Out of Step, Out of Party: Party Switching in American State Legislatures
2026-07-13
Do parties shape legislators' revealed behavior, or do they merely reflect it? Party switchers offer a rare opportunity to examine party effects because the same legislators can be observed before and after crossing the aisle. I assemble a dataset of 500 state legislative party switches across 49 states from 1993 to 2024 and use comparable ideal‐point estimates to examine two questions: what predicts switching, and how much legislators' voting records change afterward. Ideological misfit is by far the strongest predictor of switching, while district preferences add little independent explanatory power. Partisan context also matters: switching is less common under stronger conditional party government and more common for minority‐party members in lopsided chambers. After switching, legislators move substantially toward their new parties. Among major‐party switchers, the average shift is about 0.64 common‐space units, roughly half the typical gap between party medians. These shifts are largest for party‐median Democrats, consistent with the view that they faced the greatest party pressure, and they grow larger in chambers with stronger conditional party government and greater legislative professionalism. The results show that parties constrain legislators' behavior, and that the strength of that constraint varies systematically with party‐system conditions, competitive context, and institutional differences.