Examining the effects of institutional design on electoral outcomes
Williamson, Ryan Dane
MetadataShow full item record
This dissertation contains three essays that examine how institutional design— independent of the behavior of political actors—influences electoral outcomes. First, during the late nineteenth century, research has shown that existing norms and rules governing the redistricting process gave state legislatures enormous discretion in the timing and manner in which they redrew congressional boundaries. However, the Supreme Court rulings during the 1960s altered these redistricting rules. As such, a process that once fostered great competition and turnover now centers on maintaining partisan control within states. I seek to understand how representatives respond to changes in their districts in the modern era. Utilizing an experimental approach, I find that certain redistricting plans lead members of the House to move towards the extreme ends of the ideological spectrum. Second, studies of congressional elections often overlook an important component of the process—primary elections. Given the overwhelming partisanship of a large proportion of districts, many elections are essentially decided well before November. Additionally, primary election laws are not uniform across states as are those governing general elections. Therefore, I seek to understand how certain primary election laws can influence who runs and ultimately wins House primary elections. I find that imposing term limits on state legislators and increasing the diversity of who can participate in a primary produce systematic differences in quality challenger emergence and the distribution of votes among candidates. I also find substantial differences between Democratic and Republican candidates in terms of their performance and success. Third, previous research on the effect of disaster declarations on electoral politics has produced conflicting results. I attempt to reconcile these inconsistencies by reconsidering the theoretical framework used to understand disaster assistance and using unique and innovative data to test the effects of electoral value on aid received as well as the effect of assistance on the public's opinion of the president. I find that disaster declarations differ meaningfully from other actions described as unilateral powers, presidents do not consider electoral value when providing relief, and respondents' evaluations of presidential actions are almost entirely driven by partisanship.