Existing literature is in agreement on the importance of education in the determination of labor
market outcomes. Using data from South Africa’s National Income Dynamics Study, this thesis
explores this relationship. It does this firstly, by examining the effect of school quality measured
by school inputs–pupil-teacher ratio and expenditure per pupil–on educational attainment in South
Africa. Following a reduced form production function approach, a partial generalized ordered
probit is applied in the analysis. The thesis finds that both pupil-teacher ratio and expenditure
per pupil have strong and significant effects on educational attainment of African South Africans.
The ratio is more important at lower schooling levels, indicating a ‘lagged effect’ on educational
attainment. A small increase in expenditure has large effects. This suggests declining returns to
fiscal investment in education. Alternatively, it could indicate inefficiency in the use of funds, or
compensatory funding of poor schools. Secondly, we examine the extent to which wage differences
shown among district councils in South Africa can be explained by the magnitude of external
returns to education. We use an augmented Mincerian regression to investigate this, considering
the effect of district council share of college graduates on workers’ wages. The study employs
District Council’s annual average climate to instrument for the share of district council college
graduates. The results show that a 1% increase in a District Council’s share of college graduates
raises workers’ wages by 5-8%. There are also spillovers effects, with college graduates being the
beneficiaries. Thirdly, we estimate unemployment duration by gender and by competing risk, that
is, exits into employment or economic inactivity. We use the Kaplan-Meier estimator and the Cox
proportion hazard model in these estimations. The results from both estimators show factors that
influence unemployment differ by exits, and their effects vary by gender. The hazard rates show
that transition rate into employment is higher for men than for women with similar characteristics.
They show that age and race significantly influence employers’ choice between educated men and
women. However, this bias is less obvious at higher levels of education.