Migration and gender wage differences in Kenya

The division of labor in many traditional rural societies particularly those of Africa entails that men acquire more human capital investment relative to women. Therefore, men, both single and married, migrate to urban areas in greater numbers to take urban jobs. Differences in skills thus create an earnings gap between men and women. This study has two themes. The first essay views the household as the decision making unit as far as migration is concerned. The household is assumed to maximize a common von Newman Morgenston utility function defined over three time periods. The model explains the characteristics which influence the family to migrate together as a unit (joint migration) and those which motivate one member--typically the husband in his capacity as the household head--to move first and the rest of the family to remain in the rural area, to join the migrant in the urban area at some future time period (sequential migration). The imbalance in marketable skill levels results in men commanding a higher urban wage relative to women, leading to the second essay, an examination of the gender earnings differences in the urban areas. The findings strongly support the theoretical model. Namely that family migration patterns can be explained very well within the context of an inter-temporal expected utility maximizing framework of the household in the case of married couples. The evidence suggests that the uncertainty related to migration and the wife's household production opportunities are the key motivations for sequential migration. Age, family size, and the level of education of the husband influence joint migration. The widely used Oaxaca model, and the recently developed Cotton/Newmark technique are used to derive estimates of gender discrimination in earnings. The results suggest that a considerable earnings gap exists between male and female wages, and that much of it is directly attributable to discrimination. The discriminatory component of the earnings gap appears to be considerably greater than the portion of the earnings differentials that can be explained by differences in productivity characteristics of males and females.