Featured image

Nationalized elections and the 2018 U.S. Senate midterms

In the months leading up to the 2018 midterm elections, media outlets and academics have devoted considerable attention to the potential for Democratic gains in the U.S. House of Representatives. There is a general consensus that the Democrats will pick up Republican-held seats and many election prognosticators favor the Democrats to regain control of the House. The 2018 U.S. Senate elections, however, may well yield Republican gains. While Democrats have managed to capitalize on favorable electoral conditions to field a slate of Senate candidates who have made several races unexpectedly competitive—Representatives Beto O’Rourke (D-TX) and Krysten Sinema (D-AZ) are two notable examples—the Senate electoral map nonetheless favors the Republicans. The Democrats must defend 25 seats this cycle compared with the Republicans’ eight seats. These numbers are not, in and of themselves, the problem for Democrats, but instead it is the fact that 10 of these 25 contests are in states President Donald Trump won in 2016. While a Democratic incumbent is running in all 10 of these states, the power of incumbency may not be enough to ensure the Democrats hold each of these seats.

Figure 1: Presidential and Senate Elections Won by the Same Party

As Figure 1 demonstrates, the association between the partisan outcome of Senate and presidential elections has increased considerably over time. During the 1980s, only about half of presidential and Senate elections resulted in the same party winning the statewide vote. By the 1990s and 2000s, the congruence between the partisan outcome of presidential and Senate elections increased to around 60 and 70 percent respectively. By the 2012 election cycle, the proportion of races won by the same party was nearly 85 percent.

The increased correspondence between presidential and subpresidential elections is known as nationalization. As Dan Hopkins notes in his recent book, there are many ways that elections can become more nationalized. With respect to vote choice and election outcomes, however, nationalization is typically thought of as a process in which top-down forces exert greater influence on vote choice and election outcomes than do candidate-specific characteristics or local forces. By top-down forces, I mean that factors like presidential vote choice, presidential approval, and/or partisan identification drive vote choice rather than do attributes of subpresidential candidates, such as incumbency or candidate ideology.

There are a several notable consequences of increasingly nationalized congressional elections. In the context of the 2018 midterms, the most important is that the electoral benefits of incumbency have waned as more and more voters cast straight ticket votes. Indeed, the marked growth of the incumbency advantage in the 1970s was a function of voters either crossing the aisle to support a congressional candidate from the other party or supporting different candidates in presidential and subpresidential elections. In research with Seth C. McKee, we estimated the incumbency advantage in Senate elections by decade from the 1980s through the 2016 election. While political scientists have proposed multiple ways to quantify the incumbency advantage, the ultimate goal in each study to is discern the average electoral boost an incumbent candidate can expect to receive. For example, an incumbency advantage of 10 would indicate that the average incumbent performs 10 percentage points better than a similar non-incumbent candidate would perform in an open seat election.

Our findings, which are shown in Table 1, indicate that indicate that recent Senate incumbents have not performed much better than non-incumbents. From the 1980s through 2000s, however, Senate incumbents received a sizable boost relative to their counterparts in open seat races. Furthermore, while the early 2000s represent a deviation from the larger trend, the estimated benefits of incumbency have been in decline since the 1980s, which is consistent with Gary Jacobson’s findings for U.S. House elections. As a point of comparison, Table 1 also provides information about the estimated relationship between Senate and presidential election outcomes. Specifically, it indicates how well a party performed in a Senate election given their performance in the most recent presidential election. For example, an estimate of 0.50 would indicate that a 10-percentage point increase in the party’s state-level presidential vote would lead to a 5-percentage point increase in the party’s performance in a Senate. As the size of this estimate increases, which it has done since the 1980s, it means that the outcome of Senate and presidential elections are more closely linked. In sum, the decreased role of incumbency in conjunction with a greater correspondence between Senate and presidential election outcomes are all consistent with greater electoral nationalization.

These results further demonstrate the challenges Senate Democrats face in what could otherwise be a fairly successful midterm election. While there was a time when incumbents could expect a 6 to 8-point boost on Election Day, today’s more nationalized electoral environment will likely lead to Democrats losing at least a few of their more vulnerable seats. That is, of course, not to suggest that every Democratic incumbent is in an equally dire position. Indeed, five of the Democratic incumbents in states President Trump carried in 2016—Senators Tammy Baldwin (D-WI), Debbie Stabenow (D-MI), Sherrod Brown (D-OH), Bob Casey (D-PA), and Joe Manchin (D-WV)—currently have roughly double digit leads in the most recent public opinion polls. Even some Democratic incumbents in traditionally Republican states, such as Senators Jon Tester (D-MT) and Joe Donnelly (D-IN), are consistently polling ahead of their Republican challengers. In other solidly Republican states, such as Missouri, North Dakota, and Tennessee, incumbency and the emergence of unexpectedly strong challengers will likely not be enough for Democrats to counteract the strong influence of partisanship.

The rise of nationalized politics also has important consequences for legislative politics. First, as senators’ and presidential electoral constituencies become increasingly defined along partisan lines, senators face fewer incentives to compromise or work with the president. Frances Lee has chronicled how senators are increasingly motivated to deny the president a win or to use the oversight process to score political points against a presidential administration. The long-term consequence of these developments is an increase in what Lee refers to as partisan team play, which occurs when legislative behavior is informed by political considerations rather than just a legislator’s ideological position. Second, nationalized elections have the potential to undermine political accountability. When vote choice is driven by top-down forces instead of evaluations of an individual senator, it means that candidate-specific factors and office-based judgments of political performance take a back seat. Similarly, when senators have little reason to expect to be punished on Election Day, it frees them to make choices that run counter to constituent preferences.

Joel Sievert
Joel Sievert is an Assistant Professor of Political Science at Texas Tech University. He received his Ph.D. in Political Science from the University o...