WP44: Sources of ethnic identification in Africa

Introduction

Welcome to the Afrobarometer publications section. For short, topical analyses, try our briefing papers (for survey rounds 1-5) and dispatches (starting with Round 6). For longer, more technical analyses of policy issues, check our policy papers. Our working papers are full-length analytical pieces developed for publication in academic journals or books. You can also search the entire publications database by keyword(s), language, country, and/or author.

Filter content by:

Working papers
2004
44
Bannon, Alicia, Edward Miguel, and Daniel N. Posner

This paper explores the sources of ethnic identification in Africa. The first part draws on survey data from more than 14,000 respondents in nine African countries to investigate the factors that predispose individuals to identify themselves in ethnic terms. Contrary to popular assumptions that Africans are intrinsically "ethnic" people, we find that fewer than one-third of respondents rank their ethnic group as their most important associational membership. In strong support of modernization theory, we find the sources of ethnic identification lie in exposure to education, employment in non-traditional sectors, and political competition. We also find the salience of ethnicity to be negatively related to ethnic diversity -a result that challenges widely held assumptions in the literature.

The second part of the paper draws on a separate survey of nearly 1,200 respondents in two Kenyan market towns to explore why individuals identify themselves in terms of one dimension of identity rather than another. We find that this choice depends on the scope of the social arena the individual inhabits: respondents who are rural, less educated, and working in traditional occupations are must likely to identify with their sub-tribe than with their tribe. Taking advantage of the proximity of the survey to the 2002 Kenyan general election, we also provide suggestive evidence for the political sources of ethnic identification by showing that a respondent's identification with his or her tribe or sub-tribe can in part be explained by whether or not it will link the respondent with the old, discredited regime.