March 5, 2014

Public’s anti-incumbent mood hasn’t always predicted big electoral swings

anti_incumbencyThe public’s anti-incumbent mood is at record levels, but the vast majority of senators and representatives running for re-election appear to have little to worry about. What’s up with that?

The Rothenberg Political Report currently rates 383 House seats as safe and 13 as having a clear favorite; the Cook Political Report deems 358 seats safe and another 33 “likely” wins for either party. Both firms have similar estimates for the number of competitive seats in the November elections: 44 (Cook) and 39 (Rothenberg). The University of Virginia’s Center for Politics, which rates 41 races as tossups or leaning to one party or another, also sees little prospect for massive change: As senior columnist Alan Abramowitz wrote last month on the Center’s “Crystal Ball” site, the fall elections “are likely to result in minimal change in the party balance of power.”

That would seem to be at odds with recent survey results.  A Washington Post/ABC News poll released this week, for instance, found that only 22% of the public said they were inclined to re-elect their representative in Congress. The all-time low (since 1990) was 17% in August 2011, just after that year’s debt-ceiling showdown. In the latest poll, two-thirds (68%) of people said they were “inclined to look around for someone else to vote for,” also a high for the Post/ABC poll.

Last fall, a Pew Research Center survey also found record anti-incumbent sentiment, though it went about measuring it somewhat differently. We asked registered voters whether or not they’d like to see their representative re-elected in the next election. Fully 38% said no, the highest level for that response in the two decades we have been asking that question.

Similarly, only 46% of registered voters in a January Gallup poll said their representative deserved re-election, compared with 17% who said “most members of Congress” deserved re-election. Both figures were record lows.

Of course, it’s one thing to say your representative doesn’t deserve to be re-elected, or even that you’re open to voting him or her out, and actually doing so. Historically, Americans have been more favorably disposed toward their own representatives than to Congress as a whole (a phenomenon dubbed “Fenno’s Paradox,” after eminent political scientist Richard Fenno). That disjuncture that has been reflected in typically high re-election rates. On average, in each of the seven House elections between 1978 and 1990 (inclusive), fewer than 20 representatives were defeated for re-election.

DefeatedIt’s also worth noting that opinion surveys on incumbent sentiment haven’t always been the best predictors of big electoral swings. In fact, two of the biggest anti-incumbent waves in recent history weren’t preceded by surges in anti-incumbent sentiment.

In August 1992, only about half (48%) of registered voters in the Post/ABC poll said they were looking around for another candidate. That November, 43 incumbents were defeated, the most in nearly two decades. And just ahead of the 2010 elections, which unseated 58 representatives (54 of them Democrats, which President Obama called a “shellacking”), 55% of voters told Pew Research surveyers they wanted their representative to keep his or her job.

Topics: Congress, Elections and Campaigns

  1. Photo of Drew DeSilver

    is a senior writer at Pew Research Center.


  1. DrT Happy3 years ago

    Hermes states that it is one thing to say you want the incumbent out but another to actually do it because we are consistently looking at the lesser of two evils. I think that if you are tied mentally to a party, the other party will always look like the more evil. What needs to happen is for many people to go with the notion that let’s vote out the incumbents on a consistent basis and then the parties will start to understand they have to put up much better candidates.

  2. David Mehegan3 years ago

    Drew, has the opposite ever happened: a wave election predicted that did not happen?

    1. Drew DeSilver3 years ago

      Depends on how you define “predicted,” I guess. In October 1990, 57% in the Post/ABC poll said they were included to look around for someone else to vote for (almost identical to October 1994), but that year only 16 House incumbents were defeated.

  3. Richard DeSilver3 years ago

    Many incumbents run un-opposed due to gerrymandered districts. Personally, I would like to see all Senators, Representatives and the President elected for one six year term only. In this way, perhaps decisions would be based on what is good for the country rather than how they effect re-election.

  4. slk3 years ago

    haven’t voted for mccain in ages, and will never vote for flake again!!! lie to me once, you’re out!!!

  5. DonB3 years ago

    So how does these numbers break down if they were looked at for JUST Democrats and JUST Republicans?

    Are Democrats more likely to be in the “Keep in office” camp or are Republicans?

    Can that question be answered for past polls and would that change the predictive ability of these types of polls?

  6. Hermes3 years ago

    It might be one thing to say you don’t want your representative re-elected and not actually voting against them, but it is another thing to look at the alternatives and see they would be much worse. Our choices for politicians in America are typically deciding which is the lesser of two evils, when voters say they don’t want to re-elect their representative it doesn’t preclude the desire for a candidate whom one feels would actually represent their interests.

    1. slk3 years ago

      look for the ones with the smallest pockets!!!