"BARGAINING" AND GENDER RELATIONS:
WITHIN AND BEYOND THE HOUSEHOLD

Bina Agarwal

FCND DISCUSSION PAPER NO. 27

Food Consumption and Nutrition Division

International Food Policy Research Institute
1200 Seventeenth Street, N.W.

Washington, D.C.  20036-3006 U.S.A.
(202) 862-5600

Fax: (202) 467-4439

March 1997

FCND Discussion Papers contain preliminary material and research results, and are circulated prior to a full
peer review in order to stimulate discussion and critical comment.  It is expected that most Discussion Papers
will eventually be published in some other form, and that their content may also be revised.



ABSTRACT

Highlighting the problems posed by a "unitary" conceptualization of the household,

a number of economists have in recent years proposed alternative models.  These models,

especially those embodying the bargaining approach, provide a useful framework for

analyzing gender relations and throw some light on how gender asymmetries are

constructed and contested.  At the same time, the models have paid inadequate or no

attention to some critical aspects of intrahousehold gender dynamics, such as: what

factors (especially qualitative ones) affect bargaining power? What is the role of social

norms and social perceptions in the bargaining process and how might these factors

themselves be bargained over? Are women less motivated than men by self-interest and

might this affect bargaining outcomes?  Most discussions on bargaining also say little

about gender relations beyond the household, and about the links between extrahousehold

and intrahousehold bargaining power.  This paper spells out the nature of these

complexities and their importance in determining the outcomes of intrahousehold

dynamics.  It also extends the bargaining approach beyond the household to the

interlinked arenas of the market, the community, and the State.



CONTENTS

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

1.  Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2.  Bargaining and Intrahousehold Gender Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Determinants of Intrahousehold Bargaining Power . . . . . . . . . . . . . . . . . . . . . . . . 9
Bargaining for Subsistence Within the Family . . . . . . . . . . . . . . . . . . . . . . 11
Social Perceptions and Bargaining Power . . . . . . . . . . . . . . . . . . . . . . . . . 15
Do All Factors Carry Equal Weight? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Bargaining over Family Land . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Social Norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Norms as Limits to What Can Be Bargained About . . . . . . . . . . . . . . . . . 22
Norms as Determinants of or Constraints to Bargaining Power . . . . . . . . . 24
Norms and How Bargaining Is Conducted . . . . . . . . . . . . . . . . . . . . . . . . 26
Bargaining over Social Norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
What Affects Bargaining over Norms? . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Self-perceptions, Altruism, and Self-interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Do Women Suffer from False Perceptions? . . . . . . . . . . . . . . . . . . . . . . . . 35
Altruism or Self-Interest? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Altruism and Intrahousehold Coalitions . . . . . . . . . . . . . . . . . . . . . . . . . . 42

3.  Beyond the Household:  The Market, the Community, and the State . . . . . . . . . . . . 43

The Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
The Community . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
The State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Interactions:  The Household, the Market, the Community, and the State . . . . . . 53

4.  In Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

FIGURE

1. Factors likely to affect rural women's bargaining power (in relation to subsistence)
in different arenas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55



iv

ACKNOWLEDGMENTS

This paper builds on and develops some of the issues I had raised in A Field of

One's Own:  Gender and Land Rights in South Asia (Agarwal 1994a).  Initial discussions

with and subsequent comments on the paper from Lawrence Haddad, Janet Seiz,

Elizabeth Katz, Agnes Quisumbing, and Nancy Folbre were most useful, and I thank

them all for their responses. I also thank Joti Kohli for her help in locating several useful

articles.

Prof. Bina Agarwal
Institute of Economic Growth
University of New Delhi



 Studies on gender and microeconomic outcomes are too numerous to list here, but on gender and1

macroeconomics, see, especially, World Development (1995).

 Biology (pregnancy, childbearing, etc.) may have influenced the historical construction of some2

aspects of gender relations, such as the gender division of labor.  But biology cannot explain the entire gamut
of gender inequalities we observe today, nor even the perpetuation of an observed gender division of labor
(e.g., technical developments have minimized the importance of muscular strength; contraceptive technology
reduces the disability of frequent pregnancies; and a variety of possible arrangements make child care a less
binding constraint).  In any case, the considerable variation of gender relations across cultures indicates the
enormous importance of nonbiological factors.

1.  INTRODUCTION

The nature of gender relations—relations of power between women and men—is

not easy to grasp in its full complexity.  But these relations impinge on economic

outcomes in multiple ways.   The complexity arises not least from the fact that gender1

relations (like all social relations) embody both the material and the ideological.  They are

revealed not only in the division of labor and resources between women and men, but

also in ideas and representations—the ascribing to women and men of different abilities,

attitudes, desires, personality traits, behavior patterns, and so on.  Gender relations are

both constituted by and help constitute these practices and ideologies, in interaction with

other structures of social hierarchy such as class, caste, and race.  Neither uniform across

societies nor historically static (as numerous studies of different cultures, regions, and

communities bear out), they may be seen as largely socially constructed (rather than

biologically determined).   The process of this social construction, however, is2

inadequately understood, as is also how particular forms of gender inequalities are

maintained, and by what means they might change over time.

Recent growing interest among economists in intrahousehold dynamics addresses

these questions.  For instance, there has been an increasing criticism of "unitary"

household models that assume that household members seek to maximize utility on the



2

basis of a set of common preferences represented by an aggregate utility function, and a

common budget constraint.  A range of alternative household models use the game-

theoretic approach to incorporate a more complex understanding of how family

decisionmaking occurs, variously allowing for individual differences in preferences, in

budget constraints, and in control over resource use.  Although most do not explicitly

address gender asymmetries, they can accommodate such asymmetries, and thus provide

a useful approach for analyzing gender relations and point economic thinking in a new

direction.  At the same time, these formulations, consisting mostly of formal models, are

restricted in their ability to incorporate the full complexity of gender interactions within

households, and the simultaneity of various processes and forms of decisionmaking.  In

addition, most say little about gender relations beyond the household.

This paper outlines some aspects critical for understanding intrahousehold

dynamics from a gender perspective that existing household models either miss

altogether, or do not adequately address.  For instance, most models characterize these

dynamics as a form of "bargaining," but typically say little about the complex range of

factors, especially qualitative ones, that might determine bargaining power; what role

social norms and perceptions play in the bargaining process; what effect gender

differences (if such exist) in the exercise of self-interest might have on bargaining; and so

on.  Such factors can impinge crucially on the accuracy of theoretical formulations,

empirical predictions, and policy interventions, and must therefore be given cognizance in

framing hypotheses, data gathering, and analyses.

Equally, models and policies could go awry if intrahousehold dynamics are

assumed (as they often are) to exist in isolation, without examining the extrahousehold

socioeconomic and legal institutions within which households are embedded, and how

these institutions might themselves be subject to change.  The paper seeks to provide

pointers on this count as well, extending the bargaining approach beyond the household

to the interlinked arenas of the market, the community, and the State.



3

For this purpose, I both use and emphasize the usefulness of what I term "analytical

description" for capturing the complexity and historic variability of gender relations in

intra- and extra-household dynamics.  By analytical description I mean a formulation that

seeks to comprehensively spell out both qualitative and quantitative factors that might

impinge on outcomes, without being preconstrained by data limitations or by the structure

that formal modeling imposes.  A formal model could be seen as a subset of analytical

description; and empirical analysis based on the model as a further subset.  All

three—analytical description, a formal model, and empirical analysis—can in different

ways extend our understanding about how gender relations get constructed and contested

within and outside the household.

2.  BARGAINING AND INTRAHOUSEHOLD GENDER RELATIONS

Are you suggesting that women should be given rights in land? What do

women want? To break up the family?
(Minister of Agriculture to the author at an                  

Indian Planning Commission seminar on Land Reform, June 1989.)                 

Whether or not so intended, the Minister's reaction implies at least two assumptions

about the family: that its stability as an institution hinges on the maintenance of unequal

resource positions between women and men; and that economic self-interest plays an

important role in intra-family gender relations, which would be revealed with particular

starkness in gender conflict over a critical form of property such as arable land.  Such a

picture of the family is a far cry from that implicit in much of standard economic theory,

namely of the family as an undifferentiated unit governed primarily or solely by altruism. 



4

 For interesting discussions on problems associated with a unitary conceptualization of the household,3

see, among others, the writings of economists Doss (1996), Folbre (1986, 1988), Haddad, Hoddinott, and
Alderman (1994), Hart (1993), IDS Bulletin (1991), Katz (1992), Agarwal (1994a), Kabeer (1994), Nelson
(1994), Ott (1995), Seiz (1991), and Sen (1983, 1990); and anthropologists Guyer (1981), Harriss (1981),
Moore (1991), as well as others in Guyer and Peters (1987).  In addition, a critique of unitary household
models is implicit in the writings of a number of economists (mentioned below) who have formulated
alternative models.

 Some scholars distinguish between "household" and "family" (e.g., Shah 1973), but I have used them4

interchangeably, given the empirical variability of these units across regions, and their definitional variability
across the literature (Guyer and Peters 1987; Kolenda 1987).  For instance, households can be commensal and
residential units, and/or units of joint property ownership, production, consumption, and investment, or they
can constitute some intersection of these dimensions.  They also vary in membership composition from units
of single persons, to those of parents and children, to those with additional relatives: siblings, grandparents,
and so on.

 Also referred to as the common preference model, the altruistic model, or the benevolent dictator5

model.

