Measuring Women’s Empowerment

as a Variable in International Development

Anju Malhotra, PhD

International Center for Research on Women

Sidney Ruth Schuler, PhD

JSI Research and Training Institute

Carol Boender
Consultant, JSI Research and Training Institute

This paper was commissioned by the Gender and Development Group of the World Bank, in
support of an ongoing policy research effort on gender and development issues, and by the
World Bank’s Social Development Group, with funding from a Norwegian trust fund, as part
of a larger study on Empowerment and Social Inclusion.

Final Version: June 28, 2002

 Acknowledgements
The authors would like to thank Keera Allendorf for providing research support on this
paper. We would also like to thank Karen Mason, Lynn Bennett, Andy Mason, Naila Kabeer,
Vijayendra Rao, and Deepa Narayan for their helpful comments and feedback.

1
 Table of Contents
Introduction ..................................................................................................................................... 3

Section 1. Conceptualizing Women’s Empowerment .......................................4

How Should Empowerment Be Operationally Defined? ................................................................ 4
Consensus on Conceptualization..................................................................................................... 5
Sifting through Terminology........................................................................................................... 6
The Process of Empowerment ......................................................................................................... 8

Section II. Measuring Women’s Empowerment............................................. 10

Measuring Empowerment from a Universalist Perspective .......................................................... 10
Multidimensionality and Existing Frameworks ............................................................................ 11
Levels of Aggregation................................................................................................................... 14
Challenges to Measuring Empowerment ....................................................................................... 17
Empowerment Is Context Specific..................................................................................... 17
Difficulties in Measuring a “Process”.............................................................................. 19

Section III. Evidence from Empirical Studies................................................. 21

Overview ....................................................................................................................................... 21
How Has Empowerment Been Measured Empirically? ................................................................ 24
Household-Level Studies ................................................................................................... 24
Aggregate-Level Studies.................................................................................................... 29
Intermediate-Level Studies ................................................................................................ 32
How Conclusive Is the Evidence? ................................................................................................. 33

Section IV. Conclusions and Recommendations............................................. 34

Conclusions ................................................................................................................................... 34
Recommendations for Next Steps ................................................................................................. 35

Appendix A: Dimensions of Empowerment Proposed by Various Authors ................................. 37

Appendix B: Empirical Studies Where Empowerment is the Outcome of Interest ...................... 38
Appendix C: Empirical Studies Where Empowerment Affects Other Outcomes of Interest ....... 42

References Cited............................................................................................................................ 51

2
Introduction

The World Bank has identified empowerment as one of the key constituent elements of
poverty reduction, and as a primary development assistance goal. The Bank has also made
gender mainstreaming a priority in development assistance, and is in the process of
implementing an ambitious strategy to this effect. The promotion of women’s empowerment
as a development goal is based on a dual argument: that social justice is an important aspect
of human welfare and is intrinsically worth pursuing, and that women’s empowerment is a
means to other ends. A recent policy research report by the World Bank (2001a), for example,
identifies gender equality both as a development objective in itself, and as a means to promote
growth, reduce poverty, and promote better governance. A similar dual rationale for
supporting women’s empowerment has been articulated in the policy statements put forth at
several high- level international conferences in the past decade (e.g,. the Beijing Platform for
Action, the Beijing+5 declaration and resolution, the Cairo Programme of Action, the
Millennium Declaration, and the Convention on the Elimination of All Forms of
Discrimination against Women [CEDAW]).

Yet to date neither the World Bank nor any other major development agency has developed a
rigorous method for measuring and tracking changes in levels of empowerment. In the
absence of such measures, it is difficult for the international development community to be
confident that their efforts to empower women are succeeding and that this important
Millennium Development Goal will be achieved.

This paper is a first step in the attempt to outline the most promising methodological
approaches to measuring and analyzing women’s empowerment. We review major strands of
theoretical, methodological, and empirical literature on empowerment from the fields of
economics, sociology, anthropology, and demography, and attempt to summarize what we
know and don’t know about what leads to women’s empowerment, and its consequences for
development and poverty reduction. As a first effort at covering this vast and interconnected
topic, we have been selective in limiting ourselves to literature that is at the core of the
discourse on women’s empowerment and gender relations, leaving as a next step its
connection with a broader range of literatures and discourses of relevance. Based on our
analysis, we provide some concrete recommendations regarding where the field stands in
defining, conceptualizing, and measuring empowerment, and what might be next steps for
utilizing and refining existing fr ameworks, collecting data and conducting analyses, and
incorporating approaches from related literatures. Thus, this review attempts to provide the
following:

1. an indication of the different ways in which empowerment has been conceptualized;

2. a critical examination of some of the approaches that have been developed to measure
and track changes in women’s empowerment ;

3. an examination of some of the ways in which the effects of policies and programmatic
interventions to promote women’s empowerment have been measured;

3
 4. a summary of the evidence on how women’s empowerment affects important
development outcomes such as health, education, fertility behavior, income levels, etc.

The paper begins with a presentation of definitional and conceptual issues. Section II
discusses measurement issues and the challenges to operationalizing women’s empowerment
empirically. Section III reviews the empirical evidence from studies that have examined
women’s empowerment as the outcome of interest as well as an influencing agent on other
development outcomes. Section IV provides some conclusions and recommendations.

SECTION I. CONCEPTUALIZING WOMEN’S EMPOWERMENT

How Should Empowerment Be Operationally Defined?

“Empowerment” has been used to represent a wide range of concepts and to describe a
proliferation of outcomes. The term has been used more often to advocate for certain types of
policies and intervention strategies than to analyze them, as demonstrated by a number of
documents from the United Nations (UNDAW 2001; UNICEF 1999), the Association for
Women in Development (Everett 1991), the Declaration made at the Microcredit Summit
(RESULTS 1997), DFID (2000), and other organizations. Feminist activist writings often
promote empowerment of individuals and organizations of women (Sen and Grown 1987;
Jahan 1995; Kumar 1993) but vary in the extent to which they conceptualize or discuss how
to identify it.

Another line of thought in development promotes social inclusion in institutions as the key
pathway to empowerment of individuals and has at times tended to conflate empowerment
and participation. Capitalism, top-down approaches to development, and/or poverty itself are
seen as sources of disempowerment that must be challenged by bringing “lowers”—the poor
and disenfranchised—(Chambers 1997) into the management of community and development
processes. The growth of civil society and participatory development methods at both macro
and meso levels of society are usually proposed as the mechanisms by which empowerment
takes place (Friedmann 1992; Chambers 1997). For example, Narayan et al. (2000a) focus on
state and civil society institutions at both national and local levels, including informal
institutions such as kinship and neighborhood networks. Institutions at the micro level, such
as those of marriage and the household, are not considered part of the state or of civil society,
but interpersonal gender dynamics within the household are considered part of the equation of
social exclusion and in need of directed efforts at change.

Bennett (2002) has developed a framework in which “empowerment” and “social inclusion”
are closely related but separate concepts. Drawing on Narayan (2002), Bennett describes
empowerment as “the enhancement of assets and capabilities of diverse individuals and
groups to engage, influence and hold accountable the institutions which affect them.” Social
inclusion is defined as “the removal of institutional barriers and the enhancement of
incentives to increase the access of diverse individuals and groups to assets and development
opportunities.” Bennett notes that both of these definitions are intended to be operational, and
describe processes rather than end points. The empowerment process, as she characterizes it,

4
operates “from below” and involves agency, as exercised by individuals and groups. Social
inclusion, in contrast, requires systemic change which may be initiated “from above.” As
Narayan (2002) and Ravallion and Chen (2001, in their discussion of “pro-poor growth”) also
argue, systemic change is necessary to sustain empowerment over time. It is through the
process of social inclusion that the “rules of the game” are modified and institutions
transformed so that economic growth is widely shared. Bennett argues that the distinction
between empowerment and social inclusion is important because the World Bank’s
comparative advantage is at the system or policy level.

In general, women do not take a central place in much of the literature on social inclusion or
empowerment. While clearly, the broad reference to empowerment as the expansion of
freedom of choice and action, as articulated in the World Bank’s Sourcebook on
Empowerment and Poverty Reduction (Narayan forthcoming 2002), applies to women as well
as other disadvantaged or socially excluded groups, it is important to acknowledge that
women’s empowerment encompasses some unique additional elements. First, women are not
just one group amongst several disempowered subsets of society (the poor, ethnic minorities,
etc.); they are a cross-cutting category of individuals that overlaps with all these other groups.
Second, household and interfamilial relations are a central locus of women’s disempowerment
in a way that is not true for other disadvantaged groups. This means that efforts at
empowering women must be especially cognizant of the implications of broader policy action
at the household level. Third, several of the authors whose work we review in this paper
argue that women’s empowerment requires systemic transformation in most institutions, but
fundamentally in those supporting patriarchal structures (Kabeer 2001; Bisnath and Elson
1999; Sen and Grown 1987; Batliwala 1994).

Development cooperation agencies have begun to focus on the development of indicators for
assessing gender equa lity, and in some cases “empowerment,” and there is a growing body of
literature in which efforts have been made to clearly define the concept of empowerment. This
literature emerges from both activist and research writings. A diverse body of research has
emerged on measuring empowerment and relating it to other variables of interest in
international development. This research tends to take place at the interstices of the
disciplines of demography, sociology economics, anthropology, and public health; it does not
occupy a widely recognized niche in any academic field.

Consensus on Conceptualization

Given the diversity in the emphases and agendas in discussions on women’s empowerment,
we found greater consensus in the literature on its conceptualization than expected. There is a
nexus of a few key, overlapping terms that are most often included in defining empowerment:
options, choice, control, and power. Most often these are referring to women’s ability to
make decisions and affect outcomes of importance to themselves and their families. Control
over one’s own life and over resources is often stressed. Thus, there is frequent reference to
some variant of the ability to “affect one’s own well being,” and “make strategic life choices.”
For example, G. Sen (1993) defines empowerment as “altering relations of power…which
constrain women’s options and autonomy and adversely affect health and well-being.”

5
Batliwala’s (1994) definition is in terms of “how much influence people have over external
actions that matter to their welfare.” Keller and Mbwewe (1991, as cited in Rowlands 1995)
describe it as “a process whereby women become able to organize themselves to increase
their own self- reliance, to assert their independent right to make choices and to control
resources which will assist in challenging and eliminating their own subordination.”

Also appearing frequently in definitions of empowerment is an element related to the concept

of human agency— self-efficacy. Drawing mainly from the human rights and feminist
perspectives, many definitions contain the idea that a fundamental shift in perceptions, or
“inner transformation,” is essential to the formulation of choices. That is, women should be
able to define self- interest and choice, and consider themselves as not only able but also
entitled to make choices (A. Sen 1999; G. Sen 1993; Kabeer 2001; Rowlands 1995;
Nussbaum 2000; Chen 1992). Kabeer (2001) goes a step further and describes this process in
terms of “thinking outside the system” and challenging the status quo.

Kabeer (2001) offers a useful definition of empowerment that effectively captures what is
common to these definitions and that can be applied across the range of contexts that
development assistance is concerned with:

"The expansion in people's ability to make strategic life choices in a context where this
ability was previously denied to them."

We employ this definition as a reference point in the present paper. Although brief, this
definition is specific enough to distinguish it from the ge neral concept of “power,” as
exercised by dominant individuals or groups. Kabeer’s definition is especially attractive
because it contains two elements which help distinguish empowerment from other closely
related concepts (as discussed further below): 1) the idea of process, or change from a
condition of disempowerment, and 2) that of human agency and choice, which she qualifies
by saying that empowerment implies “choices made from the vantage point of real
alternatives” and without “punishingly high costs.”

Sifting through Terminology

One reason why the degree of consensus on the conceptualization of empowerment is not
readily apparent in the literature is because of the variation in terminology used to encompass
it. In this review, we do not limit ourselves to theoretical or empirical work using only the
term “empowerment.” The literature contains a range of terms, concepts, and data that may
be relevant for assessing “empowerment”; for example, various studies have aimed at
measuring women’s “autonomy” (e.g,. Dyson and Moore 1983; Basu and Basu 1991;
Jeejebhoy and Sathar 2001), “agency,” “status” (e.g., Gage 1995; Tzannatos 1999),
“women’s land rights” (e.g., Quisumbing et al. 1999), “domestic economic power” (e.g.,
Mason 1998), “bargaining power” (e.g., Beegle et al. 1998; Hoddinott and Haddad 1995;
Quisumbing and de la Briere 2000) “power” (e.g., Agarwal 1997; Beegle et al., 1998;
Pulerwitz et al. 2000), “patriarchy” (e.g., Malhotra et al. 1995), “gender equality” (World
Bank 2001a and 2000b), or “gender discrimination.” Often there is no clear demarcation

6
between these terms. Mason (1998) and Mason and Smith (2000), for example, treat
empowerment, autonomy, and gender stratification interchangeably. Similarly, Jejeebhoy
(2000) considers autonomy and empowerment as more or less equal terms, and defines both
in terms of women “gaining control of their own lives vis-à-vis family, community, society,
markets.” In contrast, other authors have explicitly argued that autonomy is not equivalent to
empowerment, stressing that autonomy implies independence whereas empowerment may
well be achieved through interdependence (Malhotra and Mather 1997; Govindasamy and
Malhotra 1996; Kabeer 1998).

Early studies of “women’s status” often covered aspects of empowerment without explicitly
labeling it as such. One of the earliest empirical studies in this area, for example, used the
more general term “women’s status” but located a nexus of gender-related power differentials
in the household, noting how important the family unit is to understanding the operation of
gender in a society (Acharya and Bennett 1981). Acharya and Bennett also highlight the links
between women’s economic roles and their control over resources and life options.

Similarly, “women’s empowerment,” “gender equality” and “gender equity” are separate but
closely related concepts. The recent policy research report by the World Bank (2001a)
employs the term “gender equality,” which it defines in terms of equality under the law,
equality of opportunity (including equality of rewards for work and equality in access to
human capital and other productive resources that enable opportunity), and equality of voice
(the ability to influence and contribute to the development process). Gender equality implies
“equivalence in life outcomes for women and men, recognizing their different needs and
interests, and requiring a redistribution of power and resources.” Gender equity “recognizes
that women and men have different needs, preferences, and interests and that equality of
outcomes may necessitate different treatment of men and women” (Reeves and Baden
2000:10).

Notwithstanding the similarities in the concepts underlying many of these terms, we think that
the concept of empowerment can be distinguished from others based on its unique definitional
elements. As discussed above, the first essential element of empowerment is that it is a
process (Kabeer 2001; Chen 1992; Rowlands 1995; Oxaal and Baden 1997). None of the
other concepts explicitly encompasses a progression from one state (gender inequality) to
another (gender equality). Much of the emphasis on empowerment as a process is found in
the conceptual literature, but this understanding is also beginning to be incorporated into the
frameworks of empirical studies. For example, even as Jejeebhoy (2000) considers autonomy
and empowerment to be fairly similar, she argues that the former is a static state—and thus
measurable by most available indicators—while the latter is change over time, and not so
easily measurable.

