Articles and Analysis


Assessing the 2010 House elections

Topics: 2010 , House of Representatives

Tom Edsall quoted me in a Huffington Post article today on the 2010 elections:

There are, however, a number of factors that suggest 2010 will be quite different from the Democratic rout of 1994 -- the so-called Gingrich Revolution. "First, 1994 was the culmination of the South moving into the Republican column; there's no equivalent regional shift trending against Democrats in this cycle. Second, the GOP brand is still in terrible shape relative to 1993-1994," says Brendan Nyhan, a political scientist at the University of Michigan.

For more, see this post on the 1994/2010 comparison from September. The statement about the Republican brand is a reference to this post, which shows that the GOP's net favorables in August 2009 were the worst since 1993 for an opposition party in the first year after a presidential election.

My assessment is roughly in line with the other political scientists Edsall quoted, Charles Franklin of the University of Wisconsin and Pollster.com and Larry Sabato of the University of Virginia:

"I'd say a loss of 20-30 seats, but not yet in the high 30s to make change of control a probable outcome," says University of Wisconsin political scientist Charles Franklin, who bases his prediction on historical precedents. "Presidential support needs to be in the low 40s to predict a very large loss of seats, based on post WWII data. Also, the GDP [Gross Domestic Product] per capita should be in decline or very small gains. At the latest revision of 2.2% in the third quarter, we are low but not as low as in worst midterms for parties."

The economy remains the crucial unknown: "If GDP grows at a three percent or so rate through the election, I think approval will turn up into the 50s, and that probably leads to Republican gains of 15 to 20 seats, which historically wouldn't be bad for the Democrats," Franklin says. If GDP begins to decline, "then approval will fall more and Democrats could be looking at 30-plus lost seats -- still a stretch for Republicans to gain control, but not out of reach."

..."There are several differences with 1993," says the University of Virginia's Larry Sabato. "First, Democrats then didn't believe it was possible for them to lose the House; now they know better and are more cautious." In addition, he says, there have been fewer retirements this year; the Democratic base after Obama's 53 percent win is stronger than it was when Clinton only won a 43 percent plurality in 1992; and the public image of the GOP was much better in the early 1990s than it is now.

For context, here's a lightly edited version of what I sent to Edsall:

As far as the House, I've seen nothing that would dramatically change what I wrote back in September. The Democrats will almost surely lose a significant number of seats, but at this point I still expect them to narrowly retain their majority. Also, there are two important differences between 2010 and 1994. First, 1994 was the culmination of the South moving into the Republican column; there's no equivalent regional shift trending against Democrats in this cycle. Second, the GOP brand is still in terrible shape relative to 1993-1994.

In terms of Obama's coalition, I don't think the decline so far has been especially dramatic (at least relative to my expectations). He started off with honeymoon levels of approval we haven't seen in some time, but now he's reverting toward where Reagan and Clinton were at this point in their term. We shouldn't have expected anything different -- Republicans and GOP-leaning independents were going to revert to disapproval of him as soon as he did anything controversial. Also, we expect him to (a) suffer from the poor economy (b) face a public that trends toward a preference for less government during a period of unified Democratic control and (c) lose seats in his first midterm like most recent presidents. Given all of those factors, I think he's in pretty good shape.

In related news, the Intrade futures market currently estimates the probability of the Democrats retaining control of the House at 66.5%: Price for 2010 US House of Representatives Control at intrade.com

(Cross-posted to brendan-nyhan.com)



I would really love to see someone do a good analysis of how accurate a predictor Intrade has been historically. I'd take a stab at it myself but I can't find where you can get historic intrade data on their site 9at least I couldn't the last few times I looked, maybe it's changed).

My suspicion is that intrade is a pretty poor predictor more than a few days ahead of an event but I'd like to move beyond the hunch level of accuracy.



Just to play devil's advocate: the overall economy (especially unemployment) was in much better shape in 1994 than so far in 2010, which may cancel out the Democratic advantages addressed by Edsall.



Democrats will lose the supermajority in the senate and some seats in the house, however, a 1994 scenario is unlikely for several reasons; Democrats have huge fundraising advantages that they did not have in 1994. I also think that the GOP was much stronger in the blue states in 1994 and now it has become a more regional party. If the job pictures continues to make a rebound, even if it isn't by a whole lot, Democrats will also have greater success.



I don't have precise historical data on Intrade, however, I've followed their results over the last couple of elections and discovered a strong pro-Republican bias, which shouldn't be all that surprising. More of the "players" in Intrade trading tend to be "small-town" stockbrokers, who are overwhelmingly pro-GOP and make small-time bets to game the system. Big financial players, who might have a more dispassionate bent toward real results since they'd have more money at stake, don't ever seem to be playing in the game.

The "volume-drive" is pretty tiny in Intrade's numbers. Sometimes, Intrade might have better predictability if you're talking about Democratic Party primaries, since the pro-GOP slant of its players doesn't have as much sway there, but in general elections, I've found that Intrade seems to have about a +6-8 point Republican bias until you get within a week of an election when the independent polling is quite predictable.



The intensity of the anti-government sentiment is being underestimated. The GOP "brand" favorables are down. So? Where are the democrats "brand" favorables?

Voters will not flock to the pols on election day. They will flock there to vote against democrats. The democrats have pissed off the electorate so bad that the electorate is going to kick the snot out of them come election day.

As much as I'd love to see the democrats lose both the house and the senate - that isn't going to happen. The republican party is too poorly run for that to happen.


Post a comment

Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.