"Doubly-Robust Difference-in-Differences Estimators." (with Pedro H. C. Sant'Anna), Journal of Econometrics, vol. 219 (1), p. 101-122, 2020.
"Campaign Spending in Senate Elections: A Structural Analysis of a Two-Stage Election with Endogenous Candidate Entry." (Job Market Paper), October 2021.
Abstract: This paper quantifies the effect of campaign spending on electoral outcomes in the U.S. Senate elections. To this end, I develop a two-stage asymmetric contest model where in the first stage potential candidates decide whether to enter, and in the second stage those that decide to enter participate in an election consisting of within-party primaries and a general election where they make campaign spending decisions to influence the probability of winning. I also specify voters’ decisions through a latent utility model. I estimate this game-theoretic model using data on the vote shares of Senate candidates in primary and general elections, data on candidates’ campaign spending and candidate-specific characteristics, as well as state-year level covariates for Senate elections between 1994 and 2018. Taking account of endogenous entry, I find that candidates with past legislative experience and incumbents spend more and run more relative to potential challengers, and that campaign spending is more effective in increasing votes in the primary compared to the general election. Moreover, I show that the primary election in the incumbent’s party is important in the sense that it can increase the electoral competition. I also conduct a counterfactual analysis and show that a public funding policy targeting at challengers can reduce the amount of campaign spending by candidates, and increase the probability of challenger victory modestly.
Work in Progress
"Nonparametric Estimation of Tullock Contest Models."