The second element of empowerment that distinguishes it from other concepts is agency—in
other words, women themselves must be significant actors in the process of change that is
being described or measured (G. Sen 1993; Mehra 1997). Thus, hypothetically there could be
an improvement in indicators of gender equality, but unless the intervening processes
involved women as agents of that change rather than merely as its recipients, we would not
consider it empowerment. However desirable, it would merely be an improvement in

7
outcomes from one point in time to another. The importance of agency in the discourse on
empowerment emerges from “bottom up” rather than “top down” approaches toward
development (Oxaal and Baden 1997; Rowlands 1995; Narayan et al. 2000a and 2000b). At
the institutional and aggregate levels, it emphasizes the importance of participation and
“social inclusion” (Friedmann 1992; Chambers 1997; Narayan et al. 2000a and 2000b) At the
micro level, it is embedded in the idea of self-efficacy and the significance of the realization
by individual women that they can be the agents of change in their own lives.

The Process of Empowerment

There are various attempts in the literature to develop a comprehensive understanding of

empowerment through breaking the process down into key components. The specific
components tend to differ depending on the orientation and agenda of the writer, but here as
well there is greater consensus than we had expected. Kabeer’s (2001) understanding of
“choice” comprises three interrelated components: “resources, which form the conditions
under which choices are made; agency, which is at the heart of the process through which
choices are made and achievements, which are the outcomes of choices.” The World Bank’s
report on “Engendering Development” (2001a) defines rights, resources, and voice as the
three critical components of gender equality. Chen (1992) describes “resources, perceptions,
relationships, and power” as the main components of empowerment, and Batliwala (1994),
characterizes empowerment as “control over resources and ideology.” UNICEF uses the
Women’s Empowerment Framework constructed by Sara Longwe, which encompasses
welfare, access to resources, awareness-raising, participation, and control (UNICEF 1994).

Resources and agency (in various forms and by various names, e.g., control, awareness, voice,
power) were the two most common components of empowerment emphasized in the literature
we reviewed. In many discussions, however, resources are treated not as empowerment per
se, but as catalysts for empowerment or conditions under which empowerment is likely to
occur. In the context of policy and evaluation, it may be more useful to think of resources as
“enabling factors;” that is, as potentially critical inputs to foster an empowerment process,
rather than as part of empowerment itself. And, in fact, many of the variables that have
traditionally been used as “proxies” for empowerment, such as education and employment,
might be better described as “enabling factors” or “sources of empowerment” (Kishor 2000a).
As our review of measurement issues and empirical evidence illustrates below, although many
empirical studies have used variables such as education and employment as proxies for
empowerment, there is a growing understanding that this equation is problematic
(Govindasamy and Malhotra 1996; Malhotra and Mather 1997; Kishor 1995; Mason 1998).

The second component, agency, is at the heart of many conceptualizations of empowerment.

Human agency is a central concept in A. Sen’s (1999) characterization of development as the
process of removing various types of “unfreedoms” that constrain individual choice and
agency. Kabeer’s (2001) essay on women’s empowerment draws on Sen’s understanding of
agency as well as his conceptualization of the links between individual agency and public
action. Among the various concepts and terms we encountered in the literature on
empowerment, “agency” probably comes closest to capturing what the majority of writers are

8
referring to. It encompasses the ability to formulate strategic choices, and to control resources
and decisions that affect important life outcomes.

Some characterizations of empowerment have included an additional component, which

Kabeer refers to as “achievements” and Longwe as “welfare.” And, in international policy
processes, women’s empowerment is implicitly equated with specific (usually national- level)
achievements such as political participation, legal reform, and economic security. In the
context of evaluation, we would argue that achievements are best treated as outcomes of
empowerment, not as empowerment per se (just as resources may be more usefully construed
as enabling factors or catalysts for empowerment). Granted, one might question whether
agency really amounts to empowerment if there is no meaningful result in terms of women’s
status, strategic position, or welfare. But whether any empowerment indicator or form of
agency (such as exercising control over decisions or resources) really amounts to
empowerment in a particular context will always be an empirical question. As we note
elsewhere, it may be possible to identify empowerment indicators that can be applied in a
wide variety of contexts, but there will always be situations in which a particular indicator
does not signify empowerment. As we will discuss later, this does not mean that the
development of empowerment indicators is a futile task, only that the complexities of
measuring empowerment must be taken into account in developing conceptual frames and
research designs.

In identifying agency as the essence of women’s empowerment, we are not suggesting that all
improvements in women’s position must be brought about through the actions of women
themselves or that empowering themselves is the responsibility of individual women. There is
ample justification for governments and multilaterals to promote policies that strengthen
gender equality through various means, including legal and political reform, and interventions
to give women (and other socially excluded groups) greater access to resources (e.g., World
Bank 2001a). National and international institutions have the responsibility for ensuring the
inclusion of disadvantaged populations socially, economically, and politically. The question is
whether it is useful to describe all actions taken toward that end as “empowerment,” and we
would suggest that it is not. The major reason for the emphasis on agency as the defining
criterion is because of the many examples in the literature of cases in which giving women
access to resources does not lead to their greater control over resources, where changes in
legal statutes have little influence on practice, and where female political leaders do not
necessarily work to promote women’s interests. Thus while resources—economic, social,
and political—are often critical in ensuring that women are empowered, they are not always
sufficient. Without women’s individual or collective ability to recognize and utilize resources
in their own interests, resources cannot bring about empowerment.

Having argued that “agency” should be treated as the essence of empowerment, and resources
and achievements as enabling conditions and outcomes, respectively, another caveat is
necessary. While distinctions such as those between “resources, agency, and achievements”
(Kabeer 2001) or “sources versus evidence” of empowerment seem clear at the conceptual
level, it is not always easy to completely separate them in developing empowerment
indicators. And too, a given variable may function as an indicator of women’s access to
resources (or an enabling factor) in one context, of women’s agency in another, and may

9
represent an achievement in still other contexts. For example, microcredit programs and
employment opportunities are often seen as resources for women’s empowerment. But if a
woman seeks to gain access to microcredit, or to get a job, then getting the job or joining the
credit program might be best characterized as a manifestation of women’s agency, and the
benefits she draws as a result—income, discretionary spending, healthcare, etc—as
achievements. In some analyses discussed in Section III, below, women’s economic
contribution is treated as an enabling factor and used to predict other outcomes such as control
over important decisions and even the outcomes of decisions such as family size or
contraceptive use. But in other contexts women’s economic contribution would be more
accurately described as a form of agency or, again even an achievement. Similarly, assets
owned could function as sources of empowerment (e.g., Kishor 2000a) but they could also
constitute evidence that empowerment had been achieved. The meaning of any empowerment
indicator will always depend on its interrelationships with other variables.

Empowerment is a dynamic process. Separating the process into components (such as

enabling factors, agency and outcomes) may be useful in identifying policy interventions to
support empowerment, and for evaluating the impact of such interventions, but it is important
to realize that a framework for research or evaluation of a specific policy or intervention will
refer to only one phase of the process. Which phase it refers to will depend on the context, the
interventions being assessed, and the outcomes of interest.

SECTION II: MEASURING WOMEN’S EMPOWERMENT

Measuring Empowerment from a Universalist Perspective

As we move from a discussion of conceptualizing empowerment to measuring it, it is

important to note that measures of empowerment must involve standards that lie outside
localized gender systems and a recognition of universal elements of gender subordination
(Sen and Grown 1987; Bisnath and Elson 1999; Nussbaum 2000). It is clear from the
literature on gender and empowerment that the role of gender in development cannot be
understood without understanding the socio-cultural (as well as political and economic)
contexts in which development takes place. The concept of empowerment has meaning only
within these specific contexts. At the same time, operatio nal definitions (e.g., definitions
embodied in indicators to be applied in the context of development assistance policies,
programs, and projects) should be consistent with the spirit of international conventions to
which countries providing international development assistance have been signatories. The
approach based in universal human rights offers the best operational framework for this task.

Local structures of gender inequality are typically experienced as “natural,” and therefore may
seem unalterable to actors in a particular social setting. Kabeer (2001) elaborates on this point
drawing on Bourdieu’s 1977 idea of “doxa”—the “aspects of tradition and culture which are
so taken- for-granted that they have become naturalized.” When women internalize their
subordinate status and view themselves as persons of lesser value, their sense of their own
rights and entitlements is diminished. They may acquiesce to violence against them, and make
“choices” that reinforce their subordinate status. For example, in her life cycle, a South Asian

10
woman may “graduate” from the comparatively subservient position of daughter- in- law to
that of mother-in- law, and in this role she may dominate her son’s wife. Based on the
“agency” criterion for describing something as empowerment, one might call this behavior
empowered. As a mother- in-law, the woman gained the ability to exercise agency (in the form
of power over another person), in a way that she could not when she was a young woman
herself. But, we would argue against such a use of the term empowerment. The mother- in- law
is acting within an inequitable gender system that severely constrains her ability to make
strategic life choices. The system lets her exercise power, but only in ways that reinforce the
system. This sort of agency is similar to what Kabeer (2001) describes as choices that reflect
women’s consent and complicity in their own subordination. When they lack agency in a
broader sense, women should not be considered to be making empowered choices.

Internalized subordination receives particular attention among writers on international

education, informed by a Freirian perspective on raising the critical consciousness of the poor
(Freire 1994[1973]). For example, Stromquist writes that empowerment includes cognitive
and psychological elements: It involves “women’s understanding of their conditions of
subordination and the causes of such conditions at both micro and macro levels of society….
It involves understanding the self and the need to make choices that may go against cultural
and social expectations” (1993:14). Thus, universal standards are necessary to identify
empowerment.

Multidimensionality and Existing Frameworks

As early as 1981, Acharya and Bennett noted that status is a function of the power attached to
a given role, and because women fill a number of roles, it may be misleading to speak of “the
status of women” (p. 3). Another early writer on the topic, Mason (1986), pointed out that the
phenomenon of gender inequality is inherently complex, that men and women are typically
unequal in various ways, and that the nature or extent of their inequality in different settings
can vary across these different dimensions (as well by social setting and stage in the life
cycle). Since that time, a number of studies have shown that women may be empowered in
one area of life while not in others (Malhotra and Mather 1997; Kishor 1995 and 2000b;
Hashemi et al. 1996; Beegle et al. 1998). Thus it should not be assumed that if a development
intervention promotes women’s empowerment along a particular dimension that
empowerment in other areas will necessarily follow. It may or may not.

Several different efforts have been made in recent years to develop comprehensive
frameworks delineating the various dimensions along which women can be empowered. In
Appendix A, we present the essential elements of the empowerment frameworks developed
by selected authors. These frameworks employ different levels of specificity. For example,
the CIDA (1996) framework includes four broad dimensions of empowerment, while Kishor’s
(2000a) framework includes broad (e.g., valuation of women, equality in marriage) as well
specific (e.g., lifetime exposure to employment) elements.

In Table 1, we synthesize and list the most commonly used dimensions of women’s
empowerment, drawing from the frameworks developed by these various authors. Allowing

11
for overlap, these frameworks suggest that women’s empowerment needs to occur along the
following dimensions: economic, socio-cultural, familial/interpersonal, legal, political, and
psychological. However, these dimensions are very broad in scope, and within each
dimension, there is a range of sub-domains within which women may be empowered. So, for
example, the “socio-cultural” dimension covers a range of empowerment sub-domains, from
marriage systems to norms regarding women’s physical mobility, to non- familial social
support systems and networks available to women. Moreover, in order to operationalize these
dimensions, one should consider indicators at various levels of social aggregation -- the
household and the community, as well as regional, national, and even global levels. In the
table we group commonly used and potentially useful indicators within various “arenas” or
spheres of life. Some of these indicators have been suggested within the frameworks
referenced above, while others are a first effort on our part to “flesh out” this schematic for
application in development assistance contexts.

12
Table 1. Commonly used dimensions of empowerment and potential operationalization
in the household, community, and broader arenas.
Dimension Household Community Broader Arenas
Economic Women’s control over Women’s access to Women’s representation
income; relative employment; ownership of in high paying jobs;
contribution to family assets and land; access to women CEOs;
support; access to and credit; involvement and/or representation of women’s
control of family representation in local trade economic interests in
resources associations; access to macroeconomic policies,
markets state and federal budgets

Socio-Cultural Women’s freedom of Women’s visibility in and Women’s literacy and

movement; lack of access to social spaces; access to a broad range of
discrimination against access to modern educational options;
daughters; commitment transportation; participation Positive media images of
to educating daughters in extra -familial groups and women, their roles and
social networks; shift in contributions
patriarchal norms (such as
son preference); symbolic
representation of the
female in myth and ritual

Familial/ Participation in domestic Shifts in marriage and Regional/national trends

Interpersonal decision-making; control kinship systems indicating in timing of marriage,
over sexual relations; greater value and autonomy options for divorce;
ability to make for women (e.g., later political, legal, religious
childbearing decisions, marriages, self selection of support for (or lack of
use contraception, access spouses, reduction in the active opposition to) such
abortion; control over practice of dowry; shifts; systems providing
spouse selection and acceptability of divorce); easy access to
marriage timing; local campaigns against contraception, safe
freedom from domestic domestic violence abortion, reproductive
violence health services

Legal Knowledge of legal Community mobilization Laws supporting women’s

rights; domestic support for rights; campaigns for rights, access to resources
for exercising rights rights awareness; effective and options; Advocacy for
local enforcement of legal rights and legislation; use
rights of judicial system to
redress rights violations

Political Knowledge of political Women’s involvement or Women’s representation

system and means of mobilization in the local in regional and national
access to it; domestic political system/campaigns; bodies of government;
support for political support for specific strength as a voting bloc;
engagement; exercising candidates or legislation; representation of women’s
the right to vote representation in local interests in effective
bodies of government lobbies and interest groups

Psychological Self-esteem; self- Collective awareness of Women’s sense of

efficacy; psychological injustice, potential of inclusion and entitlement;
well-being mobilization systemic acceptance of
women’s entitlement and
inclusion

13
When sorted by sphere or level of aggregation (which we discuss more fully below),
frameworks delineating dimensions of women’s empowerment offer potential roadmaps for
operationalizing and measuring women’s empowerment. These frameworks have been
important in highlighting the potential independence of the various areas within which women
can be empowered—i.e., women can be empowered in the familial sphere without making
similar gains in the political sphere. In terms of practical measurement, however, it is difficult
to neatly separate the dimensions. For example, many aspects of economic or social
empowerment overlap considerably with the familial dimension, as in the case of control over
domestic spending or savings, or the limitations on mobility or social activities.

Empirical research has begun to show that in given settings, some dimensions of
empowerment may be more closely interlinked than others. Kishor (2000a) conducted a factor
analysis of several variables that may impact empowerment in Egypt. Among the ten
empowerment dimensions that resulted from the analysis, some were comparatively better
correlated than others. Dimensions did not all bear equally on the survival of women’s
children (measured by child mortality and immunization); only “women’s lifetime exposure
to employment” and “family structure” (residence with in- laws, etc.) affected both child
survival outcomes. Jejeebhoy (2000) likewise found that, in India, decision- making, mobility,
and access to resources were more closely related to each other than to child-related decision
making, freedom from physical threat from husbands, and control over resources.

