regard to age differences (Casterline, Williams, and McDonald, 1986:13). Although there is very little documentation of the effects of a large age gap on a young bride and it is difficult to disentangle social-structural factors that are a consequence of a large age difference from those that are determinants (Casterline, Williams, and McDonald, 1986), it is reasonable to assume that adolescent girls with much older partners are hindered in their capacity to negotiate with their spouses about sex and reproduction as well as other aspects of domestic life.
Previous analyses of survey data in developing countries indicate that the age difference between spouses is inversely related to the wife’s age at marriage, even after controlling for education (Mensch, 1986; Mensch, Bruce, and Greene, 1998). Figure 7-2a summarizes recent DHS data on age differences between spouses by age of marriage of the wife. Note that this analysis is restricted to currently married women ages 25-34 because the question on the age of one’s spouse or partner is asked only of women who are married at the time of the survey. Note also that the analysis is restricted to those in first marriages, because higher order marriages tend to take place at older ages well after the transition to adulthood has taken place. The graph confirms that women who marry prior to age 18 are more likely to have spouses who are older. It also shows that there is as much, if not more variation among regions as within regions by age. Age differences are largest in Western and Middle Africa, where polygyny is still common and teenagers often become junior wives of older men. Age differences are smallest in the former Soviet Asia.
Figure 7-2b presents parallel data from the DHS but instead plots age differences as a function of male age at marriage for the same sample of women. As Casterline, Williams, and McDonald (1986) observed nearly 20 years ago in their analysis of World Fertility Survey data, age differences vary much more as a function of male age at marriage than of female age at marriage because of the greater variance in male age at marriage. Note, however, that no new information is actually provided by this graph, since male age at marriage is equal to female age at marriage plus the age difference. Nevertheless, these findings are broadly consistent with a model in which men accept or choose from a limited age range of marriageable women irrespective of their own age. In some regions, older grooms actually marry younger brides than do younger grooms, possibly a consequence of arranged marriage.
Given that there has been an increase in age of marriage for both men and women in most regions of the world, it is not obvious what the trends are with regard to age differences. In her analysis of data from 55 of 56 African countries17 over the last 50 years, Hertrich (2002) notes a decline in