It is well known that members of Congress have become more ideologically polarized in recent years. In new research, Connor Halloran Phillips, James M. Snyder, Jr., and Andrew B. Hall look at the influence of the pool of potential candidates for Congress on the partisanship of those who are elected. Examining partisanship among state legislators, they find that Democrats and Republicans who run for Congress have similar levels – or slightly higher for Republicans – of partisanship compared to their candidate pools. What has changed, they write, is that the ideology of candidate pools has become much more polarized, contributing to the upsurge in polarization in Congress.

The widening divide between Democrats and Republicans in Congress has made effective governance much more difficult, with the latest evidence being last spring’s protracted partisan standoff over raising the debt ceiling. In large part, this trend is due to the increasingly extreme positions taken by those running for Congress. A 2019 book authored by one of us estimates that even if every election for a seat in the US House of Representatives since 1980 had been won by the most moderate candidate in the race, the ideological distance between the parties in the House would still have grown by 80 percent as much as it did. To understand the roots of partisan polarization in Congress—and how to address it—we need to know why candidates for Congress have become more ideologically polarized.

Identifying the pool of congressional candidates

To answer this question, we need to examine the underlying group of potential candidates, US citizens who are constitutionally qualified to run for Congress (over 25, have been a US citizen for at least seven years, and live in their state of election) and possess the necessary resources, experience, and abilities to mount a viable candidacy. Otherwise, we cannot determine whether, for example, this pool of candidates has polarized as a whole or if the most extreme members of that group are now more likely to run for Congress. Although it is nearly impossible to identify the entire candidate pool in each congressional district, one particularly important segment of candidates is the state legislators whose constituencies overlap with that district.

State legislators are highly qualified potential candidates for Congress, with built-in name recognition and political experience in a geographic area. In fact, almost half of members of Congress are former state legislators. Furthermore, political scientists have used state legislative voting records along with responses to a national survey of candidates for office to develop a measure of state legislators’ ideology called NP-Scores. This means that we can identify which legislators are more extreme than others. In addition, when we compare how state legislators and those who are not state legislators running for Congress responded to that national survey of candidate positions, or the voting behavior of members of Congress who served in state legislatures and those who did not, we find that the two groups are nearly indistinguishable from each other when evaluating politicians in similar constituencies. We can conclude from this analysis that state legislators are ideologically representative of the broader candidate pool in their district.

Photo by JOSHUA COLEMAN on Unsplash

Comparing candidates to their pools

By matching state legislators to congressional districts based on their constituencies, we can explore how the ideology of state legislators who run for Congress compares to the ideology of state legislators in the same congressional district who do not run. We construct these matches using election returns, voter registration records, and US Census data for eleven different congressional elections from 1996 through 2016 and plot these differences in Figure 1 for Democrats on the left and Republicans on the right. Democrats who run for Congress are not consistently more or less liberal than their pools. Conversely, Republicans who run for Congress tend to be slightly more conservative than their pools on average, but this gap did not change during the two decades we examine, as it would have if polarization were driven by relatively extreme potential candidates becoming more likely to run over time.

Figure 1 – Average difference in ideology between state legislators who do and do not run

Note: Higher numbers represent greater conservatism on the scale we use, so a positive value indicates that candidates are on average more conservative than their pools and a negative value that they are more liberal.

What has changed is the ideological composition of the candidate pools. Figure 2 plots the mean ideology of the candidate pools for Democrats and Republicans over time (“Average of Pools”), the mean ideology for all Democratic and Republican state legislators (“National Average”), and the average ideology of pools in safe and competitive districts for each party. We define competitive districts as those in which the margin between the district-level vote shares of the Democratic and Republican presidential candidates was within five points of the national margin, whereas safe districts for each party are those in which that party’s candidate performed at least five points better than they did nationally.

Figure 2 – Changes in state legislators’ ideology over time

Across all pools of candidates, the average ideology has moved away from the center over time, with Democratic pools becoming more liberal and Republican pools more conservative. As the “National Average” lines show, this evolution has occurred in tandem with an overall shift to the left among Democratic state legislators and to the right among Republican state legislators. Accordingly, the upsurge in polarization among candidates appears to be driven by a parallel one in the underlying pools of potential candidates.

This finding does not mean that all the pools are similar. As the lines in Figure 2 show, pools in safe districts are much more ideologically extreme on average than pools overall. The proportion of districts that are safe has risen during this period, from 62 percent in 1996 to 84 percent in 2016, so the evolving geographic distribution of these extreme pools has also likely made polarization worse. Nevertheless, its contribution is relatively small: We estimate that 67 percent of the increase in the ideological gap between the average Democratic and Republican state legislator elected to Congress over this period is the result of the polarization of pools, and only three percent is due to the greater proportion of safe districts.

Implications for addressing polarization

We establish that growing ideological extremism in state legislatures has been a key driver of polarization among congressional candidates, and thus in Congress itself. The profile of state legislators who run for Congress has remained constant over the years—Democrats who are ideologically representative of their pools and Republicans who are slightly more conservative than their pools—but as the state legislators serving within each congressional district polarize, these candidates will be more extreme. A similar pattern holds for those who win election and serve in Congress, although winners are a bit more extreme on average than those who run among both Democrats and Republicans.

Because state legislators who run for Congress are ideologically similar to non-state legislators who do so, we suspect the same trends have occurred among the local elected officials, business leaders, attorneys, activists, and other engaged citizens who form the rest of the congressional candidate pool. Even if our results only apply to state legislators, however, they still reveal that the roots of polarization in Congress reach all the way into state politics. If reformers want to foster greater representation of moderate views in the federal government, they would be well-advised to start by encouraging more moderate individuals to run for lower-level offices like state legislature to begin building a pipeline of less extreme potential candidates.

• This article is based on the paper, ‘Who Runs for Congress? A Study of State Legislators and Congressional Polarization’, in the Quarterly Journal of Political Science.
• Please read our comments policy before commenting. 
• Note: This article gives the views of the author, and not the position of USAPP – American Politics and Policy, nor the London School of Economics. 
• Shortened URL for this post: https://wp.me/p3I2YF-e4n