Gender, production and well-being: rethinking the household economy 1

Transcription

1 IDS Discussion Paper 288 Gender, production and well-being: rethinking the household economy 1 Naila Kabeer May 1991 INSTITUTE OF DEVELOPMENT STUDIES Brighton, Sussex BN1 9RE ENGLAND 1 I would like to thank Diane Elson, Martin Greeley and Hilary Standing for their comments on an earlier draft of this paper. i

2 Institute of Development Studies, 1991 ISBN ii

3 Summary This paper pulls together various criticisms of the neo-classical household model, from both within and outside the field of economics in order to move towards a more empirically-relevant approach to household analysis. Starting with the pure Beckerian household, the paper utilises empirical data from a variety of cultural contexts in order to contest the neo-classical claim to transcultural explanations of human behaviour and organisation. It then goes on to examine bargaining approaches to the household which have evolved to counter the early emphasis on the conflict-free nature of household relations. The work of Amartya Sen in this area is explored through more detailed case-material from Bangladesh and the Gambia and reformulated in the light of this discussion. The final section moves towards an alternative more disaggregated framework for conceptualising the household on the grounds that it is more likely to generate culturally-sensitive research, and information that has greater relevance to policymakers. Themes: women; agricultural and rural problems iii

4 iv

5 Contents Summary iii 1 Introduction 1 2 Noises and signals: disciplinary dissensions 1 3 New househould economics: from each according to their ability; to each according to their need? 3 4 Unpacking the household unit: the problem of household boundaries 5 5 Upacking household strategies: beyond the joint welfare function 7 6 Unpacking household decision-making: bargaining, cooperation and conflict 9 7 Unpacking the cooperative conflict dichotomy: structural incentives in production and distribribution Household accounting systems Household labour processes Reformilating conflict and cooperation 14 8 Some case studies Empirical material from south Asia: the case of Bangladesh Empirical material from west Africa: the case of The Gambia Concluding comments 17 9 Rethinking interlinkages: production, reproduction and welfare Disaggregating well-being indicators Expanding concepts of well-being: the use of time Implications for the household welfare differentials Summing up Integrating gender into the household economy: analytical framework and an agenda for research 23 A Defining the household 25 B The household as a double specification 25 C The household as arena of cooperation and conflict 25 D Key dimensions for research Policy relevance of gender-informed household analysis Conclusion: overcoming conceptual and methodological dichotomies 29 References 32 v

6 vi

7 1 Introduction Economics is the science which studies human behaviour as a relationship between ends and scarce means which have alternative uses. (Lionel Robbins 1935: 16) The household has been recognised as an important area for investigation from a number of perspectives. From a theoretical viewpoint, it is widely recognised that many of the allocational activities referred to above (in what is probably the most famous definition of economics as a discipline) are organised through the intimate household relations of marriage, parenthood and family. On the practical level, development policy is frequently formulated and information collected on the understanding that the household will act as a front-line agency in implementing its objectives. The investigation of household activities and concerns are, in other words, critical to understanding the problems of production as well as of welfare the twin concerns of development efforts. Despite a broad acceptance of the importance of household analysis, however, there is considerable dissension as to how such analysis should be undertaken. This dissension takes a variety of forms. It occurs between disciplines, particularly between the universalising a priori assumptions of many economists and the more empirically-grounded approaches of anthropologists and sociologists. It is evident between those who are silent on gender in the household economy and those who see it as a primary site of gender relations. Finally, it can be discerned between those who approach gender within the household as a relation of difference and those who interpret it as a relation of inequality and power. This paper discusses alternative approaches to the household in the light of these dissensions. It focuses particularly on the linkages between household production, distribution and consumption as these are critical to understanding welfare issues within the household. The starting point of the analysis is the neo-classical economic household, which, despite trenchant criticism, survives and continues to exert a powerful influence on how policies are thought about and statistics collected. 2 The paper goes on to modify this paradigm in the light of various critiques, touching along the way on the question of method. Its objective is to provide an alternative framework for the household economy which, while retaining the central problematic of economic analysis, as outlined by Robbins, also permits a more gender-informed understanding of its processes and outcomes. 2 Noises and signals: disciplinary dissensions Speaking of students of the human sciences, a well-known astronomer suggested that they had problems distinguishing genuine from irrelevant information, that is, signals from noise : Instead of separating the noise throwing it away as the physicists do - they spend their energies chasing through every detail of the darned stuff... Students of sociology might indeed be described as the ultimate students of 2 See, for instance, comments in Kabeer and Joekes (1991) on Haddad (1991). 1