The latter picture is today difficult to defend in the face of growing evidence to the

contrary.   Households/families (I will use the two terms interchangeably ) are3         4

recognizably constituted of multiple actors, with varying (often conflicting) preferences

and interests, and differential abilities to pursue and realize those interests.  They are

arenas of (albeit not the sole determinants of) consumption, production, and investment,

within which both labor and resource allocation decisions are made.  And evidence from

many regions reveals persistent gender inequalities in the distribution of household

resources and tasks. 

A number of economists are today grappling with such complexities, within and

outside the game-theoretic format.  Most take as their starting point a critique of Becker's

(1965, 1981) unitary model of the family,  agreeing that it is problematic.  This model5

treats the household as a single entity in relation to both consumption and production.  It

assumes that all household resources and incomes are pooled, and that resources are

allocated by an altruistic household head who represents the household's tastes and

preferences and seeks to maximize household utility.  There is less agreement, however,

on alternative approaches to intrahousehold interactions.  These approaches cover a

diverse range: cooperative, noncooperative, collective, or some mix of these.  In varying



5

 See, especially, Doss (1996), Haddad, Hoddinott, and Alderman (1994), Haddad, Hoddinott,6

and Alderman (1997), Hoddinott (1991), Katz (1996), and Strauss and Thomas (1995).

 See Bourguignon and Chiappori (1992), Browning et al. (1994), and Chiappori (1988, 1992).7

degree, they seek to incorporate the social reality of the family as described in

anthropological and sociological writings.  Differences between the approaches, outlined

in several excellent review articles, do not need detailed repetition here.   Suffice it to6

mention their broad features.  Barring the "collective models"  that make no assumption7

other than that the outcome will be Pareto efficient (and that attempt to uncover the

decisionmaking rules and processes through empirical analysis), all the others

characterize household decisionmaking as some form of "bargaining."  A brief outline of

the bargaining framework is therefore warranted.

Within the bargaining approach, intrahousehold interaction is characterized as

containing elements of both cooperation and conflict.  Household members cooperate

insofar as cooperative arrangements make each of them better-off than noncooperation. 

However, many different cooperative outcomes are possible in relation to who does what,

who gets what goods and services, and how each member is treated.  These outcomes are

beneficial to the negotiating parties relative to noncooperation.  But among the set of

cooperative outcomes, some are more favorable to each party than others—that is, one

person's gain is another person's loss—hence the underlying conflict between those

cooperating.  Which outcome will emerge depends on the relative bargaining power of

the household members. 

A member's bargaining power would be defined by a range of factors, in particular,

the strength of the person's fall-back position (the outside options that determine how

well-off s/he would be if cooperation failed), also termed as the "threat point."  An

improvement in the person's fall-back position (better outside options) would lead to an

improvement in the deal the person gets within the household.



6

The early formal critiques of the unitary model used a cooperative bargaining

approach, and relaxed only the assumption of common preferences while retaining that of

income pooling (Manser and Brown 1980; McElroy and Horney 1981), adapting to a two-

person household Nash's (1950, 1953) formulation of cooperative "bargaining problems"

within game theory.  Household members bargain over the use of the pooled income, the

outcome depending on their bargaining power, determined by their respective fall-back

positions.  These depend, in turn, on what McElroy (1990) terms extrahousehold

environmental parameters (EEPs) such as parental wealth, a person's nonwage income,

and the legal structure governing marriage and divorce.  These cooperative models

assume the attainment of Pareto optimality in household decisions, enforceable and

binding contracts, and symmetrical positions between the parties in relation to

information availability and bargaining ability, and say little about the actual process of

bargaining. 

In contrast, the "noncooperative" models relax many of these assumptions,

including those of Pareto efficiency, income pooling, and enforceable and binding

contracts.  In addition to allowing differing preferences between individuals, they allow

for individual production decisions and asymmetry between the parties with respect to

information and the rules of the game (see Wooley 1988, and Kanbur and Haddad 1994,

among others).

In between are models that combine both approaches, recognizing the possibility of

"separate spheres" of activities while cooperating over, say, the production and/or

consumption of some joint goods or activities.  For instance, Lundberg and Pollak (1993),

who pioneered the notion of separate spheres, point out that for many small decisions in a

marriage, divorce is not a credible threat should cooperation fail.  Rather, people may

remain within the marriage but withdraw into separate spheres, defined by, say, a division

of labor based on socially recognized gender roles that emerge without explicit

bargaining.  This withdrawal option would constitute an "internal" threat point.  The

parties would, however, still bargain over jointly shared goods and activities, such as



7

 Some argue that since available data do not usually permit us to test hypotheses that could help us8

choose between the bargaining and neoclassical household models, the analytically simpler neoclassical model
is preferable. This is not an adequate reason for dismissing the bargaining approach, but merely strengthens
the case for further data gathering. Also, as Katz (1996, 16) notes, "Even if the same predictions can be
generated ... in a unitary framework, it may be asked why parsimony and not descriptive accuracy is the
relevant ... criterion for the choice of model." 

meals and child care, with the bargaining operating like a Nash cooperative game.  In

other words a noncooperative solution is used as a threat point in a cooperative game

(also see Ulph 1988, Jones 1983, and McElroy 1997). 

These "nested" separate spheres of noncooperation within the household may also

be defined by separate semi-autonomous loci of production and consumption (Carter and

Katz 1997; Katz 1992).  In Katz's (1992) alternative "reciprocal claims model," each

household member has a distinct income stream and makes resource allocation decisions

subject to her/his own budget constraint; any transfers of income, labor, land, and other

resources between individuals are explicitly recognized; and household resource

allocation is treated as a set of individually determined allocation choices.  Inter-member

resource transfers are bargained over between the parties concerned.

Attempts to test the relevance of alternative models (excellently reviewed in

Haddad, Hoddinott, and Alderman 1997, and Doss 1996) indicate a considerable

justification for rejecting the unitary model, or at least for "shifting the burden of proof"

to those defending it, even though there is no clear answer as to which alternative model

might be the most suitable.   Rather, as Katz (1996, 19) argues, "unitary, cooperative,8

noncooperative and `collective' decisionmaking rules may all coexist in the same

household," varying by the type of resource or expenditure.

How we characterize the household impinges not just on academic analysis but

also, critically, on policy.  Policymakers in many countries, assuming a unitary model,

have typically directed resources to male household heads, assuming equitable

intrahousehold sharing of resources or benefits thereof.  A bargaining model would

suggest that policies and resources be directed differently, taking account, say, of the



8

gender of the recipient, insofar as the welfare, efficiency, and equity implications could

differ by gender.  Also it would point policymakers (seeking to affect intrahousehold

allocations) to interventions in addition to price changes, such as legal and institutional

changes (see also Haddad, Hoddinott, and Alderman 1994).

While bargaining models have contributed in interesting ways to household

analysis, it is critical to think beyond the restrictions imposed by fully-specified models,

and to move toward a less restrictive formulation that incorporates qualitative aspects,

and greater complexity.  In other words, the bargaining perspective or approach has

particular usefulness in examining gender relations, in the application of which we need

not be restricted to game-theoretic formulations (see, also, Seiz 1991).  Some of these

aspects could no doubt be incorporated into formal models, but there would still be

limitations imposed by structure.

In any case, my purpose here is to focus, through the lens of gender, on some

dimensions that are critical to bargaining outcomes, but which most discussions of

household bargaining treat as exogenous and outside the realm of their analytical

specifications, such as the following:

• What determines intrahousehold bargaining power?  Bargaining models

define bargaining power in terms of fallback positions, but their specification

of what factors impinge on these positions tends to be rather narrow:  few

specify factors beyond income (earned or unearned); even fewer explore

qualitative aspects of power.

• Are there differences in the relative importance of factors that determine

fallback positions?  Clearly, not all factors would carry equal weight.

• What is the role of social norms in determining bargaining power and in

setting the limits to what is bargained over? Again there is inadequate

discussion on this, as there is also on the possible endogeneity of the norms,

viz., of norms themselves being the subjects of bargaining;



9

• How are bargaining processes and outcomes affected by differences in

individual perceptions (about needs, contributions, etc.) and pursuit of self-

interest? In particular, is women's relative deprivation due in part to their

failure to perceive their true self-interest, or to their being more altruistic

than men? Existing discussion on these questions is limited, and much of it

problematic.  The possibilities of bargaining on behalf of others and of

coalitions within households are also not considered.

• What are the links between intrahousehold bargaining and bargaining outside

the household (e.g., in the market, the community, and the State)? And what

determines extrahousehold bargaining power?

In the sections that follow, I examine these neglected dimensions.  Although many

of the illustrative examples are taken from rural South Asia, the issues discussed have

wider relevance.

DETERMINANTS OF INTRAHOUSEHOLD BARGAINING POWER

Any attempt to identify the determinants of bargaining outcomes must grapple with

several complexities.  First, a wide range of factors could define a person's bargaining

power:  some are quantifiable, such as individual economic assets; others are less so, such

as communal/external support systems, or social norms and institutions, or perceptions

about contributions and needs.

Second, some resources are both determinants of a person's bargaining power vis-à-

vis other resources, and themselves need to be bargained for.  Take arable land.  We

could argue that landowning women would have a stronger fallback position and

therefore greater bargaining power than landless women vis-à-vis the allocation of

household subsistence.  Yet, to gain a share in arable land may itself require bargaining,

and a somewhat different set of factors would determine women's bargaining power in



10

relation to land.  Similarly, social norms both affect the outcomes of bargaining and can

themselves be subjects of bargaining.

Third, in bargaining for something like a share in arable land, insofar as the social

or legal legitimacy of any share at all for women may first need to be established, the

outcomes of intrahousehold bargaining would be preconditioned by the outcomes of

extrahousehold bargaining with the community and the State (as elaborated later).