Because empowerment is multidimensional, researchers must use care in constructing index

or scale variables relating to empowerment. Such variables may mask differential effects of
interventions on distinct aspects of empowerment. Inappropriate combining of items relating
to gender and empowerment may also mask differential effects of the component variables on
outcomes of interest. Ghuman (2002), for example, critiques a logit regression analysis by
Durrant and Sathar (2000), which found that mothers’ decision- making autonomy on child-
related issues demonstrated a weak, statistically insignificant effect on child survival. Ghuman
points out that this finding was the result of a summative scale of items on mothers' decision-
making about child-related issues such as schooling, illness matters, and punishment for
misbehavior, but tha t these items varied greatly with respect to their individual associations
with child survival. Thus, Durrant and Sathar found a weak effect because their aggregated
scale items had different individual effects. Ghuman found that punishing children for
misbehaving (which she links to women’s reluctance to take on authority with respect to other
child-related decisions) had an important negative association with child mortality. However,
it is also true that a single indicator is not usually sufficient to measure even a specific
dimension of empowerment (Kishor 2000b; Estudillo et al. 2001). Additional information is
usually needed to interpret data on any given indicator.

Levels of Aggregation

Many writers have noted that because power relations operate at different levels, so does
empowerment (Mayoux 2000; Bisnath and Elson 1999). However, exactly how these levels
are defined varies. Economists, for example often differentiate between the micro and macro
levels, but the macro level is generally meant to include market and political systems where as
the micro level may include not only individuals and households, but also communities and

14
institutions (see Pitt and Khandker 1998; Rao 1998; Tzannatos 1999; Winter 1994; Narayan et
al. 2000a and 2000b) When sociologists and demographers refer to the micro level, they
usually mean the individual or the household while the macro level may include anything
from the community to the polity. (Gage 1995; Jejeebhoy and Sather 2001; Kritz et al. 2000;
Malhotra et al. 1995). Thus, while there is clarity at the highest and lowest levels of
aggregation, this is less the case with the intermediate levels. This also means that in
operationalizing empowerment, there is theoretical interest but less empirical attention to
aggregations that fall in the middle, especially at the community level where institutional and
normative structures such as family systems, infrastructure, gender ideologies, regional or
local market processes, etc. are most likely to affect women’s empowerment. It is often
precisely at these intermediary levels that normative changes occur and where programmatic
or policy interventions often operate.

Theoretically, the frameworks that delineate dimensions of empowerment can be

operationalized at any leve l of aggregation. For example, the legal dimension can be
measured in terms of individual women’s knowledge and/or use of their legal rights, to
representation of women’s interests and concerns in community, regional, or national laws.
Our review suggests, however, that in the studies to date, the political and legal dimensions
tend to be operationalized at fairly high levels of aggregation (regional or national), while the
familial, social, and economic dimensions are generally operationalized at the individual or
household levels, with some limited efforts at considering these at the level of the community
or institution. The psychological dimension of empowerment is rarely operationalized in
empirical research at any level.

The most cutting-edge empirical research makes efforts to measure empowerment at multiple
levels. Anthropological and qualitative studies are particularly adept at blending individual or
household situations with the “meso” level institutional structures and normative conditions
(Kabeer 1997, Mayoux 2001; Hashemi et al. 1996). Quantitative studies that have attempted
multi- level analyses of empowerment have used both aggregations of individual/household
data as well as direct measures of community levels characteristics (Kritz et al. 2000; Mason
and Smith 2000; Jejeebhoy and Sathar 2001). Generally these studies find that both
individual and community level effects are important in determining empowerment or related
outcomes. For specific outcomes, moreover, aggregate level effects may play a significantly
greater role than individual level effects (Mason and Smith 2000; Jejeebhoy 2000; Kritz et al.
2000).

Research that blends theory with empirical work tends to focus on women’s empowerment in
their relationships within the household and local community. Much of this literature focuses
on individual as opposed to collective empowerment and examines conjugal relationships and
sometimes women’s relationships with others in the household as well. The discourse on
collective forms of empowerment emerges largely from the activist literature (G. Sen 1993).
Oxaal and Baden (1997) argue that to the extent that mainstream development discourse
views empowerment as an individual rather than collective process, it emphasizes
entrepreneurship and self-reliance rather than cooperation to challenge power structures. The
discourse on social inclusion also sees the potential for empowerment in a collective form,
whether it is through political, economic, or social mobilization of groups. Narayan et al.

15
(2000a and 2000b) find that poor people’s NGOs often lack transformational power, but that
with capacity building, access to information, and increased accountability in both state and
civil-society institutions, both groups and individuals can become empowered.

Community- level social norms, such as those constraining women’s freedom of movement,
access to economic resources, and “voice” in the local community, may become more
egalitarian as a result of strategic collective action by groups of women. Examples include the
anti-arak movement in south India and the collective actions by female street vendors who
joined the Self-Employed Women’s Association (SEWA) in Ahmedabad, India. Collective
action in the social, economic, and political spheres (e.g., mobilization to end violence again
women, unionization, advocating for certain laws) can also lead to empowerment in the form
of expanded legal rights and political representation at various levels of society. Thus,
organized collective action may alter policies and practices such that the ability to make
strategic life choices is improved not only for the women who were directly involved in the
bringing about the changes, but for other women as well.

But the normative structures that limit women’s strategic life choices may also change
through processes arising from individual actions in response to new situations and
opportunities. For example, restrictions on women’s freedom of movement and interactions
with men who are nonrelatives may loosen as a result of microcredit programs or employment
opportunities, and women’s access to social networks outside the family may increase along
with their presence in the public sphere. Regardless of whether they have been intentionally
fostered through development policies (as opposed to being merely by-products of policies
with other aims) such changes may eventually become manifest as “aggregate” shifts in
women’s position within family systems, or markets, and may represent a transformation in
gender norms.

Kabeer (2001) argues that individual women may challenge structural inequalities when they
act in ways inconsistent with gender norms, but that the impact of such actions tends to be
limited. The “project” of women’s empowerment, she believes, requires collective action in
the public sphere. Notwithstanding the importance of collective action, we would argue that
individual actions may bring about significant normative changes at the community level that
can be accurately described as empowerment, depending on the social processes involved.
According to sociological theories of diffusion, new ideas and practices often spread
gradually without collective consciousness that fundamental change is occurring (Rogers
1995). At some “tipping point,” however, the prevalence of new processes becomes great
enough for there to be a revolution in norms as well as in the collective consciousness. For
example, in rural Bangladesh in the early 1990s women who took up nontraditional jobs,
adopted contraceptives, or traveled by themselves to health facilities before these behaviors
were common were pioneers in the process of diffusion (Hashemi et al. 1996; Schuler et al.
1996). Gradually, other women did the same, and while women in these communities
continue to suffer from gender inequity in many spheres of their lives, there has nevertheless
been a significant change in terms of women’s freedom of movement, their ability to exercise
control over their own fertility, and their knowledge and involvement in the public sphere.
These changes are widely recognized by men and women in the research communities. We

16
would characterize this process as “empowerment” even though the collective aspect
developed more or less organically, rather than resulting from organized action.

Challenges to Measuring Empowerment

Empowerment Is Context-Specific

One of the major difficulties in measuring empowerment is that the behaviors and attributes
that signify empowerment in one context often have different meanings elsewhere. For
example, a shift in women’s ability to visit a health center without getting permission from a
male household member may be a sign of empowerment in rural Bangladesh but not in, for
example, urban Peru. Context can also be important in determining the extent to which
empowerment at the household or individual level is a determinant of development outcomes.
It could be argued, for example, that if investments in public health systems are strong, then
women’s role as the intermediaries for their children’s health through better education or
decision- making power in the household will be less important than when this is not the case.

The variation in the nature and importance of empowerment across contexts poses a challenge
in terms of both consistency and comparability in measurement schemes. How important is
context in defining empowerment in different settings? Does the context-specific nature of
empowerment mean that we must constantly reinvent indicators to suit socioeconomic,
cultural, and political conditions? What is the role of context in determining the relationship
between women’s empowerment and development outcomes? How dependent is this
statistical relationship on the choice of indicators and whether or not they are appropriate to
the setting in question?

In the 1990s there have been a few pioneering efforts at sorting out some of these issues
through empirical research (Mason and Smith 2000; Jejeebhoy 2000; Kritz et al. 2000;
Schuler et al. 1995b; Hashemi et al. 1996). The body of work emerging from this research
unequivocally confirms the importance of context in both defining and measuring the impact
of women’s empowerment on development outcomes. Although not fully conclusive
concerning several of the other challenges that context specificity represents, this work also
sets the stage for exploring measurement and analytical schemes that can better address these
challenges.

The group of “Status of Women and Fertility” studies conducted by Mason and her colleagues
aimed for comparability in measuring women’s empowerment and its impact of reproductive
behavior across five settings in Asia: India, Pakistan, Thailand, Philippines, and Malaysia
(Mason and Smith 2000; Jejeebhoy and Sathar 2001). Although there were small variations
in wording to make each que stion appropriate to the country setting, there was an effort to
employ similar indicators across countries and within 59 community settings in the five
countries. In her 1998 analysis, for example, Mason is able to compare “economic decision-
making power in the family” based on a scale constructed from six indicators that were
collected relatively consistently across the five countries.

17
In this approach, contextual factors are brought in as important determinants at the analytical
rather than the measurement stage. Thus, analyses from this set of studies include community
level measures on family systems, marriage systems, religion and ethnicity, female
participation in the work force, rates of child mortality, etc. Kritz et al. (2000) employ a
similar approach by developing an index of the gender contexts in four communities using
eight indicators, such as mean spousal age difference, percentage of wives in modern work,
mean score on wives’ physical mobility, and percentage of wives who control how to use
income. A consistent set of findings from this approach is that the contextual factors are often
more important in determining women’s empowerment and its outcomes than individual- level
factors. At the same time, there is inconsistency in the studies’ findings on which particular
contextual conditions are most empowering to women. Mason (1998:130) summarizes:
“While our analysis suggests that the community context is very important for the
empowerment of individual women, it also makes clear that the community conditions which
empower women tend to be idiosyncratic rather than universal.”

Studies that apply indicators across cultures can be useful for making international or inter-
regional comparisons with reference to an external yardstick of power, women’s status, or
gender equity, but they raise the issue of how appropriate similar indicators are in measuring
empowerment across settings. An alternative potential approach to addressing the challenges
of context is to rely on a consistent conceptual framework for measuring empowerment and
its effects, but to allow flexibility in the specific indicators used to define the key components
of that framework across different settings. Any given context, at any given point of time can
be seen as having behavioral and normative “frontiers,” that need to be crossed for women to
be empowered along a specific dimension, within a specific arena. Specifying these frontiers
helps defines the indicators of relevance to that particular context, at that particular time.

This is the approach that Schuler et al. (1995a and 1995b) advocate. In their work on
Bangladesh, India, and Bolivia, Schuler and her colleagues relied on a common conceptual
framework in which they specified the dimensions along which women’s empowerment or its
effects could vary. In measuring the dimensions, however, they used indicators relevant to
each particular country and community setting. Their analysis also allowed for greater or
lesser weight on certain dimensions as opposed to others across contexts. Hashemi and
Schuler argue that initial groundwork through qualitative and exploratory methods,
conceptual analysis, and stakeholder consensus through participatory processes is essential to
establishing parameters that define empowerment in specific country and development project
contexts (Hashemi et al. 1996).

This approach, however, does require a balancing act between the “universalist” principles
around which empowerment must be conceptualized, and the localization of contextspecific
indicators. The underlying structures of gender inequality are often invisible to the actors in a
particular social milieu; they are often experienced as “natural” and, as such, inalterable. And
yet individuals find ways to exercise agency, and to control others, even in contexts where
they are comparatively powerless. In exploring the viability of indicators and building
stakeholder consensus through participatory processes, there is the danger that indicators may
be “too” internally defined, reflecting the limited viewpoint of the actors.

18
The contextual nature of empowerment suggests that “universal” measures may be
impossible. However further comparative research might reveal whether some empowerment
indicators are “more universal” than others.

Difficulties in Measuring a “Process”

Many writers describe empowerment as a “process,” as opposed to a condition or state of

being, a distinction that we have emphasized as a key defining feature of empowerment.
However, as “moving targets,” processes are difficult to measure, especially with the standard
empirical tools available to social scientists. In this section we discuss the major
methodological challenges in measuring the process of women’s empowerment, including:
the use of direct measures as opposed to proxy indicators, the lack of availability and use of
data across time, the subjectivity inherent in assessing processes, and the shifts in relevance of
indicators over time.

Some authors who have made efforts at empirically measuring empowerment have argued
that as a process, it cannot be measured directly, but only through proxies such as health,
education level, knowledge (Ackerly 1995). For example, Kishor (2000a) has argued that
while the end product of empowerment can be measured through direct indicators, the process
can be measured only through proxies such as education and employment. Several large-scale
studies of relationships between gender and economic or demographic change have used
proxy variables. However, an increasing body of research indicates that commonly used
proxy variables such as education or employment are conceptually distant from the
dimensions of gender stratification that are hypothesized to effect the outcomes of interest in
these studies, and may in some cases be irrelevant or misleading (Mason 1995, p.8-11;
Govindasamy and Malhotra 1996). Studies have found that the relevance of a proxy
measurement of women’s empowerment may depend on the geographic region (Jejeebhoy
2000), the outcome being examined (Kishor 2000a), or the dimension(s) of empowerment that
is of interest (Malhotra and Mather 1997).

In response, there have been increasing efforts at capturing the process through direct
measures of decision- making, control, choice, etc. Such measures are seen as the most
effective representations of the process of empowerment by many authors since they are
closest to measuring agency (Hashemi et al. 1996; Mason 1998, Mason and Smith 2000;
Malhotra and Mather 1997). It could be argued that the indicators with “face validity” (i.e.,
indicators of empowerment based on survey questions referring to very specific, concrete
actions) represent power relationships and are meaningful within a particular social context.

Ideally, the best hope of capturing a process is to follow it across at least two points in time.
Moreover, the gap in time required to measure the process may depend on the nature and
extent of change in empowerment. Depending on the dimension of empowerment, the
context, and the type of social, economic, or policy catalyst, women may become empowered
in some aspects of their lives in a relatively short period of time (say 1-3 years) while other
changes may evolve over decades. For policy and programmatic action, specifying the
aspects of women’s empowerment that are expected to change as well as the “acceptable”

19
time period for change is critical in defining success or failure. As conceptual frameworks and
indicators of empowerment become more sophisticated, however, there is an enormous
problem with regard to the availability of adequate data across time. For example, while there
is increasing agreement that measures with “face validity” are preferable to “proxy”
indicators, survey data that include “face validity” measures are often one-of-a-kind attempts,
and are not systematically or routinely collected across more than one point in time.

