Three Essays in Development Economics

Abstract

This dissertation consists of three chapters. They study two topics in the context of a developing economy: how households make health-related decisions and how firms make use of political connections. Chapter One aims to understand the role of social stigma in the HIV epidemic. Public health experts have seen the stigma as a leading barrier affecting the delivery of HIV-related health care. This chapter uses a field experiment in Mozambique to tackle this issue. To obtain local measures of the HIV stigma environment in the study sites, I conducted a baseline survey one year before the experiment. Experiment participants with excessive concerns, defined as overestimating the stigma in their communities, were randomly assigned an intervention to relieve stigma concerns. The intervention, which drew upon findings from the baseline survey, was designed to reveal the correct degree of stigma that a participant had overestimated. Analyses show that this intervention raised the HIV test uptake rate by 7.7 percentage points (or by 37 percent) from 20.7 percent under the control condition. To quantify the intervention effect, I introduced testing coupons of different values to estimate the demand curve for an HIV test. The concern-relieving intervention raised an individual’s willingness-to-pay for an HIV test by $1.30 or more than half of the daily cost-of-living in the study population. Chapter Two evaluates a prominent effort to help families cope with HIV/AIDS: a U.S. government program in Mozambique, “Strengthening Communities and Children” (or Portuguese abbreviation, FCC), that implements home visits alongside a set of complementary interventions. This chapter focuses on the primary outcome of HIV testing, and two key mechanisms: improvements in HIV-related knowledge, and reductions in HIV-related stigmatizing attitudes. Causal identification exploits multiple levels of random assignment, most prominently of entire communities to FCC program receipt or to a control group. We find that the FCC program has positive but small effects on HIV testing. Treatment effects are only one-fifth the magnitude of, and statistically significantly smaller than, the average of expert predictions elicited in advance. Likely mechanisms behind these modest effects are that the program worsened some aspects of households’ HIV-related knowledge, and also worsened HIV-related stigmatizing attitudes. Additional treatments randomly assigned at the household level during our follow-up survey further highlight the role of these mechanisms: treatments improving knowledge and alleviating stigma concerns raise the impact of the FCC program on HIV testing. Chapter Three focuses on a new context and studies the value of pollical connections in China. Inviting a government official to sit on the board is a commonly used strategy for firms seeking to become politically connected. This chapter estimates the value of this type of political connection with a nationwide, targeted policy shock in China. In October 2013, the central government announced a new policy that restricted government officials from working in firms. Firms with government-official outside directors were affected. I find that government-official outside directors do add to firm value: The stock return of affected firms is, on average, eight percentage points lower than that of the control firms in the 12 months following the policy change. The variation in treatment effects across firms suggests that firms rely on this type of political connection to different degrees. Several potential working mechanisms are explored.