Fourth, the outcomes of bargaining at one point in time, by strengthening or

weakening a person's fallback position, could affect the outcomes of bargaining at a later

point in time.  For instance, assets accumulated in one round of bargaining would affect

the threat point and, therefore, outcomes in the next round.  Such "iterative" bargaining

could be between the same parties, or between different parties (e.g., a property

settlement favoring a widow or divorcee in one marriage could strengthen her bargaining

power in a subsequent marriage); and it could apply to both the short term and the long

term (see, also, Sen 1990, for examples).

Fifth, the outcomes of bargaining need not result from an explicit process of

negotiation between the parties; they could even result from implicit differences in

bargaining power.  For instance, a man in north India rarely has to tell his sister that he

will break all contact with her if she demands her share of ancestral land.  The fact that he

can do so at low economic and social cost to himself, but at high potential cost to her,

may be enough for her to forego her claim.  Indeed, the fact that one party can get a

favorable outcome without open contestation suggests a considerable bargaining power.

Sixth, in a limited sense, relative bargaining power within the household/family

could be revealed in who participates in decisionmaking and about what.  Hence, women

who participate in decisionmaking concerning, say, agricultural production or cash

expenditure in the home may be said to have greater bargaining strength than those

excluded from such decisionmaking altogether.  But more fundamentally, relative

bargaining power is revealed in whose interests prevail in the decisions made, namely in

final outcomes: in the intra-family distribution of resources, goods, services, and tasks,



11

 Sen, in this context, does not expand on how families or individuals arrive at a certain endowment9

position.

 The category, NGOs, is broadly used here to include organizations that differ in size, the social10

backgrounds of their members, their objectives, ideological positions, issues taken up, forms of operation, and
so on.  Some might have a mass base, others small memberships. I will use the term "gender-progressive"
NGOs or groups for those whose activities are centrally or partially aimed at reducing gender inequities.  This
could include organizations with mixed (male and female) membership but with a specific gender focus in their
activities, as well as women's groups promoting gender-specific programs.  "Gender-retrogressive" implies the
opposite.

the treatment meted out by family members, the control exercised over resources, and so

on. 

Consider some of these aspects in more detail. 

Bargaining for Subsistence Within the Family

What determines a woman's bargaining power within the family in relation to

subsistence needs such as food and health care? Sen (1981), in his entitlement approach

to famine, highlights two factors as significant in determining a person's (or a family's)

ability to command subsistence goods (including food) and services: endowments (what a

person owns, such as assets, labor power, etc.) and the exchange entitlement mapping (the

exchange possibilities that exist through production and trade, which determine the

consumption set available to a person with given endowments).  Typically for rural9

families, the most important endowments would be arable land and ability to labor; and

the exchange entitlements would be determined by the possibilities of using these

endowments for production and trade (including seeking employment, in the case of

labor), and by the structure of factor and commodity prices. 

However, we could extend this list to incorporate other entitlements that do not

derive from private ownership nor, usually, from market exchange.  At least three appear

important: traditional rights in communal resources, traditional social support systems,

and support from the State and nongovernmental organizations (NGOs).  Further,10

impinging on all the factors mentioned would be social norms embodying accepted



12

notions about the division of labor, resources, etc., and social perceptions about

contributions, needs, and abilities (and therefore about who deserves what).

It could thus be suggested that a rural person's bargaining strength within the family

vis-à-vis subsistence needs would depend (given the exchange entitlement mapping)

especially on eight factors:

1. ownership of and control over assets, especially arable land;

2. access to employment and other income-earning means; 

3. access to communal resources such as those of village commons and forests;

4. access to traditional social support systems such as of patronage, kinship,

caste groupings, etc;

5. support from NGOs;

6. support from the State;

7. social perceptions about needs, contributions, and other determinants of

deservedness; and

8. social norms.

These factors would individually and interactively affect a person's ability to fulfil

subsistence needs outside the family.  The premise here is that the greater a person's

ability to physically survive outside the family, the greater would be her/his bargaining

power over subsistence within the family (although, as will be elaborated below, factors

such as norms and perceptions also affect bargaining power independently of the fallback

position).  Inequalities among family members in respect of these factors would place



13

 This approach could be adapted to the situation where persons withdraw into separate noncooperative11

spheres of activity, while maintaining some "spheres of cooperation" vis-à-vis goods and services that are
jointly produced and/or consumed.

 Some suggest that women's childbearing and child-rearing responsibilities could also weaken their12

household bargaining position (e.g., Anne Marie Goetz, personal communication, Sussex, 1992).  In my view
the effect of this factor is difficult to judge a priori.  In South Asia, for instance, on the one hand, frequent
pregnancies and caring for a large number of young children could reduce women's bargaining power, say by
limiting their employment options.  On the other hand, childlessness, equated with barrenness, could lead to
divorce (Cain 1988).  Again having sons could increase a woman's bargaining power, but having only
daughters could weaken it.  Also, the extent of women's reproductive responsibilities varies across cultures,
and much depends on what autonomy a woman can exercise with regard to how many children she has, her
access to child-care support (through relatives or the State), and her husband's contribution to child care.

some members in a weaker bargaining position relative to others.  Gender is one11

significant basis of such inequality.12

While my focus here is on the determinants of well-being if cooperation should

fail, it is important to note that many of the factors that determine a person's fallback

position also influence her/his ability to make contributions within the relationship.  If a

woman loses her job or her assets, it both worsens her fallback position and diminishes

the income she can bring into the family.  This dual effect can lead to marriage

dissolution and family abandonment in periods of severe crisis (such as a famine), as

discussed further below.

These determinants of bargaining power can complement or substitute for one

another.  The significance of the first two factors, namely command over property

(especially land) and access to employment, in strengthening a rural person's survival

ability outside the family, is self-evident; that of access to communal resources and

various external support systems needs some elaboration, as do social perceptions and

social norms.

In rural economies, village commons (VCs) and State forests are especially

important for two reasons: one, they provide a wide range of items essential for daily use,

especially to the poor whose dependence becomes critical during drought and famines

(Agarwal 1990); and two, their products, typically gathered by women and children



14

(Agarwal 1991), provide women with an independent source of subsistence unmediated

by dependency relationships on men.  Women usually have use rights in VCs by virtue of

their membership (through birth or marriage) in the village community, whereas their

access to the cash economy and (in areas of strong female seclusion) to the marketplace

itself may be constrained and dependent on the mediation of males (Agarwal 1994a).

Similarly, social support systems of patronage, kinship, caste groupings, and even

friendships (namely, various forms of "social capital") can prove critical during economic

crises.  These support systems are constituted of relationships between persons or social

groups in which, usually, considerations other than the solely economic take precedence,

falling under the rubric of "the moral economy" (Scott 1976).  They typically relate to

nonmarket transactions, such as interest-free credit from relatives or friends, or

intergenerational transfers between parents and children.  Again, these support structures

can prove especially important for women (Agarwal 1990).

Other external systems of potential support, such as the State and NGOs, can add to

a person's intrahousehold bargaining power, both by a direct provisioning of subsistence

(e.g., widow's pensions), and indirectly by increasing access to employment, assets,

credit, infrastructure, etc.  Many NGOs, in particular, have in recent years helped enhance

household subsistence possibilities in these and other ways, by helping people to

challenge biases in State laws, policies, and their implementation.

 However, these interventions could either increase or decrease women's bargaining

power, depending on their gender bias.  Organizations that enhance credit and income-

earning opportunities for women relative to men, such as the Grameen Bank in

Bangladesh and the Self-Employment Women's Association (SEWA) in India, would

strengthen women's bargaining power.  NGO interventions that enhance only men's

access to land or credit would weaken women's bargaining power.  The same holds for



15

 Most State interventions in relation to land have, in fact, tended to strengthen men's bargaining13

power:  land tends to be distributed to male heads of households in almost all land reform programs and
resettlement schemes.  NGO interventions typically replicate this bias.  Credit interventions used to be similarly
male-biased, but this has changed in recent years, with more recognition being given both by the State and by
NGOs to women's independent credit needs.

 For a definition of gender-progressive groups, see footnote 10.  For the role such groups can play,14

see, especially, Agarwal (1994a).  Folbre (1997) similarly emphasizes the role of what she terms "gender
coalitions."

State interventions.   Gender-progressive organizations, including women's groups, can13

play a particularly important role in directing State policies and laws in women's favor.14

In other words, the first six factors listed, by impinging on women's and men's

subsistence opportunities and access to resources from outside the family, would affect

their bargaining power and so their access to subsistence within the family as well.  The

last two factors, namely social perceptions and social norms, can affect subsistence

distribution both directly, in that intrahousehold allocations depend on perceptions about

deservedness and on prevailing norms of sharing within families, and indirectly by

impinging on the other six factors.  Social norms is a vast issue and is therefore discussed

in a separate section.  Consider that of social perceptions. 