Qualitative studies of empowerment make an effort at capturing the process through in-depth
interviews and case studies that follow the life changes for specific women (and men) through
retrospective narratives. G. Sen (1993) has suggested that the process of empowerment is
essentially qualitative in nature. Even indicators such as women’s participation in power
structures like the political system are still often inadequate in telling us whether
empowerment is occurring without a qualitative sense of what that representation is like or
what it means (Oxaal and Baden 1997). Kabeer’s work (1997) suggests that the assessment
of the process is not only qualitative, but subjective as well. According to Kabeer (1997 and
1998), the subjectivity of the process should also extend to measuring empowerment in terms
of women’s own interpretation; rather than relying on what is valued by the evaluators of
programs, the process of empowerment should be judged as having occurred if it is self-
assessed and validated by women themselves.

Another complicating factor in assessing the empowerment process is that the behavioral and
normative frontiers that define appropriate indicators for measuring empowerment are
constantly evolving. The “meaning” of a particular behavior within a particular socio-cultural
context (whether it signifies empowerment and whether it is influenced by empowerment) is
likely to change over time, and it may change very rapidly. As a result, the relevance of
specific indicators will change over time and according to the level of analysis. Data from the
early 199’s suggested that in rural Bangladesh empowered women were more likely than
others to use contraception (Schuler et al. 1997). Now contraceptive use is the norm—over
half of all married, reproductive age women currently use it and more than three quarters have
used it at one time or another. Once a behavior becomes the accepted norm there is little
reason to expect that it would be influenced by an individual actor’s level of empowerment.

At the individual level, the case could be made that individual empowerment should be
measured as a function of the distance between the individual’s behavior and the community
norm. This would be true of indicators such as “ability to move about one’s village” or
“ability to visit a health center without getting permission.” However, an indicator that is no
longer a good marker of empowerment at the individual level within a community may still be
a good indicator for distinguishing relative levels of empowerment between communities, as
long as some variation within the larger society persists. Table 2 illustrates the shifting
meanings and potential utility of one indicator in three different normative contexts at the
individual and community levels.

20
 Table 2: The usefulness of the indicator “individual woman moves about community
without permission” in different normative contexts.

Normative Context
Women’s mobility is Claustration norm Claustration is
Level of Analysis a longstanding norm recently changed still the norm
Individual Probably not useful Probably not useful Potentially useful
Community Probably not useful Potentially useful Potentially useful

SECTION III: EVIDENCE FROM EMPIRICAL STUDIES

Overview

Empirical research on gender issues in the fields of anthropology, sociology, demography,

and economics has acquired increasing levels of sophistication during the past three decades.
From being limited to qualitative studies in anthropology, empirical work on women’s status
began appearing in sociological and demographic studies in the 1970s. During the following
decade, the shift from women’s status to a better understanding of gender inequality and
concepts such as female autonomy and power moved the field forward. In the 1990s, the call
for better theoretical models allowing for gender differences and the accompanying data to
test them both at the macro and micro levels began to take hold within economics. While
gender issues—especially as they relate to development—are not in the mainstream of any of
these disciplines, important intersections exist across all four. Natural affinities have also
begun to develop. Thus the critique of the unitary household model and experimentation with
a range of bargaining models of power in economics has some natural affinity with
conceptual frameworks emphasizing decision- making and control within demography and
sociology. Researchers who have recognized and realized the va lue of interdisciplinary
intersections in their work have perhaps made the greatest contributions in moving this field
forward (Kabeer 1997and 1998; Mason 1998; Acharya and Bennett 1981; A. Sen 1999; G.
Sen 1993; Hashemi et al. 1996).

We reviewed approximately 45 empirical studies that have used quantitative and/or

qualitative data to understand some aspect or variation of women’s empowerment in an effort
to answer the following questions: How has existing research handled the challenges and
promises of measuring empowerment and related outcomes? What patterns prevail in the
current body of research and what conclusions can be drawn, in terms of policy relevance as
well as for future directions for operationalizing women’s empowerment empirically? Our
search encompassed all four of the disciplinary fields mentioned above, and it was extensive
but by no means comprehensive. A study did not have to use the term women’s
empowerment to be included; whether or not a study would contribute to a better
understanding of how empowerment can be measured was a subjective assessment made by
the authors. In general, we limited ourselves to studies that tested some form of
“empowerment” within a reasonably well-developed conceptual framework and that

21
incorporated an active effort at capturing at least one of the two defining elements of
empowerment described in Section I—agency and process.

It should be noted that these criteria led to the effective exclusion of most of the reports
emerging from NGO programmatic efforts at “empowering” women. Many of these reports
lack the conceptual and empirical rigor we felt was necessary for inclusion in the current
review. With the exception of the work on microcredit, we searched in vain for
programmatic assessments that were based on well-designed evaluation frameworks to test
either the impact of interventions on empowering women, or the impact of empowering
women on other outcomes. Given that the term “empowerment” has acquired a certain cachet
in the development community, there is a tendency to use the term loosely, without
embedding it in a larger conceptual framework. Thus, programmatic reports often describe
mothers as empowered merely because they bring their children in for immunization, attend a
health education lecture or other training, or participate in women’s NGOs or microcredit
groups, without examining the circumstances or outcomes of these actions. For the most
part, therefore, we ended up restricting ourselves to published works, covering studies that
were interested in women’s empowerment as an outcome of interest, as well as those that
assessed its importance as an intermediary factor in affecting development outcomes.

Table 3 provides a tabulation of key characteristics of the studies we reviewed. In 16 of the

45 studies, women’s empowerment itself was the outcome of interest while the other 29
studies were interested in examining the effect of certain social, economic, political, policy, or
programmatic factors on development outcomes through an intermediary process of
empowering women. The regional distribution of the studies was heavily biased toward Asia,
and South Asia in particular. Over half the studies (25) were focused on Asia—the vast
majority on India and Bangladesh—while only seven studies covered countries in Africa, and
only four studies covered the LAC region. A significant proportion of the work on Asia
focuses on demographic issues. The regional bias may reflect the authors’ greater familiarity
with demographic research but it may also reflect greater interest or data advantages in the
study of women’s empowerment in Asia.

22
 Table 3: Distribution of studies reviewed by region, methodology, measurement across
time, level of analysis, and data source.

“Empowerment” “Empowerment”
Characteristics of Empirical Studies Is the Outcome of Affects Other
Interest Outcomes of Interest
Region Africa 1 6
Middle East and North 0 3
Africa
Asia 11 14
Latin America and 2 2
Caribbean
Multiple regions 2 4
Methodology Qualitative 2 1
Quantitative 7 21
Qualitative and quantitative 7 7
Measurement across time 2(2) 1
Level of Analysis Individual/household 9 16
Community 0 1
District or state 0 5
Country 3 2
Multiple levels 4 5
Data source Primary data 11 12
Secondary data 5 17
Number of Studies 16 29

The majority of studies (28) use quantitative methodologies. However, a significant

proportion of the studies examining empowerment as the outcome of interest incorporate a
combination of quantitative and qualitative techniques and efforts at triangulation in their
analyses (7). Most surprisingly, however, only three of the studies (Winter 1994; Kishor
1992; Tzannatos 1999) use data from more than one point in time to assess empowerment,
whereas this scope of data and analysis is entirely missing from almost all of the other studies.
Two qualitative studies (Kabeer 1997; Mayoux 2001) are implicitly making a comparison
across time using retrospective interviews with the women in their samples. For the most
part, therefore, measurement of process as one of the essential features highlighted in the
conceptualization of empowerment has been rarely undertaken in the empirical work to date.

The range of development outcomes examined in research that focuses on women’ s

empowerment as an intermediary factor is also surprisingly limited. The heaviest
concentration is on fertility/contraceptive use (12) or child health/welfare related (8)
outcomes, with just a handful of studies focusing on broader issues of household well-being
(5), women’s health (3), or development processes (1). Similarly, the examination of the
impact of policy and program initiatives does not seem to be a high priority in empirical
research. Admittedly, there is a heavy concentration of research on the impact of microcredit
programs on women’s empowerment (mostly for Bangladesh), but few other initiatives are

23
included. At the policy level, Winter (1994) examines the impact of shifts in labor conditions
and labor laws on women’s empowerment in the labor market in Latin America and
Chattopadhyay and Duflo (2001) examine the effect of the Panchayat Raj Initiative in India.

Studies focusing on women’s empowerment itself as the outcome of interest are more likely
to rely on primary data sources (11 of 16) as opposed to those where empowerment is an
intermediary factor in affecting other outcomes. Clearly, there has been a strong effort on the
part of researchers interested in women’s empowerment at collecting the appropriate data
themselves. In contrast, studies interested in the intermediary role of empowerment on other
development outcomes rely more heavily on secondary data sources, often large-scale sample
surveys or national censuses (17 of 29). As such, the indicators used by these studies are
more likely to be proxy measures that are not directly measuring empowerment. In recent
years, the Demographic and Health Surveys in some countries have incorporated modules of
decision- making and women’s autonomy in the household, thus offering the benefits of large
datasets with the specificity of analysis at the individual level. Five studies use DHS data in
this way; surprisingly, none of these treat empowerment as the dependent variable. Other key
data sources include the Indonesia Family Life survey (used by researchers at RAND and
IFPRI), the Indian Census, the Status of Women and Fertility surveys in Asia, the Ethopian
Rural Household Survey, and the Project for Statistics on Living Standards and Development
survey in South Africa.

How Has Empowerment Been Measured Empirically?

Household-Level Studies

Table 3 indicates that empirical analyses of women’s empowerment are heavily concentrated
at the individual and household level, and our review of the literature suggests that this is the
level of aggregation at which the greatest strides in the measurement of empowerment have
been made. Given the centrality of the household to gender relations, it is not surprising that
this level of data collection and analysis has received the greatest attention. It may also be
true that the feasibility of operationalizing both the agency and process components of
women’s empowerment in a concrete manner is more readily apparent at the household level
rather than at larger levels of aggregation. And yet, our review suggests that empirical
research at this level has made greater strides in operationalizing agency rather than process.
With the exception of two qualitative studies which use retrospection and inference with
regard to change over time, none of the household level studies operationalize empowerment
by utilizing data from two points in time.

In the past two decades, researchers interested in gender inequality as it operates at the
household level have made significant efforts at better capturing the agency component of
empowerment by themselves designing and conducting household- level surveys that
interview women, sometimes their husbands, and occasionally, other family members.
Within sociology and demography, the major effort has been at measuring household
decision- making processes, financial control, and social or familial constraints directly. This
has been motivated by interest both in understanding empowerment itself and in outcomes

24
such as fertility, contraceptive use, and child health and well- being. Within economics,
household surveys have become increasingly sophisticated in capturing “exogenous”
measures that influence household bargaining power such as assets at marriage and non- labor
income as well as intra-household allocation and control of resources and income. The major
outcomes of interest among economists have been household consumption patterns and child
well-being. While such surveys and analyses have advanced efforts to measure empowerment
in many ways, an important limitation has been the overwhelming focus on the situation of
married women (or those living in unions), where empowerment is operationalized largely in
terms of relations between marital partners.

Table 4 lists the indicators used to operationalize empowerment at the individual or household
level in the empirical studies we have included in our review. The two types of indicators
used almost universally in the empirical literature are those measuring domestic decision-
making, and those measuring either access to, or control over resources. Often, these two
aspects merge since indicators on domestic decision- making tend to focus heavily on financial
and resource allocation matters.

25
 Table 4. Individual/household-level indicators of empowerment
used in empirical studies.

Most-Frequently-Used Indicators
Domestic decision-making
Finances, resource allocation, spending, expenditures
Social and domestic matters (e.g., cooking)
Child-related issues (e.g., well-being, schooling, health)
Access to or control over resources
Access to, control of cash, household income, assets, unearned income,
welfare receipts, household budget, participation in paid employment
Mobility/freedom of movement

Less-Frequently-Used Indicators

Economic contribution to household

Time use/division of domestic labor

Freedom from violence

Management/knowledge
Farm management
Accounting knowledge
Managerial control of loan

Public space
Political participation (e.g., public protests, political campaigning)
Confidence in community actions
Development of social and economic collective

Marriage/kin/social support
Traditional support networks
Social status of family of origin
Assets brought to marriage
Control over choosing a spouse

Couple interaction
Couple communication
Negotiation and discussion of sex

Appreciation in household

26
 Sense of self worth

The emphasis on such measures in the empirical literature corresponds well with the emphasis
on resources and agency in the conceptual literature, as well as with the frequent equation of
empowerment with choice, control, and power. Certainly, there is an intuitive appeal to
decision- making and control as signifying important aspects of agency. At the same time,
since data collection at a single point in time does not effectively allow for direct
measurement of long-term strategic choices, a de facto operating assumption (albeit not
always directly stated) in most household level studies is that a person’s ability to make
strategic life choices is linked with her access to, and control over, economic and other
resources and her ability to make smaller, quotidian decisions.

In our basic definition of empowerment drawn from Kabeer (2001), “strategic life choices”
would refer to decisions that influence a person’s life trajectory and subsequent ability to
exercise autonomy and make choices. Examples include decisions related to marriage,
education, employment, and childbearing. One argument is that as such strategic choices are
likely to take place relatively infrequently in a person’s life, it is often difficult to link them
with policy and program interventions unless the time frame of the research is very long.
Given the measurement constraints imposed by the infrequency of “strategic life choices” in
an individual’s life, it almost becomes necessary to consider “small” actions and choices if
measuring empowerment in the short term. Indeed, given their scope, most household- level
studies that have included indicators of women’s empowerment have not focused on
“strategic life choices” but, rather, on what might be termed “empowerment in small things.”

There is some published evidence from empirical studies that the assumption that the ability
to make strategic life choices is linked with the ability to maker smaller decisions is valid, but
results from other studies suggest that this is not always the case. It is not easy to judge from
the existing body of research to what extent the negative results are due to inadequate study
designs and imprecise measurement, due to the multidimensional or contextual nature of
empowerment, or simply the lack of implementing a research design for measurement across
time. For example, it is often not easy for researchers to know whether they have included all
the relevant small or large decisions that are likely to matter for women in specific
circumstances—the relevance of decisions is often specific to the community context, as well
as to ethnic and socio-economic status. Moreover, it is difficult to assign relative weights to
the importance of decisions that are included in an analysis: decision- making power over
cooking is unlikely to be equivalent to decision- making power over children’s schooling or
health, or marriage, but empirical studies often rely on additive indices of domestic decision-
making.

Similarly, the allocation and control of resources can be murkier than they appear at first
sight. For example, Kabeer (2001) points out a lack of conceptual rigor in many quantitative
studies in their operational definitions of access to and control over resources, both of which
are often measured based on questions about women’s involvement in decisions related to
various household expenditures and management of money. The extent to which such
decision- making reflects merely women’s implementation of the tasks relegated to them by
convention remains a question. On the other hand, studies also show that the fact that a

27
woman brings resources into the home or marriage may strengthen her position in the
household, even if she exercises little control over the resource. For example, a woman’s
assets at marriage or participation in a microcredit program may help establish her bargaining
position in the conjugal relationship even if the actual resource utilization is in the hands of
her husband (Hashemi et al. 1996).

Freedom of movement is another common indicator in empirical research at the

individual/household level, especially in studies on South Asia where women’s presence in
the public sphere is often severely constrained. In some circumstances, freedom of movement
could be seen as an empowerment resource, an enabling factor for women’s agency in other
areas of life. On the other hand, taking the initiative to work outside the home or bring a sick
child to a health center could be seen as a form of agenc y in a setting where female
claustration is the norm. Few studies have made qualitative efforts to tease out precisely how
increased freedom of movement either facilitates or reflects the process of empowerment.

