How Representatives Respond to Ideological Primary Challenges

• In Event: Voting in Primaries

Abstract

In recent decades, congressional primaries have become dominated by ideological challenges; to incumbents’ left in the Democratic Party and, especially, to their right in the Republican Party (Boatright 2013; Cowburn 2020). This paper tests how incumbent members of the U.S. House of Representatives who survive such intra-party challenges respond in subsequent congresses and elections. Using a multi-stage difference-in-differences design across eight elections between 2006 and 2020, I find that representatives responded to ideological primary challenges by adjusting both their roll-call voting behavior (one-congress-at-a-time NOMINATE) and donor base (CFscore), moving away from the center on both measures in the following congresses and election cycles.

As expected, the adaptation of incumbents’ voting behavior and donor base is conditional on challengers’ electoral performance. The new iteration of difference-in-difference literature enables the estimation of heterogeneous responses by incumbents to primary challengers with varying degrees of success. In this paper, this effect is modeled using a dichotomous treatment of any primary challenger, thresholded to multiple levels of vote share, and, most precisely, using challengers’ vote share as the varying treatment dosage. This paper presents evidence that challengers who receive a higher share of the primary vote elicit a larger adaptative response from incumbents.

Aggregated by party and election cycle, the average treatment effect on the treated (ATT) increased in both parties during the period, indicating that incumbents felt an increasing need to respond to ideological challenges from within their own party, and that they did so by adopting non-centrist roll-call voting behavior and appealing to more extreme donors. This finding suggests that representatives have become more responsive to their primary constituency (Fenno 1978), likely as a result of the proliferation of safe seats in general elections. Increased responsiveness to intra-party threats may also be a reaction to the commonly held notion that primary electorates are ideologically extreme. Incumbents who felt threatened by a challenger from their ideological flank had less-centrist roll-call voting behavior in subsequent congresses and altered the ideological make-up of their donor base in an attempt to ward off a future intra-party threat.

Across the period, adaptation by Republican incumbents was larger than their Democratic counterparts, who often face greater electoral pressure in general elections. This finding also aligns with literature indicating an asymmetric movement of the parties in Congress, with Republicans moving further to the right than Democrats have to the left (McCarty, Poole, and Rosenthal 2006; Theriault 2013) in part due to organizational differences in the parties’ coalitions (Grossmann and Hopkins 2016).

The between-election adaptative effect of primaries on representatives’ positions remains an under-studied way in which ideological nomination contests may contribute to polarization and radicalization. Given the limited empirical evidence that primary selectorates in either party are more ideologically extreme than non-primary voting general election party voters (Hirano et al. 2010; Sides et al. 2016), this effect is driven largely by—potentially incorrect—elite perceptions of electoral risk. These findings therefore have implications for mechanisms of polarization and radicalization in Congress, elite behavior, and theories about intra-party homogeneity.

Author

• Mike Cowburn, European University Viadrina