One of the most common generalizations concerning refugee populations is that they are dominated by female heads of households and children. It is claimed that men are either killed in the wars that prompt displacement or are left behind to fight. This assumption has continued to determine the policies of relief and development agencies, as well as governments in countries of asylum and return. On the basis of empirical data from UNHCR and household data from Eritrea, this article questions the validity of such a dominant assumption. The article also problematizes the concept of household headship by showing that it is a cultural construction whose meaning varies from one cultural context to another. There is, thus, no definition of headship that can apply cross-culturally. It also argues that since female heads of households (FHHs) are not socially and economically homogeneous, household headship is not an appropriate method of identifying the poorest of the poor for targeting or provision of emergency relief or for productive inputs in development programmes.
DISCOURSES ON REFUGEE AND RETURNEE HEADSHIP STRUCTURESKnowledge about returnee populations, including their demographic structures, is very much influenced by the discourses that have influenced research and practice relating to refugees and other displaced populations. Generally, any mention of refugee populations in developing countries tends to evoke a number of stereotypical assumptions. One of these is the alleged preponderance of female headship (see, for example, UNESC, 1991). Statements which postulate that women and children under five years typically represent up to 70 to 80 per cent of refugee populations permeate the literature. Up to 75 per cent of the total refugee population in the world is estimated to consist of women and girls, and 60 to 80 per cent of all refugee heads of households worldwide are thought to be women (Cohen, 1995; A debt of gratitude is due to the Swedish Authority for Research Cooperation with Developing Countries (SAREC/SIDA) for generously funding the research project on which this article is based. Thanks are also due to Joke Schrijvers, Laura Hamond and the anonymous referees of this journal for their constructive comments on an earlier draft.Development and Change 34(2): 311±337 (