Social Perceptions and Bargaining Power

There can be, and not uncommonly is, a divergence between what a person actually

contributes, needs, or is able to do, and perceptions about her/his contributions, needs, or

abilities.  In particular, a person's contributions may be undervalued because of her gender

or race.  The work women do might be labeled "unskilled" and that which men do as

"skilled," simply because of their gender, even if the tasks done by both require equal

amounts of skill.  Perceptions about contributions can also depend on how "visible" the

work is: home-based or unwaged work is often seen as less valuable than work that is



16

 See, especially, Agarwal (1986) and Sen (1990).15

 See, for example, Abdullah and Zeidenstein (1982), and Goetz (1990).16

 Sen focuses only on "contributions" as the principle underlying distribution.  In fact, as noted below,17

notions about legitimate shares can stem from a range of principles, of which contributions is but one:  see,
for example, Engle and Nieves (1993), and Farmer and Teifenthaler (1995).

physically or monetarily more visible.  Indeed, women's contributions to the household15

are typically undervalued not just by family members, but often also by policymakers and

bureaucrats implementing development programs.16

Similarly, perceptions about needs may differ from actual needs.  In many parts of

the world, women's needs are underplayed and assumed to be subordinate to or even

synonymous with the "family's" needs, while for men, the distinction between family and

personal needs is widely accepted and sanctioned.

Such perceptions affect intrahousehold allocations and bargaining power. 

Systematic undervaluation of women's contributions or needs, in a system where these are

important distributive principles, would reinforce gender-related deprivation.  Women

and girls would receive less because their contributions to the household are seen as being

less valuable than that of men or boys—what Sen (1990) terms "perceived contribution

response" —and/or because they are seen as needing less (what one could term17

"perceived need response").  Here, strengthening the fallback position of a rural woman,

by providing her better access to village commons, may have less than the desired effect,

if what she gathers (as nonmonetized items) is seen as having less value than the cash a

man brings in (even if the imputed value of the gathered items is more).  Such

undervaluation is not confined to developing countries.  The Western feminist debate on

"wages for housework" arose from the recognition that unwaged work was "invisible" and

perceived as having little value.  England and Kilbourne (1990), drawing from studies of

American households, argue that women who earn cash have more bargaining power than

those who are solely housewives, because of, among other things, the cultural devaluation

of housework.  Women's entry into wage labor could thus be one way of increasing their



17

intra-family bargaining power not just directly, but indirectly, by increasing the perceived

legitimacy of their claims (see, also, Sen 1990).

At the same time, a woman's bargaining power outside the household, say in the

labor market, is also affected by perceptions; for example, solely on account of her

gender, she may be perceived as having lesser ability or commitment, or to be only a

supplementary earner.  Gender, as is race, has been known to define perceptions about

abilities, and to lead to discriminatory hiring and payment practices.  Rural women in

many parts of South Asia, for instance, are paid less than men even for the same tasks, on

the assumption that women are less productive, although few productivity studies have

been conducted, and some that exist show the contrary (Agarwal 1983).  Incorrect

perceptions can thus reduce a woman's bargaining power in relation to family subsistence,

not only by leading to an underestimation of her needs and an undervaluation of her

waged contributions, but also by affecting her "worth" in the labor market, thus limiting

the mentioned potential advantage of her seeking waged work.

It may be noted that perceptions impinge on social norms but are not the same as

social norms.  For instance, norms might define on what principles family food is

shared—say, contributions and/or needs, but the translation of those norms into

allocations would depend not just on actual, but perceived contributions and needs. 

Social norms relate to customs that are established.  Incorrect perceptions could get

institutionalized as social norms, such as in systematically lower wage rates or lower

subsistence allocations for women than men.  But perceptions would usually be only one

among several factors influencing norms, while also affecting bargaining power

independently of norms.  Like norms, perceptions may themselves be subject to

contestation and change.

Do All Factors Carry Equal Weight?

Clearly, not all factors affect bargaining power in equal extent.  However, even the

few economic studies that list factors that might affect intrahousehold bargaining power



18

 Guyer (1997), an anthropologist among economists, does, however, implicitly prioritize wealth and18

assets.

 See Breman (1985) and Commander (1983) on eroding patron-client relationships, and Cain,19

Khanam, and Nahar (1979), Drèze (1990), Fernandes and Menon (1987), Jansen (1983), and Kabeer (1994)
on declining support from kin.

 See Agarwal (1990), Drèze (1990), Fernandes and Menon (1987), Jansen (1983), and White (1992).20

do not discuss the need for setting priorities.   Identifying the more critical factors (these18

would vary by context) is especially important for policy.

In agrarian economies, for instance, of the first six factors listed as affecting

bargaining for subsistence, effective command over landed property holds a privileged

position for several reasons.  For illustration, consider evidence from rural South Asia,

although the points made are of wider relevance.  First, the rapid decline in forests and

VCs, especially in semi-arid areas, is effectively eroding this source of supplementary

economic support for the poor, and especially for women (Agarwal 1991).  Second,

erosion is also occurring in social support systems of patronage, kinship, and caste

groupings.   The decline in kin support is especially apparent among communities that19

have become poorer over time, the effects being dramatic in tribal communities

traditionally characterized by a high degree of communal and intra-gender cooperation in

work and social life.  Among them, and elsewhere, the worst affected are usually women,

especially the widowed and aged.20

Third, the returns from wage employment and other income-earning means are

themselves often linked with access to land; for instance, rural nonfarm earnings in South

Asia are substantially greater among households with some land, relative to the totally

landless (Chadha 1992; Islam 1986), as is the probability of women finding wage

employment (Ryan and Ghodake 1980; Lipton 1983).  Families with some land also have

a higher reserve price for their labor, which can push up aggregate wage rates (Raj and

Tharakan 1983; Bardhan 1984).  Effective rights in land can thus strengthen women's



19

 For example, Lucas and Stark (1985) found this in relation to sons' remittances to parents in a study21

for Botswana.

 On this point, Dasgupta (1993, 329) misinterprets my paper on drought and famine (Agarwal 1990)22

in attributing to me the argument that a collapse of the woman's fallback position relative to her husband's in
itself leads to her being abandoned in a famine.  As I have stressed, a woman's fallback position here
diminishes simultaneously with her potential contribution to family income, since factors such as her ownership
of assets and access to employment affect both her fallback position and her ability to contribute economically
to the family's well-being.  Hence, in a severe crisis, while the sharp decline in the wife's fallback position may
improve the husband's bargaining situation, this would provide him little realizable advantage, given the
simultaneous (and severe) decline in her ability to contribute to joint well-being, so that it would still be in his
economic interest to abandon her. 

fallback position not only directly, but also indirectly, by improving returns from other

income sources.

Fourth, especially the elderly are able to use property, in particular, landed

property, to bargain for better care and support from their families (Sharma and Dak

1987; Raj and Prasad 1971).  For some, the mere fact of possessing land helps; others

may use landed property and valuables for explicit bargaining, promising favor to those

family members who serve them best, as elderly women are noted to do in northern India

(Sharma and Dak 1987).  Migrant children's remittances to parents may also be associated

positively with the latter's property status.21

 Fifth, land rights could prove crucial during severe subsistence crises, as during

drought and famine.  In such contexts, poor rural households first dispose of assets such

as jewelry, household utensils, and small animals, keeping the productive

resource—land—until the last (Agarwal 1990).  While disposing of the more liquid assets

first makes economic sense at the household level, it also has important gender

implications.  The items noted to be disposed of first are often the only ones a woman

owns, while land is typically in men's names.  As a result, women tend to be left with

both a weaker fallback position than men, and a more diminished ability to contribute to

family income.  During famines, an oft-noted outcome of this shift in bargaining power

and contributions is the abandonment of wives and children by men (whose outside

options, especially where they have land, do not deteriorate in equal degree).22



20

 Under bilateral inheritance, ancestral property passes to and through both sons and daughters; under23

matrilineal inheritance, ancestral property passes through the female line.  Under patrilineal inheritance,
ancestral property passes through the male line.  The complex workings of these inheritance systems in South
Asia are detailed in Agarwal (1994a).

 See, among others, Davidson (1988), Koopman (1991), and Quisumbing (1994) for Africa; and24

Deere (1985) for Latin America.  Also, in relation to famines, Watts (1983) notes that in Nigeria, small
livestock, typically owned by women, are among the first casualties, and Vaughan (1987) describes women's
increasing dependence on men, due to reduced outside options, during the Malawi famine.

Sixth, land rights can, over time, help women negotiate less restrictive social norms

and better treatment from husbands.  The situation of South Asian women who

traditionally had rights in land, as among communities that practiced matrilineal or

bilateral inheritance, is indicative.   They enjoyed substantial freedom of movement and23

interaction outside the home, often controlled the household food stores, and, as observed

among the matrilineal Khasis of northeast India, ate before their husbands, if the latter

were out late (Agarwal 1994a).  All this is in contrast to the patrilineal northwest, where

women's lives are circumscribed in numerous ways that limit their livelihood

opportunities.  Similarly, women who were landless, but who have, in recent years,

acquired independent plots from the government or through NGO support, report an

enhanced sense of economic security and self confidence (and therefore negotiating

ability), and improved treatment from husbands and kin (Agarwal 1994a).

In other words, command over private land could strengthen rural women's

bargaining power in ways that merely enhancing wage employment opportunities, or

stemming the erosion of common property resources, would not.  Although illustrated by

South Asian evidence, arable land carries similar weight for rural women in Africa and

Latin America.   In more urban industrial contexts, however, property other than land24

would be of greater importance.  But the general point is that prioritizing/ weighting

factors that affect bargaining power is important both analytically and for policy.

This raises an additional issue: what affects bargaining outcomes of the prioritized

factors, say of family land, which are exogenous in some contexts and endogenous in

others.



21

 For an empirical elaboration of the relevance of these factors in the context of South Asia, see25

Agarwal (1994a).

