Distributional changes in the gender wage gap in the post-apartheid South African labour market

Gender inequality in the labour market is real and it matters. This thesis examines gender wage inequality in the post-apartheid South African labour market from 1993 to 2015 and finds that there is a substantial median wage gap of between 35% and 23% since the end of apartheid. This gap is unexplained by differences in human capital characteristics and it is not declining over time. There has however, been a substantial decline in the gender wage gap at the bottom of the wage distribution from 0.47 log points (about 60%) in 1997 to 0.18 log points in 2010 (about 19%). Similarly, the wage gap at the mean narrowed during the period studied, from 0.34 log points (about 40%) in 1993 to 0.15 log points (about 16%) in 2014. An examination of the gender wage gap is important because in the wake of a new democratic government in 1994 and the introduction of anti-discrimination and affirmative action policies, we expect that the gender wage gap should have narrowed over time. Interestingly, at the median where we would expect the legislation to have been binding, the gender wage gap has stagnated. We find however, that minimum wage legislation in low wage industries (agriculture and domestic work), has had some effect of narrowing the wage gap at the bottom of the distribution. This is especially interesting because this type of wage legislation was not specifically targeted at narrowing the gender wage gap. For this analysis we utilize the Post-Apartheid Labour Market Series (PALMS) dataset from 1993-2015. To do this, we need to address several data quality issues involving how household survey data has been constructed over time. One problem we need to address is that measurement has changed over time. In chapter two of this thesis we show how measurement issues affecting data on domestic work in 1994 and 1995 led to mixed results on the gender wage gap with some researchers finding a rise in the wage gap between 1995 and 2006 and others reporting a drop depending on the choice of the base period. This analysis goes beyond pointing out the classification issues in 1994 and 1995 and takes steps to update the data and address the “missing” raw gender wage gap in these two years. In chapter three, we analyse the trends in the gender wage gap in detail first starting with a mean decomposition using the Oaxaca decomposition method. We then extend the analysis to the entire wage distribution using the Dinardo Fortin and Lemieux (DFL) re-weighting method (DiNardo et al. 1996) for the aggregate decomposition and the Re-centred Influence Functions (RIF) method (Firpo et al. 2009) for the detailed decomposition. Results show that the changes in the gender wage gap are heterogeneous across the wage distribution. There has been a substantial narrowing of the gender wage gap at the bottom of the wage distribution which we attribute to improved female human capital characteristics and minimum wage legislation. On the contrary the median wage gap which is greater than the mean wage gap has been stagnant and displays very little movement in the period studied. There was some decline in the gap at the 90th percentile in the period between 1993 and 2005 but the gap seems to be expanding in recent years due to a continually expanding unexplained component of the gender wage gap. Given the snap shot nature of cross-sectional data, results from cross sectional analysis confound life cycle, generational and period effects on the gender wage gap. The fourth chapter of this thesis involves constructing synthetic cohort data from repeated cross-sections which we use to examine dynamic aspects of the gender wage gap. Life cycle trends show that younger cohorts of men and women have experienced a rise in wages over time which we attribute to improved human capital characteristics. Younger cohorts of women have on average more education than the cohorts before them and have more education than men from the same cohort. These cohorts of women are more likely to be in a skilled profession compared to women born 30 years before them and who joined the labour market during the apartheid era. Marriage rates and trade union rates have greatly declined for recent cohorts. These cohort changes mean that more recent cohorts of women have similar labour market or better labour market characteristics than men from the same cohort. This has led to the narrowing of the gender wage gap at the mean evident in the cross-sectional analysis. The general conclusion from this analysis is that the gender wage gap persists in the South African labour market and that the experience of women at the top end of the wage distribution is different from the experience of women at the bottom end or at the median. Therefore, policies towards narrowing the gender wage gap will need to be unique to challenges women face in different parts of the wage distribution. For example, while raising the minimum wage at the bottom of the wage distribution will narrow the gender wage gap in this part of the wage distribution, narrowing the gender wage gap at the top end of the wage distribution should focus on increasing the number of women in top paying occupations. This will require among other things alleviating the disproportionate burden of care work (for example availing child care and providing creche facilities) shouldered by women to enable them to commit more time to the labour market. That non-gender specific wage legislation worked to narrow the gender wage gap at the bottom of the wage distribution is an indication that addressing the gender wage gap may require more than labour market legislation aimed at reducing gender discrimination. It may require a broader look at society and social norms that may have or may not have shifted over time