours worked per month for cash income by men and women within the household, fertility levels, the education of children as well as total consumption expenditure. These authors also concluded that households receiving loans were largely better-off than those not receiving loans. In addition, the findings that the gender of the loanee did influence the pattern of household decision-making outcomes was interpreted as evidence that women's preferences carried greater weight in determining decision-making outcomes in households where they had received a loan compared to households where either men received the loans or in households where no loans had been received. A third example of a “positive” verdict is by Hashemi, Schuler, and Riley (1996). They explored the impact of credit on a number of indicators of empowerment: (i) the reported magnitude 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 cash savings; (v) involvement in major decision-making, such as purchasing land, rickshaw or livestock for income earning purposes; (vi) freedom from family domination, including the ability to make choices concerning how their money was used, the ability to visit their natal home when desired and a say in decisions relating to the sale of their jewellery or land or to taking up outside work; (vii) political awareness such as knowledge of key national and political figures and the law on inheritance and participation in political action of various kinds; and finally, (viii) a composite of all these indicators. They found that women's access to credit was a significant determinant of the magnitude of economic contributions reported by women; of the likelihood of an increase in asset holdings in their own names; of an increase in their exercise of purchasing power; of their political and legal awareness as well as of the value of the composite empowerment index. In addition, BRAC loanees tended to report significantly higher levels of mobility and higher levels of political participation while Grameen members reported higher involvement in “major decision-making.” When women's economic contribution was used as an independent variable, the effect of access to credit on the empowerment indicators was reduced but remained significant, suggesting that one important route through which women's access to credit translated into “empowerment” was via their enhanced contribution to family income. The study also found that access to credit appeared to be associated with an overall reduction of the incidence of violence against women. Regression analysis suggested that older women, women who had sons and women with education were less likely to have been beaten in the past year (Schuler, Hashemi, Riley, & Akhter, 1996). These findings are consistent with the lower status of young wives who are relatively new in the husband's home, with the prevailing culture of son preference and with the greater agency attributed to women with education (see, for instance, Dreze & Sen, 1995). In addition, they found that membership of a credit program was associated with a statistically significant reduction in violence, but that the magnitude of women's economic contribution did not have any significant effect. They concluded that it was women's participation in the expanded set of social relationships embodied in membership of credit organizations rather than increases in their productivity per se which explained reductions in domestic violence. (c). Explaining the conflicts: methods, questions, interpretations and models These conflicting conclusions about the “empowerment” potential of credit for women are both apparent and real. What appear to be contradictory findings concerning, for instance, the extent to which credit exacerbates or lessens violence against women, enables or fails to enable them to acquire independent assets, is associated with an increase or decrease in their living standards is partly a difference in methodology. It reflects the fact that some studies relied largely on statistical data and significance tests for their findings while others relied on more qualitative, sometimes anecdotal, evidence. Consequently, some of the differences in findings relate to differences in the incidence of empirical outcomes, some findings referring to “average” and others to “nonaverage” outcomes. Thus Hashemi et al's finding that women's access to credit was associated with an overall reduction in the incidence of domestic violence is perfectly compatible with the finding that it exacerbated violence in a number of individual households reported both by them (see Schuler et al., 1996) as well as by Goetz and Sen Gupta. Conflicting conclusions about the impact of credit also reflect differences in the questions asked by different evaluations. By and large, the negative evaluations focused on processes of loan use while the positive ones focused on outcomes associated with, and attributed to, access to loans. The validity of both sets of measures depends on their conceptual clarity and on the validity of their underlying premises. There are, for instance, reasons to question whether some of the process-based measures do indeed measure what they are intended to measure. In conceptualizing the process by which women's access to resources translates into impact, Pahl (1989) had made an analytically important distinction between what she calls “control,” the ability to make policy decisions concerning the allocation of resources, “management” which relates to decisions to do with the implementation of policy and “budgeting” which merely involves keeping track of income and expenditure. Ackerly's measure of empowerment is ambiguous because it does not distinguish between women who acquired their “accounting knowledge” through an active involvement in the control and management of their loans, in the way that she appears to assume, or merely through a