Essays in Empirical Development and Education Economics

This dissertation includes three empirical essays on development economics and one on the economics of education. The first essay (co-authored by Stephan Klasen) is a contribution to the debate surrounding the Millennium Development Goals (MDGs) and the on-going debate about what international development goals should be adopted in the future. The main contribution of this paper is to assess whether the newly proposed child mortality goal, which calls for a reduction in the child mortality rate to a minimum level of 20 deaths per 1,000 live births by 2030, is feasible. We use dynamic panel data methods which allow for a tendency of rates of reduction to revert to a long-run mean. Since African countries were benefiting from an unusually favorable set of circumstances in recent years, assessments based on extrapolating from recent reductions result in predictions that are overly optimistic. Hence, the new goal would again set up African countries for failure. At the same time, we find that a ‘re-newed’ MDG4, which calls for a reduction in child mortality by two-thirds within 25 years, seems more realistic today. A relative target would also maintain the advantage of being relevant to countries that have attained a comparatively low level of under-five mortality already today. The second essay (also co-authored by Stephan Klasen) investigates the potential for proxy means-tests (PMTs) in Bolivia. PMTs are one way of overcoming problems with data availability in developing countries when the goal is to target social programs to the poor. Results show that the proportion of poor that receive the program is low when the program is narrowly targeted to the poorest but much higher when PMTs are used to target a broad share of the population. Despite these problems in identifying the poorest households, PMT-based transfers have a large poverty effect only when targeted narrowly to the poorest. This points to a trade-off that has received little attention in the literature so far: high accuracy in targeting does not automatically translate into a large effect on poverty. The third essay (co-authored by Malte Reimers) investigates whether rural households in developing countries sell livestock in response to economic hardship. As previous empirical studies have been inconclusive, we re-investigate this question expoliting two large panel datasets on rural households that cover Burkina’s 2004 drought. These are combined with data on annual rainfall at the province-level. Our main argument is that conflicting results in past studies can be explained by differences in empirical approach: while some authors focus on the number of livestock sold, others focus on changes in the value of net sales. We show that there is unambiguous evidence for an increase in sales following adverse shocks. However, we also show that prices decline as many households resort to sales at the same time. These two effects tend to offset each other so that the net value of sales does not vary significantly with rainfall. In addition, we document a large role for savings in the form of grain storage. Nevertheless, we also find evidence that households' consumption expenditure varies positively with transitory income. The fourth essay (co-authored by Marten von Werder) analyzes the causal effects of tracking, the practice of grouping students by ability, on educational outcomes and the intergenerational transmission of education. While proponents of tracking argue that the practice increases efficiency in educational production, opponents point out that tracking potentially aggravates initial differences between students and that miss-classification is often rife. We find that de-tracking in Germany during the 1970s had no effect on educational outcomes on average. We document a sizeable effect of this reform on the inter-generational transmission of education for males yet not for females. As a consequence of the reform, the gap in years of education between males with educated and those with uneducated parents decreased by about 1.2 years, roughly 50 percent of the initial gap. While years of education increased by about 0.6 years for male students with uneducated benefits, they decreased by about the same amount for male students with educated parents. These results are robust to various empirical specifications and alternative samples.