8 noise... (Hoyle and Hoyle 1971: ; cited in Sprey 1990: 25). To this, Sprey responds, This is a somewhat naive, but also worthwhile observation. Naive, because after all, one person s noise may be another s signal; worthwhile, because it alerts us to the fact that the world of marriage and the family is a noisy one indeed. (p. 26). Although these diverging viewpoints come from the natural and human sciences, they capture remarkably well the epistemological differences between neo-classical economists (still the mainstream within economics) and others within the field of social sciences. Mainstream economists claim simplicity, parsimony and elegance for their approach; they achieve it through a process of theoretical and methodological reductionism. The bewildering and contradictory complexity of everyday lived reality is compressed into a few, tightly bounded concepts which lend themselves to quantitative approximation, econometric manipulation and determinate outcomes; the rest is discarded as so much noise. The simplifying core of economic theory is the assumption of rational choice, viz. human agents are essentially and universally motivated by self-interest so that all human behaviour can be explained as the attempt to maximise individual utilities in the face of economic scarcity. The economic problem is thus reduced to non-prescriptive or positivist predictions of how scarce resources are likely to be allocated between alternative uses to meet this maximisation criteria. The pre-eminence of positivist thinking within economics also explains the privileged status given to value-free quantitative verification of theoretically-derived hypotheses. The reductionism inherent in this approach to social reality is pithily summarised in Jodha: the first step is to measure whatever can be easily measured...the second step is to disregard what can t be measured...the third step is to presume that what cannot be measured easily is not very important...the fourth step is to say that what cannot be easily measured really does not exist. (Jodha 1985: 1) Critiques of this tradition of economics have existed almost as long as the approach itself. Refuting its claims that rational choice theory is applicable to all human behaviour, is indeed the universal grammar of social science (Hirschleifer 1985:53 cited in Farmer 1989: 16), critics have cast doubt on its explanatory power even within its own preferred domain: In modern economics, that which is sacrificed is an understanding of the principles of market process and of the relationship of this process to the institutional setting in which persons choose (Buchanan 1986: 15; cited in Farmer 1989: 23). Criticisms of formal economic theory are directed at its aggregating, averaging and universalising tendencies and at its impoverished construction of social reality. Structural constraints, institutional processes and cultural variations the noise which neo-classical economists discard have all been invoked to challenge and deconstruct the tightly bounded concepts which constitute the neo-classical model of the universe. Other disciplines have, of course, their own limitations. Sociologists, as we saw, have been described as the ultimate students of noise. Similarly, the relentless focus on ethnographic minutiae of many anthropologists lays them open to charges which reverse those levelled at economists, namely, an 2

9 antipathy to generalisable rules and preoccupation with exceptions ; unease with certainties and emphasis on uncertainties; indifference to the larger picture and immersion in the details of daily existence. Too much noise, in other words, and not enough signals. These characteristics have frequently limited the usefulness of anthropological enquiry in offering operationalisable policy recommendations (Conlyn 1985). While it is indeed the case that social reality is infinitely diverse, uncertain and contradictory, theoretical paradigms need to possess boundaries and limitations if they are to be useful. The economist s concern with the central features and tendencies of resource allocation may be justified if it can be shown that exceptions to the rule are random, accidental or anomalous and have negligible social significance. If, however, the exception carries valuable information, for instance, regarding sub-groups who are adversely affected by a largely positive policy intervention (Conlyn s example) or who represent the first evidence of changes in average behaviour, then to preclude further study of contrary cases and non-average behaviour diminishes our understanding of the very phenomena which economic investigation purports to explain. Clearly, exceptions to the rule deserve investigation before being rejected as noise. 3 New household economics: from each according to their ability; to each according to their need? Bearing in mind these methodological dilemmas, this section outlines the dominant economic approach to the household, a relatively recent area of enquiry in the discipline. While neo-classical economic theory has long dealt with social collectivities firms and households, in particular it tended to treat them as individual decision-making units, whose internal relations were not part of the economic explanandum. Consequently, household behaviour was generally depicted by the behaviour of individuals in their capacity as either consumers (the demand for goods and services) or workers (suppliers of labour). This was rectified with the emergence of the new household economics, whose contribution was to integrate the production and consumption aspects of the household economy and extend maximisation principles to its internal workings. It is important to mention here that there are dissensions among economists as to whether intrahousehold behaviour properly constitutes the domain of economic analysis. At one extreme are those who declare On our preferred interpretation the economist continues to search for differences in prices or incomes to explain any differences or changes in behaviour (Stigler and Becker 1977). Other economists, however, have preferred to regard the threshold of the household as a demarcating boundary between economic and non-economic behaviour, beyond which they will not venture. For this reason, it is the Beckerian household which has dominated discussions of economic approaches to the household. According to Becker, households are formed by men and women who come together in marriage because of their complementarity in biological reproduction (Becker 1974). Once formed, the household unit behaves as if it was maximising a joint welfare function. It combines purchased goods and services with domestically available inputs, primarily family labour, to produce a welfare-maximising bundle of Z-goods, the objects of final consumption (e.g. health, nutrition, child services, warmth). The 3