budget-keeping role of the kind pointed to by Pahl. Goetz and Sen Gupta's index of managerial control is similarly ambiguous. It essentially conflates “control” and “management,” making no distinction between decisions about loan use and decisions related to implementation. But policy decisions about how loans are to be used are separate from, and indeed prior to, decisions relating to the management of the enterprise to which the loan is assigned. Since the authors offer no information on the decision-making processes by which the loans were allocated, we have no way of knowing the extent to which the observed allocations reflected a sound economic calculus on the part of women, the specific individual circumstances of their household or the blatant exercise of male power. Indeed, it is in principle possible (though in practice unlikely) that, with the exception of the unknown number of the 22% of women in their “no control” category who did 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 the causality implicit in process-based indicators. The possession of accounting knowledge or exercise of managerial control do not, on their own, suffice as evidence of empowerment. To be persuasive as such evidence, we would need to know more about their relationship to other valued achievements, perhaps of the kind outlined in Ackerly's description of the ideal-typical “empowered woman.” Indeed, the assumption that managerial control over loan use is a necessary condition for women to be empowered by their access to loans is explicitly rejected by Hashemi et al. In their study, they classified all the women loanees in their sample according to the “control” categories developed by Goetz and Sen Gupta and confirmed that large percentages of women loanees in their sample had indeed “lost control” over their loans by this criteria. This did not, however, prevent a significant proportion of them from achieving a range of other valued impacts, although, as we noted, the likelihood of these positive impacts was strengthened if women used at least part of their loans to increase the value of their own economic contributions. As far as outcome indicators are concerned, their validity depends on how well they capture changes in the structures of gender inequality within the household and community, not merely on how well they capture changes in household living standards or even in children's welfare. One of the strengths of the study by Hashemi et al. is that their indicators meet this criteria. They can all be seen as valued outcomes in their own right as well as being linked to the structures of constraint which give rise to gender inequality in Bangladesh. By contrast, the study by Pitt and Khandker is undermined by the absence of any obvious rationale for the particular decision-making outcomes selected for their study. Their findings are consequently not always easy to interpret. The only outcomes with relatively unambiguous theoretical links with women's empowerment are (i) women's ownership of nonland assets, an increase in which could be interpreted as a strengthening of their fall back position, and (ii) the gender gap in education, a reduction in which could be seen as addressing a longstanding gender inequality in the value given to children. Their own interpretations of some of their findings tend to be somewhat ad hoc and open to other equally plausible and very differing interpretations. This takes us to yet another factor behind the conflicting conclusions we have been discussing which is the proclivity to “read” empirical findings in the light of preconceived notions about loan impact so that the same findings are given extremely contradictory interpretations. Thus, Pitt and Khandker take their finding that loans to women led to an increase in their market-oriented work to indicate an empowerment effect. By contrast, all three negative evaluations warn against the intensification of women's workloads and fatigue. Pitt and Khandker interpret the higher level of household consumption expenditure associated with loans to women as evidence of the greater weight given to women's preferences in household decision-making; Montgomery et al. suggest that such findings demonstrate that loans to women are `heavily compromized by the persisting responsibilities of women to cover the consumption needs of the family' (Montgomery et al., 1996, p. 168). Similarly, the increase in women's welfare levels as a result of their access to credit is linked to their enhanced role in household decision-making by Rahman, but given a much more passive interpretation by Goetz and Sen Gupta who suggest that women give up their loans to men “in exchange for the right to have greater expenditures on their own or their children's clothing and health.” In short, there are differing judgements embodied in these evaluations as to what kinds of changes constitute evidence of empowerment, differences which in turn reflect the differing models of households, and the power relations within them, which these evaluations draw on. While both positive and negative evaluations accept the premise of gender inequality in intrahousehold relations, they vary considerably in the significance and meaning attached to cooperation and conflict between men and women within the household and consequently to autonomy, dependence and interdependence within the household. By and large, the negative evaluations tend to be negative because they stress gender antagonism within the household and discount the significance of cooperation. Thus, for Montgomery et al., reports of “jointness” in the management of household enterprise and income are merely examples of disguised male dominance; only the exercise of autonomous female authority is counted as evidence of empowerment. Goetz and Sen Gupta's discussion of the circumstances under which the investment of women's loans in the purchase