At the individual and household levels, other important indicators of empowerment have been
used, but much less frequently in the empirical literature we reviewed. Within the domestic
domain, for example, the relative value of a woman’s economic contribution is used much
less often than the simple fact that she brings in an income or has control over resources.
Kabeer (1997) discusses the shifts in women’s importance in the family because of the weight
of their earnings in her qualitative study of factory workers in Bangladesh. Similarly, despite
the extensive literature on the importance of time use and the domestic division of labor for
defining women’s life options and domestic power in developed country settings, these
indicators are rarely incorporated in research on empowerment for developing country
settings. Acharya and Bennett (1983) demonstrate a relationship between time spent in
market versus non- market activities and women’s decision- making power. In addition, using
the Indonesia Family Life Survey, Frankenberg and Thomas (2001) are able to incorporate
time use in their recent analysis of domestic decision- making and power, mainly due to the
unusually rich data available through this source.

Inclusion of indicators on couple communication has been limited largely to studies on

contraceptive use, while efforts at measuring sexual negotiation and communication have
only begun to gain legitimacy with emerging research on HIV/AIDS. Wolff et al.’s (2000)
analysis of condom use in Uganda considers women’s ability to negotiate and discuss sexual
relations. In the same vein, it is only recently that studies on empowerment have started to
include measures on physical violence or threat, even though it is clear that physical or sexual
intimidation is of critical importance defining ones ability to make strategic life choices. Rao
(1998) finds wife beating to be a key determinant of children’s caloric intake in India.
Qualitative studies (Kabeer 1997 and 1998) often find physical violence and threats of
abandonment to be central elements in processes which shape women’s disempowerment, but
Schuler et al.’s (1996) work in Bangladesh and Jejeebhoy’s (2000) study of women’s
autonomy in India represent the limited quantitative efforts at incorporating this element
within a comprehensive conceptual framework of empowerment.

Similarly, there are valiant, but only sporadic efforts in the literature at capturing
empowerment indicators for social capital and support, or women’s engagement in public

28
spaces and processes (economic, social, and political), again emerging more from qualitative
rather than quantitative studies (Mayoux 2001). Although several household surveys measure
contextual indicators at the community level, few consider the possibility of measuring
individual women’s engagement in community or political processes. Hashemi et al. (1996)
include women’s political and legal awareness and political participation, while Kabeer
(1998) includes confidence in community interactions in their separate analyses of
microcredit and women’s empowerment in Bangladesh. Although not thoroughly reviewed
here, qualitative studies have delved into the emotional and psychological spheres by asking
women about their sense of self worth or value to others (Kabeer 1997 and 1998).

Aggregate-Level Studies

Empirical measurement of women’s empowerment at the aggregate level has not progressed
as substantially as has household- or individual- level measurement. Conceptual frameworks
of how women’s empowerment should be operationalized at the macro level are less well-
developed, and the indicators utilized in studies are less sophisticated, with continued reliance
on proxy measures. Table 5 lists the indicators used to measure women’s empowerment by
empirical studies at the aggregate level. Clearly, capturing either process or agency becomes
much more difficult at higher levels of aggregation; most of the indicators are one step
removed and tend to measure the enabling factors or conditions for empowerment in terms of
labor force participation, labor laws, literacy, education, characteristics of marriage and
kinship, and political representation by women. This inherent difficulty with measuring
women’s empowerment at the aggregate-level analysis is compounded by the lack of
consistent and adequate data on a comprehensive set of even proxy measures or enabling
factors for most developing countries.

29
 Table 5. Aggregate level indicators of empowerment used in empirical studies.

Labor Market
Female labor force participation (or female share, or female/male ratios)
Occupational sex segregation
Gender wage differentials
Child care options
Labor laws
Percentage of wives/women in modern work
Ratio of female/male administrators and managers
Ratio of female/male professional and technical workers
Women’s share of earned income

Education
Female literacy (or female share, female/male ratio)
Female enrollment in secondary school
Maternal education

Marriage/Kinship system
Singulate mean age at marriage
Mean spousal age difference
Proportion unmarried females aged 15-19
Area of rice cultivation
Relative rates of female to male migration
Geographic region

Social Norms and Practices

Wives’/women’s physical mobility

Health/Survival
Relative child survival/Sex ratios of mortality

Political and Legal

Ratio of seats in parliament held by women
Women’s legal rights
Questions, complains, requests from women at village council

In effect, the limited number of studies we were able to select in this category is in itself an
indication of the relative lack of conceptual attention given to measuring women’s
empowerment in empirical work at the aggregate level. In sifting through the large number of
studies that have related the above (and similar) indicators to various development outcomes,
we found only a handful that satisfied our criteria of either addressing agency or process
directly, or having some semblance of a conceptual framework with regard to women’s

30
empowerment. For a substantial portion of this literature, the equation of specific
indicators—eithe r individually, in some combination, or through construction of indices—
with gender inequality, and occasionally empowerment, is often based on arbitrary choices
rather than on conceptual frameworks. The indicators most frequently used in this manner
are women’s education, employment, and legal rights. For example, in their analysis of
gender inequality and economic growth, Dollar and Gatti (1999) decide that women’s
empowerment is one of four relevant dimensions of gender inequality, and measure it using
the percentage of women in parliament and the year when women earned the right to vote,
without discussing why their four specific choices should comprise gender inequality, why the
specific indicators used for each are the appropriate choices for those dimensions, or why
parliamentary representation and the year the right to vote was gained are adequate measures
of women’s empowerment. Similarly, Boone (1996) uses data on rights to construct an index
of “gender oppression,” in order to measure women’s autonomy, without justifying why legal
rights are adequate for measuring either of these concepts.

Our review suggests that single indicators or even composite indices such as the Gender
Empowerment Measure (GEM) are inadequate to the task of measuring wo men’s
empowerment at the aggregate level (Bardhan and Klasen 1999; Oxaal and Baden 1997).
Elson (1999) and others have made an eloquent case for why aggregate- level indicators such
as women’s participation in paid work do not necessarily capture empowerment. In large part,
this is because it is not clear that women are always able to exercise agency as a result of
employment. Similarly, a strong case can be made with regard to the inadequacy of measures
on legal or political rights in reflecting women’s actual ability to exercise rights. For
example, Meinzen-Dick et al. (1997) and Rocheleau and Edmunds (1997) argue that titling,
group political organization for equitable access and control of land, and the actual
distribution of use and ownership at the local level are critical factors in determining if land
rights for women are realizable or not. Similarly, Kabeer (2001) stresses the failure of many
empirical studies in taking full account of the dynamic between de jure rights to land and
other resources, and de facto rights. She cites examples showing that cultural taboos often
make women reluctant to claim their legal inheritance rights. Empirical work at documenting
de facto rights would certainly involve attention to intermediary levels of aggregatio n—at the
community level. Similarly, assessment of women’s empowerment in the labor market would
require analyses at the industry or sectoral levels.

As indicators of control and decision- making are difficult if not impossible to measure at the
macro level, we suggest that the task of inference from macro data would be considerably
facilitated by evaluating a strong cluster of aggregate- level indicators on a specific dimension
of empowerment across at least two points in time. The studies by Tzannatos (1999) and
Winter (1994) provide important examples of this approach. In each case, the authors are
interested in one specific dimension of empowerment: shifts in women’s labor market
position over time. Tzannatos focuses on a range of countries across the world, while
Winter’s work is limited to a group of countries in Latin America. Instead of a single
indicator or an index, both authors use a cluster of labor market indicators—including female
labor force participation, gender wage differentials, occupational sex segregation, etc—to
make their assessment. As Winter’s study is limited to a smaller number of countries, she is
able to use a broader set of indicators, covering labor laws and child care options, thus giving

31
greater weight to her conclusion that women’s position in the labor market in these countries
has improved.

This approach can be considered a form of triangulation: rather than relying on a single
indicator or index to represent a specific dimension, multiple indicators within the same
dimension provide opportunity for confirmation as well as contradiction. Based on the
cumulative evidence, one could infer whether or not women’s ability to make strategic
choices within this specific dimension has increased over time, and the reliability of the
inference would depend on the number and adequacy of the indicators chosen. As mentioned
before, a major difficulty in implementing this approach is the lack of adequate gender-
disaggregated data for developing countries on most indicators that are likely to be relevant in
formulating a cluster.

Intermediate-Level Studies

Our review makes it clear that the lack of studies addressing levels of aggregation
intermediate between the individual/household and the district/state/nation is one of the most
significant gaps in efforts at empirically measuring women’s empowerment. Community,
institutional, and normative factors are considered in studies that take a multilevel approach
(Jejeebhoy 2000; Kabeer 1998; Kritz 2000; Mason 1998), but in general how women in
specific communities may be empowered through shifts in norms, marriage systems, political
processes, etc. is not assessed in the literature. Studies on collective action are one area of
potential contribution to the analysis of women’s empowerment at the local or community
level. Our review of the literature on collective action and women’s empowerment, however,
mostly identified work that is either descriptive or prescriptive in nature. Thus, although there
are excellent examples of social movements and collective action empowering women,
especially in South Asia—e.g., the anti- liquor movement, SEWA, Sonagachi—most studies
document rather than analyze the processes involved (Carr et al. 1996; Wedeen 1996; Bhatt
1989).

We have also noted that efforts at conducting and documenting scientifically sound
evaluations of programmatic and policy interventions are lacking in the literature relevant to
women’s empowerment. It is precisely at the community, institutional, and normative levels
that such interventions tend to play out, and if women’s empowerment is to be a goal of such
initiatives, then some potential measurement schemes need to be developed for capturing the
process at this level. The one study in our review that points to some directio n in this area is
the (rough) assessment of the initiative to increase female representation in village panchayats
in India by Chattopadhyay and Duflo (2001). These authors consider a different range of
questions than the ones addressed by the studies referred to thus far; they examine women’s
participation in village councils, their engagement in the political process, and their
preferences for development investment.

32
How Conclusive Is the Evidence?

Given the strengths and limitations of indicators used to measure empowerment, what can we
say about the existing evidence regarding the factors that empower women, and whether or
not the empowerment of women results in positive development outcomes? As we discuss
below, the vast majority of such studies do not measure empowerment effectively enough for
this question to be answered adequately at this stage. Most studies capture some possible
slice of empowerment rather than empowerment itself. Within this limited range, the record
based on the empirical literature we reviewed seems to be heavily weighted toward positive
relationships. Most studies conclude that enabling factors such as education, employment,
positive marriage or kinship conditions, or programmatic interventions such as microcredit
lead to women having more choice, options, control, or power over their life conditions.
Similarly, studies examining the intermediary role of empowerment also conclude that
women’s control of assets, income, household decision- making, etc. leads to better outcomes
for their families, improved child well-being, and reduced fertility rates.

At the same time, it is important to emphasize that the results are not unequivocally positive,
and that in fact, considerable subjective judgment is involved in the types of analyses
conducted, and the results that are highlighted. For example, if we consider the one
programmatic intervention that is most studied in the literature, microcredit, we find some
conflicting results depending on the authors’ orientation and emphasis. Hashemi et al. (1996)
and Kabeer (1998) conclude that microcredit participation is empowering for women in
Bangladesh while Goetz and Gupta (1996) and Ackerly (1995) conclude that it is not. Kabeer
(1998) argues that the negative results for the latter authors reflect a narrow operationalization
of empowerment. It would also be correct to say that the discrepancy reflects differential
expectations and assessments. Hashemi et al. and Kabeer acknowledge that larger patriarchal
structures were not altered by microcredit, but that women had incrementally more power and
control over their lives within the familial domain. Goetz and Gupta, and Ackerly, on the
other hand, were hoping for shifts in women’s position within the market, an agenda perhaps
too ambitious for the scope of single- focus interventions such as microcredit programs.

The empirical research also indicates contextual differences in the impact of microcredit
programs. For example, Khandker (1998) and Pitt and Khandker (1998) find a very
substantial impact on household expenditure of women rather than men borrowing in
Bangladesh. Schuler et al. (1994 and 1997) also find that it leads to increased contraceptive
use within the Bangladeshi setting. However, Schuler and her colleagues did not find positive
results regarding the impact of participation in microcredit programs on contraceptive
behavior in the Indian and Bolivian settings. Similarly, Mayoux (2001) finds that in
Cameroon, the potential for channeling credit through traditional social networks limits
benefits to women.

The evidence with regard to child health and well-being is perhaps even more equivocal.
Several studies argue that the impact of women’s empowerment on child health and well-
being is generally positive. For examp le, Basu and Basu (1991) and Kishor (1993) base this
conclusion on findings using Indian census data showing that women’s labor force
participation is negatively associated with female-to- male child mortality ratios. At the

33
household level, Rao (1998) concludes that children’s caloric consumption is favored by
women’s education, income, and freedom from domestic violence. At the same time, Kishor’s
(1992) analysis of the Indian census across time (1961 to 1981) suggests that development
may actually allow parents—including women—better opportunities for gender
discrimination by practicing son preference. That factors that may empower women are
possibly emphasizing the disadvantage to their daughters is also apparent in two household-
level studies in Cote d’Ivoire (Haddad and Hoddinott 1994) and Indonesia (Thomas et al.
1997). Both these studies find that women’s relative advantage in assets or income share
leads to health benefits for sons, but not necessarily daughters.

SECTION IV: CONCLUSIONS AND RECOMMENDATIONS

Conclusions

As a review of existing theoretical and empirical literature, this paper is a first step toward
identifying the most promising methodological approaches to measuring and analyzing
women’s empowerment. Our review indicates that although additional iterations of this
process are clearly necessary in order to develop an approach that is viable for the World
Bank and other multilateral organizations, the existing literature provides important guidance
and direction in moving this task forward. While clearly, women’s empowerment is a
complex concept that poses many challenges in conceptualization and measurement, these
challenges are probably no greater than is the case for other complex development concepts
such as poverty reduction or social inclusion. As has been the case with these other concepts,
sustained efforts at analysis and refinement are necessary for moving the measurement agenda
forward.

Our review revealed a number of important strengths in the existing work on women’s
empowerment that provides the foundation for further progress on measurement. The most
important of these is the fact that despite the confusion in rhetoric and terminology, there is
greater consensus in the theoretical literature on what empowerment means and how it should
be conceptualized, and even operationalized, than we had expected. We find that there is
substantial agreement on “process” and “agency” as being essential to women’s
empowerment, and in distinguishing it from related terms such as gender equality. Moreover,
considerable groundwork has already been done in developing frameworks that specify the
dimensions of empowerment, its contextual nature, and the various levels at which it could be
measured. Although no existing framework stands out above others as the one to be adopted,
taken as a whole, the existing frameworks provide the essential raw materials for developing a
workable roadmap for measuring women’s empowerment.

Empirically as well, a number of studies from a range of disciplines have attempted to

measure various aspects of women’s empowerment, either as the outcome of interest, or as
the intermediary factor affecting other development outcomes. Efforts at data collection and
analysis, especially at the household and individua l level, have become more common and
sophisticated in recent years, and although they continue to have limitations, they provide
important guidance for future efforts at measuring women’s empowerment.