Bargaining over Family Land

Contestation over subsistence resources, given one's endowments, is only one level

of bargaining.  A second, more basic one, involves bargaining over the endowments

themselves.  At this level, the factors determining women's bargaining power become

even more complex.  For instance, a daughter's ability to successfully claim a share in

parental landed property (assuming she is not voluntarily given it) is likely to depend

especially on the following factors:25

• the existing inheritance laws;

• the social legitimacy of her claim, that is, whether the claim (even if legally

valid) is perceived as socially valid by her community;

• her educational status and legal literacy;

• her access to government officials who administer land-related matters,

register land inheritance claims, etc;

• her access to economic and social resources for survival outside the support

systems provided by contending claimants such as brothers or kin; and

• her economic and physical access to legal machinery.

In other words, individual women's attempts to acquire a share in family land could

require interlinked contestation outside the household as well, such as contestation with

the community to establish social legitimacy for women's claims to independent land

rights, contestation with the State to make inheritance laws gender equal and to ensure

their better implementation, and so on.

Gender differences in intrahousehold bargaining power are thus linked with the

person's extrahousehold bargaining power, such as with the community and the State. 



22

This would be especially so in contestation over landed property, since control over

arable land helps define (and is also defined by) wider access to economic, social, and

political power (Agarwal 1994a).  What factors might affect women's bargaining power

in the market, the community, and the State are discussed later.

SOCIAL NORMS

In the literature on intrahousehold economics, only a few authors explicitly

recognize the importance of social norms and model them.  Of those that do, some treat

them as exogenous (Lundberg and Pollak 1993), while a few recognize the possibility of

their being endogenous (e.g., Agarwal 1994a; Folbre 1995, 1997; Hart 1993; Katz 1996). 

But major gaps remain in spelling out the nature and specifics of social norms in

particular contexts, and how they may affect bargaining.  This section seeks to fill some

of these gaps.

Norms could impinge on bargaining in at least four ways:

• They set limits on what can be bargained about.

• They are a determinant of or constraint to bargaining power;

• They affect how the process of bargaining is conducted, e.g., covertly or

overtly, aggressively or quietly.

• They constitute a factor to be bargained over, that is, social norms can be

endogenous in that they can themselves be subject to negotiation and change.

Norms as Limits to What Can Be Bargained About

To begin with, norms set the limits to bargaining.  They can define which issues

can legitimately be bargained over, and which fall in the arena of the uncontestable.  At

any given time, for a given society, some decisions would fall in the realm of what the

French sociologist Bourdieu (1977: 167-70) terms "doxa"—that which is accepted as a



23

 Within "the field of opinion," Bourdieu further distinguishes between orthodoxy and heterodoxy.26

He does not fully spell out this distinction, but implies that orthodoxy would be at one end of the spectrum and
heterodoxy at the other, the former representing one dominant system of beliefs and the latter representing
several alternative systems of beliefs.

natural and self-evident part of the social order, which goes without saying and is not

open to questioning or contestation—the "undiscussed, unnamed, admitted without

argument or scrutiny."  A good deal of what is justified in the name of "tradition" would

fall in this category: "the tradition is silent, not least about itself as tradition."  In contrast

to doxa is the "field of opinion, of that which is explicitly questioned," "the locus of the

confrontation of competing discourses."26

In the present context, doxa could include widely accepted norms and practices. 

Social norms enter virtually every sphere of activity.  They may define what category of

persons cannot intermarry (e.g., most patrilineal north Indian Hindus forbid marriages to

close kin or within the village); the gender division of labor within the home (e.g.,

housework and child care are usually seen as women's responsibilities); the gender

division of labor outside the home (task specification and occupational segregation is

common in both rural and urban employment, in most countries); whether women should

work outside the home (female seclusion norms restrict this among some Hindu and

many Muslim communities); who can participate most in household decisionmaking

(e.g., older daughters-in-law who have sons may have more say than new brides); by what

criterion society's resources should be shared, e.g., "to each according to ability" or "to

each according to need"; and so on.

These limits to bargaining may favor group over individual interests, or favor some

groups over others (say men over women), or favor some individuals over others (say

older women over younger).  They can draw legitimacy from religious or other beliefs. 

And they can reflect the dominant perceptions of the needs and rights of people prevailing

in a community.



24

 In a study of food distribution among Guatemalan households, Engle and Nieves (1993) found27

considerable variation among families in the principles underlying observed distribution.  See, also, Farmer
and Teifenthaler (1995) for a discussion on some of the other criteria.  The principles underlying
intrahousehold distribution of subsistence are not only of academic interest, but have a bearing on public policy
interventions, such as child subsidies and school feeding programs.

 Of course, even equality as a criterion could translate into several possible allocations:  Farmer and28

Teifenthaler (1995) suggest at least six different ways by which food might be allocated to children, all in
keeping with particular notions of equality that the parents might hold.  For instance, parents may care about
equitable food inputs that affect health or about equitable health outcomes, and each of these could be
measured in terms of absolute equality, proportionate equality, or equality of shortfalls.  But these choices
would typically reflect individual judgments, rather than socially-established norms.

Somewhat less restrictively, norms may set limits to bargaining by admitting

something as contestable but restricting the range of contestation: for instance, that males

should eat before females and get the better quality food may be a norm in some societies,

but negotiations may be possible over what allocational criterion— contributions, needs,

equality, or investment (especially in children) —is used to make allocations among the27

males, or among the females.28

The complexity of bargaining is less if the social legitimacy of a person's claim to

some share of the contested item is recognized and only the size of that share is in

dispute, than if the legitimacy of claiming any share at all is questioned.  For instance, the

rights of female family members to basic subsistence are usually not in dispute, although

the proportionate shares may be, while any share for them in ancestral land might be

opposed in some communities and placed outside the arena of bargaining.

The overarching nature of norms suggests that, for a start, contestation would be

necessary to enlarge the range of issues that can be bargained over, and even to admit

specific persons as legitimate contestants.

Norms as Determinants of or Constraints to Bargaining Power

The kinds of restrictions to bargaining outlined above, in turn, affect women's

bargaining power within the household.  Consider, for instance, bargaining for better

subsistence.  Social norms (as suggested earlier) often define by what principles family



25

 See Goldin (1990) for an interesting historical account of women workers in the United States being29

fired upon marriage, and married women not being hired by many manufacturing firms in the 1940s.  She
notes:  "Social consensus had been formed on the necessity for married women to remain at home with their
children and on the need for their husbands to support them" (p.  6).

food may be shared.  Food allocation norms (quantity or quality of food, or who eats first)

favoring males, found, for instance, in northern South Asia, would limit women's and girl

children's ability to bargain for better shares.  Or the criteria justifying distributions—say,

individual contributions to family income or well-being, individual needs, or investment

in a person's future earning capacity—may themselves be gender-neutral, but if (as

suggested earlier) perceptions about contributions, needs, etc., are biased against females,

this again would reduce women's bargaining ability.

Social norms can also weaken women's intrahousehold bargaining position (over

subsistence and other resources) by restricting their earning possibilities in various ways,

such as by discouraging (or even preventing) them from working outside the home,

limiting the range of tasks they may perform, institutionalizing lower wages for them than

for men, restricting their presence in public spaces and thus their access to markets and

the marketplace, defining child care as their responsibility and so limiting their mobility

and job options, ideologically constructing them as dependents and men as

breadwinners,  and so on.29

Similarly, norms can restrict a woman's bargaining power in relation to family land

by providing justifications for upholding gender-unequal property laws; by predefining

men as household heads and thus as the appropriate recipients of land under land reform

and resettlement programs; by necessitating that women be married into distant villages

and thus limiting their ability to claim and manage any share they inherit from parents; by

purdah practices that restrict their access to legal, administrative, and economic

institutions, and so on (for elaboration, see Agarwal 1994a).

Norms also affect bargaining power by defining the extent of voice a person has

within the household (as illustrated further below), and by impinging on the possibility of



26

 On "voice" or "exit" as ways of expressing discontent within an organization, see Hirschman (1970)30

and footnote 54.  See Agarwal (1994a), Carter and Katz (1997), and England and Kilbourne (1990) for
applications of these concepts to intrahousehold bargaining.

"exit."   For instance, women's exit options in marriage would depend not only on their30

economic prospects outside marriage, but on the social acceptability of divorced women,

and their possibilities of remarriage (their worth in the "marriage market").  Divorced and

widowed women, older women, and women with children are typically less "eligible"

than men with these characteristics.

Norms and How Bargaining Is Conducted

Mediated by gender, age, and marital status, social norms often define how

household members should conduct themselves.  In many societies, behavior that is

assertive and loud is much more tolerated in boys and men than in girls and women.  And

among women, assertiveness is more accepted from older women than younger ones,

from mothers-in-law than young daughters-in-law, and from daughters than daughters-in-

law.

Gendered norms thus set the stage for the form that bargaining can take, even

within the marketplace.  Fish trading in South India, for instance, true to its proverbial

reputation, is typically associated with loud haggling and aggressive marketing.  The

women who earn a livelihood by this means risk being dubbed as "masculine" and being

summarily rejected as role models by their educated daughters (Ram 1989).  At times, the

behavior of women fish vendors is even seen by the village men as sexually provocative,

inviting verbal or physical abuse (Roy and Dewan 1988).

Within the household again, the cultural construction of appropriate female

behavior affects their ability to bargain.  For instance, Tibeto-Burman women of Nepal,

who are not subject to purdah, enjoy considerable freedom of movement and are

significant and visible participants in all types of economic activity, including agriculture

and trading.  But even they are subject to subtle aspects of gendered behavioral norms. 



27

These norms impinge, among other things, on women's ability to negotiate their rights,

including property claims within the family.

Consider, for illustration, March's (1988, 19-20) description of the response of a

Tamang (Tibeto-Burman) woman, Nhanu, when her family property was being divided. 