10 allocation of family labour time between competing uses is determined rationally by the principle of comparative advantage so that each household member specialises in those activities which give them the highest relative returns. 3 However, the household can only be portrayed as a welfare-maximising unit if the distribution of consumption resources is also demonstrated to maximise the joint welfare of its members. Here a problem arises. Unlike factors such as household labour time and productive assets, which are potentially, if not actually, marketed, individual welfare cannot be given a price tag since it is not a marketable commodity. Nor is it possible in the liberal economic tradition to make direct inter-personal comparisons of individual welfare. Economists resolve the problem of welfare maximisation in household distribution by assuming that this is precisely what happens. In other words, they posit either total altruism within the household so that the welfare of each member of the family is normally integrated into a unified family welfare function (Schultz 1973: S6) or else they posit partial altruism, with household welfare maximised through the benevolent dictatorship of the household head (Becker 1981). This does not rule out welfare differentials in consumption patterns, if these are considered to maximise joint household welfare (Rosenzweig 1986; Rosenzweig and Schultz 1982). 4 But it does rule out welfare inequities in the sense of unequal weighting being given to individual utility within the joint welfare function. Welfare differentials are thus depicted as freely chosen and accepted by individual household members in the interests of the joint welfare of all. This representation of household decision-making as one of comparative advantage in production and altruism in distribution has been compared to the primitive communist rule governing the allocation of the work-load and the distribution of consumption viz. from each according to their ability; to each according to their need. (Saith and Tankha 1972: 351 cited in Ellis 1988). Although certain modifications to the model are introduced in adapting the model to Third world peasant economies, 5 the logical structure of the model is retained intact because it is considered to have universal applicability. As Becker claimed, economic theory may well be on its way to providing a unified framework for all behaviour involving scarce resources, non-market as well as market, non-monetary as well as monetary, small group There are conflicting interpretations of this point among writers who otherwise subscribe to rational choice analysis. Some, such as Becker, appear to take women s comparative advantage in childbearing and rearing as the independent variable, explaining why women fail to invest in acquiring human capital qualifications for the labour market. Others, such as Rosenzweig and Schultz (1982), appear to hold that the observed differentials between male and female earnings in the market place explain lower investments in marketable human capital credentials for women and their specialisation in child care and domestic labour. As Sawhill points out, the neo-classical school is guilty of a certain circularity in reasoning:...women earn less than men because of their special role within the family, but their special role within the family and indeed the desirability of marriage and children are importantly related to the economic status of women (Sawhill, 180: 133). My thanks to Diane Elson for this point. See, for instance, Philippines Economic Journal 1978, Low 1986 and Ellis

11 as well as competitive (Becker 1974: 299). A similar claim for the choice-theoretic approach is made by Crouch (1979: vi): a general theory of human behaviour based on choice theory...whose explanatory power is not restricted in time or space (Crouch 1979: vi, cited in Winrich 1989: ). Summarising the insights of the neo-classical household model, it integrated for the first time the production, distribution and consumption aspects of the household and it identified labour time, paid or unpaid, as an important constraint on the household economy. While the actual gender division of labour and consumption was obviously treated as a matter for empirical investigation, the neo-classical approach did assume that the division of labour between household members could be explained by rational comparative advantage. Such a claim can be tested through empirical information on the relative returns to labour and household endowments in alternative uses. It also predicted that intra-household welfare distribution was guided by altruistic principles. However, since individual utility is difficult to measure and compare, the model is silent on the empirical operationalisation of this claim. If welfare differentials occur within the household, they are inferred to be freely chosen in the interests of joint welfare maximisation. The rest of this paper examines selected criticisms of this approach in an attempt to unpack the aggregating assumptions by which it has excluded serious considerations of gender differentials in labour, resources and power from its explanandum. It takes as its point of departure the useful discussion in Evans (1991) and attempts to draw up an alternative framework for household economic analysis. The main fallacies of aggregation which will be addressed are: the aggregation of production, consumption and residence into a unified and bounded entity the aggregation of different interests within the household into a joint welfare function the aggregation of different individuals within the household economy into a single welfaremaximising agent. 4 Unpacking the household unit: the problem of household boundaries A fundamental criticism of the household comes from those who question its very validity as a concept. They challenge the way in which economic theory unreflectingly collapses units of residence, consumption and production together into a fixed and unchanging entity: the domestic group is often seen as monolithic, clearly bounded and unchanging, except in so far as the domestic cycle produces changes in personnel (Pittin 1987: 26). In place of the single, bounded unit, these critics posit a range of household forms: shifting, flexible structures in which boundaries are difficult to discern...a diversity of family and household composition and social relations, mediated through marriage and kinship, creating a variety of conjugal and residential arrangements. Variations in marital practices, inheritance, property owning rules and authority structures within and outside the domestic group clearly create an empirical diversity in household forms which cannot be easily accommodated within an economics which claims transcultural relevance (cf. Crouch 1979: xv, cited in Winrich 1989: 178). 5

