Abstract
This dissertation studies how policy and property-rights institutions shape the allocation of production and natural resources, with consequences for environmental outcomes and aggregate efficiency. Across three essays, I use microdata and quasi-experimental designs to quantify how regulatory authority, ownership structure, and enforcement regimes reallocate activity across economic units and space, which could generate unintended spillovers that standard place-based evaluations miss.
Chapter 1 examines within-sector industrial subsidies in China and asks how uneven subsidy incidence across firm types affects the allocation of capital and production. Leveraging variation in provincially promoted sectors as an instrument for subsidy exposure, I find that subsidies expand firm scale, raising revenue and capital, with heterogeneous responses across ownership and geography. The preferential treatment of state-owned enterprises (SOEs) highlights that targeted support can shift resources within narrowly defined industries toward specific firm groups, raising the possibility of allocative-efficiency costs when transfers disproportionately expand inputs in SOEs with lower returns.
Chapter 2 tests the Pollution Haven Effect (PHE) within a single country by studying China’s Key Cities Air Pollution Control (KCAPC) program. Using a synthetic difference-in-differences (SDID) design to address targeted policy placement, I estimate effects on sulfur dioxide (SO2) emissions and industrial composition in treated cities and their non-treated neighbors. I find limited net reductions in SO2 but strong evidence of spatial reallocation: treated cities shift production toward cleaner sectors, while neighboring cities expand pollution-intensive output and capital, consistent with leakage. A further analysis on the underlying mechanisms indicates that province-level coordination through state-owned enterprises is a primary channel, complemented by firm-level product switching and firm entry/exit.
Chapter 3 is a joint work with Peter Debaere, Tianshu Li, Swatantra Sharma, and Mesfin Mekonnen. Using microdata on diversion structures along the Colorado River, we study the allocative efficiency of the prior-appropriation system, which prioritizes access to water by seniority. We focus on "marginal users", who withdraw in excess of their rights during periods when enforcement is effectively suspended. Using a probit framework with instrumental variables that isolate plausibly exogenous variation in agricultural productivity arising from differential exposure to binding scarcity (i.e., periods when excess withdrawals are not feasible), we show that the most secure rights are unlikely to be excess users. At the same time, the raw positive correlation between excess use and productivity suggests that the prevailing allocation is not fully efficient, leaving room for efficiency improvements.
Together, these essays show that institutional details governing targeting and enforcement can reallocate activity across firms and jurisdictions, producing leakage, distributional consequences, and efficiency losses that are central to policy evaluation.
