|
 |
 |
 |
 |
|
GENDER ANALYSIS: ALTERNATIVE PARADIGMS Carol Miller and Shahra Razavi (May 1998) The DPU Framework (Gender Plannning) Feminist Economics: Taking Gender Analysis to the Macro-Economic Level Neo-Classical Feminist Economics: Gender and Adjustment Gender and Adjustment: the Models Neo-Classical and Feminist Critical Economics: Convergence and Divergence Appendix: Alternative Analytical/Training Frameworks Training Packages and Material Carol Miller is Research Associate in the Gender Programme at UNRISD in Geneva, where she has assisted in the research activities of a UNDP-funded inter-regional project in Bangladesh, Jamaica, Morocco, Uganda and Vietnam, 'Technological Co-operation and Women's Lives: Integrating Gender into Development Policy'. She is the author of numerous articles and papers on gender and development, including conceptual shifts in the women and development discourse and studies of gender mainstreaming in different institutions. As Associate Lecturer at the Open University, a long-distance learning institution based in the UK, she tutors second year students on Third World Development and provides academic counselling for continuing students. Shahra Razavi is Project Co-ordinator at UNRISD in Geneva, where she is primarily responsible for conceptualizing new areas of research, notably the 'Gender, Poverty and Well-being' inter-regional project, and writing conceptual papers on gender and poverty. She also co-ordinates the UNDP-funded inter-regional project 'Technical Co-operation and Women's Lives: Integrating Gender into Development Policy.' A specialist in agrarian and development issues, she is the author of numerous publications on gender and development, must recently Agrarian Change and Gender Power, forthcoming from Macmillan. Ms. Miller and Ms. Razavi work together frequently on gender issues. Together they supervised the conceptualization and publication of an Occasional Paper series for the Fourth World Conference on Women, covering themes such as gender and micro-enterprise, national level women's policy machinery, and industrialization and export promotion. Their most recent collaboration, a book entitled Missionaires and Mandarins: Feminists and Development Organizations, is forthcoming from ITU Press in London. They also write regularly for the Gender Programme's annual newsletter FOCUS on the politics of integrating gender into development policy. Gender analysis, once confined to the margins of development theory, has over the last ten years penetrated both the thinking and the operations of international development institutions, transforming not only the practice but the objectives of development planning and programmes. Over this period, gender analysis has grown and developed; it is no longer a single perspective, but several. This study reviews the growing body of work on gender analysis, including recent approaches, explicating their main lines of convergence and difference and assessing the results of their incorporation into training packages and programmes. The principal conceptual paradigms that have framed gender analysis over the past decade include: the Gender Roles Framework (GRF); the University College-London Department of Planning Unit (DPU) Framework ('triple roles model'); the Social Relations Framework developed by the Institute for Development Studies (IDS), Sussex; and the approaches of feminist economics, including both neo-classical and critical feminist economics. These approaches differ in terms of their conceptualization of gender, scope of institutional analysis, implications for development, and issues of social and organizational change. The GRF looks at access to and control over income and resources, highlighting the incentives and constraints under which men and women work in order to anticipate how projects will impact their productive and reproductive activities as well as the responsibilities of other household members. Its approach to gender mainstreaming can be considered 'integrationist' rather than 'transformational'; viewing development as a process from which everyone can benefit, it argues that for women to do so, projects must more adequately incorporate women and gender issues into the development plan. It limits its analysis to the household, rather than extending it to markets, firms and the state. It also tends to assume that women are a homogenous-and unchanging- category, overlooking the ways in which gender cross-cuts hierarchies of race and class in different circumstances. The DPU Framework also examines women's roles, going beyond the household to look at women's triple roles-in production, reproduction, and community management-and the implications of these for their participation in development. This model distinguishes between gender practical needs, arising out of the division of labour in society, and gender strategic needs, arising out of differential power relationships between women and men. Like the GRF, its focus on roles does not fully illuminate the nature of gender relations in society, touching only tangentially on the social relations through which resources such as status and authority are produced. The institutions that perpetuate gender inequality, including markets and the state, are not addressed. Moreover, the distinction between practical and strategic needs focuses only indirectly on the conflictual nature of gender relations. The Social Relations Framework is informed by an analysis of the social relations within the family, market, state and community that illuminates the ways in which gender and other inequalities are created and reproduced. It examines the social processes through which human needs are met as well as the institutions through which inequalities are constructed and reproduced. It regards gender relations as both conflictual and collaborative-a process of bargaining and negotiation. Focusing on gender roles as well as gender differences in access to and control over resources, it also shows the web of relationships, including class, ethnicity, age, religion, etc, in which gender is embedded. It locates women's subordination within the process of market-led production and distribution and argues that women are not left out of the development process but integrated into that process on unequal terms. It also looks at the infrastructure needed for the process of women's empowerment to take place. The body of work known as feminist economics examines gender hierarchies in production and reproduction, focusing on macro-economic policy and legislative reform. Neo-classical feminist economics looks at how gender biases undermine structural adjustment policies by causing the misallocation of resources between export and non-export sectors, preventing women from participating on equal terms in the production and sale of exports. Policy recommendations focus on improving women's ability to participate in different markets while assuming that women's work in the social reproduction of the labour force will continue. Feminist critical economics argues that women's role in social reproduction is not 'natural,' but a function of power relations. It examines the link between economic and human resource indicators and the need to redefine efficiency to include unpaid resource use in the social reproduction and maintenance of human resources. It looks at the politics of development and who controls it, emphasizing the role of women's constituencies in promoting gender-aware development policy. Several themes emerge from this review. First: the institutional focus of gender training frameworks has widened over time-from the household to other institutions through which gender inequality is reproduced. Second, the type of development interventions the frameworks target has moved from the project level to sectoral and macro-economic policies. And finally, there is an ongoing tension between so-called 'integrationist' and 'agenda-setting' approaches to development and how the training frameworks implicitly or explicitly embrace one or the other of these. Finally, the extent to which the frameworks view development institutions themselves as 'gendered' has changed over time, raising issues of organizational change for each of them. This study offers a comparative analysis of alternative paradigms, or conceptual frameworks used for analyzing gender issues within the development context. Some of these frameworks have been translated into gender-training modules which are being used by a wide spectrum of development organizations (donor agencies, Southern states and non-governmental organizations) to provide their staff with gender analytical skills. Others are drawn from selected writings in the field of gender and development (e.g., the gender dimensions of structural adjustment policies). While the latter have not yet been distilled into distinct training modules they may nevertheless influence the ways of thinking and 'doing' of development practitioners. The study seeks to evaluate the strengths and weaknesses of alternative paradigms. To do this, it examines various aspects of social reality (e.g., co-operation between genders; price incentives; gender conflict; social relations of production) which are highlighted in each approach, and those that are left out. Our assessment is based on four essential dimensions. The first dimension concerns each framework's understanding of 'gender'; in other words, aspects of gender relations that are emphasized, and those that are left out. In general, most approaches include a discussion of gender-differentiated roles-a descriptive account of what men and women do. In fact, for some, gender analysis is fundamentally a matter of disaggregation. Less prominent is a focus on the relations of subordination and domination that underpin gender as a power relationship. The second dimension relates to each framework's approach to development-whether market-led growth is seen as a process that is essentially benign or as one that is polarizing and socially unsustainable. The first approach highlights the 'constraints' that need to be removed in order to ensure women's active participation in the process of growth initiated by liberalization policies, and is less concerned about changing the content of those policies. Based on a different reading of social and economic forces, the second approach emphasizes the need to re-think economic policies (and agendas) and to restructure resource allocations along lines that will place human well-being before production and monetary targets (e.g. growth, profits, accumulation of wealth, etc). From a theoretical point of view the former approach is very often based on a liberal understanding of the world (wherein the rational individual with pre-determined preferences and tastes exists prior to society), while the latter takes a more structuralist view of society and economy (class structures, social relations). Jahan (1995) refers to the distinction between these two approaches as 'integrationist' versus 'agenda setting', or transformatory. The third dimension concerns the extent to which each framework can offer insights into different institutions-household, community, markets, states, civil society. This is an important aspect of the assessment because much of the early gender and development (GAD) literature focused on the household, where gender relations are most visible. This is no doubt an important arena, given the pervasive inequalities that mark intra-household gender relations. But recent feminist scholarship also shows the extent to which gender hierarchies are reproduced through the workings of other institutions-markets, firms and state institutions being some of the more critical sites. It is therefore important that gender analysis extends beyond the confines of the household. The final dimension of this assessment is the level of development intervention at which each framework is pitched: projects; programmes; sectoral and macro-policies. Again, as various critics have pointed out, much of the pioneering work on gender in the development field was based on insightful evaluations of development projects. These early studies of project 'misbehaviour' were critical in showing the gender biases that underpin development practice. But in response to the developments of the 1980s-in particular, the negative outcomes for women of structural adjustment policies-it became clear that it was equally important to scrutinize macro-economic and sectoral development interventions from a gender perspective with a view to removing the gender biases that inform them. It should be noted that the literature from which we draw our frameworks is not homogeneous: some, such as the gender roles framework, are incorporated in existing gender training modules, while others, such as 'gender and adjustment', are more diffuse, based on our reading of a set of analytical writings which share a number of basic premises. This allows us to assess the relative merits of some new approaches to gender analysis, not yet translated into concrete training modules, which may provide the basis for creating new and innovative training modules. A final point of comparison among these frameworks is the entry points through which they attempt to facilitate change: technical or personal. As will become clear, most of the training frameworks reviewed here introduce gender as a technical and/or analytical issue-taking a 'problem-solving' approach. The trainee is thus encouraged to apply his/her new skills to his/her professional tasks-such as project planning, for example, or policy analysis. Given the nature of the subject, however, questions are periodically raised about issues that will affect the planner in a personal way-gender relations within the household, or gender bias within development institutions (see Kelleher et al. 1996; Goetz 1992; Staudt 1990; IDS Bulletin 1995). Less prominent within development institutions are approaches that introduce gender as a 'personal' issue-encouraging the trainee to re-think his/her attitudes, experiences and social behaviour. Some of the gender-training modules developed in the South take 'the personal' as their entry point. Their marginalization from mainstream development institutions may in part be a reflection of the threatening nature of their message (an issue that touches on the persona of the male planner). But it may also reflect the difficulty of applying what is learnt by the trainee to his/her professional day-to-day tasks-drawing up project proposals or commenting on macro- or sectoral policies. The gender roles framework (GRF) derives from insights and concerns of the early WID approach, in particular the above-mentioned project appraisals.1 Developed by researchers at the Harvard Institute of International Development in collaboration with the Women in Development Office of USAID, the framework has become widely used in other mainstream development institutions. Its theoretical underpinning is provided by 'sex role theory' that informs liberal feminism (Connell 1987). Many of the shortcomings that critics have identified in the gender roles framework are thus a reiteration of those directed at sex role theory. The framework takes as its starting point the view that the household is not an undifferentiated unit with a common production and consumption function. Households are seen as systems of resource allocation themselves (Sims Feldstein and Poats 1989:10). Gender equity is defined in terms of individual access to and control over resources; women's (actual and potential) productive contributions provide the rationale for allocating resources to them. Gender equity and economic efficiency are thus synergistic. GRF: Conceptualization of Gender What does gender analysis entail? 'Gender analysis' is described as a diagnostic tool for planners to overcome inefficient resource allocation (see Overholt et al. 1985; Sims Feldstein and Poats 1989). While GRF focuses on what men and women do-the gender division of labour-it does not restrict itself to a discussion of activities/roles alone. It also highlights gender differences in access to and control over income and resources. It then considers the implications of these divisions and differences for project design. In sum, it aims to highlight the key differences between the incentives and constraints under which men and women work. The insights gained from this analysis are then used for tailoring planned interventions (credit, education, training, etc.) in such a way as to improve overall productivity. An example of the GRF approach is Cloud's (1985) analytical framework for agricultural projects. The first step involves developing an activity profile for the individuals performing different productive activities by asking an open-ended series of questions about the division of labour within the household. The second step overlaps with the first but places more focus on access to and control over resources (land, technology, labour, capital, etc.) and benefits (income, assets, etc.). In each phase, planners are encouraged to ask questions about the impact of project components on women's time availability as well as on their access to and control over productive resources and benefits. If, for example, a project places new labour demands on adult women, planners are encouraged to ask about how that will impact on their existing productive and reproductive activities, their leisure time, and on the responsibilities of other household members. Might daughters, for example, be withdrawn from school in order to take over their mothers' reproductive responsibilities as a result of adult women's increased agricultural work burden? The emphasis on women's time constraints also figures prominently in the Development Planning Unit (DPU) framework, which is also based on the concept of gender roles. The gender roles framework thus provides important data on the distribution of roles and resources within the household. The systematic enquiry into men's and women's activities attempts to overcome the ideologies and stereotypes that render invisible women's work. Armed with such information, planners and policy-makers are in a position to avoid some of the mistakes that resulted in project failures. While the attempt to differentiate activities and resources along gender lines thus goes a long way in meeting the need for a gender-sensitive planning methodology, the framework falls short in several respects, most of which stem from the way in which the gender division of labour is conceptualized. What important gender issues are left out? vBy treating the gender division of labour primarily as a relationship of separation, the gender roles framework tends to neglect its 'social connectedness' (Kabeer 1992: 14). While the gender division of labour involves men and women undertaking different activities, it also entails an intricate and changing system of co-operation and exchange - one that is potentially conflictual. In fact, the allocation of responsibilities for household maintenance is as important a family process as the allocation of resources. Even in the sub-Saharan African context where women and men engage in a limited degree of independent farming, women's outputs have often been seen as a source of accumulation, as well as a buffer for fluctuations in men's incomes (Guyer 1988). Moreover, while the management of responsibilities has tended to be gender-specific (e.g., men pay for children's school fees while women buy foodstuffs), specialization has never been complete. Rather, responsibilities tend to oscillate according to each sex's ability to cope with its own sphere, and its ability either to tap into the other or to shift the responsibilities. For example, where the revenues from men's cash crops have dropped, women have had to intensify their productive activities (e.g., beer brewing, trading) to assume many of men's traditional responsibilities. Conversely, women's enhanced earning capacity very often means that they will end up making a more significant contribution to the household budget as men's contributions are re-directed to other uses (personal consumption, productive investment, etc.). By neglecting the concrete relations between men and women, the framework fails to raise questions about how change is brought about in men's and women's roles in production and in the division of responsibilities between them.2 vThe GRF makes a useful distinction between access to resources and control over them (Overholt et al. 1985), directing the planner's attention to issues of power and control as an important aspect of gender relations. In so doing, it raises a question about development interventions which seek to direct resources to women (e.g., credit schemes): namely, will women be able to exercise any meaningful control over them?3 However, by restricting its analysis of power to tangible resources (e.g., land, credit, cash), the GRF gender analysis fails to highlight the role of 'intangible' resources-contacts, information, political clout-which also impact on power relations. Often women's extra-household relations-through informal associations or grassroots organizations that pursue empowerment-strengthen their authority over decisions that affect their lives (vis-à-vis their husbands, vis-à-vis employers). These 'intangible' resources are important to take into account if development interventions are to be informed by a full understanding of the factors contributing to gender inequalities in power. GRF: View of Development The key assumption underpinning GRF is that women have not been able to benefit from 'development' because planning efforts have failed to recognize fully women's actual and potential contribution to the development process. Logically then, if planners' prejudices (or ignorance) are rectified and women are 'integrated' into development interventions, they-like men-can benefit from the process. GRF thus views development as a process that is inherently benign-one from which everyone can benefit.4 Moreover, according to GRF, if gender issues are correctly identified and incorporated within development interventions, then project effectiveness will increase; 'both economic growth and social justice call for increased attention to the integration of women into the development process' (Overholt et al. 1985:4). There are two main problems with this approach. First, the assumption that economic development has been generally beneficial, except for its negative impact on women, ignores the wealth of evidence documenting the polarizing outcomes of development interventions