The 2010 anti-incumbency meme that refuses to die.

Who's winning, who's losing, and why.
Aug. 4 2010 6:37 PM

Anti-Anti-Incumbency

The 2010 election meme that refuses to die.

Sen. Arlen Specter. Click image to expand.
Incumbent Sen. Arlen Specter lost his primary

BREAKING: A historic wave of anti-incumbency is sweeping the nation.

Angry voters first threw Sen. Bob Bennett, R-Utah, out of office in a dramatic Tea Party coup. Then Sen. Arlen Specter, D-Pa., went down to Rep. Joe Sestak, D-Pa., in the primary. On Tuesday, Rep. Carolyn Cheeks Kilpatrick, D-Mich., lost a primary. "Incumbents beware," warned the Associated Press. "Another lawmaker just bit the dust." If you ask Fox News, this has been "a primary season fraught with anti-incumbent furor." Whose head will the anti-incumbent hordes claim next?

But let's put this in perspective. So far this year, 282 federal-level incumbents have been up for re-election. Of those, only six have lost their seats—four in the House and two in the Senate. (Aside from Bennett, Kilpatrick, and Specter, there's Rep. Alan Mollohan, D-W.Va.; Rep. Parker Griffith, R-Ala.; and Rep. Bob Inglis, R-S.C.) That's 2 percent of all incumbents. If you count only the 119 incumbents who have faced primary challengers, the proportion who were defeated goes up to 5 percent.

Is that a lot? Historically, no. While it's hard to game out exactly how many incumbents had been defeated at this point in election years past, this year isn't a huge anomaly, according to University of Virginia's Larry Sabato, who has been trying to tamp down the anti-incumbency myth since May. "This is about average," Sabato says. "Usually there's one Senate seat lost, we're at two; about five House seats, we're at four." Looking at the rest of the primaries, he says, "I can tell you at most three or four more will lose, total."

The six incumbents who have lost are largely anomalies. Bennett, for example, was ousted by a small Utah convention dominated by the party's conservative base. Sure, all primaries skew toward the party's base, but this was a caucus of 3,500 Republican activists. Had the whole state of Utah voted, Bennett might well have survived. Specter, too, was a special case. After switching parties, he was asking for the votes of Pennsylvania Democrats who had spent their lives voting against him. Same with Parker Griffith, who lost to another Republican soon after becoming one. Kilpatrick's loss was abnormal, too: She barely won the primary in 2008, when her son, Detroit Mayor Kwame Kilpatrick, was embroiled in a sex scandal. In 2010, after he was sentenced to five years in prison, she lost.

That's not to say a lot of incumbents won't lose in the general elections. They probably will. But that's not because of anti-incumbent sentiment so much as anti-Democrat sentiment. Many of the moderate Democrats who joined in the 2006 rebellion against the party in power are now vulnerable as the country rebels against … the party in power. Some of them, fearing defeat, have chosen not to run. So goes the pendulum.

There is no proven relationship between incumbent performance in the primaries and in the general. (If an incumbent loses in a primary, after all, he or she usually doesn't compete in the general.) In both 2004 and 2006, for example, two members of the House lost their seats in the primaries. But whereas 2004 turned out to be a fairly status-quo general election, 2006 saw a Democratic surge. Looking at the primary results, though, you never would have known. Doomsayers in 2010 could point to proxies for incumbent losses—a bad economy, low presidential approval ratings—but again, those factors tend to work against the incumbent party, not so much incumbents across the board.

The fact is, even an "anti-incumbent" year isn't that bad for incumbents. The average rate of re-election for members of the House since 1964 has been 93.3 percent. (Over the last decade, it's been 96 percent.) In the Senate, the average since 1964 has been 81.6 percent. So what's the worst-case scenario for this year? In June, George Washington University political scientist John Sides compared House re-election rates from years past with historical polling on whether voters would prefer candidates who have been in Congress or those who haven't. He found a slight correlation and extrapolated that, based on polling in 2010, the re-election rate would be somewhere around 87 percent. Compared with past years, that's pretty low. But if a year in which 87 percent of incumbents win re-election is considered "anti-incumbent," that says as much about the power of incumbency as it does about the year itself.

Rebecca Kaplan contributed research to this article.

Like Slate on Facebook. Follow us on Twitter.

Christopher Beam is a writer living in Beijing.