of a rickshaw would, and would not, constitute exercise of “control” also reveals this individualized notion of empowerment. Rickshaw pulling in Bangladesh is a purely male activity so that the purchase of a rickshaw, an extremely common use of loans to poor women, represents investment in an activity to which women are unable to contribute any labor. While such women would automatically be classified in the “little” or “no” control category by Goetz and Sen Gupta's criteria, they suggest that a woman could still be classified as exercising “significant control” if the rickshaw was licensed in her name and if she established a contractual rental relationship with the rickshaw puller. In the context of rural Bangladesh, however, this would constitute extremely anomalous behavior on the part of a woman who had an unemployed son or husband in the family who was able and willing to pull the rickshaw and to take responsibility for loan repayment. The more positive evaluations, by contrast, are positive partly because they do not privilege individualized over joint forms of behavior. Pitt and Khandker attempt to capture possible increases in the weight given to women's preferences in a series of household decisions following their access to loans, but do not rule out joint decision-making. Both Hashemi et al., as well as Rahman explicitly incorporate some “jointness” on interests within the household into their indicators of empowerment. In the final analysis, the plausibility of one or other set of conclusions about the transformatory impact of credit for women will rest on the credence attached to the models of power which inform the analysis. Despite their differences, however, both sets of evaluations share in common an absence of testimonies by women loanees themselves as to the impact of credit on their lives. Obviously, in the context of evaluation studies where valued resources are at stake, personal testimonies on impact have to be interpreted with caution, given that there may be a strong incentive among beneficiaries to present impact in a positive light. At the same time, participatory impact assessments can help to enrich academic theorisations of gender subordination by providing important insights into inequality as a “lived experience.” In the rest of the paper, I want to report on the findings of my own evaluation of a rather different credit program in Bangladesh in which I sought out the testimonies of 50 female and 20 male loanees as to the impact of loans on their lives. In addition, I also carried a quantitative survey of 700 households to provide basic descriptive statistics on the loanees, their households, their patterns of loan utilization as well as on some of the impacts identified in the evaluation literature. I will be drawing on the loanees' testimonies as a different vantage point from which to contextualize and assess the findings of the various evaluations discussed here as well as to consider what the perspectives of women loanees themselves can add to our understanding of the transformatory potential of credit targeted at women. I will be concluding with some general conceptual and methodological comments on the evaluation of women's empowerment.

Let me conclude by making a general point about microfinance and women. While the recent questioning of the empowerment potential of loans to women helps to counter the earlier, somewhat single-minded preoccupation with “repayment rates,” the recommendations which come out of the more negative evaluations cited in this paper carry the danger of overloading microfinance organisations with empowerment-related goals to the extent that their ability to deliver effective and sustainable financial services is likely to be seriously undermined. This point is made more generally by Rutherford (2000) who suggests that many NGOs promoting microcredit in the South Asian context have failed to develop effective financial services for the poor “because they are not primarily interested in financial services but in much wider social issues” (p. 9). There are multiple rationales for lending to women, apart from empowerment. The fact that women are much more likely to share their loans with male household members than men are with women, in my view, merely strengthens the argument for lending to women. The entire family is much more likely to benefit economically, and women are much more likely to benefit personally and socially, when loans are directed at women rather than men. Loans to men do little to challenge the internal gender inequalities of households, and indeed appear to reinforce them by giving men an affordable base from which to prevent their wives from engaging in their own income-earning activities. There are other arguments as well. It is one of the injustices of the way that society is organized in Bangladesh that extremely able women, even those from better-off households, are unable to realize their entrepreneurial potential because their gender acts as a barrier to gaining access to the necessary resources. Men, even poor men, have always had more choices in terms of accessing economic opportunities than women from an equivalent class. Women's higher repayment records do not merely reflect their socialized compliance in the face of the instrumentalist authority of NGO or government officials, as the more negative evaluations tend to suggest, but also the compliance which comes with having few choices. If purposive interventions can help to direct resources to women, thereby overcoming past barriers which have led to the suppression of their entrepreneurial potential, then they must be welcomed on grounds of efficiency and equity. If greater efficiency and equity help to lay the grounds for women to tackle other aspects of injustice in their lives, then we will have found a different and perhaps more sustainable route to women's empowerment.