High School Dropout and Youth Labour Market outcomes in Sub-Saharan Africa: Evidence from South Africa

Chapter 2: Youth unemployment and high school dropout: The second chapter of this thesis provides a general description of the nature of high school dropout and youth unemployment in South Africa. This chapter is set as a motivational chapter for subsequent chapters. I defined a high school dropout in the data as an individual who was enrolled in high school at the beginning of the survey but failed to complete the last year of high school education (i.e., grade 12 in South Africa) before terminating their enrolment. A preliminary description of the data shows a high dropout rate of about 50% on the average. Upon estimation, I find that individuals who dropped out of high school have higher continuous spells of unemployment. In addition, I identify the following individuals as those more that are more vulnerable to high school dropout: black students, young students, students with record of poor academic performance, students who live close to school, and students with less educated fathers. On the other hand, factors that are associated with successful transition from school to employment include good academic performance, being white or coloured, age, and having an educated father. Chapter 3: Assessing the Patterns of Self-selection into High School Drop-out and Graduation. The third chapter of my thesis focuses primarily on assessing the patterns of self-selection into high school dropout and high school graduation. Specifically, I adopt a flexible framework that allows me to quantify selection on unobservables between dropouts and graduates and establish to what extent this selection matters. To do this, I write down a flexible discrete-time conditional competing risks model that allows for a general correlated risks across destinations. I treat individual differences that may exist between dropouts and graduates as both observed and unobserved. I then use the structure of my model to estimate a reducedform measure of self-selection into dropout and graduation. As a reduced-form measure of self-selection, I propose to consider the correlation coefficient of the bivariate distribution of unobserved heterogeneity across two possible exit states: high school dropout and high school graduation. I make no parametric assumption about the distribution of unobserved individual heterogeneity. Rather, I used a non-parametric approach, which relies on the data in estimating the distribution of unobserved differences. Results suggest that individuals who choose to drop out of high school are not systematically different from those who complete high school education in terms of unobservables. This runs contrary to typical findings in developed countries, which suggest that high school dropouts are individuals with relatively low ability, low expectations, and a set of negative preferences. Chapter 4: High School Dropout and its Wage Consequences The fourth chapter of my thesis focuses on the wage consequences of dropping out of high school. Specifically, I am interested in the severity and time pattern of the wage disadvantage dropouts face in the labour market, if any. This knowledge is important because lack of information or incomplete information about the worth of a high school diploma in the labour market could influence the decision to drop out of high school. Prior to estimating the wage equation, I check for self-selection both in the decision to drop out of high school and in the decision to participate in the labour market. Results from IV estimations show that OLS estimates are upward biased, even after controlling for measures of ability. Also, I find that dropouts earn less in the first year of labour market experience, and their wages progressively declined in subsequent years. This suggests that high school diploma may act as a signal of productivity, both immediately after graduation and in subsequent years.