34
At the same time, important challenges remain. It is apparent that most of the empirical
studies we reviewed utilize indicators and analyses of empowerment that do not effectively
operationalize the consensus-based definition and conceptualization of empowerment that we
outline in the first half of this paper. The vast majority of empirical studies are not measuring
the process element of empowerment. Additionally, macro- level studies are especially weak
on measuring agency and often do not employ a relevant conceptual framework. Household-
level studies have made significant progress in conceptualizing broader, context-specific
frameworks and in specifying indicators that can be said to capture aspects of agency, but
considerably more work is required in this area. The lack of empirical research at “meso”
levels presents an important gap, as does the relative lack of rigorous research on policy and
programmatic efforts.

Data limitations have also presented an important constraint in efforts to measure women’s
empowerment. Macro- level studies are especially limited in the extent to which they can
operationalize innovative approaches to this task by the lack of gender-disaggregated data
from developing countries on a vast majority of relevant indicators. Our review of the
empirical literature suggests that in many cases researchers recognize the data constraints, but
have had to adapt their studies to the limitations. Others have collected primary data, but
often in isolation from each other, and never through longitudinal or panel studies.

Recomme ndations for Next Steps

We propose the following next steps for moving forward the agenda on measuring women’s
empowerment:

1. Development of a framework of domains or dimensions that can be applied across settings

would be the natural next step for building on the strengths of the existing literature on the
conceptualization of empowerment. Procedures for determining indicators for each
domain, at different levels of aggregation and across contexts, should also be developed.
This effort would move the measurement of women’s empowerment agenda forward
considerably by allowing for greater specification of exactly what aspect of
empowerment—i.e., which dimension—is of interest, and realistic specifications of the
type of change that can be expected over a specific period of time, and given specific
interventions. It would also move forward efforts to develop context-specific measures
that more closely resemble what they are meant to measure and reduce the reliance on
proxy measures.

2. Better, more coordinated efforts at data collection are needed. For example, the process
component of women’s empowerment cannot be effectively captured in any measurement
scheme without the availability of data across time. Attention to process also requires a
discussion of the appropriate time periods for data collection of various types of
indicators. At the aggregate level, a broader range of more sophisticated, gender-
disaggregated data is needed with regard to the labor force, market conditions, legal and
political rights, political and social processes. At the household level, data need to be

35
 more frequently collected for important but relatively under-utilized indicators such as
time use or violence against women.

3. Greater attention to measuring women’s empowerment at meso levels is required along

with efforts at documenting the impact of program and policy interventions. For
programmatic and policy evaluation, existing models of monitoring and evaluation that
are effective need to be tapped, and their adequacy for women’s empowerment as an
outcome or intermediary process should be assessed. At a minimum, quasi-experimental
evaluation designs and the collection of baseline and endline data must be considered in
implementing programs aimed at empowering women. Measurement of institutional and
normative change in communities requires new and innovative approaches. One approach
to consider is the business school model of case studies. Documentation through
narratives which are then analyzed using qualitative techniques would be another option.
Exploration of the work on collective action may also provide further guidance. This is
clearly an area where a review of lessons learned from related efforts and cross-
disciplinary approaches would be helpful.

4. Greater interdisciplinary engagement is necessary to develop indicators and approaches

that capture the key elements of women’s empowerment, have scientific merit, and have
acceptability among important stakeholders. Although at this stage we have drawn only
from literature that has been at the core of the discourse on women’s empowerment, it is
clear that continued efforts at moving this work forward would benefit from drawing on a
wide range of disciplines. Moreover, based on what we reviewed from sociology,
demography, economics, and anthropology, it is clear that there is overlap but not much
interaction across disciplines. Further interdisciplinary engagement would greatly
facilitate the task of translating the current consensus on conceptualization to the actual
measurement of women’s empowerment.

36
 Appendix A: Dimensions of Empowerment
Proposed by Selected Authors

CIDA 1996 Legal empowerment

Political empowerment
Economic empowerment
Social empowerment

Jejeebhoy 1995 Knowledge autonomy

Decision- making autonomy
Physical autonomy
Emotional autonomy
Economic and social autonomy and self-reliance

Kishor 2000a Financial autonomy

Participation in the modern sector
Lifetime exposure to employment
Sharing of roles and decision-making
Family structure amenable to empowerment
Equality in marriage
(lack of) Devaluation of women
Women’s emancipation
Marital advantage
Traditional marriage

Schuler and Hashemi 1993 Mobility and visibility

Hashemi et al. 1996 Economic security
Schuler et al. 1996 Status and decision-making power within the household
Schuler et al. 1997 Ability to interact effectively in the public sphere
Participation in nonfamily groups

Stromquist 1995 Cognitive

Psychological
Economic
Political

A. Sen 1999 Absence of gender inequality in:

Mortality rates
Natality rates
Access to basic facilities such as schooling
Access to professional training and higher education
Employment
Property ownership
Household work and decision- making

37
 Appendix B: Empirical Studies Where Empowerment Is the Outcome of Interest

Independent and Indicators of

Study and Location Sample and Design Intermediary Empowerment Findings
Variables (as dependent variable(s))

Acharya and Bennett 1983 Surveyed and interviewed 478 Market labor versus unpaid Women’s role in farm Bringing women into the
Nepal women and 443 men in 7 family labor management, domestic, and market economy positively
villages resource allocation decisions affects their influence in
resource allocation and
domestic decision-making.

Ackerly 1995 Analyzed data on 826 loans to Loan characteristics and Women’s accounting Women gain knowledge and
Bangladesh 613 women in 3 locations women’s involvement in knowledge: ability to provide empowerment through
market activities funded by information on input costs, market access, but it rarely
loans product yield, and occurs.
profitability of the loan-
funded activity

Frankenberg and Thomas 2001 Analyzed data on 5,168 Relative status of husbands Women’s role in household Status influences financial
Indonesia couples from the decision- and wives at marriage, e.g., decision-making: control over arrangements and decision
making module in the 1997-8 education, age, and social cash, spending, and time use making power.
Indonesia Family Life Survey, status of family of origin
including data from 4 focus
groups

Goetz and Gupta 1996 Interviewed 253 women and Loan characteristics, e.g., size Women’s versus men’s Men often control loans given
Bangladesh 22 men loanees in 5 regions of loan and investment managerial control of loan: to women. Thus, microcredit
activity reported control of productive programs are not necessarily
process, marketing, and empowering women.
inputs

Grasmuck and Espinal 2000 Surveyed 126 men and 75 Women’s versus men’s Women’s role in household Both gender ideology and
Dominican Republic women micro-entrepreneurs financial contribution to expenditure decision-making reliance of households on
and conducted 20 in-depth household individuals’ income are
interviews in Santiago important to autonomy.

38
 Independent and Indicators of
Study and Location Sample and Design Intermediary Empowerment Findings
Variables (as dependent variable(s))

Hashemi et al. 1996 Surveyed 1,248 women Microcredit participation and Empowerment in household Microcredit empowers
Bangladesh following ethnographic women’s contribution to and community spheres: women (in all but two
research in 6 villages household mobility (e.g., number of empowe rment domains) by
places woman goes alone); giving them greater economic
economic security (e.g., value to their families.
women’s investments);
decision making power (e.g.,
ability to make large
purchases and history of
domestic violence); political
and legal awareness (e.g.,
knowledge of name of gov’t
official); participation in
public protests and political
campaigning

Jejeebhoy 2000 Surveyed 1,842 women in two Women’s and household Role in economic decision Some dimensions of
India districts in each of two states characteristics, e.g., district, making (e.g., having major autonomy are more closely
and conducted focus groups religion, education, say in purchase of jewelry); related than others. Proxies
participation in waged work, role in child-related decision are good estimates of
dowry size, marriage making (e.g., having major autonomy for some indicators
endogamy, spousal age say in what to do if child falls in some regions only.
difference, household ill); mobility; freedom from
economic status threat (e.g., whether woman
fears man); control of
resource (e.g., whether
woman has savings)

Kabeer 1997 Interviewed 60 women and 30 Factory wage work and Women’s perceived status in Women reported greater
Bangladesh men garment factory workers women’s contributions to the the household (e.g., status in the household as a
in Dhaka household appreciation, domestic result of factory work, but
violence, input in decision- men reported that women
making) factory workers are low
status.

39
 Independent and Indicators of
Study and Location Sample and Design Intermediary Empowerment Findings
Variables (as dependent variable(s))

Kabeer 1998 Surveyed 696 women loanees Women’s involvement in Perceived changes in Microcredit has decreased the
Bangladesh and conducted in-depth market activities funded by women’s self-worth, agency, trade-offs that women have to
interviews with 50 women and loan contribution to the household, make between dimensions of
20 men and confidence in community their well-being.
interactions

Malhotra and Mather 1997 Surveyed 577 women in Women’s and husband’s Women’s role in household Work for pay and education
Sri Lanka Kalutara district and characteristics, e.g., decision making: control over increase decision- making
conducted a series of focus education, participation in money matters and other input in financial, but not
groups waged work, life course important household matters social and organization
stage, family structure matters in the household.

Mason 1998 Surveyed probability samples Social context in terms of Women’s role in household Social context has indirect
Pakistan, India, Malaysia, of women in 26 clusters of gender and family systems, expenditure decision-making and direct effects on women’s
Thailand, the Philippines villages or urban and women’s and household economic power.
neighborhoods characteristics, e.g., land
assets, participation in waged
work, wife’s rank relative to
husband

Mayoux 2001 Conducted 13 focus groups Microcredit participation and Women’s individual income, Using existing forms of social
Cameroon and in-depth interviews with social capital, i.e. kinship, control over income, and capital to channel microcredit
women in 4 provinces neighborhood, and market development of collective limits benefits to women,
networks social and economic activities especially the poorest
women.

Schuler et al. 1996 Surveyed 1,248 women Microcredit participation Incidences of domestic Domestic violence is less
Bangladesh following ethnographic violence common in communities
research in 6 villages. where microcredit for women
is available.

40
 Independent and Indicators of
Study and Location Sample and Design Intermediary Empowerment Findings
Variables (as dependent variable(s))

Tzannatos 1999 Analyzed International Labor Economic growth—change Women’s labor market Evaluating the most
Multiple countries and regions Organization data from 1950s over time position: labor force representative labor market
to 1990s for available participation rates, indicators in developing
countries—most comparisons occupational sex segregation, countries during the last few
for the 1980s and gender wage differentials decades shows that there has
been a rapid improvement in
women’s labor market
position.

UNDP Human Development Synthesized national-level Gender Empowerment The GEM reflects economic
Report 1995 and 1998 data from a variety of sources Measure (GEM): and political decision
Worldwide ratio of seats in parliament making—that is, women’s
held by women; ratio of ability to take advantage of
female administrators and capabilities.
managers; ratio of female
professionals and technical
workers; women's share of
earned income

Winter 1994 Analyzed employment and Employers’ policy Women's labor market Overall, women's labor
Brazil, Chile, Colombia, Costa earnings data from a variety of interventions in women’s position: labor force market position has
Rica, Honduras, Venezuela national household surveys in formal sector work, e.g., participation rates, improved, although there are
the 1980s hiring and wage behavior occupational sex segregation, still significant personal costs
gender wage differentials, to women. Protective laws,
child care accessibility, and maternity protection laws,
labor laws and laws on child care may
raise the cost of hiring
women. Equal pay provisions
are often ineffective and
discrimination explains much
of the gender wage gap.

41
 Appendix C: Empirical Studies where “Empowerment” Affects Other Outcomes of Interest

Indicators of
Study and Location Sample and Design Empowerment Dependent Variable(s) Findings
(as independent or
intermediary variable(s))

FERTILITY AND CONTRACEPTIVE USE

Abadian 1996 Analyzed data from a variety Women’s relative autonomy Total fertility rate Autonomy has a negative
54 countries of UN and World Bank in household (e.g., singulate impact on fertility.
surveys mean age at marriage, mean
spousal age difference, and
female enrollment in
secondary school)

Dyson and Moore 1983 Analyzed Indian census data Women’s social and Fertility Kinship patterns have a
India from 1971 and secondary economic autonomy (e.g., strong influence on women’s
sources of ethnographic study female labor force autonomy and fertility levels.
participation, percentage of
women practicing purdah,
female literacy, percentage of
births medically attended,
index of son preference)

Gage 1995 Analyzed data on 3,360 Women’s individual Modern contraception use Women’s control over choice
Togo women from the 1988 Togo socioeconomic status and of spouse and access to cash
Demographic Health Survey autonomy (e.g., control over (women’s autonomy)
choice of spouse, and increases contraceptive use.
participation in waged work)

42
 Indicators of
Study and Location Sample and Design Empowerment Dependent Variable(s) Findings
(as independent or
intermediary variable(s))

Govindasamy and Malhotra Analyzed data on 7,857 Freedom of movement, Current contraceptive use and Freedom of mobility and
1996 women from the 1988 Egypt weight of wives’ opinion in preference for joint versus influence in nonreproductive
Egypt Demographic Health Survey household, and preference for independent fertility decision dimensions result in higher
who should control household making contraceptive use. Women
budget who prefer joint decision-
making are not less likely to
use contraceptives. The use of
education and employment as
proxies for autonomy should
be questioned.

Jejeebhoy 1995 Reviewed multiple studies on Education, kinship structures, Fertility Kinship structures affect
Worldwide the relationships between and women’s autonomy educational expansion.
education and fertility. Education decreases fertility
by promoting women’s
autonomy.