12 However, there is an important distinction between denying the usefulness of the household concept as defined in economic textbooks and abandoning the concept itself. The empirical significance of household relationships in the routine management of resource entitlements in a variety of contexts (Humphrey et al. forthcoming 1991) suggests a more useful approach would be to reconceptualise, rather than jettison, the notion of the household. Moreover, as Moore (1988: 55) points out: households are important in feminist analysis because they organise a large part of women s domestic/reproductive labour. As a result, both the composition and organisation of households have a direct impact on women s lives, and in particular on their ability to gain access to resources, to labour and to incomes. Moore (1988: 55) The challenge then is to find generalisable approaches to kinship-mediated economic flows which can accommodate the empirical diversity of household forms. Rather than abandoning the notion of households altogether, it is necessary to abandon the cultural baggage and universalising assumptions by which the new household economics achieved its formal elegance. That is, we need to replace an a priori solidary, discrete and bounded household entity with, to use Friedmann s term (1979), a dual specification of households as internally diverse organisations, embedded within, and shaped by, wider structures. We need to start our analysis of households by specifying the encompassing structures of market and nonmarket forms of resource allocation, legal and customary entitlements and community norms and values, since these jointly constitute the context within which the household economy survives, reproduces or transforms itself. The other step is to specify the boundaries between households and the rest of the economy. Anthropological research has been useful here in uncovering the wide variation globally in the boundaries of the domestic group; but it has also pointed to the existence of certain kinds of regional uniformities. It has documented, for instance, the higher incidence of corporate forms of householding, organised around the conjugal bond, in the regions of North Africa, South Asia and the Middle East compared for instance with the weaker cohesiveness of the conjugal unit in the Caribbean, parts of Latin America and in sub- Saharan Africa (Caldwell 1982; Cain 1984; Goody 1976; Guyer 1981; Youssef 1974). As a general principle, therefore, it may be an acceptable procedure in many South Asian contexts to assume a considerable degree of corporateness in householding arrangements. Cain, for instance, suggests that the conjugally organised nuclear family is a useful unit for empirical analysis in much of the Indian subcontinent (op. cit.). In sub-saharan Africa, on the other hand, an anthropological consensus seems to be that the solidary, discrete household is a particularly inappropriate construct, given the widespread prevalence of non-conterminous units of production, reproduction, consumption or residence (Guyer and Peters 1987; Whitehead 1990). In such contexts, Guyer and Peters usefully summarise the two main forms of relationship between domestic units: overlapping memberships, where some but not all of the members of the minimal unit belong to a single, encompassing unit; and nesting memberships, where each minimal 6

13 unit is totally assimilated to a larger unit (p. 206). They suggest that analysis in these situations should proceed by tracing the composition, activities and relationships between the various units where these various activities are centred. Although different from the a prioristic tradition in economics, such a procedure could still illuminate its central concerns: the institutions and processes through which valued scarce resources are allocated between competing ends. In the final analysis, the notion of the household is an analytical construct and its boundaries may have to be drawn pragmatically, in response to the concrete research questions being investigated. The nuclear family, centred on the conjugal relationship, might be a useful unit for analysing production, consumption or reproductive behaviour in many parts of South Asia; the extent to which the extended family is considered relevant would largely depend on the nature of the precise research question. In the sub-saharan Africa context, Caldwell (1982) suggests that the lineage group might be the relevant unit of analysis for the study of reproductive behaviour, although conjugal relationships would clearly form a secondary level of analysis. A study of household economic activities in Africa might, on the other hand, focus on the compound as the main analytical category, described in Pittin s study of the Hausa community as the one unit with clear spatial parameters, which did not assume prima facie any form of dependency, responsibility, joint production, or joint consumption (1987: 40). Alternatively, a study of budget flows and consumption arrangements might begin with the mother-child unit or hearthhold because of its relative autonomy and separate identity : It is the house of the house property complex classically described for southern and central Africa (Guyer and Peters 1987: 207). This contrast between household forms in sub-saharan Africa and those in South Asia is not intended to suggest fixed regional categories of the kind found in Boserup (1970) and Goody (1976). Households are historically constituted in response to social and economic forces; changes in extrahousehold relations can transform household relations and produce greater regional commonalities in household forms. The greater commercialisation of agriculture or changing land distributions in sub- Saharan Africa, for instance, might be associated with the emergence of proletarianised peasant household economies which bear a striking resemblance to those found in South Asia. Bearing in mind these important caveats about the idea of the household, we will use the term in the rest of this paper in a non-prescriptive way to refer to the bundle of relationships in a society through which reproductive activities are organised, recognising that these frequently involve principles of kinship and co-residence. The economy of the household then refers to the activities, relationships and processes by which households produce, acquire and distribute valued resources (for use, investment and exchange) essential to the survival, reproduction and prosperity of its members. 5 Unpacking household strategies: beyond the joint welfare function A second fallacy of aggregation in the new household economics stems from the way in which it collapses individual interests within the household into a joint welfare function through the assumption of total or 7

