The Rise in Union Wage Premia in South Africa

The change in the premia earned by African, urban, male, union members, relative to earnings of other, comparable workers in South Africa, between 1985 and 1993, is examined using two independent, national, sample surveys. The interval from 1985 to 1993 was part of a period of rapidly rising union membership for African workers, following the Industrial Conciliation Amendment Act of 1979 which extended the definition of an employee to encompass African workers for collective bargaining. Membership now reported by the unions suggests one of the highest rates of unionization of wage employees among the developing countries. Meanwhile, the simple average earnings of African employees has risen substantially, in real terms, despite extremely high rates of unemployment. In this context, our sample data show that the premia in mean earnings of African urban, male, union members relative to other African, urban male regular employees rose 18.5 percent from 1985 to 1993. Clearly this rise could reflect changes in composition of regular employees and of union membership in particular during this episode of rapidly rising membership. Four apporaches to estimation of the union premia are applied to examine this possibility: earnings equations whith a union dummy; separate earnings regimes for members and non-members; a switching model with separate earnings regimes and endogenous membership; and a multinomial logit model which extends the standard switching model to encompass sample selection into employment as well as into regular employment among the employed. A series of nesting hypotheses are conducted to explore the significance of differences between these approaches. The levels of estimated premia differ widely across the various approaches. However, there is uniform agreement that the union earnings premia, based on either the mean observed characteristics of union or non-union members, had risen by 1993 and that the change in measured characteristics of union and non-union members between sample years actually tended to diminish the premium. This suggests that the observed rise in union premium must either reflect an increase in rents to union members as compared to identical non-member workers or some change in the composition of union members which we are unable to detect in our data.