Kishor 2000a Analyzed data on 7,123 Women’s role in household Contraceptive use Decis ion-making and
Egypt women from the 1995-6 Egypt decision making (about 7 freedom of movement have
Demographic Health Survey topics); freedom of different effects on unmet
movement (e.g., number of need. Empowerment needs to
places a woman goes alone or be measured directly and in
with children); education; conjunction with proxies.
participation in waged work

Kritz et al. 2000 Analyzed data on 4,870 Gender context of community Desire for children and Gender equity by province
Nigeria women from a 1991 survey of (e.g., index of mean spousal contraceptive use positively affects
women’s status and fertility in age difference; percentage of reproductive behavior, and
4 provinces in each of 5 states wives in modern work; wife’s SES has strongest
wives’ physical mobility); impact in contexts of low
women’s role in household gender equity.
decision-making; individual
socioeconomic status

43
 Indicators of
Study and Location Sample and Design Empowerment Dependent Variable(s) Findings
(as independent or
intermediary variable(s))

Malhotra et al. 1995 Analyzed data on populations Three dimensions of Total fertility rate Each dimension of patriarchy
India of 358 districts from the 1981 patriarchy: active has a relationship to fertility,
Indian census discrimination towards but together they do not
women (e.g., sex ratio of completely explain fertility
mortality); marriage system levels.
(e.g., female migration);
economic value of women
(e.g., area of rice cultivation)

Mason and Smith 2000 Surveyed probablity samples Women’s empowerment in Desire for children and Gender stratification does not
Pakistan, India, Malaysia, of women and conjugal economic and reproductive contraceptive use influence spouses’ agreement
Thailand, Philippines couples in 26 clusters of decision making (e.g., who about number of children to
villages or urban decides the number of have but does influence use
neighborhoods children to have); mobility; of contraception, so that in
freedom from threat (e.g., highly gender stratified
whether wife is beaten); communities, husbands’
couple communication (e.g., preferences have a greater
whether couple has discussed effect than wives’.
family planning)

Schuler and Hashemi 1994 Surveyed 1,248 women Empowerment in household Contraceptive use Microcredit empowers
Schuler et al. 1997 following ethnographic and community spheres: women. Women who are
Bangladesh research in 6 villages mobility (e.g., number of empowered are more likely to
places woman goes alone); use contraceptives. Credit
economic security (e.g., participation and
investments); decision empowerment have
making power (e.g., ability to independent effects on
make large purchases and contraceptive use.
history of domestic violence);
political and legal awareness
(e.g., knowledge of name of
gov’t official); participation
in public protests and
political campaigning

44
 Indicators of
Study and Location Sample and Design Empowerment Dependent Variable(s) Findings
(as independent or
intermediary variable(s))

Schuler et al. 1995a Surveyed 363 women vendors Empowerment in household Contraceptive and modern Microcredit has no effect on
Bolivia and producers who received and community spheres: e.g., health services use empowe rment (except
microcredit and 295 who did whether woman holds office leadership in trade
not, and conducted in-depth in trade association, receives associations), decision
interviews with 30 women and household help from making, contraception or
8 men on contraception husband, tolerates violence, modern health services use,
and participates in traditional perhaps because these are not
support networks) the right empowerment
indicators for Bolivia.

Schuler et al. 1995b Analyzed 50 life histories of Empowerment in household Contraceptive use SEWA is empowering
India self-employed women in and community spheres women, but it does not
Ahmedabad, including 32 (same as for Schuler and translate into greater
members of SEWA, a Hashemi 1994, above, except contraceptive use
women’s NGO that indicators of mobility
were replaced with indicators
of sense of self and vis ion of
future, e.g., saving for the
future, and self-efficacy)

CHILD HEALTH AND WELL-BEING

Basu and Basu 1991 Analyzed state-level data from Women’s labor force Child mortality and sex ratio Women’s employment leads
India the Indian Census on participation in child mortality to decrease in female child
employment and child survival mortality compared to male
from 1981 along with child mortality.
qualitative data from 1988

Desai and Alva 1998 Analyzed data on women with Education Infant mortality, height for Child health does improve
22 developing countries children under 5 years old age, and child immunization with maternal education, but
from 22 Demographic Health status mostly because education is a
Surveys proxy for SES, not because it
empowers women.

45
 Indicators of
Study and Location Sample and Design Empowerment Dependent Variable(s) Findings
(as independent or
intermediary variable(s))

Haddad and Hoddinott 1994 Wives’ versus husbands’ Children’s height for age and Increasing wives’ income
Cote d’Ivoire share of cash income weight for height share leads to better height
for weight outcomes for sons
but not daughters.

Kishor 1992 Analyzed data on 309 districts Women’s economic worth Relative child survival, Increases in development are
India from the Indian census of (e.g., labor force measured by the difference associated with decreases in
1961 and 1981 participation); kinship between log of 1981 and log relative female survival,
structure (e.g., relative female of 1961 juvenile sex ratios perhaps by enabling parents
migration); social to exercise preferences for
stratification (e.g., percentage sons. Greater female labor
of landless farm workers) force participation mitigates
the impact of exogamy on
female survival.

Kishor 1993 Analyzed data on 352 districts Women’s economic worth Gender differences in child Kinship structure (culture)
India from the Indian census of (e.g., labor force mortality and female labor force
1981 participation); kinship participation (economy) are
structure (e.g., relative female both important to gender
migration); social differentials in early
stratification (e.g., percentage childhood mortality.
of landless farm workers)

Kishor 2000b Analyzed data on 3,783 32 indicators of behavioral Infant mortality and child Important to measure all the
Egypt women who had a birth in the and attitudinal factors immunization status components of
last 5 years from the 1995-6 grouped into 10 dimensions empowerment, as different
Egypt Demographic Health of empowerment dimensions are relevant to
Survey different development
indices.

46
 Indicators of
Study and Location Sample and Design Empowerment Dependent Variable(s) Findings
(as independent or
intermediary variable(s))

Rao 1998 Surveyed all 177 women Women’s characteristics Children’s caloric Wife-beating negatively
India potters and 130 of their (e.g., education, number of consumption affects children’s caloric
husbands in three villages, and living children, individual intake. Wife’s income,
interviewed 70 women and 30 income) and household education, and greater
men characteristics (e.g., number of male children
incidences of wife-beating, reduce wife-beating.
and net dowry)

Thomas, Contreras, and Analyzed data on 5,168 Assets brought into marriage Gender differentiation in Sons of women with higher
Frankenberg 1997 couples from the decision- by husbands and wives child illnesses assets at marriage are less
Indonesia making module in the 1997-8 likely than their sisters to
Indonesia Family Life Survey, experience respiratory
including data from 4 focus disorders
groups

HOUSEHOLD CONSUMPTION AND WELL-BEING

Hoddinott and Haddad 1995 Wives’ versus husband’s Household consumption Wives’ share of cash income
Cote d’Ivoire share and control of cash increases budget share spent
income of food and decreases budget
share spent on clothing, meals
eaten out, alcohol and
cigarettes.

Pitt and Khandker 1998 Surveyed 1,528 households in Amount of microcredit to Household consumption Credit to women but not men
Khandker 1998 87 villages women and men and increases non-land assets held
Bangladesh women’s control of resources by women, male and female
labor supply and boys’ and
girls’ schooling. The impact
of female borrowing on total
per capita expenditure is
twice as large as the impact
of male borrowing.

47
 of male borrowing.
Indicators of
Study and Location Sample and Design Empowerment Dependent Variable(s) Findings
(as independent or
intermediary variable(s))

Quisumbing and de la Briere Surveyed 826 households Women’s assets at marriage Expenditure shares of food, Wife’s assets have a positive
2000 from 47 villages in 3 sites and current assets clothing, children’s education and significant effect on the
Bangladesh share of expenditures on
children’s clothing and
education while husband’s
current assets have a positive
effect on food expenditure
share.

Quisumbing and Maluccio 1999 Analyzed data from IFPRI Women’s assets at marriage Expenditure shares of food, When women control more
Bangladesh, Indonesia, surveys of 826 households in education, health, children’s resources expenditures on
Ethiopia, South Africa Bangladesh and 114 clothing, alcohol/tobacco use, education increase, but not
households in Indonesia. Also and child schooling equally for girls and boys
analyzed data on 1500 across nations. Effects on
households from the 1997 other expenditures vary by
Ethopian Rural Household region.
Survey and data on 500
households from the 1998
Project for Statistics on Living
Standards and Development in
KwaZulu-Natal

Thomas 1997 Analyzed data on 55,000 Male and female nonlabor Expenditure shares, nutrient Women’s income is spent
Thomas 1990 households from the Estudio income, total income, and intakes per capita in more on human capital
Brazil Nacional da Despesa Familiar women’s control of income household, and child investments and is associated
(ENDEF) Survey anthropometric outcomes with greater nutrient intake
and better child health.

48
 Indicators of
Study and Location Sample and Design Empowerment Dependent Variable(s) Findings
(as independent or
intermediary variable(s))

REPRODUCTIVE HEALTH

Beegle, Frankenberg and Analyzed data on about 2,000 Women’s characteristics Prenatal care and hospital Individual assets, education,
Thomas 1998 couples from the 1997-8 (e.g., individual assets, delivery and social status of a woman
Indonesia Indonesia Family Life survey education; social status of increase her chances of
family of origin, and getting prenatal and delivery
education of father) care.

Blanc and Wolff 2000 Surveyed 1,356 women and Negotiation and discussion of Condom use The influence of marriage
Wolff et al. 2000 their stable partners and sex between partners and women’s work varies by
Uganda conducted 34 focus groups district, but education and
with women and men in 2 urban residence consistently
districts enhance women’s negotiating
abilities.

Hindin 2000 Analyzed data on 1861 Women’s role in household Women’s anthropometric Women with no control in the
Zimbabwe women who had given birth in decision-making on major status (BMI and Chronic household are more likely to
the last 3 years from the 1994 purchases and participation in Energy Deficiency) have a lower BMI. Thus, lack
Zimbabwe Demographic waged work of control can compromise
Health Survey women’s health.

INVESTMENT IN DEVELOPMENT

Chattopadhyay and Duflo 2001 Surveyed 1/3 of all women Women’s participation in Public goods investment in Women are more likely to
India councilors in 161 village village council (e.g., roads, drinking water, fuel participate if the leader of the
councils and interviewed questions, requests, and equipment, education, and council is a woman and invest
villagers in one village from complaints from women at health more in infrastructure that is
each of 3 village council areas the village council) directly relevant to rural
in Birbhum district, West women’s needs (water, fuel,
Bengal health, roads, etc.); men
invest more in education.

49
 References Cited
Abadian, S. 1996. “Women’s Autonomy and Its Impact on Fertility.” World Development
24(12):1793-1809.

Acharya, Meena, and Lynn Bennett. 1981. “Rural Women of Nepal: An Aggregate Analysis and
Summary of 8 Village Studies.” The Status of Women in Nepal. Volume II, Part 9: Field
Studies. Centre for Economic Development and Administration. Kathmandu: Tribhuvan
University.

Acharya, Meena, and Lynn Bennett. 1983. Women and the Subsistence Sector: Economic
Participation and Household Decision-making in Nepal. Working Paper Number 526.
Washington: World Bank.

Ackerly, Brooke A. 1995. “Testing the Tools of Development: Credit Programmes, Loan
Involvement, and Women’s Empowerment.” IDS Bulletin 26(3):56-68.

Agarwal, Bina. 1997. “Bargaining and Gender Relations: Within and Beyond the Household.”
Food Consumption and Nutrition Division Discussion Paper No. 27. International Food
Policy Research Institute.

Bardham, Kalpana, and Stephan Klasen. 1999. “UNDP’s Gender-Related Indices: A Critical
Review.” World Development 27(6):985-986.

Basu, Alaka Malwade and Kaushik Basu. 1991. “Women’s Economic Roles and Child Survival:
the Case of India.” Health Transition Review 1(1): ???.

Batliwala, Srilatha. 1994. “The meaning of Women’s Empowerment: New Concepts from
Action.” Pp. 127-138 in Population Policies Reconsidered: Health, Empowerment and
Rights. G. Sen, A. Germain, and L.C. Chen, eds. Cambridge, MA: Harvard University
Press.

Beegle, Kathleen, Elizabeth Frankenberg and Duncan Thomas. 1998. “Bargaining power Within
Couples and Use of Prenatal and Delivery Care in Indonesia.” Studies in Family Planning
32(2):130.

Bennett, Lynn. 2002. Using Empowerment and Social Inclusion for Pro-poor Growth: A Theory
of Social Change. Working Draft of Background Paper for the Social Development
Strategy Paper. Washington: World Bank.

Bhatt, Ela. 1989. “Toward Empowerment.” World Development 17(7):1059-1065.

Bisnath, Savitri and Diane Elson. 1999. Women’s Empowerment Revisited. Background paper,
Progress of the World’s Women. UNIFEM. Available on- line at
www.unifem.undp.org/progressww/empower.html.

50
Boone, Peter. 1996. Political and Gender Oppression as a Cause of Poverty. CEPR Discussion
Paper 294. London: London School of Economics, Center for Economic Performance.

Carr, Marilyn, Martha Chen, and Renana Jhabvala (eds). 1996. Speaking Out: Women’s
Economic Empowerment in South Asia. London: IT Publicatons on behalf of Aga Khan
Foundation Canada and UNIFEM.

Chambers, Robert. 1997. Whose Reality Counts? Putting the First Last. London: ITDG
Publishing.

Chattopadhyay, Raghabendra and Esther Duflo. 2001. “Women’s Leadership and Policy
Decisions: Evidence From a Nationwide Randomized Experiment in India.”

Chen, Marty. 1992. Conceptual Model for Women’s Empowerment. Unpublished.

Canadian International Development Agency (CIDA). 1996. Guide to Gender-Sensitive

Indicators. Ottawa: CIDA

Desai, S. and S. Alva. 1998. “Maternal Education and Child Health: Is here a Strong Causal
Relationship?” Demography 35(1): 71-81.

DFID. 2000. Poverty Elimination and the Empowerment of Women: Strategies for Achieving the
International Development Targets. London: DFID

Dollar, David and Roberta Gatti. 1999. “Gender Inequality, Income, and Growth: Are Good
Times Good for Women?” Background paper for Engendering Development.
Washington: World Bank.

Durrant, V.L. and Z.A. Sathar. 2000. Greater Investments in Children Through Women’s
Empowerment: A Key to Demographic Change in Pakistan. Paper prepared for
presentation at the annual Population Association of America meetings, March 2000, Los
Angeles.

Dyson, Tim and Mick Moore. 1983. “On Kinship Structure, Female Autonomy, and
Demographic Behavior in India.” Population and Development Review 9(1):35-60.

Elson, Diane. 1999. “Labor Markets as Gendered Institutions: Equality, Efficiency and
Empowerment Issues.” World Development 27(3):611.

Estudillo, Jonna P, Agnes R. Quisumbing, and Keijiro Otsuka. 2001. “Gender Differences in
Land Inheritance, Schooling and Lifetime Income: Evidence from the Rural Philippines.”
Journal of Development Studies 37(4):23-48.

Everett, Jana. 1991. The Global Empowerment of Women. Summary of the annual conference of
the Association for Women in Development, November 17-19, 1989, Washington.

51
Frankenberg, Elizabeth, and Duncan Thomas. 2001. Measuring Power. Food Consumption and
Nutrition Division Discussion Paper No. 113. Washington: International Food Policy
Research Institute.

Friedmann, John. 1992. Empowerment: The Politics of Alternative Development . Cambridge,

Mass.: Blackwell Publishers.

Freire, Paulo. 1994[1973]. Pedagogy of the Oppressed. Translated by Myra Bergman Ramos.
New York: Continuum.

Gage, Anastasia J. 1995. “Women’s Socioeconomic Position and Contraceptive Behavior in

Togo.” Studies in Family Planning 26(5):264-277.

Ghuman, Sharon J. 2002. Women’s Autonomy and Child Survival in Five Asian Countries. PhD
dissertation in Demography, University of Pennsylvania.

Goetz, Anne Marie, and Rina Sen Gupta. 1996. “Who Takes the Credit? Gender, Power, and
control ver loan Use In Rural Credit Programs in Bangladesh.” World Development
24(1):45-63.

Govindasamy, Pavalavalli, and Anju Malhotra. 1996. “Women’s Position and Family Planning
in Egypt.” Studies in Family Planning 27(6):328-340.

Grasmuck, Sherri, and Rosario Espinal. 2000. “Market Success or Female Autonomy? Income,
Ideology, and Empowerment among Microentrepreneurs in the Dominican Republic.”
Gender and Society 14(2):231-255.

