Women as Policy Makers: Evidence from a Randomized Policy Experiment in India

What changes when women gain authority over a public budget, rather than merely being consulted about their needs? Raghabendra Chattopadhyay and Esther Duflo’s 2004 article studies the reservation of village council leadership positions for women in India. It examines whether representation affects public-goods provision. The question concerns access to decision-making institutions, not whether women possess an innate disposition to care for a community.

The final Econometrica article uses data from 265 village councils in West Bengal and Rajasthan. Its institutional setting reserved one third of council head positions for women through random allocation; residents still elected leaders. The researchers compared infrastructure in reserved and unreserved councils. Random reservation is the basis for identifying the policy’s effects, not a random assignment of individual women to office.

A council’s responsibilities make representation material. Decisions concern local public goods and the resources devoted to them. The researchers use the kinds of formal requests submitted by women and men as proxies for expressed priorities. These requests are informative about what reaches the council, but they are not transparent records of every resident’s need or an inventory of permanent female and male preferences.

Water in both states, roads in different directions

The comparison of roads is especially revealing. In West Bengal, women raised drinking water and roads more often than men, and reserved councils invested more in both. In Rajasthan, women raised water more often but roads less often; reservation increased water investment and reduced road investment. The same policy did not produce an identical infrastructure package in the two settings.

That variation supports a structural reading of gendered priorities. Everyday responsibilities and local conditions can make an amenity matter differently to people positioned differently. The result should not be converted into a claim that women naturally prefer one public good. Nor does changing the allocation settle whether every woman benefits equally: differences within the category remain relevant to evaluating public provision.

The authors also investigate whether inexperience, disadvantaged backgrounds or expectations about re-election explain the findings, and report no evidence that these alternatives drive the policy differences. This strengthens their interpretation of representation within the studied institutions. It does not turn a policy experiment into an isolated biological effect of sex, or establish that any female politician in any political system will produce the same priorities.

A further question follows from the design: whose preferences become publicly expressible? Formal requests are made under conditions that may themselves be unequal. Comparing them with budgets reveals an important part of representation, while leaving room to ask about people who cannot attend, complain or influence a representative. This is an analytical extension of the measurement choice, not a new result claimed by the paper.

Read with Agarwal’s work on gendered access to environmental resources, water and roads become questions about labour, access and authority. Kabeer on empowerment measurement helps keep a distinction between holding office, changing expenditure and expanding people’s effective choices. The article offers evidence for taking representation seriously while preserving the local and institutional conditions of the finding.