Title: Ethnic Identity, Collective Action and Conflict: An Experimental Approach
Name: Macartan Humphreys

Associate Professor
Department of Political Science
Columbia University
Center for the Study of Development Strategies

Year: 2003
Type: Research Grant

A large literature shows that ethnically homogenous communities often do a better job than diverse communities of producing satisfactory schools and health care, adequate sanitation, low levels of crime, and other essential outcomes that depend on community- wide cooperation. This project seeks to find out why. The research indicates that the principal obstacle to cooperation in diverse groups is not ethnic favoritism or a lack of consensus on what should be done but rather the stronger expectations of reciprocity that exist within than across ethnic communities. The results offer important lessons for policymakers committed to improving the welfare of people living in diverse communities.

The study was conducted in a neighborhood of Kampala, Uganda that has both high levels of diversity and low levels of public goods provision. The researchers used behavioral games to explore how the ethnicity of the person with whom one is interacting shapes social behavior. Hundreds of local participants interacted with various partners in strategic games involving the allocation of money and the completion of joint tasks. Each game was designed to capture a different channel through which ethnic diversity might affect social cooperation. Many of the subsequent findings debunk long-standing explanations for diversitys adverse effects.

Contrary to the prevalent notion that shared preferences facilitate collective action within ethnic groups, differences in goals and priorities among participants were not found to be structured along ethnic lines. Nor was there evidence that subjects favored the welfare of their coethnics over that of non-coethnics. When given the opportunity to act altruistically in an anonymous fashion, individuals did not choose to benefit coethnics disproportionately. Yet when anonymity was removed, subjects behaved very differently. With their actions publicly observed, subjects gave significantly more to coethnics, expected their partners to reciprocate, and expected that they would be sanctioned for a failure to cooperate. This effect was most pronounced among individuals who were otherwise least likely to cooperate.

These results suggest that what may look like ethnic favoritism is, in fact, a set of reciprocity normsstronger among coethnics than among non-coethnicsthat make it possible for members of more homogeneous communities to take risks, invest, and cooperate without the fear of getting cheated. Such norms may be more subject to change than deeply held ethnic antipathiesa powerful finding for policymakers seeking to design social institutions and promote development in diverse societies.

Bibliography: Humphreys, Macartan. Coethnicity: Diversity and the Dilemmas of Collective Action (with J Habyarimana, D Posner, and J Weinstein). New York: Russell Sage Press, 2009.