Essays in Development and Public Economics

Abstract

This dissertation contains three papers that together engage with the impact agenda in the economics of human development. Chapter 1 is an effort to conduct meta-research on how causal evidence is sourced and used. Particularly in public finance, there have been few opportunities to translate program evaluation evidence from impact evaluations into welfare comparisons. I use established social cost-benefit frameworks to develop general criteria for welfare comparisons, and take advantage of previously missed opportunities to conduct meta-research. I conduct a comparative welfare analysis of 40 policies implemented in low and middle-income countries since 1997, which were previously evaluated using experimental and quasi-experimental methods. In all geographic regions, marginal value of public funds/cost-benefit ratio point estimates are clustered between approximately 0 and 2. I also offer partial adaptations to the canonical MVPF approach to account for the particular public finance position of developing countries, and provide evidence on the fitness-for-purpose of current priorities in impact evaluations. Chapter 2 applies data science methods to questions in global health economics. This paper rigorously plots the relationship between health system coverage in India and the goals of a health system, most directly population access to key health services. This project combines a wide range of publicly available data, in many cases at village level, in order to create a novel dataset of health system coverage in two states. This enables me to test for correlations with health outcomes at a more granular geographic level than has been possible in literature to date. I combine border discontinuities between select states - exploiting the 2014 division of Andhra Pradesh state and creation of Telangana - with estimates of the physical distance to health facilities and concentration of the health workforce. This study provides evidence that greater distance to care is not reliably associated with lower uptake of services, although the exact findings vary by type of facility. Furthermore, I use census and geographic data to descriptively show the extent of undercoverage in a large proportion of subdistricts and villages, pointing to potential misallocation of health services in rural areas.

Chapter 3 also focuses on health systems in India, using a mixed-methods case study approach. This paper draws on an array of public data to map changes in different components of health system strengthening processes and outcomes in two Indian states, during and after the division of Andhra Pradesh and creation of Telangana state. The findings paint a consistent picture of incremental health system development in Telangana in the decade since it was created, with strong performance against national indices. It appears from both the qualitative and quantitative analysis that new statehood has not spurred radically different policy, but has also not held Telangana and AP back from consistent health system performance.