Description: Department Seminar

Location: 368A Heady Hall

Contact: Katie Harris-Lagoudakis

Title: What Drives Voting: Estimates from a Multilevel Model

Abstract: We study identification and estimation of a two-candidate discrete choice voting model when micro data in various forms are available. In particular, we consider sub-market level vote shares, individual voting turnout, and individual characteristics. We demonstrate the value of micro data in the voting context across multiple dimensions. We estimate the model using 2016-2022 voter files from across the entire United States, paired with campaign finance, voting laws, and polling place locations. Preliminary estimates indicate that voting costs are higher in Democratically leaning areas such that improvements to voting access would not only increase turnout but could also slightly change election outcomes.  In addition, the sorting of demographically similar individuals into certain regions and district-boundary effects appear to give an advantage to Republicans. Finally, we show that the micro data can help refine standard district bias measures like Partisan Bias.