Nhanu had left in her parent's home an expensive bronze drinking bowl, purchased from

the profits of her trading expeditions.  After her father's death, when the brothers were

dividing the family property, she described the event as follows:

I sat there quietly, without saying a word, just sitting and watching as they

each took their separate shares of the family property.

[Whispering] The bronze drinking bowl that I had bought that time in

Kathmandu was given out in my younger brother's—Busru's father's—share.

Well! While they were dividing the shares, I thought to myself, "Oh dear!

My bronze drinking bowl, the one I bought from the efforts of my trips to

Kerong and Kathmandu, has been given out in Busru's father's share!" But I

continued to sit there quietly.

[Loudly] Then well! my second younger brother came up to get his share. 

He said, "That bronze drinking bowl must be given to Elder Sister! That's the

one she bought with the gallon measure of salt she was given after going to

Kerong! The only thing that she bought from that salt was that bronze

drinking bowl; that bowl's hers! She didn't waste even one paisa on that trip."

And then, right then!, he reached out and in a single sweep of his arm, Lo! he

grabbed that bronze drinking bowl back and set it in a separate pile for me. 



28

  In the above quotations, the insertions are as given in March's paper.31

 For South Asia, see Arens and Van Beurden (1977), Bennett (1983) Mandelbaum (1988), and Pastner32

(1974).  For Mexico, see Roldan (1988).  For the United States, see Zelizer (1994), who notes that in the early
twentieth century, when domestic money was still a husband's property, "a wife's chances of additional cash
were limited to ... asking, cajoling, downright begging, and even practicing sexual blackmail.  If these
techniques failed, there was also a repertoire of underground financial strategies, ranging from home pocket-
picking to padding bills" (page 141).

Since he spoke up, they gave it to me and I took that bronze drinking bowl

away with me [laughing].31

Nhanu's reticence (which could well have cost her the bowl) contrasts with the

volubility of her brothers, and underlines expected differences in male and female

behavior even in communities where women are not explicitly constrained from asserting

themselves.

In cultures or contexts where social norms stifle explicit bargaining or voice,

women may be pushed to using implicit forms of contestation.  Persistent complaining,

pleading ill-health, playing off male affines and consanguines against each other,

threatening to return to the natal home, withdrawing into silence, and withholding sex

from husbands, are all means by which women are noted to bargain within the family, not

only in South Asia, but elsewhere, as in Latin America and the United States.   These32

can, however, prove less effective in many contexts than more explicit forms of

bargaining.

Bargaining over Social Norms

Conventionally, economists assume that individual preferences and social norms

are exogenously given.  In bargaining models, too, when social norms have been

introduced as factors, as by Lundberg and Pollak (1993), they have been assumed to be

exogenous, although rare analytical descriptions recognize their endogeneity (Folbre

1995; Agarwal 1994a).  In fact, social norms are not immutable and are themselves



29

subject to bargaining and change, even if the time horizon for changing some types of

norms may be a long one.

Indeed, a good deal of what is socially passed off as natural and indisputable,

including women's roles and modes of behavior, may be the outcomes of past ideological

struggles.  To shift what has long been taken for granted by a community into the arena of

contestation and discourse (from "doxa" to "heterodoxy") may therefore itself require

bargaining.  Gaining acceptance for the idea that the inequities women suffer are not

biologically rooted but socially constructed would be a part of this process, as would

proposing how gender relations can be differently constructed.

In relation to bargaining over social norms, there are three points of particular note:

one, the role of economic factors in pushing people to challenge norms; two, the role of

groups (as opposed to individuals) in enhancing people's ability to challenge norms; and

three, the interactive nature of bargaining within and outside the household in effectively

challenging social norms.  (Virtually by definition, the arena of bargaining over norms

has to extend beyond the household, since for ideas and practices to become "norms"

requires their acceptance beyond the individual household.)

All these aspects are revealed when we examine contestation over purdah norms in

the predominantly Muslim societies of Bangladesh and Pakistan.  In these societies, as

indeed in many other parts of South Asia, women caught in the poverty trap face

conflicting choices between survival needs and social status within the community. 

Many resolve this dilemma by taking up income-generating work, some within the home,

others outside it.  The implications vary accordingly.  Women working at home in

individual isolation are unable to challenge purdah (Shaheed 1989), while those working

in urban factories are beginning to do so.  Garment workers in Dhaka city argue:  "The

best purdah is the burkah [veil] within oneself, the burkah of the mind.  People only say

that working violates purdah in order to keep women down" (Rahela, in Kabeer 1991,

16).



30

It is, however, the collective challenge posed by women as part of an organized

group that is found to best facilitate this process of change, not least because those whom

they need to oppose are often gender-retrogressive groups seeking to maintain, even

define, specific norms.  In Bangladeshi villages, women members of the Bangladesh

Rural Advancement Committee (BRAC) (an NGO seeking to improve the economic and

social position of the rural poor through group schemes for credit and income-generation)

are beginning to question the legitimacy of purdah itself, as something that is defined by

the elite:

When the women from rich households need to go to the town to appear in

court, even to remain in town for 3-4 days at a time, this is sanctioned as

[within the norms of] purdah.  When women from a BRAC-organized group

want to go ...  even for a day, to attend a workshop or meeting ...  [t]heir

action is condemned as bepurdah.  The norms of purdah that may be relaxed

for the wives of the rich can, just as easily and quickly, be clamped down on

the women of other households (a BRAC woman, cited in Chen 1983, 73).

Further: 

The mullahs or religious leaders have some say on what is purdah or

bepurdah but it is village leaders who have the final say.  The mullahs, upon

request from the elders, will start the rumour that such-and-such action or

behavior is bepurdah....  In this way, the rich and elders (through the

religious leaders) can determine what work is suitable or not suitable for

women to perform (a BRAC woman, cited in Chen 1983, 73).



31

On the one hand, economic necessity has created the impulse to challenge these

restricting norms:  "We do not listen to the mullahs anymore.  They do not give us even a

quarter kilo of rice" (Chen 1983. 175-177).

On the other hand, group solidarity within BRAC has clearly strengthened women's

ability to bargain both within the household and with the community.  Some BRAC

women narrate their experiences as follows (Chen 1983, 177, 165):

Now nobody talks ill of us.  They say 'they have formed a group and now

they earn money.  It is good.'

Before the village elders and union-council members abused and threatened

us for joining the group, now they are silent....  Before we did not understand

our ways, now we understand profit and loss....  Before we did not know our

rights to rations and medical services, now we are conscious and exert

pressure to receive our due....  Before we did not go outside our homes, but

now we work in the field and go to the town....  Before our minds were rusty,

now they shine.

Women also report that, as a result of their economic contributions and group strength,

their husbands are now less opposed to them joining BRAC, and are also less physically

and verbally abusive, more willing to allow them freedom of movement, and more

tolerant toward their interaction with male strangers in work contexts.  In other words,

there has been a loosening of restrictive social norms both within the home and outside it.

Economic analysis that treats purdah norms as exogenous, for instance in

specifications of female labor supply functions, would thus be inadequate insofar as

women who undertake income-generating work are able to contest and redefine the

norms, thereby highlighting their endogeneity.  Contestations around other social norms

similarly warrant examination.



32

What Affects Bargaining over Norms?

We can surmise from the above that the ability of persons to challenge norms that

go against their self-interest would depend on at least three factors: their economic

situation; the link between command over property and control over institutions that

shape gender ideology; and group strength.  The latter two aspects need further

elaboration.

First, those who own and/or control wealth-generating property can exercise

substantial direct and indirect control over the principal institutions that shape ideology,

such as educational and religious establishments and the media (defined broadly to

include newspapers, TV, radio, film, theater, as well as literature and the arts).  These can

influence social norms in either gender-progressive or gender-retrogressive directions.  

At the local level, BRAC women's observation that the rich and the village leaders

are able to define purdah norms is also striking.  It links economic and political power

with the ability to bend religious ideologies and social norms to one's own purpose. 

Some writers, while recognizing that social norms can be contested, locate the

contestation perhaps too much in ideology and give inadequate weight to the links

between gender ideologies and economic inequalities, or to economic inequalities as a

significant (although not sole) determinant of relative male-female power within (and

beyond) the household.  Moore (1991, 8-9), for instance, notes: "[T]he relations of

domination and subordination which are at the base of gender inequalities within the

household cannot be explained as a simple outcome of economic inequalities," and

further that "bargaining and negotiation between women and men...  are often about

definitions and interpretations, and it is for this reason that gender relations are always

involved with power." But if power is not to be seen as a thing in itself, we do need to

ask: of what is this power constituted, and what is its source?  Here the interactive effect

of the economic and political appears crucial.  Economic inequalities, while not the only

influence, do usually play a critical role in structuring power relations, by giving some

people greater authority over definitions and interpretations than others.  Here we might



33

 For the Grameen Bank, see Hossain (1988) and Rahman (1986); and for SEWA, see Rose (1992).33

 It has been argued by some (e.g., Bourdieu 1977) that the interest of the dominant groups would be34

to maintain the space of the undiscussed, and that of the dominated to reduce it by exposing the arbitrariness
of the taken-for-granted, and by bringing issues hitherto seen as "private" into the public domain—that is, by
redefining the boundaries of doxa.

 Fraser also elaborates on how discourses about women's needs tend to be structured by the power35

relations between women and men.

 For India, see Bhasin and Agarwal (1983), Chakravarti (1983), Kalia (1979), and Krishnan and Dighe36

(1990).

also link women's lesser command over property with the shaping of norms that

disadvantage them.

