Public Debt Levels and Real Interest Rates: Causal Evidence from Parliamentary Elections
with Gabriel Ehrlich and Aditi Thapar
2026. Forthcoming, American Economic Review: Insights
Abstract: We use close elections in parliamentary democracies as natural experiments to estimate public debt levels’ effects on real interest rates. An election in which no party achieves a majority causes the debt-to-GDP ratio to increase by 17 percentage points over the following five years relative to an election in which one party barely secures a majority. Real interest rates rise by a relative 99 basis points following such an election, implying that a one percentage point increase in the debt-to-GDP ratio causes a 5.8 basis point increase in real rates. That effect is larger than most previous estimates.
Transmission Lowers US Generation Costs, But Generator Incentives Are Not Aligned
with Dasom Ham and Catie Hausman
2026. Proceedings of the National Academy of Sciences, 123(9).
Abstract: The US electricity grid is rapidly evolving with the entry of low-cost renewable electricity. As a result, new supply is not spatially matched to demand, and the transmission network has become more strained. Better market integration could thus lower US generation costs. We document that eliminating interregional constraints would have reduced electricity generation costs across the lower US 48 states by $5.8 to 7.1 billion in 2022 and $3.4 to 5.0 billion in 2023. But market integration creates winners and losers among generation companies, and we show that producers in some regions have incentives to delay or block grid integration despite the overall system benefits.
Time-Limited Subsidies: Optimal Taxation with Implications for Renewable Energy Subsidies
with Michael Ricks
2025. Journal of Political Economy, 133(12): 3801-3845.
Abstract: Pigouvian subsidies are efficient, but output subsidies with uncertain or limited durations are not Pigouvian. We show that optimal "time-limited" policies must also subsidize investment to correct externalities generated after the output subsidy ends. Furthermore, an output subsidy's optimal duration is characterized by the change in production when it ends. In the wind-energy industry, we find that power generation decreases by 5-10% after the end of facilities' ten-year eligibility for the Renewable Energy Production Tax Credit. This behavioral response has implications for energy transitions and highlights how time limits could cause larger distortions in more elastic industries.
Pollution Taxes and Clean Subsidies in an Open Economy
Revise and resubmit, Journal of Public Economics
Abstract: In open economies, the effectiveness of taxing a global pollutant is diminished by "pollution leakage," where some polluting activity shifts abroad as a result of the tax. This paper shows that the same economic conditions that lead to pollution leakage enhance the efficacy of clean subsidies. As a result, the optimal policy in an open economy is a combination of a pollution tax and a clean subsidy, the balance of which depends on the leakage rate. Furthermore, efficient corrective policy sets the sum of the tax and subsidy rates, a measure of policy ambition, equal to the marginal damages from pollution, and therefore does not depend on the leakage rate.
Processing Power: The Effect of Data Centers on Wholesale Electricity Markets
with Robert Reaser and Reid Taylor
Under review
Abstract: Data centers are reversing two decades of flat U.S. electricity demand. Using an hourly, unit-level dispatch model of 23 regional markets linked by constrained transmission, we estimate that data centers raised average wholesale prices 6.8 percent nationally and 16.5 percent in Northern Virginia in 2024–2025. Under our preferred scenario, the national effect reaches 13.5 percent by 2028. Trade raises prices even in regions hosting few data centers. Roughly half the increase in generator payments is a transfer from consumers to inframarginal generation. We also estimate effects on carbon emissions and derive conditions under which load growth lowers retail rates.