Out of Step and Still in Congress? Electoral Consequences of Incumbent and Challenger Positioning Across Time

Recent research suggests that the penalty congressional candidates pay for ideological extremism declined abruptly in 1994, when the House majority became competitive for the first time in decades. We reexamine congressional accountability in light of this evidence, first evaluating the centrality of 1994 as a turning point and then allowing that voters may not weigh incumbents’ and challengers’ positions equally. Several findings emerge. Even when the penalty for extremism is constrained to be equal for challengers and incumbents, accountability does not abruptly decline in 1994 but instead decreases gradually from 1980 through recent elections. Furthermore, once incumbent and challenger ideology are examined separately, the results on incumbents do not match those for challengers. Depending on the specification and ideology measure, incumbent accountability may stay similar, decrease, or even increase over time. By comparison, the relationship between challenger ideology and vote share consistently declines across electoral cycles. These results suggest that analyses treating incumbents and challengers identically will be prone to find decreased congressional accountability, even when the evidence on incumbents does not merit such a conclusion. ∗We are grateful to seminar participants at Columbia, Princeton, and Stanford GSB for helpful feedback. In addition, we thank David Brady, Adam Bonica, Gabriel Borelli, John Cogan, Gary Cox, Bob Erikson, John Kastellec, Eric Manning, Adam Meirowitz, and Nolan McCarty for valuable comments and Gary Jacobson for graciously providing his data on challenger quality and congressional elections. †Brandice Canes-Wrone is Donald E. Stokes Professor of Public and International Affairs, and Professor of Politics, Princeton University. bcwrone@princeton.edu. Ph: 609-258-9047. ‡Michael Kistner is a joint Ph.D. candidate in Politics and Social Policy, Princeton University. mkistner@princeton.edu. Ph: 309-531-0494. An enduring question in political science is the extent to which the ideological positions of congressional incumbents and challengers affect electoral outcomes. Among theories that predict some level of policy accountability, various hypotheses exist regarding whether and how candidates’ positions matter. Most empirical analysis has lacked measures of challengers’ ideological positions and accordingly, analyzed the roll call or announced positions of incumbents (e.g. Erikson and Wright 1993; Canes-Wrone, Brady, and Cogan 2002; Nyhan et al. 2012). This body of work offers considerable evidence that incumbent accountability exists, with those who are sufficiently “out of step” for their district faring worse at the voting booth. Recently, the development of candidate ideology estimates based on campaign contributors’ behavior has revolutionized the capacity to analyze the relationship between challengers’ ideological positions and electoral outcomes. The most widely used are the Bonica (2014) CFscores, which characterize candidates’ as well as contributors’ ideological locations. Although not the first estimate of challenger ideology, the scores are particularly attractive because they encompass the vast majority of general election challengers for over three decades. By comparison, alternative measures based on candidate surveys (Erikson and Wright 1989; Klinger, Hollingbaugh, and Ramey 2019), Twitter followers (Barberá 2015), or expert surveys (e.g., Joesten and Stone 2014) are much more limited in scope in terms of the sets of candidates and/or elections. Leveraging the benefits of the CFscores, multiple notable studies examine whether candidate accountability has changed over time and find that it has dramatically declined and/or no longer exists in a significant way. For instance, Bonica and Cox (2018) analyze congresThe literature on elections uses the term candidate accountability to refer not only to incumbent positions but also challenger positions (e.g., Bonica and Cox 2018). We recognize that challenger positions may be cheap talk, particularly if a challenger lacks experience in any political office. Still, we follow in the tradition of using the term accountability to refer to both types of candidates.