School quality and the performance of disadvantaged learners in South Africa

In South Africa, school quality within the public school system is heterogeneous and highly stratified along race, socio-economic status and geographic location. Because of the lingering effect of racial segregation, schools which historically served the white minority and accordingly received a much higher endowment of inputs are still out-performing schools which historically served the black population, 20 years after the end of apartheid. Under-privileged black children who select into these former white schools are typically from richer households than their counterparts who remain in the former black part of the school system, although significantly poorer than their white peers. In this paper, I use longitudinal data from the National School Effectiveness Study which collected test scores and background information on children in grades 3, 4 and 5 in both school systems in order to estimate the effect of attending a historically white school on the numeracy and literacy scores of black children. The models are estimated using a value-added approach in order to control for unobserved child-specific heterogeneity in the form of individual ability by controlling for lagged test scores. In addition, the various household covariates available in the data are used to control for household-level differences among children. I find a slightly larger effect for attending a former white school in South Africa than has previously been estimated for private schools in India and Pakistan and assess the validity of the estimates using various robustness checks. I also discuss the potential bias which may remain.

I22, I24, I25, I28, J13, O15

education, school choice, South Africa, value-added models, National School Effectiveness Study (NSES)