Recommendations: 1) develop a framework, 2) better, more coordinated data collection, 3) greater attention to meso level indicators, 4) more interdisciplinary work
Power over – may not be appropriate for women in LMIC settings
N ranges from 479 (freedom from violence) to 4421 (minor HH expenditures)
Analysis of DHS data in Donald et al. shows similar patterns (non-negligible disagreement)
Pro-WEAI is made up of 12 indicators of empowerment. An individual is considered adequate in an indicator if s/he achieves a certain threshold.
Each indicator receives equal weight (1/12), and an individual is considered empowered if s/he is adequate in at least 75%, or 9/12 of the indicators.
WEAI was going viral – it was piloted in 3 countries, then rolled out to 19 FTF countries, and it has now been used by over 75 organizations in 50 countries. Projects were adopting and adapting it, but we had no handle on comparability.
WEAI was developed as a population-based index, but many organizations were using it at the project level.
At the same time, projects wanted a more streamlined measure that could measure areas of empowerment that are important for many projects – such as nutrition and health, mobility, intrahousehold harmony, domestic violence.
What was really interesting was that in our qualitative work, we found that women’s and men’s own understandings of empowerment echoed these three dimensions. So we thought this framework was really in line with what we’re hearing from the communities themselves.
One key insight that we draw from these preliminary findings was that empowerment was not just about the woman herself, instead an empowered woman is someone who can take care of others.
So it’s not just about the woman, it goes beyond her.
Discriminatory social institutions are defined as the formal and informal laws, attitudes and practices that restrict women’s and girls’ access to rights, justice and empowerment opportunities. These are captured in a multi-faceted approach by SIGI’s variables that combine qualitative and quantitative data, taking into account both the de jure and de facto discrimination of social institutions, through information on laws, attitudes and practices. The variables span all stages of a woman’s life in order to show how discriminatory social institutions can interlock and bind them into cycles of poverty and disempowerment.
Measurement of Empowerment: Critiques and Innovation from Development Economics
Empowerment Workshop April 17-18th 2019: Toronto
MEASUREMENT OF EMPOWERMENT:
Critiques & Innovation from Development Economics
MEASURING ‘EMPOWERMENT’: HOW FAR HAVE WE COME?
Malhortra, Schuler, Boender (2002): “Measuring Women’s
Empowerment as a Variable in International Development”
“Vast and interconnected” . . .
“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”. . .
Household Community Broader Arenas
THE CONCEPTUAL CHALLENGE: ARE WE REALLY
“Expansion in people’s ability to
make strategic life choices in a
context where this was
previously denied” (Kabeer 2001)
Should be distinct from gender
equality, equity and status
Should operate at different
levels: 1) power within (intrinsic
agency), 2) power to
(instrumental agency), 3) power
with (collective agency)
HOW DO WE OPERATIONALIZE AGENCY? (Donald et al. 2017)
• Intra-household decision-making (e.g. DHS)
• Autonomy; Relative autonomy index (RAI)
• Capacity to set & achieve goals
• Locus of control
• Self-assessed life freedom & control (e.g. WVS)
• Voice (individual & collective) & participation
DECIDING OVER WHAT?
1. What type of decisions are ‘strategic’ life choices in different
• Are purchases for household daily needs strategic?
• Is visiting family or friends strategic?
• Are decisions across domains appropriately captured (social, economic,
2. Are decisions specific enough?
• Instead of asking “Who usually makes decisions about healthcare for
• Consider asking: “If you ever need medicine for yourself (for a
headache, for example), could you go buy it yourself?
An evaluation in Bangladesh found different responses to general
vs. specific question; suggests specific questions are easier to
accurately answer (Glennerster et al. 2017).
HOW DO WE INTERPRET
Is joint decision-making “better” or
“worse” than sole decision-making?
How do we minimize implicit
judgement about what is preferred?
Analysis shows that women and
men (in Ghana & Bangladesh)
associate different domains more
and less strongly with autonomous
Sole DM matters for women in
Ghana select domains; Women in
Bangladesh prefer joint DM
Women in Ghana (N=1785)
• Should we be measuring ‘how much’ input, rather than
‘with whom’ (by whom)?
• Alternative from the WEAI: How much input do you have into
decision around [X]?