14 partial altruism among household members. Criticism of this assumption has had a long history both within the economic discipline and outside. Galbraith, for instance, indicts neo-classical theory for burying the subordination of the individual (woman) within the household (1974: 35) and suggests that the household, in the established economics, is essentially a disguise for the exercise of male authority (p. 36). Sen points out the absurdity of believing that only the household head s view of the welfare of his members should count in welfare analysis and asks why the views of subordinated and subjugated members regarding their own and family welfare should be given no status (1984: 373). And Folbre points to the untenable paradox at the heart of household economics which juxtaposes naked self-interest as the guide to individual behaviour in the market context with altruism guiding individual behaviour in the household context (Folbre 1986a). This does not imply that the notion of the household as moral economy is entirely redundant. The greater significance of altruism in household relations than elsewhere explains, for instance, the remarkable convergence between the neoclassical depiction of intra-household relations as akin to primitive communism (Saith and Tankha, op. cit.) and the conventional Marxist depiction of the household, to use Folbre s words, as a miniature utopian socialist society, untroubled by internal conflict (1986a: 6). However, the growing literature on gender and age discrimination within the family, which results in health, nutritional and survival inequalities between members (cf. Folbre 1986b; Sen 1990) does challenge the universality of altruistic welfare maximisation within the family. Up to a point, distributional asymmetries can be explained within the neo-classical paradigm as a matter of individual preferences and tastes. Preferences do not enter explicitly into the paradigm because they are taken to be exogenously given, stable and randomly distributed (Folbre 1986a; Amsden 1980) so that, in general, inexplicable behaviour due to tastes is believed to be either trivial or idiosyncratic (Amsden: 13). It is therefore perfectly feasible that individual members of the family will choose to deprive themselves of welfare resources in the interests of joint welfare maximisation. However, when, in certain contexts, (a) the subordination of personal welfare in favour of the welfare of others appears systematically to be the property of one social category of individuals (women) while the beneficiaries of such preferences appear systematically to belong to another social category (men) and (b) the consequences are life-threatening levels of health and nutrition for the former, then the notion of individual preferences governing household distribution becomes patently absurd (Kabeer and Humphrey 1991). Clearly, the relationship between distributional processes and individual welfare outcomes must be empirically explored rather than theoretically inferred. If the choices of individuals or of benevolent patriarchs lead to systematic welfare inequities within the household, there are strong grounds for believing that political economy considerations must also be brought to bear on analysis of household behaviour. If differential returns to household labour reflect social-structural forces rather than individual productivities, then some members are likely to be more privileged than others. We need to consider what these forces might be and how they shape household preferences, decision-making processes and outcomes. 8

15 Household strategies can encompass a wide variety of individual as well as collective objectives which are rendered invisible within the unitary and tautological joint welfare-maximisation formulation: riskaversion, satisfying, status enhancement, safety-first, capital accumulation, altruistic cooperation, income maximisation, dictatorial benevolence or bounded rationality. However, all members are not equally empowered to pursue the full array of objectives encompassed within the household collectivity: the membership is internally differentiated by material constraints which may strongly influence which objectives, and more importantly, whose objectives are actually realised. At this point, however, another disaggregation clearly becomes necessary. The idea of household strategy and household decision-making conveys a rather anthropomorphic representation of the household. In reality, of course, the household can neither decide nor think, since analytical constructs are not so empowered. Rather, certain people within the household make decisions. One or more persons with enough power to implement them makes decisions and other less-empowered household members follow them (Wolf 1990: 60) Thus a disaggregation of the welfare function to allow for an analysis of the distribution of individual welfare outcomes within the household points inevitably in the direction of a further disaggregation: that of the decision-making unit itself and of the distribution of decision-making power within it. 6 Unpacking household decision-making: bargaining, cooperation and conflict Feminist analyses of gender asymmetries within the household as a manifestation of power relations have begun to find their way into the economic literature. This is evident, for instance, in Folbre s rejection of the notion of pure altruism within the household and her suggestion that the household s objective function should be conceptualised as a function of the relative bargaining power of its members (Folbre 1986 a). Her insight of the household as moral and political economy has been formalised in Sen s model of the household as a domain of cooperative conflict (1990). Distribution within the household takes place according to principles of both legal and accepted legitimacy ( extended entitlements ) but also reflects power relations within the household. A summary of this model is presented here because it is a serious attempt to meet feminist criticisms of economic approaches which ignore the power dimension of household relations. In general, cooperation will take place as long as it leads to outcomes which are preferable to those that would prevail should cooperation break down. The bargaining problem arises from conflicting interests in the choice between alternative cooperative outcomes. When certain choices clearly favour one member (or category of members) over another, the final decision between one of the cooperative outcomes or the break-down position will be the outcome of a bargaining process between the parties concerned. In the context of household bargaining, Sen says: 9

