Measuring earnings inequality in South Africa using household survey and administrative tax microdata

Overall income inequality in South Africa is very high, and inequality generated in the labour market is a key driver of inequality. In this paper, I use the Post-Apartheid Labour Market Series, the General Household Surveys, and administrative tax microdata to describe earnings inequality in South Africa. I estimate Gini coefficients, the variance of log earnings, and various percentile ratios to document changes in earnings inequality. I show that earnings inequality estimates from the Quarterly Labour Force Surveys are unreliable, most likely as a result of the earnings imputations in the publicly available data from Statistics South Africa. I also use the tax microdata to document the contributions of within- and between-firm differences to overall earnings inequality.