for men. The introduction of new technologies, for example, has often worked to the advantage of some groups-frequently those with assets and political clout-while alienating others from their sources of livelihood. Similarly, the intensification of commercial agriculture has often increased class differentiation; in many contexts the small farmers have been squeezed by rising input prices and land alienation. It is typically the more affluent social groups-those with secure access to land, irrigation water and marketing outlets as well as political contacts-who have been in a position to reap the benefits of growth. A better recognition of these social processes would cast some doubt on GRF's optimistic assumption of the benefits of 'integration' for women. Linked to this first pitfall is the tendency of this framework to overlook the complex ways in which gender and class hierarchies cross-cut each other. GRF often assumes that women constitute a homogeneous category with shared interests which planners can easily identify and act upon. While there are a large number of issues on which women's interests may converge (e.g., child custody; sexual harassment), there are enough points of divergence to make the task of the planner using GRF a difficult one. There is in fact very little in the GRF literature which prepares the planner for such conflicts of interest. A good example is the introduction of post-harvest rice milling technology in Bangladesh which eased the unpaid work burdens of female family members, but in so doing jeopardized one of the few employment opportunities open to women from landless households seeking wage labour (Whitehead 1985). Finally, the above example points to another major oversight in GRF: the neglect of political factors underpinning policy-making. Will planners alter their policy priorities (e.g., curtail the spread of electric rice mills) if they have accurate information (e.g., that a significant number of poor, landless women are losing their only employment option)? There is nothing in GRF to explain the political forces informing policy-making, since the entire policy process is explained as a rational, technical exercise executed by 'essentially benign and neutral agents' (Kabeer 1994a: 290). Landless female labourers have very little political weight in most national contexts (even if they are organized, which they rarely are). It is thus very unlikely that they can exert enough pressure on policy-makers to shape policies in a way that meets their interests. GRF: Institutional Scope The analytical and conceptual focus of GRF is 'intra-household dynamics'-the gender division of labour and gender differences in access to and control over resources. Underpinning the understanding of the household is a bargaining model which offsets the bias towards concepts of a unitary household and male heads as sole decision-makers and sources of information. While Cloud's (1985) more detailed formulation of the GRF framework makes reference to gender bias in land, capital, technology, and wage markets, these tend to be ad hoc or circular, as for example (p. 39):
In the context of wage labour markets, a brief reference is made to 'distinct male and female segments' (p. 41), but no explanation is offered as to how these distinct segments are created and maintained. Issues of power and control which were drawn upon to explain the intra-household inequalities do not seem to enter the domain of 'impersonal' markets. The attempt to tackle the issue of gender bias within markets lacks a clear analytical framework and conceptual underpinning. Similarly, GRF makes no attempt to understand the gendered nature of development institutions. As noted, the planner is assumed to be a rational individual-'essentially benign and neutral'. Planning and policy-making institutions (governments, international agencies) remain outside the framework and no reference is made to the complex ways in which male bias has become institutionalized within these organizations (IDS Bulletin 1995). Nor is any emphasis placed on the need to change the working procedures and organizational goals and mandates of these institutions to make them more gender-egalitarian. The problem of male bias in policy-making is essentially one of error, lack of information, or short-sightedness. In short, the GRF framework offers no guidelines or analytical tools for extending the gender analysis from the household to other institutions. As a result, issues of gender bias and female subordination become almost synonymous with the household, while other institutions escape scrutiny. GRF: Development Interventions The gender roles framework is distinctly pitched at the project level. The attempt to put together this framework was in fact a direct response to project failures extensively documented in critical WID evaluations of development projects. The task that GRF set out to accomplish was to provide project staff with 'a new set of conceptual and analytical perspectives and skills in order to deal explicitly, effectively and efficiently with women-related issues in the spectrum of projects in which they become involved' (Overholt 1985:xii, our emphasis). One of the main weaknesses of the GRF, highlighted by development agencies that have used the framework, is its focus on community-based projects. According to CIDA, 'the insights and tools that GRF offered were difficult to adapt to the policy and program level' (Woroniuk 1995:14). While this may have been true, several analyses of gender and macro-economic policy (e.g. Collier 1989; Palmer 1991) have been developed using an understanding of gender very similar to the GRF approach. In these models gender analysis is essentially about disaggregating existing macro-economic variables by gender: women, for example, specialize in non-tradables while men specialize in tradables (activities which are being prioritized under adjustment). The disaggregation is meant to highlight the way in which labour reallocation problems, caused by rigidities in gender roles, can constrain adjustment policies (see Feminist Economics below). GRF: Applications Organizations which have used the GRF methodology or have had training in the methodology include: USAID, CIDA, UNDP, IDRC and the World Bank. Illustrations of GRF can be found in the Kumarian Press handbooks, which draw on USAID-funded projects and have been used for training by USAID and other agencies: Gender Roles in Develop-ment Projects (Overholt et al. 1985), and Working Together: Gender Analysis in Agriculture (Sims Feldstein and Poats 1989). Moreover, elements of GRF, such as its access and control profile, have been taken up by agencies which do not embrace the entire training framework (UNICEF 1994). Organizations which have relied from the beginning on GRF, such as USAID and CIDA, have in recent years tried to expand the analytical scope of their training modules beyond the project, endeavouring to provide practitioners with tools that can facilitate gender analysis at the programme and policy levels. As in the case of GRF, planning is seen as an essentially rational, top-down process; equipped with information (both quantitative and qualitative), planners (disembodied, rational agents) are in a better position to design and implement development interventions. In the case of USAID's Gender Analysis Tool Kit (1994) there is no mention of participatory methods of data gathering and development planning. USAID's Gender Analysis Tool Kit includes several tools that help practitioners analyze national-level quantitative data in order to identify 'relevant gender issues that have an impact on development goals'. Gender analysis is essentially about disaggregating national data sets by gender (e.g., sex ratios, life expectancy, infant/child mortality), and analyzing their possible implications for economic growth, human resource development, population and health policies. Within this framework gender issues impact development goals in three ways: a) to contradict one or more goals; b) to act as a constraint on the achievement of one or more goals; or c) to act as a catalyst to help achieve one or more goals. For example, fertility reduction is identified as a primary goal. Thus national demographic and health surveys are analyzed (disaggregating average ideal family size of husband and wife, decision-making powers of husband and wife) in order to devise policies that can make family-planning strategies more effective. The emphasis on disaggregating data by gender parallels GRF's emphasis on disaggregating activities and access/control by gender. In both cases gender analysis is meant to facilitate the achievement of pre-determined development goals and priorities. In both cases gender issues become relevant to development if they are potentially affected by development interventions, or could act as either constraints or catalysts in the accomplishment of development goals. The DPU Framework (Gender Planning) The DPU Framework, so-called because it was developed by Carolyn Moser at the Department of Planning Unit at London University, is widely used within development institutions. While an analysis of women's triple roles constitutes its key analytical building block, any assessment also must include a consideration of its two other components: practical and strategic gender needs; and the schematization of policy approaches to women in the development context. The lack of an overarching conceptual/theoretical framework to integrate its different components renders the training module something akin to an eclectic 'tool kit' rather than a coherent framework. Nevertheless, what gives the tool kit some unity is its political message: the need to use development interventions to transform unequal gender relations. The DPU Framework, like GRF, takes as its starting point the fact that roles are gender-differentiated. In order to highlight the multiple demands on women's time in low-income households, distinctions are made among women's productive, reproductive and community managing roles-hence, triple roles. By using categories that are easy to understand and employ (if not analytically robust), Moser provides a useful mental grid that enables planners to include all that women do-even if the activity is 'invisible' because it is not valued in the market place (e.g., reproductive work) or because it is not culturally accepted (e.g., productive work). The framework's second major component is the distinction between practical and strategic gender needs. This distinction draws the attention of the practitioner to the transformatory potential of development planning. By legitimizing the use of development interventions as a means of transforming unequal gender relations (reducing women's subordination), it highlights the political (rather than technical) nature of the planning exercise. The third component of the DPU framework is a schematization of policy approaches to women (welfare, equity, poverty, efficiency and empowerment). Despite its various shortcomings, one of the main strengths of this categorization is that it makes explicit the extent to which different approaches prioritize the amelioration of women's subordination as a policy objective-a political project involving the transformation of power relations.
