Working Papers (Updated March 2026)

1. Non-Monotonic Convergence of Policy Designs: Evidence from U.S. State-level Agritourism Statutes.

   • Measures the content of every state statute governing commercial agritourism with a structural topic model and traces how regulatory designs diffuse over time. States converge toward an "optimal" design non-monotonically, a pattern consistent with learning in a multi-armed bandit or Callander's (2011) model of search.

2. Visibility by Design: Evidence from Security Breach Notification Laws in the United States, with Andrew Whitford.

   • Uses text-as-data methods to show that (1) the allocation of visibility (when a public or private entity must disclose a breach) varies by state and entity type, and (2) statutes governing the public sector are consistently thinner than those governing private entities.

3. The Marginal Ad in the Marginal District: Nonlinear Effects of Campaign Advertising, with Jamie Carson.

   • Uses spectral clustering to build profiles of House candidates' TV advertising strategies. In unfavorable districts, candidates who switch ads frequently perform better electorally than those who don't. The same association does not appear in competitive districts.

4. Strategic Compliance with State-Level Age Verification Laws for Adult Content, with David Mustard.

   • Tracks whether ~6,200 live adult websites comply with state age-verification laws and models compliance as a function of site characteristics, including institutional sophistication and the type of content offered. Less-visible sites and sites offering extreme or abusive content are much less likely to comply.

5. Estimating Executive-Legislative Bargaining Dynamics: Evidence from Vector Autoregression Models.

   • Extends Hassell and Kernell (2016) with a vector autoregression to show that presidential influence over the legislative agenda increased substantially from 1985 to 2008.

6. Money and Vulnerability: Electoral Shocks and the Endogeneity of Incumbent Spending in U.S. House Races, with Jamie Carson.

   • Uses a staggered difference-in-differences design to estimate how electoral information shocks change incumbent spending. 

7. The Transmission Problem: A Nationalization-Weighted Forecast for the 2026 Midterm Elections, with Jamie Carson. PS: Political Science & Politics, [forthcoming / 2026].

   • Forecasts the 2026 House elections with a nationalization-weighted Monte Carlo simulation. Accounting for 2024 presidential vote share substantially widens forecast uncertainty.

8. A Random Walk Toward Creative Ideas. Work in progress.

   • Develops a formal model of abductive ideation as Brownian motion driven by discrete experiences drawn from a person-specific stochastic process. Curiosity excites the process and generates more ideas. The model has implications for whether current LLM systems can generate abductive ideas.

9. Procurement and Backlash in Strategic Contracting: Evidence from Flock Safety. Work in progress.

   • Uses memorandum of agreement signing data to predict local police departments' Flock Safety contracting decisions from administrative capacity, and public backlash from local political economy. Firms selling to governments must market themselves both to the buyer and to the buyer's constituents.