• Response options:
• (a) no input or input into a few decisions,
• (b) input into some decisions,
• (c) input into most decisions,
• (d) input into all decisions,
• (e) no decision made.
• (e) allows respondents to report (opt out) if a decision in that
particular domain has not been made recently or is not
relevant to them.
DO RESPONSE OPTION CONSTRAIN (MISLEAD)
• Couples agree 6-64%
of the time
(Bangladesh) & 67-
82% of the time
• Women tend to report
decisions as joint, men
tend to report as sole
Couples in Bangladesh (2011-2012)
WHY ARE DECISIONS MADE?
The Self-determination continuum (Seymour & Peterman 2018)
ASK ME WHY? VINETTES TO UNDERSTAND
DECISION-MAKING MOTIVATION (Bernard et al. 2018)
• Vignettes around production of
dairy & consumption of dairy in
Senegal = 5 typologies of
• Men most likely to be deciding in
dictator households; Women
most likely in separate sphere
• WHO decides does not affect
outcomes, but WHY they decide
does matter (norms worst off,
informed best off).
• Capturing preferences To what extent do you feel you can make
your own personal decisions regarding [X] if you want(ed) to? (WEAI)
• Social desirability bias Observational (field experiments) or use
indirect methods (list randomization), solicit willingness to pay estimates
-- however these methods may be difficult in national surveys
• Empowerment as a process Take measures over time and analyze
• Disempowerment can heighten data collection challenges When
women face structural inequalities this can compound measurement
challenges. Create space for private interviews, same sex/locally
recruited to encourage truthful answers (Glennerster et al. 2017).
INDEX EXAMPLES: WOMEN’S EMPOWERMENT IN
AGRICULTURE INDEX (WEAI) (Malapit et al. 2019)
• Objectives: Developed by USAID, IFPRI, OPHI in
2012 to measure inclusion of women in agriculture
• Survey based: Constructed using interviews of
primary male and female in the same household
• Versions & iterations:
1. Abbreviated WEAI (a-WEAI)
2. Project-level WEAI (pro-WEAI) Health &
3. WEAI for Value Chains (WEAI4VC)
4. WE in Livestock Index (WELI)
5. WE in Fish Index (WEFI) 12 indicators of empowerment
WEAI QUALITATIVE WORK: HOW COMMUNITIES
UNDERSTAND EMPOWERMENT (Meinzen-Dick et al. 2019)
Synthesis of Pro-WEAI data across 8 projects (SA & SSA)
Empowerment = emancipated, admired, dignified, lift up, enable
Each component of empowerment is relational, not just about
the individual (other women as well as men. . . )
RESOURCES AGENCY ACHIEVEMENTS
• Economic means
• Help with labor
• Taking care of oneself
• Taking care of family needs
• Taking care of others
• Active, hard work
• Good decisions, joint decisions
• Exercising ownership
• Following social norms
• Financial status
• Well dressed
• Good appearance
MACRO-LEVEL INDEX EXAMPLE: SOCIAL INSTITUTIONS
& GENDER INDEX (SIGI) (4th edition, 2019: OECD, https://www.genderindex.org/)
“Discriminatory social institutions are defined as the formal and informal laws, attitudes and practices
that restrict women’s and girls’ access to rights, justice and empowerment opportunities.”
01. Women’s decision-making often the
primary (& only) indicator of women’s agency
operationalized—however large gaps in
understanding what is being captured & how
to interpret measures
02. Women’s & men’s goals/preferences
(values) should be reflected (captured)
alongside agency to achieve these goals.
03. We should be capturing multiple
measures of agency, over different domains
to triangulate outcomes
01. Measuring empowerment is complex—let’s
resist simplification & strive to measure it fully
(rather than related concepts, e.g. women’s
status or wellbeing)
02. Much to be gained from inter-disciplinary &
mixed method work.
03. While seeking to distill commonalities, we
cannot ignore context and individual specificity
of empowerment across different types of
agency—for women and men / girls and boys.
This presentation draws on research
of colleagues at IFPRI, including
Greg Seymour, Hazel Malapit,
Agnes Quisumbing, Ruth Meinzen-
Dick and the WEAI team, Cheryl
Doss, Melissa Hidrobo and others.