Women's triple roles Moser (1989; 1993) developed a 'triple roles' schema to draw planners' attention to the way in which the gender division of labour places a range of demands on women's time and impacts their ability to participate in planned interventions. The 'productive role' comprises market production and subsistence/home production, both of which generate an income (in cash or kind) and have an exchange value. The 'reproductive role' refers to the childbearing/rearing responsibilities and domestic tasks done by women that are required to reproduce and maintain the labour force. The 'community management role' involves women's activities at the community level, as an extension of their reproductive role, in order to ensure the provision and maintenance of collective consumption resources (e.g., water, health care, education).5 The triple roles schema claims to be more than a description of what men and women do. Deriving from an extensive literature and research on gender relations, the framework takes the view that the gender division of labour embodies and perpetuates women's subordination (Moser 1993:28). While women and men within the household engage in all three areas of activities (productive, reproductive and community management), the division of labour is rarely even. Although women are also involved in productive work, reproductive and community management work falls heavily on them because of their gender-ascribed roles. Whereas work carried out by men is valued, either directly or indirectly through paid remuneration, increased status or political power, women's community management and reproductive roles are not valued because they are seen as 'natural' and non-productive. This means that a large part of women's work is made invisible, unrecognized by men in the community and by planners who are assigned the task of assessing women's needs. By providing a relatively simple schema which draws attention to multiple roles of women, the framework enables planners to understand that women's needs are often different from men's, and that meeting those needs may entail specific planning requirements. The triple roles schema goes beyond GRF's focus on the household to recognize the way in which many aspects of household survival and reproduction are located in the community. This is a particularly important and timely insight. With cutbacks in state provision for basic social services such as water and health, it is often women who take on the responsibility at the community level for the allocation of scarce resources to ensure household survival. It is often also women who put pressure on local or state institutions for the provision of infrastructure. However, by focusing on roles, the triple roles schema, like GRF, appears to be an imprecise tool for illuminating gender relations. As regards productive and reproductive roles, the emphasis in the framework remains on what women do and what they produce (income and human resources). It is only with the definition of community roles that the framework begins to draw attention to 'intangible resources' (claims, status, authority) and the social relations through which such resources are produced (community rather than household based). Even so, as Kabeer (1994a:278) points out, the triple role schema pays insufficient attention to 'the fact that most resources can be produced in a variety of institutional locations (households, markets, states and communities) so that the same resources may be produced through very different social relations'. The point here is that it is not just what women do, but how they do it-through which social relations and institutional contexts-that should inform the planning process. For instance, women's reproductive work has different implications in terms of access to and control over resources when it is performed as unpaid family labour than when it is performed in community-based associations. Moreover, since the schema looks at social relations only in connection with community roles, it does not capture how relations of power and authority operate at the household level (or in other institutional contexts). There is some evidence to suggest that women's bargaining power within the household (as in the community) is strengthened through their participation in extra-household networks, which a focus on roles rather than relations fails to illuminate.6 Practical gender needs and strategic gender needs Another important part of the DPU framework's conceptualization of gender is the distinction made between practical and strategic gender needs. According to Moser (1989:1803) practical gender needs are those needs deriving from the concrete conditions women experience based on their position within the gender division of labour. Practical gender needs thus arise and are articulated by women themselves in response to some immediate perceived need, based on the existing gender division of labour, for such basic items as food, shelter, health care and water. Such needs relate to women's triple roles in that they arise out of the multiple roles assigned to women through the gender division of labour (food provisioning, child care, community management of basic services, etc.). In contrast, strategic gender needs refer to those needs deriving from an analysis of women's subordination to men and the formulation of an alternative, more equal organization of society. Examples include: abolition of the sexual division of labour, establishment of political and economic equality, freedom of choice over childbearing, an end to male violence against women, etc. As Moser points out (1989:1804), meeting women's practical gender needs may help to lessen the burdens on women deriving from their triple roles, but it does not challenge the gender division of labour. Because planners identify and respond primarily to practical gender needs, they are often responsible for preserving and reinforcing existing gender inequalities. The distinction between practical and strategic needs can help planners to think about more challenging interventions. For example, whereas the provision of dressmaking training can meet a practical gender need for employment, training women in areas traditionally identified as 'men's work' can help to break down existing occupational segregation, a strategic gender need. While this conceptualization of practical and strategic gender needs moves beyond the GRF in suggesting possibilities for a more transformatory planning agenda, it nevertheless tends to downplay the conflictual aspect of gender relations.7 Men, too, have practical and strategic gender needs, often opposed to those of women, which may account for male resistance to transformatory efforts. As noted below, such concerns have been incorporated into some of the training packages built around the DPU approach. DPU Framework: View of Development It is difficult to categorize the way in which the DPU framework views development, since it seems to draw neither on a liberal reading of society and economy-as does the gender roles framework-nor on a structuralist understanding of social processes and social change-as in the social relations framework. However, a number of assumptions seem to be implicit in the DPU framework, which suggest a 'populist' and eclectic approach to development. These are: v growth fails to 'trickle down' to the poor; in the name of 'efficiency', structural adjustment policies have imposed severe social costs on the poor v 'poor' and 'low-income' households are those development should serve v the reproductive economy plays a major role in achieving human well-being v bottom-up and participatory planning are beneficial to development interventions. There is no discussion of rules, resources, practices and hierarchies of command and control that constitute households, markets and development agencies as social institutions through which social inequalities are constructed and reproduced. The discussions of poverty and subordination are thus somewhat 'abstracted' from the institutional contexts in which they take place. To some extent, as critics have pointed out, the framework's failure to provide an understanding of the social processes through which women experience subordination and poverty can be attributed to the framework's preoccupation with roles (as opposed to social relations). Similarly, the framework's attempt to differentiate social groups seems to lack analytical rigour. The category 'low-income households'-identified as the target group for development interventions-is not well-defined. The emphasis on 'income' as the criteria for identifying the poor suggests a conventional market approach to poverty, which contains both measurement and institutional biases. Yet, the criticism of structural adjustment policies (that they shift costs from the paid to the unpaid economy) suggests an awareness of the limitations of market-led growth (Moser 1993:70). Finally, the emphasis on 'women in low-income households' also reflects continuing gender bias; as feminist scholarship has pointed out, women's well-being cannot be easily read off the overall status of the households to which they belong. DPU Framework: Institutional Scope The focus of this framework is at the household, and to some extent, the community level, with little analysis of the institutions through which gender disadvantage is perpetuated. The discussion of strategic gender needs within state institutions is limited to pointing out that the state has failed to meet women's strategic gender interests: no analysis of the gendered nature of the state is developed. Similarly, while the recognition that gender power relations impact on how development institutions operate is implicit in the DPU framework (see Moser 1993:180; Levy 1991), there is no attempt to analyze the gendered nature of development institutions themselves. The description of the different policy approaches to women (welfare, equity, anti-poverty, efficiency and empowerment), however, draws attention to the fact that different types of policies have appealed to different types of institutions, a point that other analyses have explored further (see Kardam 1991; Razavi and Miller 1995). These gaps can be attributed in part to the conceptualization of the DPU framework. According to Moser (1993:176), its basis is gender planning, rather than gender analysis. Methodological tools are constructed around the language and procedures of planners. Despite the assertion that the success of gender planning depends on the participation of organized women's groups, the framework says little about how planning processes can be made more participatory or how a process of 'negotiated debate' can be initiated. Overall, the focus remains on 'top-down' interventions, albeit more gender-responsive. More recently, however, efforts have been made to translate DPU's commitment to participatory gender planning into diagnostic and operational tools. The aim is to help ensure that a cross-section of community groups or 'political constituencies' is represented in the formulation, implementation and evaluation of policies, programmes and projects (see Levy 1996). Unlike some of the training approaches based upon it, the DPU framework distinguishes among the 'professional', the 'personal' and the 'political', describing gender planning training as a 'professional' activity. It focuses on gender training as a means to acquire 'technical skills' needed to improve professional competence. Overall, the approach and language are designed to diffuse hostility of those who find 'feminism' an unacceptable encroachment on 'objective' planning processes (the framework's focus on 'low-income' women can also be seen in this light).8 But through the different tools, particularly the distinction between practical and strategic gen der needs, the framework also hopes to initiate a re-examination of 'personal' as well as 'political' positions on the issue of women's subordination. DPU Framework: Development Interventions Like the GRF, the DPU framework focuses primarily on project level interventions; however, it goes beyond the GRF in looking at sectoral level interventions as well. Through the case study method, for example, the DPU framework demonstrates that by collecting data on the basis of the triple role schema and distinguishing between practical gender needs and strategic gender needs, it is possible to identify the planning needs of women in various sectors, including transport, employment and training, and housing (see Levy 1992; Moser 1993). DPU Framework: Applications The DPU methodology has been used by the African Technical Department of the World Bank, Christian Aid, Oxfam UK; ODA, NORAD, FAO, ILO, UNIFEM, Sida and CUSO. Two examples of adapting the framework to meet the needs of specific organizations are the Oxfam Gender Training Manual and Gender Training Analysis, by Save the Children, America. The Oxfam Gender Training Manual. Like many of the training packages developed in recent years, the Oxfam Gender Training Manual (Williams 1994) is eclectic, combining a range of training approaches. Although