Research

Google Scholar profile

Published Papers

  1. The Structure and Growth of Ethnic Neighborhoods (with Tianran Dai), Journal of Urban Economics, September 2023
    final (page proof) version; MPRA working paper; project page; maps; replication archive
  2. Delivery in the City: Differentiated Products Competition among New York Restaurants (with Jacob Cosman and  Tianran Dai),  Journal of Urban Economics, March 2023
    final (page proof) version; MPRA WP ; data and code repository for replication
  3. Reducing Frictions in College Admissions: Evidence from the Common Application (with Brian Knight), American Economic Journal: Economic Policy, February 2022
    on AEJ website; NBER WP; Revised WP
  4. The Common Application and Student Choice (with Brian Knight), AEA Papers and Proceedings, May 2021
    on AEAPP website
  5. The Out-of-State Tuition Distortion (with Brian Knight), American Economic Journal: Economic Policy, February 2019
    on AEJ websitefinal WPNBER WP
    Public universities typically charge much higher tuition to non-residents. We first investigate the welfare implications of this tuition gap in a simple model. While the social planner does not distinguish between residents and non-residents, state governments set higher tuition for non-residents. The welfare gains from reducing the tuition gap can be characterized by a sufficient statistic relating out-of-state enrollment to the tuition gap. We estimate this sufficient statistic via a border discontinuity design using data on the geographic distribution of students by institution.
  6. Cities and Product Variety: Evidence from Restaurants, Journal of Economic Geography, November 2015
    -Winner of Urban Land Institute Prize for best 2015 paper in Journal of Economic Geography
    on JoEG websiteworking paper versionULI presentation; 中文摘要
    This paper measures restaurant variety in US cities and argues that city structure directly increases product variety by spatially aggregating demand. I discuss a model of entry thresholds in which market size is a function of both population and geographic space and evaluate implications of this model with a new data set of 127,000 restaurants across 726 cities. I find that geographic concentration of a population leads to a greater number of cuisines and the likelihood of having a specific cuisine is increasing in population and population density, with the rarest cuisines found only in the biggest, densest cities. Further, there is a strong hierarchical pattern to the distribution of variety across cities in which the specific cuisines available can be predicted by the total count. These findings parallel empirical work on Central Place Theory and provide evidence that demand aggregation has a significant impact on consumer product variety. Click here for a longer discussion.
  7. Spatial Competition and Cross-border Shopping (with Brian  Knight), American Economic Journal: Economic Policy, November 2012
    on AEJ websiteworking paper version
    This paper investigates competition between jurisdictions in the  context of cross-border shopping for state lottery tickets. Our theoretical model, in which consumers consider both travel costs and lottery payoffs, predicts that per-resident sales should be more responsive to prices in small states with densely populated borders. Using weekly sales data from US lotteries and drawing identification from the rollover feature of jackpots, we estimate this responsiveness and find large effects that vary significantly across states. Using these estimates, we show that competitive pressures from neighboring states may lead to substantially lower optimal prices. 
  8. Momentum and Social Learning in Presidential Primaries (with Brian Knight), Journal of Political Economy, December 2010
    on JPE website; working paper version
    New York Times coverage
    This paper investigates social learning in sequential voting systems. In the econometric model, candidates experience momentum effects when their performance in early states exceeds expectations. The empirical application uses daily polling data from the 2004 presidential primary. We find that Kerry benefited from surprising wins in early states and took votes away from Dean. Owing to these momentum effects, early voters had up to five times the influence of late voters in the selection of candidates, and this helps to explain the distribution of advertising expenditures. Finally, we use the estimated model to conduct two counterfactual experiments.