Welcome! I’m a PhD candidate in Economics at Columbia University. My areas of interest include Political Economy, Labor Economics, and the Economics of Education. My job market paper is titled, “A structural model of gerrymandering.”
I am on the market this year and will be available for interviews at the 2020 ASSA/AFA Meetings in San Diego.
Legislative maps differ along dimensions of proportionality (the extent to which parties' seat shares match their vote shares) and competitiveness (the probability of close contests). Courts and map-makers want to predict how maps will perform on these dimensions in future elections. Doing this is difficult because future elections will differ from past elections due to changes in demographics and as a result of electoral shocks to preferences and turnout costs. State-of-the-art papers quantify the uncertainty in a map's performance in future elections using statistical models of the across-election variance of aggregate vote shares. By directly modeling aggregate vote data, these papers suffer three limitations: (i) they cannot incorporate all the data sources that are now available on elections, including rich individual-level covariates and individual turnout and survey data; (ii) they may suffer from ecological bias and thus do not permit counterfactuals involving individuals, such as predicting the effects of changes in demographics; (iii) they do not provide information on the preferences of non-voters. In this paper, I contribute a structural model of the voting decision process that resolves these limitations. I then develop a method for measuring gerrymandering that involves simulating counterfactual elections using draws of the structural parameters. I apply the method to rich data from the 2008 to 2018 general elections in North Carolina. Before conducting simulations, I show that the model has strong predictive power within elections for precinct-level vote shares, individual-level turnout decisions, and preference and turnout choices for survey respondents. Moreover, the strategy correctly characterizes uncertainty in excluded elections. Substantively, I find that a variety of recently used maps in North Carolina pack Democratic-leaning voters into uncompetitive districts and generate disproportionate seat shares for Republicans. Consistent with prior literature, I find that a map's geographic compactness does not reflect its proportionality or competitiveness. By contrast, I find that, at least in North Carolina, simply recording quantities related to proportionality and competitiveness in observed elections can be highly informative.
The relative importance of value added and prestige in school choice: evidence from a field experiment in Romanian high school markets (with Rajeev Dehejia, Cristian Pop-Eleches, and Miguel Urquiola)
Abstract: In choosing schools, households may consider both value added (the difference between incoming and outgoing student quality) and prestige (the level of incoming or outgoing quality). The relative weight that households place on these factors influences schools’ incentives and modulates the effect of competition on school productivity. In Romania, students choose schools under an incentive-compatible assignment mechanism. We find that a school’s popularity correlates closely with prestige and only moderately with value added on passing the national baccalaureate exam. We then conduct an intervention to disentangle whether households have limited information on value added or simply care more about other school characteristics.
The effects of the competitiveness of legislative districts on turnout and partisanship: evidence from redistricting in North Carolina
Work in progress: I examine the effects of the competitiveness of legislative districts on turnout, party registration, and predicted partisanship. I compare outcomes over time for individuals in North Carolina who get redistricted into more or less competitive districts. In various specifications, I control for rich covariates, including turnout and party history, and use a border discontinuity design. I take advantage of the large sample size to explore treatment effect heterogeneity. I pay particular attention to the effects of being redistricted into one of North Carolina’s majority-minority districts, which are highly uncompetitive but provide increased local political influence to racial minorities.
Supplementary local school funding and residential sorting of high-income families: private donations and school district parcel taxes in California
Work in progress: Due to restrictions on property taxes following passage of Proposition 13 in 1978, California school funding declined relative to other states. Beginning in the 1990s, well-off districts increasingly relied on supplementary local sources of funding, including tax-exempt private donations and a non-ad valorem tax on property parcels. I chart the growth and distribution of this local funding. I then use a dynamic regression discontinuity design, comparing districts that barely pass and fail to pass the parcel tax, to quantify the effects of local funding on school spending and on the district’s demographic composition. I examine heterogeneity in treatment effects by whether the tax is a lump sum per parcel or varies with square footage.