essentially an adaptation of the DPU gender-planning methodology, it contains a number of distinctive features. The discussion of practical and strategic gender needs raises questions about the gender needs and interest of men: Is it in their strategic interests to maintain the status quo and their position of dominance? Or is it more strategic to work for a more egalitarian society, for the empowerment of women, for more balanced relationships, etc.? Such questions help to draw attention to gender conflict (as well as co-operation) and to the need for changing attitudes at the personal level. In fact, the structure of the handbook moves from 'gender awareness and self-awareness' activities designed to promote a re-thinking of personal attitudes and behaviour to activities designed to improve planners' professional analytical skills in project formulation and monitoring. The Oxfam approach seeks to concretize the goal of 'participatory', 'bottom-up' development. Examples and case studies are drawn from the range of development and humanitarian relief activities in which the organization and its affiliates participate.9 Its audience is seen as non-governmental organizations (NGOs) and grassroots organizations, which also impacts on its methodology. Thus the manual emphasizes participation in project formulation and monitoring. Analyses of data-collection methods, such as Participatory Research Analysis (PRA), stress that data gathering must be gender-sensitive to be truly participatory. Similarly, a component on working with men and women is designed primarily to help NGOs facilitate women's involvement in needs assessment at the grassroots level. Implicit in the activities is an understanding of gender power relations which make it difficult for women to articulate their needs. Gender Relations Analysis, Save the Children, America. The Gender Relations Analysis Guide (Parker et al. 1995) also emphasizes the 'personal' and the importance of challenging attitudes and stereotypes about gender differences held by decision-makers within organizations. The main objective appears to be overall gender awareness sensitization. The concept of women's triple roles and the distinction between practical and strategic needs are seen as tools to collect gender-disaggregated data and to develop more gender-sensitive projects. Although the training model does not assume a project context, and takes as one of its guiding principles the goal of the equal capacity of women and men to influence and change current economic, political and social norms, the gender-sensitization activities in the training package provide few tools to understand how gender inequalities are constituted and reproduced within institutions beyond the household and community levels. Added to the Gender Relations Analysis is a component on the conceptualization of power. Following CIDA's four-fold classification of power, the aim to is move from an understanding of power as a zero-sum game to non-conflictual understandings, including power to (empowerment at the individual level), power with (collective use of power), and power within (internal strength). Yet unlike the Oxfam manual, the package makes no reference to the hostility and resistance (from both men and women) likely to be encountered in challenging existing gender roles and relations. The training provides guidelines on evaluating projects to promote men's and women's knowledge and capacity to negotiate for rights and to change existing decision-making patterns, but there is no mention of how beneficiaries can be brought into the project formulation process and how that process can be made more participatory, nor does there appear to be provision for advocacy training for grassroots women and men. The objective is to enable technical and management practitioners to identify the needs of beneficiaries. The social relations framework (SRF) describes an analytical approach developed at the Institute for Development Studies, Sussex (UK), derived from social relations analysis. Many of its assumptions appear to have come out of the Subordination of Women (SOW) Workshop in the mid-1970s, which criticized the Women in Development (WID) approach on several grounds: rooted in liberal individualism it tended to isolate women as a separate and often homogeneous category, it was 'predominantly descriptive', as well as 'equivocal in its identification and analysis of women's subordination' (Pearson et al. 1981:x). The SOW Workshop set out an alternative approach, informed by a Marxist analysis of social change (without succumbing to the structural determinism of many Marxist accounts) and a feminist analysis of patriarchy (Young 1993:134). Its work on the social relations of gender, along with the scholarship and practice of feminists from developing and developed countries, provided the conceptual framework for what is often referred to as the Gender and Development (GAD) approach. The social relations framework is constructed around many of the insights of this work. It takes into account both the relations of production and the interrelated range of relations through which human needs are met, 'the social relations of everyday life' (Pearson et al. 1981: x). This ensemble of relations governs the processes of production, reproduction, distribution and consumption, and operates through a broad range of institutions: the household, the community, the market and the state (Kabeer 1994a). Gender relations refer specifically to those dimensions of social relations that create and reproduce differences in the positioning of men and women in social processes. In this perspective, gender is always intertwined with other social relations: gender, class, ethnicity, age, religion, caste, etc. cross-cut one another. Thus 'a gender analysis must be embedded within a broader social-relations framework' (Kabeer 1994a). This means that in any given context, different axes of inequality assume different levels of analytical importance. Not only must similarities and differences among women be taken into consideration; in some contexts gender may not be the key factor of analysis (see Goetz 1989).
By drawing attention to the range of social processes and institutions through which social inequalities are constructed and reproduced, the SRF avoids 'abstracting' women from the institutional contexts in which they experience their lives. From the perspective of the SRF, the rules, resources, practices and hierarchies of command and control of any institution must be analyzed in order to uncover how gender is constituted as a relationship of inequality within it (see Kabeer 1994a:87). Like the gender roles framework, such an analysis highlights gender-differentiated roles and responsibilities as well as gender divisions in access to and control over resources. However, several assumptions distinguish the SRF from the gender roles framework. First, whereas the gender roles framework conceptualizes the gender division of labour as essentially a form of social separation (assigning women and men different roles and responsibilities), the SRF sees the gender division of labour as a form of social connection, making it essential for women and men to engage in relationships of co-operation and exchange (Young 1993:140; Kabeer 1992:14; Whitehead n.d.). This interdependence is not, however, symmetrical. The unequal division of resources and responsibilities means that gender relations involve conflict as well as co-operation (Kabeer 1992:19). Second, and related to the first, because conventional economic planning considers primarily individualized production and material resources, it undervalues the more fluid social or 'relational' resources such as rights, obligations and claims. The 'moral economy' often involves relationships that entail loss of status and autonomy in exchange for some measure of security (Kabeer 1992:12). For women, the gender relations of family and kinship frequently embody a trade-off between security and autonomy. Within the SRF, then, culturally constructed rules about the different capacities and aptitudes of men and women not only determine gender differentiated roles and responsibilities, they also underpin their differentiated claims, rights and obligations. In order to understand women's economic decision-making patterns and their use of labour time, it is necessary to determine how women perceive their interests and to consider how these relate to their position within the family and the household-insights that cannot be read off a disaggregation of the gender division of roles (Whitehead 1990). Where resources are directed to women, for example, an effort must be made to consider whether, in subverting existing gender relations, targeted interventions will ultimately enhance women's status or, by undermining certain familial or community rights to which they were traditionally entitled, place them in a more vulnerable position. Third, rather than downplaying the 'political' dimension of gender, the SRF brings it to the core of its analysis of gender relations, framing the relationship as one of male domination and female subordination. In other words, the unequal distribution of (tangible and intangible) resources and responsibilities within different institutional contexts and the cultural rules that uphold inequalities in distribution mean that men have more authority and control than women and thereby a greater capacity to mobilize a variety of economic and social resources in pursuit of their own interests. As such, ending women's subordination is more than a matter of reallocating economic resources. It involves redistributing power. Following from this, the SRF tends to characterize the redistributive process as a zero sum game (Kabeer 1994a:97). Men will have to relinquish some of their economic, political and social power for women to gain power. Finally, the SRF takes a dynamic view of gender relations, recognizing that the conflictual and collaborative aspects of gender relations involve men and women in a constant process of bargaining and negotiation. This view implies potential for transformation where those with a stake in change are prepared to challenge existing gender relations. Thus, in addition to supporting the view that information about existing gender roles and relations is an important prerequisite for gender-aware policy and planning, proponents of the SRF take a normative view of the need to change existing gender relations. SRF: View of Development Analyses put forward by the SOW in the 1970s provide a framework for understanding the view of development which is implicit in the SRF. Unlike WID scholarship, which characterized women as being left out of thedevelopment process, according to members of the SOW 'it is precisely the relati ons through which women are "integrated" into the development process which need to be problematized and investigated' (Elson and Pearson 1981:19). In other words, this approach challenged the benign, universalistic, vision of development inherent in the WID approach and attempted to locate gender relations and women's subordination within the processes unleashed by market-led development. Through a structuralist analysis of class and gender relations, attention was drawn to the highly varied, and often polarizing, implications of development, especially in the context of the spread of the global market economy. For example, building on a Marxist analysis of women's position as a reserve army of labour, such an approach pointed to the exploitative nature of the new international division of labour, whereby women in developing countries were being recruited into multinational factories as cheap, docile workers. The analysis also highlighted the role women's unpaid work played in underpinning market-led development (through the production and maintenance of human life), despite the failure of traditional economic approaches to value reproductive work. In its view of development, SRF thus problematizes both the narrow focus on economic growth and the hierarchy of production over reproduction within mainstream development discourse. The framework takes the final goal of development to be the achievement of human well-being, which is taken to encompass certain basic goals: survival, security and autonomy. From this perspective, production is defined as encompassing all activities that produce the means by which these final goals are met, including the reproduction and care of human beings (Kabeer 1994a:280). The redistributive implications (in terms of power and resources) of achieving the final goal of development (defined as human well-being), and the likelihood of resistance to redistributive policies, are also made explicit in this approach. The logic of the framework entails a shift in terrain from the primarily technical concerns of policy makers and planners to the political arena per se. The centrality of the power dimension of gender relations brings the importance of women's 'empowerment' to the forefront (Kabeer 1992; Young 1993). From this perspective development interventions need to be designed in ways that have transformatory