Desires for an additional child in developing countries: the role of son preference

The paper considers role of son preference in formation of desires to have an additional child in countries of South and Central Asia, Middle East, North Africa and Balkans. Although son preference has been shown to influence actual fertility and to be present in fertility ideals in many of those countries, no systematic cross-country study of its influence on desires for the next child yet exists. Discovering the role of son preference in shaping these desires is important because they are highly predictive for actual fertility behavior. At the same time, there are reasons to assume that for desires concerning the next child son preference effects can be observed in a less distorted way than they are observable for actual parity progressions. Using Demographic and Health Surveys administered between 2010 and 2020, the relationship between number of living sons to desires concerning the next child was studied with the help of linear probability models for women from 13 countries of the selected regions. The analysis was applied to women having at least one living child and gave three main results. First, the desire to have another child showed a regular negative relation to having at least one living son. Second, son preference also regularly affected desires concerning timing of the next child bearing, as the desire to have the next child within the nearest two years was more probable among women without a living son in most of the countries. The third result, unlike the first two, has indicated considerable cross-country differences. In some countries desires for the next child were weaker associated with son preference in those groups of women in whose families less pronounced gender asymmetries were expected (urban women, highly educated women and women employed outside their household). In other countries, however, effects of son preference did not differ significantly across these groups, contrary to expectation. Possible explanations of these cross-country differences are discussed.