16 The prosperity of the household depends on the totality of various activities getting money incomes, purchasing or directly producing (in the case of say, peasants) food materials and other goods, producing edible food out of food materials, and so on... the members of the household face two different types of problems simultaneously, one involving cooperation (adding to total availabilities) and the other conflict (dividing the total availabilities among the members of the household (p. 129). The solution which is then implemented - i.e. the particular pattern of division of fruits that emerge from such cooperation will reflect the bargaining power of different members within the household (p. 131). Differences in bargaining power between members (or categories of members) are the product of interlocking asymmetries, including the range of options facing members, should household cooperation break down; the perceived significance (illusory or otherwise) of their contributions to household prosperity; the degree to which members identify their self-interests with their personal well-being (and therefore the extent to which they are prepared to subordinate their own well-being to that of others) and finally the ability of some members to exercise coercion, threat or violence over others. This formulation of intra-household decision-making as a bargaining process between unequals has obvious differences and advantages over the neo-classical model. A fundamental improvement with this version is that it recognises the existence of more than one decision-maker within the household and the potential for conflicting objectives and strategies. Furthermore, the cooperative conflict formulation introduces the idea of unequal power into the process of decision-making, thus raising problems and issues which were missing, when only altruistically-unified memberships or benevolent patriarchs made decisions. While its focus is mainly on bargaining between adult women and men, it refers to, and can accommodate, cooperation and conflict in the context of pluralistic household forms. It is not, in other words, tied to a culturally and historically specific household form or limited to specific categories of household members. The other important feature of Sen s model is that it extends the necessary empirical base for analysis well beyond the Beckerian one of prices and incomes and encompasses some of the more qualitative and cultural dimensions of social reality. Both intra- and extra-household information (the social technology of production relevant to household members, their perceptions and attitudes, the options facing them should household cooperation fail) are brought into the analytical framework. Secondly, convention, custom and ideology are recognised as important dimensions in shaping the gender division of labour and thus introducing rigidities into it. This is in contrast to the Beckerian analysis where the gender division of labour within and outside the household was envisaged essentially as one of rational economic choice, in the face of individualised material constraints. While it is still a formal representation of the household, it opens up the black box of the neoclassical household through more empirically-grounded analysis of its operations, utilising information from within and outside the domestic domain. However, in examining the usefulness of Sen s model in 10

17 explaining empirical phenomena, a number of limitations become apparent. The two that will be discussed here concern the particular conceptualisation of cooperative conflict and of well-being which appear to inform his analysis. While these limitations do not invalidate the core insights of Sen s model, they do point to a need for reformulating it. 7 Unpacking the cooperative conflict dichotomy: structural incentives in production and distribution There is a strong sense in which Sen s exposition equates cooperation with production and conflict with distribution, although this equation is not in fact essential to his analysis. As the quotation cited above suggests, household members cooperate in the matter of adding to their total availabilities : through, for instance, wage labour or direct production to acquire food and other materials and then processing such materials into consumable form. This cooperation in production is induced by the interdependence implicit in the social organisation of production, while bargaining appears to apply only to the issue of distribution: the technological interdependences make it fruitful for the different parties to cooperate, but the particular pattern of division of fruits which emerges from such cooperation reflects the "bargaining powers" of the respective parties (p. 131). A possible objection to this formulation is that interdependence need not be seen as automatically cooperative; it can also carry the potential for conflict, bargaining and negotiation. Equating cooperation with production and conflict with distribution is thus unduly restrictive, foreclosing as it does on possibilities of either occurring at any stage of the household production and distribution. To some extent, Sen s cooperation/conflict formulation is effectively a reversal of the efficiency-altruism dichotomy of the neo-classical household. Instead of efficiency in production being juxtaposed with altruism in distribution, Sen s model juxtaposes cooperation in production adding to total availabilities with conflict in distribution dividing total availabilities. This unduly restrictive formulation may stem from the fact that, while Sen uses the contrast between gender relations in Asian and African households to provide empirical substantiation of his analytical insights, he relies more heavily on examples from the Asian context. Consequently, while he picks up on the now well-documented significance of women s outside work and earnings in the sub-saharan context, other dimensions of its social relations, which have important implications for his model, are missed out on. Pointers from the anthropological literature suggest that issues of separability, control and autonomy within the household economy are likely to constitute one such dimension. In the rest of this section, we discuss the significance of these issues in two areas: in household organisation and in local labour processes. 11