Equally, group solidarity and collective action appear critical for contesting social

norms, as is also apparent from the purdah-related examples.  In fact, contestation over

norms may emerge as a by-product of forming groups for the more effective delivery of

economic programs.  The experiences of the Grameen Bank and BRAC in Bangladesh, of

SEWA in India, and of many other groups, support this conclusion.   One might say,33

group organization "empowers" women to confront existing sources of inequality,

including those embodied in representations.   As Fraser (1989, 166) notes:  "[N]eeds34

talk appears as a site of struggle where groups with unequal discursive (and

nondiscursive) resources compete to establish as hegemonic their respective

interpretations of legitimate social needs."   Feminist critiques of school and university35

curricula and texts, of the images and messages of modern media (radio, TV, film), of

religious myths and mythologies, and so on, are a part of this effort to redefine how

women's capabilities, needs, and rights are represented.   But the influence of these36

critiques is likely to depend on both the economic strength and the group strength of the

critics (see, also, Folbre 1994).  



34

SELF-PERCEPTIONS, ALTRUISM, AND SELF-INTEREST

Formal household models assume fully aware, and typically self-interested,

individuals participating in the bargaining process.  But what if some household members

do not act in their own interest and therefore do not bargain to their best advantage?

For instance, Sen (1990) argues that the outcome of bargaining will be less

favorable to a person the less value s/he attaches to her/his own well-being relative to the

well-being of others ("perceived interest response"), and that this tends to be especially so

in "traditional societies" such as India, where women may tend not to think in terms of

self-interest or of their individual well-being:

[I]nsofar as intrafamily divisions involve significant inequalities in the

allotment of food, medical attention, health care, and the like (often

unfavorable to the well-being—even survival—of women), the lack of

perception of personal interest combined with a great concern for family

welfare is, of course, just the kind of attitude that helps to sustain the

traditional inequalities.  There is much evidence in history that acute

inequalities often survive precisely by making allies out of the deprived.  The

underdog comes to accept the legitimacy of the unequal order and becomes

an implicit accomplice (Sen 1990, 126, emphasis mine).

Some other scholars argue that women are, on average, more altruistic than men because

they have a less "separatist" self, or are socialized such that they are less willing than men

to drive hard bargains (see literature reviewed in England 1989); or that women are more

oriented toward fulfilling collective (especially children's) needs and men more oriented

toward personal goods (Beneria and Roldan 1987).

Are women less able to perceive their self-interest or are they more inclined toward

altruism than men? Would they therefore strike weaker bargains?  Consider, first,

women's perception of self-interest.



35

  See Scott (1985) for an elaboration of this term.37

 See Abdullah and Zeidenstein (1982), Jansen (1983), Lindholm (1982), Luschinsky (1962), Nath38

(1984), and White (1992).

Do Women Suffer from False Perceptions?

The idea that women tend to have a less sharp perception of their individual

interests in societies such as India, that is, that they may suffer from a form of "false

consciousness" (in effect, making them complicit in perpetuating their unequal position)

is interesting, but debatable.  The empirical evidence on this, while limited, points more

to the contrary.

Observationally, it is difficult to infer from people's overt behavior whether they

are conforming to an unequal order because they fully accept its legitimacy, or accept it

partially, or out of fear, or because they believe they have no other options.  For

understanding women's perceptions about the inequitable nature of gender relations, we

therefore need to examine not only their overt acts of resistance, but the many covert

ways in which they express their disaffection.  Empirical work that probes women's

covert responses, by recording their views in contexts where they can express themselves

freely, or by using participant observation methods to penetrate their "subculture,"

provides diverse examples of women's "everyday resistance"  to intrahousehold37

inequalities in resource distribution and control, and to their double work burdens.  For

instance, there are numerous cases of South Asian rural women living under norms of

seclusion, covertly trying to get some cash that they can independently control, by secretly

undertaking income-earning activities, or by clandestinely selling small amounts of

household grain to safeguard their earnings from husbands and in-laws.   Abdullah and38

Zeidenstein (1982, 47), summarizing their many interviews in Bangladeshi villages,

observe:



36

Women told us usually what other women have done.  For example, one

woman stocked rice in another woman's house so her husband could not

know she had it.  Another woman had a neighbour raise a goat for her so her

husband would not know about it.  Yet another woman has opened a pan

business with her young son and has told him to keep their earnings a secret

from the husband.  Most women say that they hide their savings in holes in

the bamboo, in the roof, or under piles of cloth.

In the North-West Frontier Province (NWFP) (Pakistan), Lindholm (1982, 201) notes:

"The husband considers this [i.e., a wife secretly selling grain] theft, but the wife

considers it her just dues for her work."  In Sri Lanka, women coir workers "usually hide

their money in different parts of the house, so that, after a beating [the woman] can

disclose one place, thereby giving [the husband] the illusion she has handed all her

savings to him" (Risseeuw 1988, 278).

Although most women spend the money they so control on family subsistence,

some also spend it on their own needs (Abdullah and Zeidenstein 1982), or buy gifts for

family members to win their support and affection (Luschinsky 1962)—in other words, to

build "social capital."  Yet others invest in goats and cows, which they keep in their

parental homes.  This is especially typical of women living with their in-laws, where

earnings have to be shared with the extended household (Nath 1984).  Unequal food-

sharing in a joint family may also be circumvented by women in ingenious ways,

including by holding clandestine picnics with women friends (Enslin 1990), or feigning

spirit possession to extract food items otherwise denied them (Khan 1983; also personal

observation in north India).

Equally, there are many indications from sociological studies, based on interviews

with peasant women, that women by no means readily accept the unequal gender division

of labor as legitimate, whether they covertly resist it or merely lament about it.  White

(1992, 318), for instance, recounts how village women in Bangladesh may serve tea



37

 Although my focus here is on women's resistance to intra-family gender inequalities, there are also39

examples from Asia of women's covert resistance in the workplace.  See, for instance, Ong (1983) on women
electronic factory workers in Malaysia claiming spirit possession to resist strict factory discipline; and
Gunawardena (1989) on frequent absenteeism, tardiness, and irregular work hours during peak cultivation
seasons among Sri Lanka's women plantation workers.  Based on her 18 months of fieldwork, Gunawardena
notes:  "Rural women ... simply did not comply to the dominant forces operating in their lives, but devised
means by which to skirt, side step, and bend the system, so to speak, to their advantage whenever possible."
She calls this strategizing for maximizing self-advantage.

without milk to their husbands' friends so that the men would "not think she had nothing

better to do than make tea for them all day and should be discouraged from returning." 

Peasant women in north India comment:

Agricultural labourer men help Jat men in the fields, but for Jat women it

only means more work.  We have to cook more food and feed the labourers

as well....  Women should also have fixed hours of work (Horowitz and

Kishwar 1982, 17).

We women stay at home and do back-breaking work even if we are feeling

ill or if we are pregnant.  There is no sick leave for us.  But we do not have

any money of our own and when the men come home we have to cast our

eyes down and bow our heads (Sharma 1980, 207).

All these examples, in different ways, challenge any simple notion that women in

rural South Asia (or indeed elsewhere) have accepted the legitimacy of intrahousehold

inequality.   The overt appearance of compliance ("cast our eyes down") need not mean39

that women lack a correct perception of their best interests; rather, it can reflect a survival

strategy stemming from the constraints on their ability to act overtly in pursuit of those

interests (e.g., "we do not have any money of our own").  Hence, although I agree with

Sen (1990, 126) that "it can be a serious error to take the absence of protests and

questioning of inequality as evidence of the absence of that inequality," I would add that



38

  Sen (1990), while recognizing that deprived groups may comply for many different reasons—habit,40

hopelessness, resignation, etc., sees this as resulting in their willingness to accept the legitimacy of the
established order rather than in their covertly resisting that order.  He writes (1990, 127):  "Deprived groups
may be habituated to inequality, may be unaware of possibilities of social change, may be hopeless about
upliftment of objective circumstances of misery, may be resigned to fate, and may well be willing to accept the
legitimacy of the established order."

Bourdieu's (1977, 167-170) notion of "doxa" and Gramsci's (1971) characterization of "hegemony" are
also of interest in this context, but neither writer explicitly addresses or resolves this issue, although Gramsci's
writings suggest an emphasis on consent via internalization.

it can equally be an error to take the absence of overt protest as the absence of a

questioning of inequality.  Compliance need not imply complicity.40

It is also likely that while on some issues women articulate or even believe in

ideologies that benefit men—for instance, maintaining that child care is women's

responsibility, on other issues there is observable opposition, such as towards family

authority structures, male control over cash, and domestic violence.  Class factors might

also affect to what degree women see their self-interest as congruent to that of the

household.  In northern South Asia, it is middle and rich peasant women (who benefit

from their husbands' properties and face greater social restrictions on outside

employment), rather than women agricultural laborers, who more typically insist that it is

important to have sons for continuing the lineage, and who have a more negative attitude

toward daughters (Horowitz and Kishwar 1982; Gardner 1990; personal observation).

In explaining gender inequalities, I would therefore place much less emphasis than

Sen does on women's incorrect perceptions of their self-interest, and much more on the

external constraints to their acting overtly in their self-interest.  Or, to put it another way,

what is needed is less making women realize they deserve better, than having them

believe they can do better (by building their self-confidence, providing information, etc.),

and by helping them to, in fact, do better, through strengthening their bargaining position. 

Grassroots organizing experience among women, in South Asia and elsewhere, also bears

this out.



