The 2014 primary season is winding down. Although there are still 13 states that will hold primaries over the upcoming month, there are no incumbents in any of those states who have any reason for concern over whether they will be on the November ballot. The alleged “Republican civil war” of the 2014 primaries is over, and it is time for the postwar tribunals to begin. On the one hand, only three House incumbents – Ralph Hall, Kerry Bentivolio, and, of course, former Leader Eric Cantor – lost their primaries. The fate of Scott DesJarlais is still uncertain, pending a recount. No Republican Senate incumbents lost. The Cantor defeat aside, this does not look like an impressive yield for insurgent tea party candidates, and each of these defeats likely says more about the failings of the incumbent than about the appeal of his challenger.
On the other hand, 2014 has so far featured 36 races in which challengers to Republican House incumbents received more than a quarter of the vote (not including top-two primaries) and four races with Senate challengers of similar stature. This year has certainly featured more spending for and against incumbents than has been the case in any previous election. The U.S. Chamber of Commerce spent heavily to ensure the victories of several embattled Republican moderates, and several primaries were decided by razor-thin margins. And as Nate Silver recently pointed out, the average primary vote share of Republican Senate incumbents fell to a historic low. Perhaps this is a victory for tea party: moderates have been served notice that they will face opposition unless they change their ways, and the bumper crop of candidates is a sign that 2016 could feature more Cantor-style upsets. For some conservatives, the fault lies not in the insurgent threat, but in the poor campaigns of some of the challengers.
So who has won? It’s all in how you look at the data. Let’s consider some different ways to interpret what has gone on in the Senate primaries. On Aug. 6, Silver noted that Republican Senate incumbents had received an average of 73 percent of the primary vote, as compared to 78 percent in 2010 and 2012 and 89 percent over the three elections before that. Silver actually undersells his case; according to my data, the average Republican incumbent’s vote share (which has now fall even further following Lamar Alexander’s 50 percent share of the Tennessee primary vote last Thursday) is actually the lowest, by far, in the past 60 years. Consider two different ways of looking at Republican incumbents’ primary election fortunes. When one looks at a scatterplot of the results, it surely seems that something unprecedented has happened.
On the other hand, if one considers the percentage of incumbents who have actually lost, it’s clear that the Republican Party’s perfect score is nothing unusual.
This is a dispute that can only be resolved by using sports analogies. Silver argues that for predictive purposes, it is more important to focus on vote share, a continuous variable, than on binary outcomes such as winning or losing. If a football team wins a string of games by the margin of a field goal or so, it is not really that good of a team and its luck is about to run out soon. In his words, “close calls count.”
This raises the question, however, of what a close call is. Last week Alexander and Pat Roberts each received less than 50 percent of the vote, and both won by about 8 percentage points. Are these close calls? Probably.
But many other Republican primaries this year were not. John Cornyn and Lindsay O. Graham received slightly over 50 percent of the vote, but both won by over 30 percentage points, and their victories were evident long before the voting. To provide a competing sports analogy, consider the final moments of last year’s NBA Finals. With a 20 point lead and minutes to go, the San Antonio Spurs rested their starters, allowing the team’s benchwarmers a chance to play. Miami closed the margin slightly after that point, but the Heat never came close enough for the Spurs to doubt the eventual outcome.
This is a standard trope in sports: when the outcome is clear, it’s “garbage time.” The same is true in elections. There are many good reasons for incumbents to be indifferent about whether they win primaries by 20 points or 40 points, especially if running up the primary vote will cost a lot of money or will compromise their chance of winning the general election.
Ironically, this is a point that was lost on many election analysts in 2012. In that election, Silver predicted the Electoral College outcome with uncanny accuracy, noting that Obama had small but consistent leads in nearly all of the battleground states, and that the apparent dead heat in the popular vote did not portend a close election. The Obama campaign’s confidence that it would narrowly win some states allowed it to focus on as many states as possible rather than running up the numbers in states it was assured of winning.
So margins do matter, but only up to a point. If we set some sort of cutoff for what a “close call” is, this year looks mildly interesting, but not revolutionary. As the graph below shows, if we assume that a “close” race is one decided by 10 percentage points or less, we still have some evidence that this year is unusual, but it’s clear that we’re really only talking about a few idiosyncratic races.
Do Republicans have reason to be worried, then? It depends on what one means by “worried.” It does look like there is now a residual anti-incumbent vote share of 20 percent or more in many Republican districts. If the worry is that Republicans will have to raise money just in case something goes wrong, then yes, they should be worried – although few congressional incumbents are particularly bad at raising money. If the worry is that incumbents will lose any more often than is the norm, or that incumbents who have not done something particularly disastrous will lose, then such concerns are misplaced, as they have been for decades.
What is most interesting – and most difficult to measure – is how Republicans are campaigning in the primaries. Both Alexander and Thad Cochran won while also reportedly defending their brand of Republicanism, while many reports alleged that Graham had tacked to the right to ward off a more serious challenge. It is important to have an estimate of primary competition, but it important to use the presence (or the threat) of competition as a starting point for looking at elections and not to get hung up on either the winning and losing or the exact percentages involved.