potential, in other words, that help build up the enabling infrastructures necessary for the process of women's self-empowerment (Kabeer 1994 a:301). One example given in this context is the provision of space, resources and time to allow women to articulate their own interests (through social action groups and grassroots movements) rather than having them identified and met on their behalf by planners. Much of this hinges on the openness of development institutions to the 'bottom-up' flow of information (participatory planning) and their receptivity to transformative ideas about gender-again raising the issue of institutional change implicit in the SRF. It should be noted, however, that the SRF also seems to reflect a certain 'agnosticism' about the market. The implication is that there is a need to evaluate the market both from the point of view of the possibilities and problems it raises for women, particularly for different groups of women-insights that a social relations analysis can help to illuminate. A case in point concerns interpretations of women's employment in export-oriented industries. While much feminist scholarship has focused on the exploitative nature of women's employment in export industries (in terms of wage discrimination, labour relations, working conditions, not to mention factors accounting for women's entry into wage labour), a recent study of the impact of women's wages on intra-household power relations in urban Bangladesh offers some interesting findings on the empowering potential of waged employment for young women garment workers, enabling them to renegotiate the terms ofunsatisfactory relationships or to walk out of, or not enter into, relationships which undermine their agency (Kabeer 1995:35). SRF: Institutional Scope In the SRF perspective, 'gender-awareness in policy and planning requires a prior analysis of the social relations of production within relevant institutions of family, market, state and community in order to understand how gender and other inequalities are created and reproduced through their separate and combined interactions' (Kabeer 1994a: 280). Since power relations between men and women are conceptualized as the product of institutional practice, the SRF takes the position that overcoming gender-based inequalities involves institutional transformation at all these levels. To take the example of credit programmes, an analysis of gender relations within different institutional contexts would show how supposedly gender-neutral policies of state and financial institutions, through collateral requirements and the need for literacy and time to understand complicated loan procedures, etc., make it difficult for poor women to get access to credit. Where institutional practice is changed to give women access to resources, a social relations approach would point to the complicated dynamics of control when women borrowers bring credit into the household: even where they are unable to retain control over resources allocated to them (Goetz and Sen Gupta 1994), there may be more indirect gains to their status and autonomy by virtue of their newly acquired loans (see Kabeer and Subrahmanian 1996). Another salient feature of the SRF institutional focus is that development agencies are seen as implicated in the reproduction of gender and other social inequalities (see Goetz 1992; 1994). Thus, not only do policy-makers need to be 'conscientized' to critical dimensions of gender relations and interests, but more importantly, unless women empower themselves to identify and establish their gender needs and interests as legitimate policy concerns within the bureaucratic planning process, these concerns are unlikely to be met. Again, this reflects the view that planning is a political as well as a technical process. SRF: Development Interventions The social relations framework suggests that all planning, from macro-economic policy to micro-level interventions, must be informed by an understanding of the social relations through which production is organized and human needs are met. As a planning tool, the SRF provides no 'quick fixes' for overcoming gender inequality, nor much guidance on the specific types of interventions that would promote gender equity. What it does provide is a framework for a sensitive reading of the intricate social relations through which women and men live their lives, the insights of which can aid development planners in tailoring interventions more appropriately. SRF: Applications As noted above, the social relations framework refers to a training approach developed by researchers and trainers at the Institute for Development Studies, the main features of which have been described in Kabeer (1992; 1994a and b) and Kabeer and Subrahmanian (1996, see also the IDS training modules entitled Gender and Third World Development.) Elements of the social relations framework have also been picked up by CIDA in its Gender and Development Briefing Module (Woroniuk 1995). While this analytical framework attempts to assist planners in developing policy interventions, it is a much more complex framework than training methodologies that focus primarily on data disaggregation or checklists of activities. Although a seven-point 'gender audit for development interventions' has been developed, presumably to be used at all levels of development interventions (from the macro to the micro), the audit itself is a good example of the challenge of capturing the analytical complexities of social relations analysis in a checklist: many of the questions parallel those that are used in the GRF and the DPU approach. Feminist Economics: Taking Gender Analysis to the Macro-Economic Level The body of work commonly referred to as 'feminist economics' is only just beginning to emerge and define itself as a discipline within economics (Cagatay et al. 1995). Some of those who define themselves as feminist economists, such as Ingrid Palmer, work within the neo-classical paradigm, while others, such as Diane Elson and Nancy Folbre, draw on a variety of critical approaches (structuralist, Marxist and political economy) which challenge the neo-classical perspective. Neo-Classical Feminist Economics: Gender and Adjustment The literature examined here is drawn from economic analyses of adjustment and gender (Collier 1989; Palmer 1991; 1992). The novelty of this genre lies in the way it conceptualizes gender using neo-classical tools of analysis (factor market rigidities, informational biases, market distortions), juxtaposed against an appreciation of intra-household inequalities and bargaining (similar to the GRF). More significantly, the analysis of gender relations is systematically linked to the design of structural adjustment programmes. Rather than a focus on how structural adjustment programmes have affected the welfare of women and children (see UNICEF's Adjustment with a Human Face), their concern is to show how gender biases and rigidities affect adjustment policies, and can ultimately frustrate them. 'Gender analysis' is thereby taken well beyond the project focus of the GRF and into the realm of macro-economic policy-making. Gender and Adjustment: the Models What is the main argument here? Adjustment, neo-classical economists would agree, is essentially about inducing the inter-sectoral flow of resources (especially labour and credit) from non-tradables to tradables in order to alter what are seen as original misallocations of resources between sectors.11 If resources, however, are not mobile between sectors, raising prices of tradables is not likely to bring about the desired flow ('supply response'). Palmer and Collier point out that resources (especially female labour) may be allocated between sectors in a skewed and inefficient manner due to various constraints which arise from gender roles and inequitable gender relations. The same constraints, they add, may reduce the mobility of resources between sectors, and leave them 'stuck' in the production of non-tradables, thereby reducing the desired supply response. Collier (1989:8) locates four distinct processes, based on underlying 'social conventions', that account for why women face differential constraints upon economic activity. The first is discrimination outside the household-in labour and credit markets. The second is that role models (in production) are gender-specific, i.e., girls copy women, while boys copy men. If a new economic opportunity is initially taken up by men, therefore, it may be diffused over the male population by a mechanism that will not transmit it to the female population. The third is that within the household rights and obligations between husband and wife are asymmetric, such that women have little incentive to increase their labour input. The final element is the 'burden of reproduction', with its attendant demands on women's health and time. Palmer (1991:11-15) identifies similar processes, although intra-household markets and bargaining are more prominent in her analysis, while 'role copying' is largely absent. Within households, she argues, at least in the African context, there are markets - albeit rigged ones. The social aspects of gender impose their own definitions of correct exchange, which tend to reflect bargaining power and status, and inevitably mean that the terms of trade are biased against women. This can be described as an asymmetry of obligations and responsibilities between women and men. What are the implications for public policy? According to Collier (1989:10), if private sector labour and credit markets and private processes of information dissemination all make it likely that women will be less mobile than men, existing public mechanisms should have an offsetting bias. Ifthese rigidities could be overcome through public interventions (e.g., by redressing the male bias in financial markets, in education and in extension services, and re-targeting payments made by public marketing channels), women would be able to participate more fully in the production of tradables, the booming sector under structural adjustment. This would reduce gender inequity, as well as reducing economic stagnation-hence the convergence between gender equity and economic efficiency. Palmer (1991:155-157) also recommends legislative and institutional reform of the public sector in the short to medium term. But she takes the efficiency argument a step further and questions the presumed efficiency of the smallholder that underpins adjustment policy support for the 'small farm' (via land titling and agricultural services). If male small-holders were to hire female labour, she asserts, their presumed superior efficiency would become questionable. Thus, encouraging alternative rural employment for women (e.g., in plantations) can be a desirable means of freeing their labour from family production where it is used lavishly, and of forcing a proper costing of labour on the family farm manager. Palmer's recommendations for achieving long-term dynamic efficiency also include opening up the non-biological elements of women's reproductive labour to market forces ('a calculus of costs'): 'tap water and electrification of homes means a paid workforce in public utilities. Creches mean professional, paid child-minders' (1991: 165). This, she argues, would not only reduce the amount of unpaid work women have to do, but would also increase efficiency in the reproduction of human capital. Gender and Adjustment: Conceptualization of Gender Gender and adjustment models share a number of key assumptions with the GRF in terms of their understanding of gender. v In both, an important aspect of gender analysis involves looking at what men and women do-gender-differentiated roles. Collier's model combines the disaggregation of agents by gender with a sectoral disaggregation of activities-between tradables and non-tradables-which is critical to structural adjustment. v There is a clear recognition in both that women's 'burden' of reproductive work acts as a constraint on their ability to engage in productive activities. The gender and adjustment literature, like the GRF, highlights constraints on women's time as an important factor that planners/policy-makers need to take into account if they want to achieve their desired policy objectives. As noted below, feminist economics offers quite a different interpretation of reproductive work. v Both models highlight intra-household gender differences in access to and control over resources and income. In both, these differences are the basis for analyzing the incentives and constraints under which men and women work; women, for example, may withhold labour from their husbands' agricultural enterprises because they have no access to the proceeds. v Both GRF and the gender and adjustment literature attempt to theorize gender using economic tools of analysis: bargaining models are used to explain the processes of intra-household decision-making for labour and resource allocation. The recognition of bargaining and conflict within the household, in particular, goes a long way in meeting the feminist demand that power relations be explicitly recognized in analyses of gender.