Haddad, Lawrence, and John Hoddinott. 1994. “Women’s Income and Boy-Girl Anthropometric
Status in Cote d’Ivoire.” World Development 22(4):543-53.

Hashemi, Syed M., and Sidney Ruth Schuler. 1993. Defining and Studying Empowerment of
Women: A Research Note from Bangladesh. JSI Working Paper No. 3. Washington: John
Snow, Inc.

Hashemi, Syed M., Sidney Ruth Schuler, and Ann P. Riley. 1996. “Rural Credit Programs and
Women’s Empowerment in Bangladesh.” World Development 24(4):635-653.

Hindin, Michelle J. 2000. “Women’s Power and Anthropometric Status in Zimbabwe.” Social
Science and Medicine 51:1517-1528.

Hoddinott, John, and Lawrence Haddad. 1995. “Does Female Income Share Influence
Household Expenditures? Evidence from Cote d’Ivoire.” Oxford Bulletin of Economics
and Statistics 57(1):77-96.

Jahan, Rounaq. 1995. The Elusive Agenda: Mainstreaming Women in Development. Atlantic
Highlands, N.J.: Zed Books Ltd.

52
Jejeebhoy, Shireen J. 1995. Women’s Education, Autonomy, and Reproductive Behaviour:
Experience from Four Developing Countries. International Studies in Demography,
IUSSP. Oxford: Clarendon Press.

Jejeebhoy, Shireen J. 2000. “Women’s Autonomy in Rural India: Its Dimensions, Determinants,
and the Influence of Context.” In Women’s Empowerment and Demographic Processes:
Moving Beyond Cairo. Harriet Presser and Gita Sen, eds. New York: Oxford University
Press.

Jejeebhoy, Shireen J. and Zeba A. Sathar. 2001. “Women’s Autonomy in India and Pakistan:
The Influence of Religion and Region.” Population and Development Review 27(4):687-
712.

Kabeer, Naila. 1997. “Women, Wages and Intra-household Power Relations in Urban
Bangladesh.” Development and Change 28:261-302.

Kabeer, Naila. 1998. ‘Money Can’t Buy Me Love?’ Re-evaluating Gender, Credit and
Empowerment in Rural Bangladesh. IDS Discussion Paper 363.

Kabeer, Naila. 2001. “Reflections on the Measurement of Women’s Empowerment.” In

Discussing Women’s Empowerment-Theory and Practice. Sida Studies No. 3. Novum
Grafiska AB: Stockholm.

Khandker, Shahidur R. 1998. Fighting Poverty with Microcredit: Experience in Bangladesh.

Washington: World Bank.

Keller, B. and D.C. Mbwewe. 1991. “Policy and Pla nning for the Empowerment of Zambia’s
Women Farmers.” Canadian Journal of Development Studies 12(1):75-88 [as cited in
Rowlands, Jo. 1995. “Empowerment examined.” Development in Practice 5(2):101-107].

Kishor, Sunita. 1992. All ‘Devis’ but Not All Wanted: A District-Level Analysis of Female
Discrimination in India 1961-1981. Paper prepared for presentation at the annual meeting
of the Population Association of America, April 30-May 2, 1992, Denver, Colo.

Kishor, Sunita. 1993. “May God Give Sons to All: Gender and Child Mortality in India.”
American Sociological Review 58(2):247-65.

Kishor, Sunita. 1995. Autonomy and Egyptian Women: Findings from the 1988 Egypt
Demographic and Health Survey. Occasional Papers 2. Calverton, Md.: Macro
International Inc.

Kishor, Sunita. 2000a. “Empowerment of Women in Egypt and Links to the Survival and Health
of Their Infants.” In Women’s Empowerment and Demographic Processes: Moving
Beyond Cairo. Harriet Presser and Gita Sen, eds. New York: Oxford University Press.

53
Kishor, Sunita. 2000b. Women’s Contraceptive Use in Egypt: What Do Direct Measures of
Empowerment Tell Us? Paper prepared for presentation at the annual meeting of the
Population Association of America, March 23-25, 2000, Los Angeles, Calif.

Kritz, Mary M., Paulina Makinwa-Adebusoye, and Douglas T. Gurak. 2000. “The Role of
Gender Context in Shaping Reproductive Behaviour in Nigeria.” In Women’s
Empowerment and Demographic Processes: Moving Beyond Cairo. Harriet Presser and
Gita Sen, eds. New York: Oxford University Press.

Kumar, Radha. 1993. The History of Doing: An Illustrated Account of Movements for Women’s
Rights and Feminism in India, 1800-1990. New York: Verso.

Malhotra, Anju and Mark Mather. 1997. “Do Schooling and Work Empower Women in
Developing Countries? Gender and Domestic Decisions in Sri Lanka.” Sociological
Forum 12(4):599-630.

Malhotra A., R. Vanneman, and S. Kishor. 1995. “Fertility, Dimensions of Patriarchy, and
Development in India.” Population and Development Review 21(2):281-305.

Mason, Karen. 1986. “The Status of Women: Conceptual and Methodological Issues in
Demographic Studies.” Sociological Forum 1(2):284-300.

Mason, Karen. 1995. Gender and Demographic Change: What Do We Know? Location: IUSSP.

Mason, Karen Oppenheim. 1997. “Gender and Demographic Change: What Do We Know?” IN
The Continuing Demographic Transition. Gavin W. Jones, Robert M. Douglas, John C.
Caldwell, and Rennie M. D’Souza, eds. Pp. 158-182. Oxford: Clarendon Press.

Mason, Karen. 1998. “Wives’ Economic Decision- making Power in the Family: Five Asian
Countries.” Pp. 105-133 in The Changing Family in Comparative Perspective: Asia and
the United States. Karen Oppenheim Mason, ed. Honolulu: East-West Center.

Mason, Karen and Herbert L. Smith. 2000. “Husbands’ versus Wives Fertility Goals and Use of
Contraception: The Influence of Gender Context in Five Asian Countries.” Demography
37(3):299-311.

Mayoux, Linda. 2000. Micro-finance and the Empowerment of Women: A Review of the Key
Issues. Geneva: International Labor Organization. Available on- line at
Ilo_data/public/english/empoyment/finance/download/wp23.wpd.

Mayoux, Linda. 2001. “Tackling the down side: Social Capital, Women’s Empowerment and
Micro- finance in Cameroon.” Development and Change 32:435-464.

Mehra, Rekha. “Women, Empowerment, and Economic Development.” The Annals of the
Academy 554:136-149.

54
Meinzen-Dick, Ruth, Lynn R. Brown, Hilary Sims Feldstein, and Agnes R. Quisumbing. 1997.
Gender, Property Rights, and Natural Resources. Food Consumption and Nutrition
Division Discussion Paper No. 29. Washington: International Food Policy Research
Institute.

Narayan, Deepa (ed). 2002. Empowerment and Poverty Reduction: A Sourcebook. Washington:
World Bank.

Narayan, Deepa, Robert Chambers, Meera K. Shah, and Patti Petesch. 2000a. Voices of the
Poor: Crying Out for Change. World Bank Series. Oxford: Oxford University Press.

Narayan, Deepa, with Raj Patel, Kai Schafft, Anne Rademacher, and Sarah Koch-Schulte.
2000b. Voices of the Poor: Can Anyone Hear Us? World Bank Series. Oxford: Oxford
University Press.

Nussbaum, Martha. 2000. Women and Human Development: The Capabilities Approach. New
York: Cambridge Press.

Oxaal, Zoe, with Sally Baden. 1997. Gender and empowerment: Definitions, Approaches and
Implications for Policy. Bridge Report No. 40. Sussex: Institute of Development Studies.

Pitt, Mark M. and Shahidur R. Khandker. 1998. “The Impact of Group-based Credit Programs on
Poor Households in Bangladesh: Does the Gender of Participants Matter?” Journal of
Political Economy 106:958-96.

Pitt, M.M., S.R. Khandker, S.M. McKernan, and M. Abdul Latif. 1999. “Credit Programs for the
Poor and Reproductive Behavior in Low- income Countries: Are the Reported Causal
Relationships the Result of Heterogeneity Bias?” Demography 36(1):1-21.

Pulerwitz, Julie, Steven L. Gortmaker, and William DeJong. 2000. “Measuring Sexual
Relationship Power in HIV/STD Research.” Sex Roles 42(7/8):637-660.

Quisumbing, Agnes R., and Benedicte de la Briere. 2000. Women’s Assets and Intrahousehold
Allocation in Rural Bangladesh: Testing Measures of Bargaining Power. Food
Consumption and Nutrition Division Discussion Paper No. 86. Washington: International
Food Policy Research Institute.

Quisumbing, Agnes R. and John A. Maluccio. 1999. Intrahousehold Allocating and Gender
Relations: New Empirical Evidence. Policy Research Report on Gender and
Development, Working Paper Series No. 2, World Bank Development Research
Group/Poverty Reduction and Economic Management Network. Washington: World
Bank.

Quisumbing, Agnes R., Ellen Payongayong, J.B. Aidoo, and Keijiro Otsuka. 1999. Women’s
Land Rights in the Transition to Individualized Ownership: Implications for the
Management of Tree Resources in Western Ghana. Food Consumption and Nutrition

55
 Division Discussion Paper No. 58. Washington: International Food Policy Research
Institute.

Rao, Viajyendra. 1998. “Wife-abuse, Its Causes and Its Impact on Intra-household Resource
Allocation in Rural Karnataka: A ‘Participatory’ Econometric Analysis.” In Gender,
Population and Development. Maithreyi Krishnaraj, Ratna M. Sudarshan, and Ausaleh
Shariff, eds. Delhi, Oxford, and New York: Oxford University Press.

Ravallion, Martin and Shaohua Chen. 2001. “Measuring Pro-poor Growth.” Policy Research
Working Paper no. WPS2666. Washington: World Bank.

Reeves, Hazel and Sally Baden. 2000. Gender and Development: Concepts and Definitions.
BRIDGE Development-Gender Report No. 55.

RESULTS. 1997. The Microcredit Summit February 2-4, 1997 Declaration and Plan of Action.

Rocheleau, Dianne and David Edmunds. 1997. “Women, Men and Trees: Gender, Power and
Property in Forest and Agrarian Landscapes.” World Development 25(8):1351-1371.

Rogers, E.M. 1995. Diffusion of Innovations, Fourth edition. New York: Free Press.

Rowlands, Jo. 1995. “Empowerment Examined.” Development in Practice 5(2):101-107.

Schuler, Sidney Ruth, Syed M. Hashemi, Amy Cullum and Mirza Hassan. 1996. “The Advent of
Family Planning as a Social Norm in Bangladesh.” Reproductive Health Matters 7:66-
78.

Schuler, Sidney Ruth and Syed M. Hashemi. 1994. “Credit Programs, Women’s Empowerment,
and Contraceptive Use in Rural Bangladesh.” Studies in Family Planning 25(2):65-76.

Schuler, Sidney Ruth, Syed M. Hashemi and Harshida Pandit. 1995a. Beyond Credit: SEWA’s
Approach to Women’s Reproductive Lives in Urban India. JSI Research Report.

Schuler, Sidney Ruth, Syed M. Hashemi and Ann P. Riley. 1997. “The Influence of Changing
Roles and Status in Bangladesh’s Fertility Transition: Evidence from a Study of Credit
Programs and Contraceptive use.” World Development 25(4).

Schuler, Sidney Ruth, Syed M. Hashemi, Ann P. Riley, and Shireen Akhter. 1996. “Credit
programs, patriarchy and men’s violence against women in rural Bangladesh.” Social
Science and Medicine 43(12):1729-1742.

Schuler, Sidney Ruth, Ann Hendrix Jenkins and John W. Townsend. 1995b. Credit, Women's
Status and Fertility Regulation in Urban Market Centers of Bolivia. JSI Working Paper
No. 8. Washington: John Snow, Inc.

Sen, Amartya. 1999. Development as Freedom. Oxford: Oxford University Press.

56
Sen, Gita. 1993. Women’s Empowerment and Human Rights: The Challenge to Policy. Paper
presented at the Population Summit of the World’s Scientific Academies.

Sen, Gita and Caren Grown. 1987. Development, Crises, and Alternative Visions: Third World
Women’s Perspectives. New York: Monthly Review Press.

Stromquist, Nelly P. 1995. “The Theoretical and Practical Bases for Empowerment.” Women,
Education and Empowerment: Pathways Towards Autonomy. Carolyn Medel-Anonuevo,
ed. Report of the International Seminar held at UIE, January 27-February 2, 1993,
Hamburg, Germany. Paris: UNESCO.

Thomas, Duncan. 1990. “Intrahousehold Resource Allocation: An Inferential Approach.”

Journal of Human Resources 25:635-664.

Thomas, Duncan. 1997. “Incomes, Expenditures, and Health Outcomes: Evidence on

Intrahousehold Resource Allocation.” In Intrahousehold Resource Allocation in
Developing Countries: Models, Methods, and Policy. Lawrence Haddad, John
Hoddinott, and Harold Alderman, eds. Ba ltimore, MD: Johns Hopkins University Press.

Thomas, Duncan, Dante Contreras, and Elizabeth Frankenberg. 1997. Child Health and the
Distribution of Household Resources at Marriage. Santa Monica, Calif: RAND, and Los
Angeles: UCLA Department of Economics.

Tzannatos, Zafiris. 1999. “Women and Labor Market Changes in the Global Economy: Growth
Helps, Inequalities Hurt and Public Policy Matters.” World Development 27(3):551-569.

United Nations Children’s Fund (UNICEF). 1994. The Women's Equality and Empowerment
Framework. Available on- line at www.unicef.org/programme/gpp/policy/empower.html.

United Nations Children’s Fund (UNICEF). 1999. Human Rights for Children and Women:
How UNICEF Helps Make Them a Reality. Available on- line at
www.unicef.org/pubsgen/humanrights-children/index.html.

United Nations Development Programme (UNDP). 1995. Human Development Report 1995:
Gender and Development. New York and Oxford: Oxford University Press.

United Nations Development Programme (UNDP). 1998. Human Development Report 1998:
Consumption for Human Development. New York and Oxford: Oxford University Press.

UN Division for the Advancement of Women (UNDAW), Department of Economic and Social
Affairs. 2001. Empowerment of Women Throughout the Life Cycle as a Transformative
Strategy for Poverty Eradication. Report of the Expert Group Meeting, 26-29 November.
New Delhi, India.

Wedeen, Laura. 1996. Women's Empowerment in Nepal. Kathmandu: USAID.

57
Winter, Carolyn. 1994. Working Women in Latin America: Participation, Pay, and Public
Policy. The World Bank LAC Region Technical Department.

Wolff, B., A.K. Blanc, and A.J. Gage. 2000. “Who Decides? Women’s Status and Negotiation of
Sex in Uganda.” Culture, Health and Sexuality 2(3):303-322.

World Bank. 2001a. Engendering Development: Through Gender Equality in Rights, Resources,
and Voice. World Bank Policy Research Report. Oxford: Oxford University Press.

World Bank. 2001b. World Development Report 2001: Attacking Poverty. New York: Oxford
University Press.

58