Economic mobility in South Africa: evidence from household survey data

Abstract

High levels of inequality, poverty and unemployment are some of the most substantial challenges facing post-apartheid South Africa. Most of the research addressing these questions has used micro datasets to compare snapshots of welfare over time. Although these studies are both interesting and useful, they have been unable to extend their analysis into a nationally-representative dynamic setting, due to the lack of available data. The paucity of large longitudinal datasets has also limited the number of studies of economic mobility, which allows researchers to track the welfare measures of the same individuals over time. This means that while we know a great deal about how South Africans are doing at a particular point in time, we know far less about how they are faring dynamically. Understanding how and why economic mobility happens in South Africa is therefore a question that demands attention. From both a distributive justice as well as a policy point of view, the distinction that arises when we drop the assumption of anonymity and move from a cross-sectional measure of welfare to a dynamic one is important. This is because many of the conclusions about longer-run welfare are dependent on the level of economic mobility present in society. This study contributes to the body of work on welfare in South Africa by addressing three different aspects of economic mobility. The first of these is about how a particular kind of measurement error in household surveys is best detected, and what effect its presence has on the understanding of labour market mobility. The second is about how best to model money-metric poverty dynamics in South Africa in order to better understand who escapes poverty and who enters poverty over time. The third is about how the persistence of intergenerational earnings should be calculated in a society with high unemployment, and what the role of education is in shaping these mobility dynamics.