Gender and adjustment models share a number of problems in the way gender is conceptualized. v Collier's emphasis on role copying as an explanation of gender segmentation in the labour market (and the gender division of labour more generally) is problematic. Calling segmentation a voluntary process (girls copy women, boys copy men) provides a misleading picture of how these gender-differentiated roles (and segments of the labour market) are created and maintained; there is no mention, for example, of the ways in which the more powerful agents construct and maintain theboundaries between male and female sectors-sometimes using brute force. v Similarly, by explaining gender asymmetries in rights and obligations in terms of 'social convention' the model provides an extremely static picture. These asymmetries are not simply given by society, as the term 'social convention' seems to imply. They are in fact constantly created and re-created through the everyday relations between men and women; they are actively mobilized by those with power against those without power. Women's obligation to use their incomes/savings to feed the family in some social settings is not just a social given. It is through conjugal conflicts underpinned by power asymmetries that the more powerful household members (adult men) are able to shift these responsibilities onto women while they maintain personal control over other income flows. Gender and Adjustment: View of Development In general, according to this literature the policy agenda being promoted through structural adjustment programmes is inherently benign-one from which everyone can benefit. In other words, markets are a reliable means of mobilizing resources for production, and (implicitly) an effective means of meeting needs. In this regard, the approach adopted is, to use Jahan's (1995) terminology, 'integrationist'. There are obvious problems with this approach. v One of the main shortcomings of Collier's work is that it does not explicitly deal with the non-monetized, reproductive sphere, i.e., the social and material processes through which the human population is maintained and renewed on a daily and inter-generational basis. Reproductive work enters the equation only insofar as it constrains women's ability to contribute labour to the production of tradables. It therefore goes along with conventional macro-economic analyses which ignore the un-monetized sphere of human reproduction. In much of feminist macro-economics, by contrast, social reproduction is seen as a necessary activity which underpins both the production of tradables and the promotion of human well-being (Elson 1995). v The export sector is not always booming, as the adjustment literature seems to assume. Low returns to export crops often persist despite adjustment measures because of rising input prices, credit shortages and falling international prices (Elson 1995). For example, the cocoa boom in Nigeria and Ghana burst long ago with the collapse of international prices, environmental problems and rising input costs. Moreover, expansion of traditional exports has limitations from a long-term price perspective. Many researchers and agencies such as UNICEF and UNCTAD believe that the fall in commodity prices in the 1980s was partly linked to the expansion in agricultural output that accompanied the introduction of the new incentive structure. v Another criticism is that markets are seriously limited in their ability to meet human needs and achieve well-being. While markets do provide opportunities (for some groups), they also entail considerable risk, given fluctuations in demand and supply. Making more and more people dependent on the market for their basic food requirements may have disastrous consequences. Were female labour to switch from the production of non-tradables to the production of tradables in parts of rural Africa, for example, the shift may jeopardize household food security. Gender and Adjustment: Institutional Scope Gender analysis is focused at the household level (using bargaining models); while there are references to gender bias within markets, the literature does not provide a robust analytical framework for understanding the gender hierarchies within markets. Other institutions (states, development agencies) escape gender scrutiny entirely. Households The Gender and Adjustment literature (especially Palmer's work) draws heavily on anthropological/feminist writings which look at intra-household dynamics. The model of the household that informs Palmer's and Collier's work is not a single unit with a joint utility function. Rather, like the GRF, it rests on an analysis of gender relations that includes ge nder asymmetries in rights (e.g., access to crop revenues) and obligations (e.g., labour exchanges between husband and wife). Nor is the 'burden' of reproduction equally shared between husband and wife (although for Collier, this does not seem to pose a problem in itself; it is a 'constraint' only insofar as it impedes women's ability to engage in production of tradables). Markets When it comes to middle-level market institutions, the Gender and Adjustment model is less analytical and more descriptive. Collier, for example, refers to gender bias in credit and land markets, but offers no analysis (or models) for understanding how these biases are created and maintained, beyond the observation that 'role copying' plays a critical role in creating male and female segments within the labour market. The conceptualization of markets is one of the main problems in Palmer's work, too. While her critique of gender discrimination within private credit markets, for example, appears to describe how 'real' credit markets operate, some of her policy recommendations are clearly framed in the realm of 'abstract' markets (i.e., idealized models), as for example her endorsement of large-scale commercial agriculture. The solution to the under-valuation of women's labour in small-holder agriculture, in this analysis, is to encourage female employment in the large-scale commercial sector since it would provide a 'proper costing' of their labour. This, however, ignores the fact, pointed out by feminist critical economics (see below), that definitions of labour (female, migrant, skilled, etc.) are built into the way markets operate. Thus even though a female labourer does not work on a commercial estate in her capacity as 'wife', her labour is still defined as 'female', thereby carrying her subordinate status with her into the workplace. While gender hierarchies can be challenged via political action, the assumption that markets can automatically provide a 'proper costing' of female labour ignores the need for public action to regulate them. Development Agencies Finally, a rather benign view of 'public policy' and development agencies is implicit in these models. Both Palmer and Collier rely on public policy to off-set biases in the private sector (households, markets), failing to consider that these public agencies themselves tend to institutionalize the power asymmetries attached to gender difference, and are poor instruments for implementing gender policies (Goetz 1992). The framing of adjustment as a case of 'misguided policies' (i.e., policy-makers working with the wrong conceptual guides) and the excessive reliance on top-down planning to 'set things right' ignores the fact that policy-making institutions are themselves 'gendered' and that they often respond to political pressures. Gender and Adjustment: Development Interventions Despite their shortcomings, the Gender and Adjustment models represent an innovative departure in the literature on women and development. They provide a systematic way of thinking about gender using economic tools of analysis, as well as a way of assessing the impact of gender differences on macro-economic concerns. The policy implications are therefore pitched at the macro-economic level, which opens up a new level of development practice to gender scrutiny. Gender and Adjustment: Applications While it is perhaps too soon to look for applications of the gender and adjustment literature, given its relative recent appearance, a number of embryonic efforts are currently being made-primarily in the form of 'technical notes' and 'advisory notes'- to bring this type of gender analysis into the realm of policy-making. The obvious example of such efforts is the Africa Department of the World Bank's Paradigm Postponed: Gender and Adjustment in Sub-Saharan Africa (Blackden and Morris-Hughes 1995). Adopting elements of Palmer's and Collier's approaches, the study draws attention to the major gender biases instructural adjustment packages-biases that result from the invisibility of women's work; from asymmetrical rights and obligations between men and women defined by the gender division of labour; and from the linkages between the paid and the unpaid economic sectors. The study does not contest the benefits of market-led growth-the policy goal which adjustment is meant to facilitate-but highlights the gender biases that need to be removed if structural adjustment policies are to achieve their intended goals (supply response in the tradable sector). While not yet adopted as official World Bank policy, the study has been used by gender advocates in their discussions with mainstream economists in the institution. Although feminist critical economics is becoming more familiar to policy-making institutions, it has not yet been translated into a discrete gender-training framework. Since it provides a number of useful analytical tools which policy-makers can use for gender analysis at the programme and policy levels, we summarize some of its main analytical threads and examine points of convergence and divergence with neoclassical approaches. The neo-classical paradigm is criticized for its male bias at the micro-, meso- and macro-levels, which it treats as fully integrated: the macro-level views the economy in terms of total marketed output and total expenditure; these monetary aggregates are the coherent result of the activities