We estimate the gender wage gap for Africans in post-apartheid South Africa over the 2001 to 2007 period. Separate male and female earnings equations yields no significant decline in the conditional wage gap, regardless of whether we correct for selection into the labour force and employment or not. Notwithstanding this, the data appear to reveal a decline in the “explained” proportion of the gap with no significant change in the “unexplained” proportion of the gap. Nevertheless, the “unexplained” proportion or discrimination accounted for 71 percent of the gap in 2007 when using the uncorrected estimates (and the male wage structure as the non-discriminatory norm) thus highlighting the presence, arguably, of substantial discrimination against African women in the post-apartheid South African labour market.

We note though that the assumption that the “unexplained” component accounts for discrimination has been criticized for a number of reasons, including the fact that women may self-select into certain types of jobs, the impact of gender-based pre-labour market factors as well as omitted variable bias. Finally, we find that using the either the male or pooled wage structure as the non-discriminatory wage structure provides similar results when undertaking the decomposition. In turn, using the female wage structure results in the harshest results as far as gender discrimination is concerned.

Written by Haroon Bhorat and Sumayya Goga, Development Policy Research Unit

Acknowledgements: The authors would like to thank Dorrit Posel for comments on earlier versions of this study.

The DPRU is a university-recognised research unit within the School of Economics at the University of Cape Town.

It conducts policy-relevant socioeconomic research in the fields of labour markets, poverty and inequality.