39

 See Beneria and Roldan (1987), Blumberg (1991), and Mencher (1988).41

 See Agarwal (1986), Drèze and Sen (1989), and Kabeer (1994).42

 See Papanek (1990), Sharma (1980), and White (1992).43

 Patrilocal: the wife takes up residence with the husband and (with or near) his patrilineal kin.44

Matrilocal: the husband takes up residence with the wife and (with or near) her matrilineal kin.

 Women in traditionally matrilineal and bilateral communities also often openly challenge their45

husbands in ways that women in patrilineal communities usually do not (Agarwal 1994a).

Altruism or Self-Interest?

Unlike the notion of false perceptions, altruism (like self-interest) implies self-

awareness.  But altruism, too, can affect bargaining outcomes.  The question is:  are

women more altruistic than men?

At one level, some of women's actions within families appear to support this view. 

For instance, developing country evidence shows that poor women spend the income they

control largely on family needs rather than on personal needs.   Again, women in South41

Asia usually forfeit their inheritance claims in land in favor of brothers (Agarwal 1994a);

peasant women in north India and Bangladesh often eat last and least, while feeding the

best food to their sons and husbands;  and so on.  A number of writers explain such42

actions in terms of women being socialized into acting more responsibly or more

altruistically than men.   This may well be part of the explanation, but does not appear to43

be all of it.

First, there are significant differences in such behavior patterns across regions and

communities.  For instance, in contrast to women in the patrilineal-patrilocal northwest,

women among the matrilineal-matrilocal Khasis of the northeast (as noted)  do not wait44

for late-returning husbands before eating their evening meal.   While this could reflect45

differences in socialization, there are also clear variations in women's material conditions

in these two contexts, with the women from patrilineal families in the northwest being



40

 See, especially, Papanek (1990) and Cain (1988) on the economic insecurity that north Indian and46

Bangladeshi rural women, in particular, face in the absence of a son.  Sons improve a woman's bargaining
power in her conjugal home with both the husband and his kin.  Maher (1984) describes a similar situation in
Morocco.

much more dependent on male members than those from matrilineal families in the

northeast.  

Second, with limited outside options, women might well seek to maximize

"family" welfare, because it is in their long-term self-interest (even if it reduces their

immediate well-being), insofar as women are more dependent on the family for their

survival than are men.  This dependence can be both economic and social.  Socially, for

instance, where female seclusion is strong, women need male mediation to deal with

outside-family institutions; or widowhood may carry social disabilities (as in India) that

widowerhood does not.  Also, women's dependence on the family can be longer lasting

than men's, given women's higher life expectancies.  In the circumstances, women may

well have, or believe they have, no other option than favoring family members over

themselves.

Third, a woman investing more in sons than in daughters, as in northern South

Asia, appears to be acting more out of self-interest than altruism, when read in the light of

prevailing male advantage in labor markets and property rights, of women's need for male

mediation in the community, and of their dependence on sons in widowhood or old age.  46

Indeed, one might ask: would altruism be so obviously sex-selective? 

Thus, if women forego their claims in family assets in favor of sons, brothers, or

the extended family, or give gifts to kin to secure their affection, these could be

interpreted as ways by which women with a weak resource position seek to strengthen

their family ties in order to ensure economic and social support when they need it,

sacrificing their immediate welfare for future security.

A similar interpretation could fit the observation of an increasing number of

Bangladeshi village women today asserting (or proclaiming their intention to assert) their



41

 See also England (1989), Folbre (1994), Lucas and Stark (1985), and Sen (1982), the last47

especially on the need to accommodate "commitment" as a part of behavior, commitment defined in terms of
"a person choosing an act that he believes will yield a lower level of personal welfare to him than an alternative
that is also available to him" (p.  92).

land rights, while a generation ago, their mothers gave up those rights in favor of

brothers.  Here the mothers' behavior may appear altruistic and that of the daughters self-

interested.  But given that, today, kin support structures are eroding, while earlier such

support was more readily forthcoming, both actions would be congruent with self-

interest.

In other words, if women expend their energies and earnings on the family and

extended kin, this appears to be as consistent with self-interest as with altruism.  Or both

altruism and self-interest might be operating,  although it is difficult to say in what47

relative measure, and with what variation by context.

Stark (1995), in fact, argues that even altruistic behavior may be specifically

cultivated by parents ("preference shaping") out of self-interest, to ensure that their

children look after them in old age.  To inculcate in the children "an internal enforcement

mechanism," parents may use the demonstration effect, by behaving in an altruistic

manner toward others in front of their children.

Finally, to the extent that both altruism and self-interest motivate behavior, this mix

need not be limited to women, although notions of self-sacrifice, nurturance, and so on,

are usually more emphasized for women than men (Papanek 1990).

The recognition that women, like men, may be motivated by self-interest (rather

than only or mainly by altruism), and that both women and men may be concerned with

individual as well as family welfare, even if in differing degrees, and even if their overt

actions place them on different sides of the spectrum, also focuses attention more directly

on the material constraints that shape women's behavior.  It cautions against explanations

that are biological in their thrust ("women are by 'nature' more self-sacrificing"), or that



42

  Folbre and Hartmann (1988) note that by virtue of their association with the family and home,48

women have come to be portrayed as relatively "non-economic," naturally altruistic creatures.  This
portrayal has been used to justify women's lower wages and limited job opportunities, but "women's
commitment to family is not necessarily a function of their preferences or their productivity.  It is often
constrained by the reluctance of other family members to help with housework and child care responsibilities"
(p.  195).

presume deficiencies in women's self-perception or economic motivation,  just as it48

cautions against assuming that self-interest is the only motivating factor.

In terms of governmental as well as grassroots interventions, this recognition would

call for a strengthening of women's fallback position, so that they are less economically

and socially dependent on sons, husbands, or brothers, rather than locating solutions

primarily in raising women's awareness of what constitutes their well-being.

Altruism and Intrahousehold Coalitions

Bargaining models assume that each person will bargain on her/his own behalf

within the household.  However, just as there can be interest coalitions outside the

household, so there can be interest coalitions within it, say, between co-wives in a

polygamous household, or between mothers and children, or between mothers and sons.

This also impinges on the question of bargaining and self-interest in an interesting

way.  Even women who may be willing to sacrifice their own interest for that of family

members out of altruism may strike a hard bargain with their husbands on behalf of their

children.  Indeed they may do so more overtly than if they were acting solely on their own

behalf; that is, women's concern with "family needs" need not include the husband's

needs.  Some evidence from South Asia and Africa does suggest that women, not

uncommonly, see their interests as congruent to those of their dependent children and



43

 In many African societies, the mother-child unit has a "relative autonomy and separate identity"49

(Guyer and Peters 1987, 207).  In India, Ross' (1961) grading of emotional closeness in 11 types of
relationships among Hindu joint families in Bangalore City put the mother-son and brother-sister
relationships in the top two positions, and husband-wife as second to last.  And Maher (1984, 115-116) notes
that in Moroccan villages, "women look on husbands and fathers as potential enemies and sons and brothers
as potential allies in the struggle they engage in to mitigate the power of the former over the conditions of their
existence."

  Folbre (1997), in her discussion of "gender-specific environmental parameters," and Agarwal50

(1994a) appear to be among the few who, in different ways, have engaged this question.

potentially antagonistic to those of their husbands.   In such contexts, there need be no49

necessary contradiction between women's possible altruism on behalf of children and

their bargaining hard with their husbands.

In light of the above, the idea of women's false perceptions appears to have been

overstretched, as also are the arguments concerning the gender-specificity of altruism.  At

the same time, it cannot be assumed (as most bargaining models do) that women and men

are motivated solely by self-interest.  To the extent that women as well as men might be

motivated by both altruism and self-interest (as appears realistic to assume), but in

degrees that cannot be specified a priori, it would be difficult to predict outcomes (see,

also, England 1989).  The implications of intrahousehold coalitions in determining

bargaining power could, however, be examined empirically.

3.  BEYOND THE HOUSEHOLD:  THE MARKET,

THE COMMUNITY, AND THE STATE

We have noted at various points in the paper that women's bargaining power within

the home is clearly associated with their situation outside it.  Although, as mentioned

earlier, some discussions of household bargaining recognize that "extrahousehold

environmental parameters" (McElroy 1990) impinge on intrahousehold bargaining power,

we need to go beyond mere recognition to examine how such parameters can themselves

be bargained over.50



44

 See, for example, Barrett (1980) and Kumar (1989).51

Outside the household/family, gender interactions take place in several arenas, of

which three are especially important: the market, the community, and the State.  The

bargaining approach can usefully be extended to characterize gender interactions in these

arenas as well.

THE MARKET

Unlike the theoretical ambiguity about motivation that surrounds characterizations

of gender relations within the household, market relations are unambiguously depicted in

economic analysis as guided by self-interest.  In this arena, bargaining takes its most

explicit form and has been focused on widely, especially in the context of labor markets

and trade unions.  Given the attention that market-related bargaining has already received,

it is not detailed here.  What does need mention, however, is that women's ability to

bargain in the market, as in other arenas, is mediated by gendered norms and practices;

and that owning and controlling property (especially landed property, in agrarian

contexts) is important for strengthening women's bargaining power in the market as well. 

Both aspects were mentioned earlier, but the former needs some elaboration.  

Consider the labor market.  Bargaining may occur over wages, the duration and

intensity of work, work conditions, and so on.  But women's bargaining power in the

workplace (in comparison with men's) would be constrained not only by gender gaps in

skills, information, and education, but also by women's domestic responsibilities, which

reduce their job options; by employers' assumptions (which, as noted, may be quite

erroneous)  regarding women's abilities, work commitment, efficiency