RESEARCH
Works in Progress
+/- Abstract
When voters approve a regulation that raises the price of a good they regularly buy, their ballots reveal something their shopping carts cannot: how much they value the policy's public benefits, net of its private costs. I develop a structural framework to measure these publicgood preferences from referendum data and to decompose voter willingness to pay into use and non-use components using only market and voting behavior-no stated-preference surveys required. The identification rests on a simple equilibrium observation: market purchases reveal Nash-equilibrium private valuation, while referendum votes reveal Lindahl-equilibrium total valuation, and the gap is exactly non-use value regardless of preference non-separabilities. Applying the framework to California's Proposition 12 (2018)-a ballot initiative mandating cage-free housing for laying hens, approved by 63% of voters despite raising egg prices-I estimate consumers' marginal willingness to pay for the cage-free attribute from a logit demand model on retail scanner data, then recover the voter WTP distribution from a structural voting model at the precinct level. A supply-side instrument for local cage-free prices-retailer chain assortment decisions that shift local cage-free availability-corrects for the endogeneity of the policy's consumer cost. I find that over 85% of the policy's total value to voters-roughly $17 out of $20 per capita per year-reflects public-good benefits that private markets systematically underprovide. The full-population WTP distribution predicts that 64% of Californians had positive net valuations for the policy, closely matching the actual 63% Yes vote. The framework generalizes to any referendum in which the voter's cost is a commodity price change rather than a tax, and speaks directly to the long-standing pocketbook-versus-sociotropic debate in the political economy of voting.
(with Zach Freitas-Groff and Carl Meyer)
[Revision Requested: European Economic Review]
+/- Abstract
Over the past two decades, U.S. meat consumption has remained high while purchases of plant-based alternatives have grown, raising questions about whether consumer behavior is shifting. Answering this question is difficult: surveys overstate meat avoidance due to social-desirability bias, and aggregated data obscure heterogeneity across households and regions. To address these challenges, we use a nationally representative household panel (2004–2020) linked with ingredient-level product data and develop a machine learning–based classification of grocery purchases. We show that, despite modest growth in aggregate meat purchases, the share of households buying no meat rises by about 10\% and the share buying no animal products nearly doubles--revealing growing polarization in dietary behavior. These patterns predate the introduction of modern plant-based meat alternatives, whose limited market share cannot explain the observed changes. Demographic analyses indicate that growing meat- and animal product-avoidance is driven largely by population turnover rather than behavioral change within existing consumers. Our findings reconcile persistently high aggregate meat consumption with the increasing visibility of meat avoidance.
Media coverage: Vox
Wildfire Zoning and Heterogeneous Responses to Wildfire Events: Evidence from the California Housing Market
(with Matthew Wibbenmeyer)
+/- Abstract
Do wildfires alter local demographics? Employing event study and difference-in-differences strategies, we study the effect of 192 large wildfire events (2016-2018) on nearby house transaction prices and quantities to identify heterogeneity in real estate market responses across Fire Hazard Severity Zones (FHSZ) mapped by the state in 2011. We find that post-wildfire sellers in "very high" FHSZ have significantly lower incomes than before, whereas post-wildfire sellers outside of any FHSZ (non-FHSZ) have the same income levels as before. Meanwhile, post-wildfire buyers have lower incomes than before regardless of zone. Controlling for house characteristics, home prices in "very high" FHSZ are unaffected by a nearby fire, however the types of homes sold in these areas after a fire are of a significantly lower quality relative to before. In contrast, home prices in non-FHSZ areas decrease after a nearby fire by more than 14 percent and remain low for at least a year, but the quality of homes sold in these areas remains the same as before. We rationalize these findings by calibrating a model of seller/buyer behavior that incorporates decreasing relative risk aversion and treats wildfires as information shocks.
No Ethical Consumption Under General Equilibrium? Evidence from the US Meat Market
(with Christoph Semken)
+/- Abstract
Many consumers engage in ethical consumption, altering their consumption choices so as to mitigate their personal contributions to negative externalities such as climate change, animal welfare, or pollution. But do individual consumption choice have any impact on overall externalities in a competitive market? In this paper, we estimate the effect of consumption changes on externalities in general equilibrium for the U.S. meat market. We derive sufficient statistics based on the demand and supply cross-price elasticities of all goods with externalities. We estimate the cross-price elasticities using home scanner data and farmer surveys, respectively, and derive the effect of ethical meat consumption on animal welfare. We further show that the effect on other externalities can be estimated using bounding assumptions on substitution patterns. Using such assumptions, we additionally estimate the effect of ethical meat consumption on overall greenhouse gas emissions
Published Work
(with Sharon Pailler, Jon McFadden, Sharon Raszap, Zach Raff, Kevin Kuruc)
Food Policy, 2025
+/- Abstract
Public interest in improving farm animal welfare has increased in recent years, but research on implementation lags behind. This review examines the incentives and barriers to adoption from the perspective of key stakeholders: farmers and other animal product producers. We perform a machine-learning aided scoping review of the academic literature studying how different rearing practices influence outcomes for producers, providing direct evidence on the (dis-)incentives of adopting the practices studied. This allows us to (1) identify existing consensus and (2) highlight research gaps on this question. Operating costs emerge as a near-universal disincentive for welfare-improving practices. Conversely, improved indoor environment shows potential benefits for animal health and productivity, suggesting the possibility of interventions that have only negligible impacts on overall profits. These takeaways are drawn from a relatively sparse literature, underscoring important research gaps. Addressing these gaps can inform evidence-based policies that align practice with public expectations for farm animal welfare while being cognizant of producer constraints and incentives.
(with Alex Hoagland)
Contemporary Economic Policy, 2018
Other Work
(with Ethan Ligon)
GiveWell "Change Our Mind Contest" Honorable Mention, 2022