18 7.1 Household accounting systems Variations in the degree of separateness within household structures can be illustrated by citing two idealtypical household systems, (identified for instance in Kandiyoti 1988) on the basis of cross-cultural review of empirical literature. The first, based on findings from sub-saharan Africa, is summarised in Palmer: There is strong evidence in many African countries of what might be called a gender division of management of crops (or fields). This is not to be confused with a gender division of labour. Management here is defined as responsibility for mobilising the variable factors of production to be applied to a piece of land... If inputs are purchased the manager pays for them with cash, labour, produce or any other means. Consequently the manager has the right to dispose of all or the bulk of the returns to the outlay. This is what is meant by women and men within the household drawing on the household-level stock of resources to form their own economic accounting units. Palmer (1988: 7) Separate economic accounting systems are generally associated with economic transactions, often in the form of commodity exchange, between women and men and, frequently, duolocal residence within the conjugal unit. We will provide some examples below to suggest that such systems also have a greater empirical association with overt bargaining and conflict in production and distribution. At the other extreme are the ideal-typical household arrangements characterising what Kandiyoti (1988: 20) has described as the belt of classic patriarchy encompassing North Africa, Middle East and South Asia... where the total control of economic resources is generally vested in the person of a titular male head who may allocate resources from the common fund differentially to household members, according to the positions they occupy in terms of age and sex. A similar description is found in Bardhan for women s position in north Indian peasant households: They have to depend on male sanction and mediation for the work they do, and its remuneration...even for the security of claim to a share of the subsistence they work to generate... The crucial reason for this is that men control property (land, equipment and other productive assets), cash flows and the vantage points in the marketing chain, and women even when forced to hard work for a living must constantly be guarded in conduct so as not to risk incurring severe penalties from family and society... Destitution, women cut off from the male-mediated order, is an ever-present spectre... Bardhan (1985: 2208) Within such systems, there are strong incentives supporting (possibly enforced) cooperation in most aspects of household decision-making. As Kandiyoti points out women s strategies to maximise their 12

19 security always involve gaining and keeping the protection of their men; through the adoption of attitudes of submissiveness, propriety and self-sacrifice. Bardhan also notes the structural constraints which reward female compliance and inhibit open conflict and struggle over household objectives (p. 2265). 7.2 Household labour processes The nature of interdependencies within the labour process is another related dimension in shaping the likelihood of cooperation or conflict within household decision-making. An important distinction made by Whitehead (1985) concerns the extent to local labour processes are, by convention, sex-sequential or sex-segregated. The former requires labour inputs from each sex at different times to produce a single product; the latter refers to processes in which one or other sex performs all (or most) of the operations necessary to produce a given product. In general, it can be hypothesised that where women contribute labour to specific tasks in a crop cycle which also requires critical inputs from or initiation or completion by male labour, there is a less clear cut basis for allocating the proceeds to labour. It is therefore more difficult for a woman to increase her share of the rewards as a result of extra effort. Women s greater knowledge or statistical importance (Whitehead 1985) in decision-making in female segments of the production process may count for little, if critical distributional decisions about the work load or the proceeds from production are associated with different and male segments of production. In such situations, women have little scope for resisting male decisions, without challenging the basis of the conventional division of labour. On the other hand, where production processes are, by and large, gender-segregated, women may have greater scope to respond independently to production incentives and to bargain with men as to how the household inputs and welfare resources are to be allocated. Different forms of interdependence in the gender division of household labour are likely to carry different potentials and costs for gender conflicts in production. Such segregation can take the form of: separate fields: e.g. the Gambia (Dey 1981) where according to customary practices among the Mandinka, women and men are allocated separate own-account fields. Women often refused to work on men s own-account fields if it conflicted with the labour cycle of their own fields. Some men were forced to buy in the labour of their wives. separate crops: e.g. south east Nigeria (cited in Leach 1990) where men and women cultivate different crops in the same fields, with men planting trees and women planting food crops such as melon and cassava. Conflicting incentives meant that women s farming practices (ineffective weeding or early planting) frequently led to failures of men s tree crops. separate sectors: e.g. in Malawi, women are active in the subsistence crop sector while the fishing sector is regarded as a separate male sphere of activity (cited in Harrison 1990). Potential conflict may arise (unless women are properly remunerated) if new aquaculture projects are developed which require the maize produced by women to be used as fish-feed by men. 13

