UC Berkeley

This dissertation combines three studies from East Africa, each investigating a different way in which individuals respond to institutional changes, and how this response shapes the broader impact of these policy levers. Each paper develops and explains one or more theories of individual behavior and combines data from a wide range of sources to test hypotheses generated from the original theories.

The first chapter measures the impact of provision of ARVs on risky sexual behavior among young women in Kenya, and estimates how this response will shape future HIV infection rates. Access to antiretroviral (ARV) drugs in Sub-Saharan Africa has rapidly expanded - from fewer than 10,000 people treated in 2000 to more than 8 million in 2011. To measure the impact of this expansion, it is necessary to identify the behavioral response of individuals to drug access. This chapter combines geocoded information about the timing of introduc- tion of ARVs in all Kenyan health facilities with two waves of geocoded population surveys to estimate the impact of proximity to an ARV provider on risky sexual behavior. Using a difference in differences strategy that matches survey clusters geographically across waves, I find a relative increase in risky behavior as reflected in pregnancy rates (increase of 82%) and self-reported recent sexual activity (increase of 40%) among young women in areas in which ARVs were introduced between 2004 and 2008. The full impact of ARV access on new infections is estimated through a simulation procedure that combines estimated behavioral responses to ARVs with medical evidence regarding HIV transmission. An increase in ARV drug access is predicted to reduce the rate of new infections despite the induced increase in risk-taking.

The second chapter looks at the impact of corruption on the effectiveness of antiretroviral drugs in preventing deaths due to HIV and the potential channels that generate this relationship. This is based on a unique panel dataset of countries in sub-Saharan Africa, which combines information on all imported antiretroviral drugs from the World Health Organiza- tion's Global Price Reporting Mechanism with measures of corruption and estimates of the HIV prevalence and the number of deaths in each year and in each country. Countries with higher levels of corruption experience a significantly smaller drop in HIV deaths as a result of the same quantity of ARVs imported. This is followed up with a single case-study from Kenya to illustrate one potential mechanism for the observed effect, demonstrating that disproportionately more clinics begin distributing ARVs in areas that are predominantly represented by the new leader's ethnic group.

The third chapter uses new data on participation to examine how local economic conditions shaped within-country variation in willingness to participate in violent activities during the Rwandan genocide. It discusses and tests the predictions of three sets of theories about the causes of violence. The data provide strong evidence that higher rates of both unemployment and education among Hutu are associated with increased participation. I find no evidence that the employment or education of the Tutsi population reduce participation rates. I also find suggestive evidence of a positive association between violence and the interaction of Hutu unemployment and education both at the commune level and at the individual level. These results are consistent with theories of opportunity costs discouraging violence, and they provide additional evidence of a connection between education, unemployment, and violence.