of millions of individuals (at the micro-level) integrated by the institutions of the meso level (markets, firms). While male bias at the meso- and macro-level arise from omitting gender, at the micro-level male bias is rooted in the way gender is conceptualized (Elson 1994).12 The micro-level A central premise of neo-classical economics is its characterization of individuals as utility maximizers with well-defined choice sets and preference orderings. Differences and inequalities between males and females can be accommodated as long as they can be conceptualized as a matter of differences in preference and resource endowments. Thus differences in women's and men's labour allocations (where women allocate more time to housework and less time to the wage labour market) are explained in terms of the lower returns to women's labour market activities and their 'preference' for housework. Similarly, inequalities between men's and women's leisure time and access to food and medical care are explained in voluntary terms-differing preferences and women's greater taste for altruism. There is no space in this model for conflict, power and preferences that are socially-constructed rather than exogenously given-'the socially entrenched asymmetries in rules and resources that enable some categories of individuals to constrain and shape the options and actions of others' (Kabeer 1994a:133). The meso-level Gender bias at the meso- and macro-levels takes on a different form: since the meso-level structures (markets, firms) and monetary aggregates (GNP, exports, etc.) are not seen as 'male' or 'female', gender analysis is considered to be irrelevant. This in fact is a premise that neo-classical economics shares with critical economic approaches (e.g., structuralist approaches). Looking at the meso-level institutions first, neo-classical economics sees markets as essentially gender-neutral and ultimately empowering for women. Markets can become contaminated with male bias as a result of extraneous factors such as prejudice. In the labour market, for example, neo-classical economics does admit the possibility that 'pure' discrimination might exist-as an irrational 'taste' on the part of employers-but sees it as self-correcting within the logic of markets and so offers no guidelines for its eradication (Joekes 1995a). Feminist critical economics starts from the understanding that market institutions are not 'intrinsically gendered' in the way that family and marriage are, but nevertheless become 'bearers of gender'; gender relations shape the relations between men and women as they engage in market transactions (see Whitehead 1979). In other words, while commercial relations between buyers and sellers, or employers and employees, are not intrinsically gendered in the way that conjugal and kinship relations are, the social positionings of employer and employee, or buyer and seller, vis-à-vis each other are conditioned by gender hierarchies and norms. Even though a female labourer does not work in a commercial estate or factory as 'wife' (in the way that she does on her husband's farm), her labour is defined as 'female', thereby carrying her subordinate status with her into the workplace. The literature on female employment both in the commercial agricultural sector and in industrial establishments demonstrates how social norms about 'femininity', 'women's work' and 'wifely dependence' operate within these so-called impersonal markets to create separate male and female spheres and distinct gender hierarchies (Humphrey 1987; Mackintosh 1989). Similarly, even if women entrepreneurs enjoy the same access to capital and technical skills as their male counterparts, they may still face a disadvantage in the market place due to their exclusion from the business-social networks through which vital information is exchanged and 'good will' is built (see Elson 1994).13 The macro-level Macro-economics is concerned with monetary aggregates, the output of human effort that money can mobilize which gets counted in the gross national product, in savings and investments, in exports and imports and in public expenditure and public revenue. Absent from this picture is the entire area of production referred to as the 'reproductive economy'-the unpaid production of human resources-which underpins the paid economy and is in turn dependent on it. It is now widely appreciated that the failure to take account of the unpaid reproductive labour of child care, housework and care for other adults both in measuring the level of aggregate output (GNP calculations, for example) and in analyzing the determinants of aggregate output renders much of women's work invisible (Elson n.d.). 'Male bias' at the macro-level thus stems from the way the macro-economy is conceptualized, and the activities and values that are left out. But this omission is not simply due to complexities of measurement; it also reflects some assumptions built into the model, with serious implications for the formulation of macro-economic policies. One of the most dangerous assumptions that is implicit within the neo-classical view of the macro-economy is that the reproductive economy will continue to function irrespective of the changes in the rest of the economy. In other words, it is assumed that women will continue to supply all the labour needed to ensure the satisfactory reproduction of human beings, compensating for all the shortfalls in purchased inputs required to sustain human resources. If, for example, public sector investment in infrastructure is cut, causing deterioration in rural water supplies, the assumption is that the members of the household (usually women) will undertake the arduous work of carrying water from distant bore holes to their homes in order to ensure that the needs of their households are met. Or if wages are reduced, the assumption is that the household (again, mainly women) will produce some of the products and services that were previously purchased (health care, repairing clothes) to compensate for the shortfall in cash resources available to the household. Feminist economists have argued that women's labour is not infinitely elastic; the breaking point might be reached where the household is no longer able to reproduce itself, with serious implications for the paid economy. If daughters are withdrawn from school to help with the reproductive work of their mothers, or if women are unable to contribute labour to their paid activities because of the urgent need to attend sick children, this is likely to have a negative affect on the balance of payments-through shortfalls in human skills and through the diversion of public resources to the tasks of policing and repairing the damaged fabric of society. The key issue at the macro level is thus the interdependence between the output that macro analysis does take into account-marketed output-and the output that macro analysis does not take into account-the unmarketed services that are critical for human resources production. This is a gender issue because responsibility for coping with this interdependence is socially constructed as women's much more than men's. The male bias in macro theory results from ignoring this gendered interdependence, with outcomes that are biased against women, and detrimental to development. (Elson n.d.:5) Feminist Critical Economics: Conceptualization of Gender Feminist economics challenges the conceptualization of the gender division of labour in neo-classical micro-economics, particularly 'new household economics', as the outcome of free choices of economic agents specializing according to their innate comparative advantage (Elson n.d.:5; Evans 1989). From the neo-classical perspective, gender differentiation is depicted as generally rational, efficient, and mutually advantageous to men and women, with women undertaking unpaid domestic labour because of their preferences and skills. The neo-classical perspective does not attempt to explain how such skill differentiation comes about. Like the Social Relations Framework, feminist economics seeks to move beyond a gender disaggregation of roles and activities to emphasize the power relations underpinning the gender division of labour (Folbre 1986; Kabeer 1994a). Questions are asked about how social institu- tions and norms shape or constrain individual 'choices' in ways that perpetuate gender inequalities. A bargaining model of the household draws attention to the household as a site of 'co-operation' and 'conflict': women and men may co-operate to increase the household capabilities as a whole, but there is likely to be conflict over the control and allocation of resources, with women in a weaker bargaining position vis-à-vis men. For feminist economics, then, gender differentiation derives from structural inequalities rather than free choices of economic agents: 'it may be seen as the profoundly unequal accommodation reached between individuals who occupy very different social positions with very different degrees of social power' (Elson 1993a:9). Feminist Critical Economics: Institutional Scope As noted, feminist critical economics applies a gender analysis across a range of institutional contexts (households, communities, markets and states). Its approach to gender analysis of economics at the micro-, meso- and macro-level is particularly useful. Its analysis of how markets act as 'bearers of gender' helps to illuminate women's disadvantaged position in market transactions. The conceptualization of the interdependence between the productive and the reproductive economy (and the male bias in measuring aggregate outputs which renders women's work invisible), is a major contribution to the effort to take gender analysis to the macro-level. An understanding of the 'gendered' nature of state institutions and development agencies also informs the feminist critical economics analysis; specifically, it rejects a 'benign' view of public policy and emphasizes the role to be played by women's constituencies within civil society in promoting gender-aware development policy. |