20 Other form aspects of the labour process which Whitehead suggests might have additional significance in analysing power between the genders concern relations of command and control in the division of labour: these are likely to vary depending on whether women enter the production process as unpaid family labourers, wage labourers, as clients or patrons in patronage networks, or as independent, ownaccount workers: clearly the very notion of labour for others, even within family relationships, carries with it the potentiality for the control and command of work to lie with the user of the labour rather than its provider a potentiality over and above the technically implied control discussed above (Whitehead p.44). Moreover, a status such as unpaid family labour on male-owned farms can only be assessed in the context of other aspects of the household economy. The bargaining power of women working on their husband s land as unpaid family labour is likely to vary, as Sen points out, according to whether or not they have, in addition, access to independent resources outside those acquired through cooperation. 7.3 Reformulating conflict and cooperation These examples suggest that the incentives for cooperation and conflict between individual household members must be sought in their relationships to each other and to the wider economy. They cannot be assumed for specific stages of decision-making. The decision-making processes within the household are better understood through an analysis of those specific aspects of household relations including property rights, labour processes and extra-household resources which are likely to engender conflict or cooperation between interdependent members. It is highly likely that the same structural incentives which determine the conflictual or cooperative nature of interdependence in the production process will operate similarly in the distributional process. Sen appears to partly recognise this in his concluding comments on the togetherness aspect of household relations, but it is largely absent from the substance of his analysis: Not only do the different parties have much to gain from cooperation; their individual activities have to take the form of being overtly cooperative, even when substantial conflicts exist...serious conflicts of interests may be involved in the choice of social technology, the nature of the family organisation requires that these conflicts be molded in a general format of cooperation, with conflicts treated as aberrations or deviant behaviour.(p. 147). In the next section, empirical examples from Bangladesh and the Gambia will be used to demonstrate some important differences in the structure of household resource management and illustrate their differing implications for gender conflicts and cooperation in household resource management (some of these issues are discussed in Aziz and Kabeer 1989; Kabeer 1989). 8 Some case studies 8.1 Empirical material from South Asia: the case of Bangladesh Rural Bangladesh typifies many aspects of classic patriarchy. Purdah norms, property rights and familial hierarchies coalesce within Bangladeshi domestic groupings to produce corporately organised, patriarchal household forms. Men tend to control most of the household s material resources, including the labour of 14

21 female and junior members of their households, and also to mediate women s relations with the nonfamilial world. Women are socially constructed as passive and vulnerable, dependent on male provision and protection for their survival. Women s well-being tends to be tied to the prosperity of the household collectivity and their long-term interests best served by subordinating their own needs to those of male family members. The social norms of male breadwinner/female dependent are reflected in men s privileged, albeit class-differentiated, access to land and waged opportunities. Women are excluded from extra-household productive activities, including field-based agriculture, and tend to be confined to the precincts of the homestead where they are mainly involved in post-harvest processing of crops and the cultivation of homestead plots. There is thus a basically gender-sequential process of production in agriculture, with men in control of certain critical stages of production its initiation and its marketing. Although women do have greater control over certain forms of production homestead crops and smaller livestock, sometimes marketing them locally these are usually not substantial enough to transform the core relations of male power/female dependency within household structures. There are minimal incentives for women to engage in open conflict over household production or distributional decision-making. They have few if any independent resources outside familial networks that would give them a basis for bargaining within their households. Indeed when cooperation breaks down, as in times of economic crisis, it is men who are most likely to abandon their female dependents; conjugal and parental cooperation becomes unviable if the costs of dependency are perceived of as too high. The high costs associated with patriarchal risk (Cain et al. 1979) (referring here to the insecurities experienced by women as a result of their extensive dependence on men) lead women to adopt cooperative strategies which maximise their longer-term security, often at the cost of their personal wellbeing. This involves, among other things, submission to the authority of their guardians (fathers, husbands, brothers or sons), deferential behaviour and conflict avoidance, especially with their husband and in-laws. When forced into income-earning activities, women tend to favour those which do not challenge purdah norms as long as they can. Post-harvest processing of crops and domestic service in the homes of wealthy families, while not particularly well-paid, is a common example of such an activity. Incentives for cooperation also extend to distribution. There is little evidence of overt conflict over distributional processes; instead there appears to be a gender hierarchy of decision-making, needs and priorities, a hierarchy to which both men and women appear to subscribe. Thus, while there is widespread evidence of marked gender inequities in household distribution, women may acquiesce to these discriminatory practices in the interests of their longer-term security. Women are not entirely powerless, of course, but their subversion of male decision-making power tends to take clandestine or unacknowledged forms (Kabeer 1990; Abdullah and Zeidenstein 1982). The use of trusted allies (relatives or neighbours) to conduct small businesses on their behalf, the secret lending and borrowing of money and negotiations around the meaning of gender ideologies of purdah and motherhood (Kabeer 1990 and forthcoming 1991) are some of the strategies by which women have resisted male power. 15