2026 Theses Doctoral

Essays in Climate and Environmental Economics

Hogan, Dylan

This dissertation comprises three essays in climate and environmental economics. The first two chapters study how climate change and conservation policy shape the natural resources that underpin agricultural production and biodiversity. The third chapter examines data quality challenges in the empirical climate-economy literature. Together, these essays combine climate data, satellite remote sensing, and other datasets with econometric methods to provide evidence on the economic and environmental consequences of climate change and the effectiveness of environmental policy.

The first chapter, The Global Freshwater Resource Costs of Climate Change Accounting for Agricultural Land Use, with Tamma Carleton, Jonathan Proctor, Ram Fishman, and Maren Ludwig, provides the first globally comprehensive and empirically derived estimates of the impacts of climate change on local water availability, accounting for the role of agricultural land use. We leverage satellite-based measures of total water storage from the GRACE mission alongside global data on weather, agricultural land use, and vegetation. We find that extreme heat has large and statistically significant effects on water storage, with locations irrigated from groundwater losing approximately 55% more water on hot days than locations with rainfed agriculture. In rainfed landscapes, temperature-driven water losses attenuate within six months as natural recharge processes restore storage; in irrigated landscapes, losses persist as groundwater withdrawals are not replenished within this time horizon. Over the last 20 years, climate trends have increased the share of arable land losing water by 4 percentage points globally. Looking forward, future warming will roughly double the rate of water loss by mid-century. We estimate that warming will reduce the global value of irrigation water by $6.5 billion per year by mid-century, with cumulative losses of $73.4 billion, and that the economic burden falls disproportionately on regions with high agricultural productivity and deep groundwater.

The second chapter, The Efficacy of Conservation Expenditures: Evidence from Local Ballot Measures, with Eyal G. Frank, Joséphine Gantois, and Anouch Missirian, estimates the biodiversity benefits from local conservation spending in the United States. We study over 1,600 local ballot referenda that allocate funds toward conservation goals between 2003 and 2023, implementing a dynamic regression discontinuity design that compares jurisdictions where ballots narrowly passed to those where they narrowly failed. We find that conservation spending modestly increases bird abundance, with an increase of $25 million in conservation expenditures leading to approximately 0.2 standard deviation higher local bird abundance. Green vegetation increases by five percent following successful ballots financed through tax measures. The effects are concentrated in New Jersey, which accounts for roughly 20 percent of all ballots and has a state-wide program that provides institutional support and matching funds for local conservation. Our findings highlight that conservation spending can produce detectable biodiversity gains, but that institutional context is important for determining effectiveness.

The third chapter, Data Anomalies and the Economic Commitment of Climate Change, with Tom Bearpark and Solomon Hsiang, shows that data anomalies in a single country, Uzbekistan, substantially bias the projected climate damages reported in Kotz, Levermann and Wenz (Nature, 2024). The original study estimates that global GDP would be lowered by roughly 62% by 2100, an impact three times larger than similar prior estimates. We show that removing the anomalous Uzbekistan data reduces projected losses to 23%, aligning closely with previous literature. Additionally, correcting the statistical uncertainty to account for spatial correlations yields a 90% uncertainty interval of -68% to +121%, indicating that the original results are not statistically distinguishable from zero at any time this century.