Essays on the Effects of Education Policy Reforms in Developing Countries

In three essays, this dissertation explores the short-run and long-run effects of education policy reforms in Sub-Saharan African countries. The first chapter examines the impacts of mother tongue education on schooling and child labor outcomes by exploiting a language policy reform in Ethiopia that took place in the early 1990s. The reform mandated the use of students' mother tongue languages as a medium of instruction in primary school instead of the country's national language, Amharic. Using cross-language and cross-cohort variations in exposure to the policy, I employ a difference-in-differences identification strategy to evaluate the effects of the reform. My results suggest that among non-Amharic language speakers in rural areas of the country, there was a 6 percentage points decline in enrollment and about 0.4 years of reduction in years of schooling as a result of full exposure to the reform. I also find that the reform caused a statistically significant increase in the probability of children as young as 5 years old engaging in some form of work. Boys were 2.3 percentage points more likely to participate in paid work in response to the reform, while girls were 4.1 percentage points more likely to participate in unpaid family work. Qualitative evidence suggests that poor implementation of the policy without proper training of teachers who are proficient in non-Amahric languages and inadequate supply of pedagogical materials may partly explain the unintended effects of the policy during this transition period.

In most developing countries, especially in Sub-Saharan Africa, majority of workers do not partake in the formal labor market, but instead are engaged in self-employment, unpaid family work or agricultural work. However, most studies that investigate the returns to education in these countries mainly focus on formal labor market outcomes. The second essay attempts to fill this gap by examining the relationship between increased schooling by household heads on households' consumption in Malawi. In order to obtain consistent estimates, I employ a two stages least squares (2SLS) strategy, which exploits the early 1990s free primary education (FPE) program of Malawi as an exogenous source of variation in schooling. The first stage result shows that full exposure (i.e. eight years exposure) to the FPE program led to about 2 years increase in schooling. The 2SLS estimates suggest that the value of households' weekly food consumption increased by about 13 percent due to an additional year of schooling by the household head. I also find increases in the consumption of non-food items as well as ownership of durable goods.

In a third essay, I investigate the impact that higher women's educational attainment has on fertility and child mortality, using nationally representative data sets from Malawi. In order to find consistent estimates, I use an instrumental variables (IV) strategy based on the FPE policy of Malawi. My first stage results show that full exposure to the FPE policy in high-intensity district leads to about 1 year of additional schooling. Furthermore, my IV estimates show that increased schooling due to the reform led to the postponement of first birth as well as to significant reductions in total fertility. Total fertility by age 25, for example, was reduced by almost a third of a child due to an additional year of schooling. I also find evidence that shows that delayed sexual activity, increased use of modern contraceptions and changes in spousal characteristics may help explain these negative effects of schooling on fertility. Moreover, I find significant reduction in the fraction of children who died, due to increased schooling by the mother. I attempt to provide evidence for several potential channels through which schooling might affect child mortality. For example, I show that an additional year of schooling led to 1.3 and 7 percentage points increases in the probability of having more than 12 months and 24 months gap between the first two births respectively. There is also evidence that suggests that maternal education has a positive and statistically significant effect on early childhood vaccination.