This PhD thesis focuses on determinants of labour market outcomes in development economics with a special interest in South Africa and China.

After an introduction in chapter 1, the key chapter 2, Ethnic Diversity and Labour Market Outcomes: Evidence from Post-Apartheid South Africa joint with Sara Tonini, investigates how ethnic diversity amongst black South Africans affects their employment opportunities in the post-Apartheid era. We find that ethnic diversity has a positive impact on the employment rate of the black South Africans, and it only affects ethnic groups with relatively large population size. To address the endogeneity of ethnic composition, we explore the location of historical “black homelands” and argue that districts more equally distant to multiple homelands are more ethnically diverse. In our instrumental variable regressions, a one standard deviation increase in ethnic diversity index increases employment rate by 3 (5) percentage point in 1996 (2001), which is around 8% (13%) of the average employment rate. We then propose a model of a coordination game to explain these findings. A more ethnically diverse place requires a higher rate of inter-ethnic communication to maintain social connection. As inter-ethnic communication requires more skills than intra-ethnic connection, people in ethnically diverse districts are motivated to invest more in social skills to be able to communicate with those outside their own group. The acquisition of these social skills makes them better equipped for the labour market.

The remaining two chapters look into the intergenerational transmission of socio-economic status in South Africa and China. Chapter 3, Returns to Education, Marital Sorting and Family Background in South Africa joint with Patrizio Piraino, applies the model of Lam (1993, JPE) which combines intergenerational transmission of ability and assortative mating to investigate the relative explanatory power of father-in-law’s and father’s background for male wages. In the empirical analysis, after correcting for potential measurement errors in earnings and education, we find that father-in-law’s schooling is more correlated with male workers’ labour market earnings, employment rate and labour force participation than own father’s schooling in contemporary South Africa. This difference is more obvious when parental educational levels are higher.

Chapter 4, Higher Education Expansion and Intergenerational Mobility in Contemporary China, studies how higher education affects the upward mobility of people from relatively disadvantaged families. Intergenerational occupational mobility is stimulated when children from different social classes end up in similar occupations. Whether or not they have similar occupational status depends not only on their level of education but also the occupational returns to education. Given there is already a convergence in educational achievements between children from different social classes in contemporary China, in this paper, I focus on their occupational returns to education. Occupational status is measured by the widely-accepted ISEI scaling system ranging from 16 to 90 points with large number indicating higher occupational status. I take advantage of an exogenous college expansion policy in 1999 as a natural experiment and find that one additional year of education increases the occupational status of their first job by 2.243 (2.774) points on average along the ISEI scale in OLS (IV) regressions. And children from upper-class families do not necessarily have higher returns to education than children from other social classes. The average occupational returns to education are higher for the most recent job than the first job, but the difference among social classes is still not significant.