The Senate Looks Like a Toss-Up
9:29 AM, Nov 2, 2010 • By JEFFREY H. ANDERSON
As we begin Election Day, the Democrats not only appear poised to lose between 60 and 75 House seats (I’ll defer to Jay Cost for the exact number), but they appear to have a roughly 50-50 chance of losing control of the Senate as well. To be sure, this is not what most people are saying, and it’s not what the polls — on their surface — are showing. But the Senate polls appear to be under-sampling independents and over-sampling Democrats — and thereby understating the Republican Senate candidates’ likelihood of victory.
Senate majority leader Harry Reid
Without a doubt, it’s amazing that the Senate is in play — given that the Democrats can hold onto it by simply going 10-27 (27 percent) in the 37 Senate races that are on the board. But it remains to be seen whether they can post that winning percentage or whether the Republicans can post the 76-percent winning percentage (28-9) that, in combination with their expected wins in the House, would give them control of both congressional chambers.
Voter turnout will go a long way toward resolving these unknowns. As I mentioned last week, the gap in party identification between Democrats and Republicans is now almost identical to that of 2004. That year, exit polls showed equal turnout between the two parties, at 37 percent apiece (joined by 26 percent independents). In contrast, Democrats in 2008 enjoyed a turnout advantage of 7 percentage points: 39 to 32 percent (with 29 percent independents).
Since the Democratic advantage in turnout was 7 points higher in 2008 than in 2004, and since party identification is now essentially identical to 2004, a reasonable guess is that this year will be about 7 points worse for Democrats — in terms of turnout margin — than 2008 was. Therefore, by taking the turnout margin in 2008 in each state and adjusting it by 7 points in the Republicans’ direction, we can approximate the likely party split in turnout this year.
However, at least two things are in play this year that differ from 2004. First, the number of independents has increased. By increasing the projected percentage of independents by 10 percent in each state (from, for example, 30 percent to 33 percent), we can approximate their increased impact.
Second, Republicans currently enjoy a clear enthusiasm advantage. In a survey released over the weekend, the Pew Research Center found that “Fully 70% of Republicans have given a lot of thought to this election, the highest figure recorded among either Republicans or Democrats over the past five midterm election cycles. And the differential between Republicans and Democrats [whose tally was 55%] is larger than ever previously recorded.” Furthermore, the survey showed, 61 percent of Republicans “are more enthusiastic than usual about voting,” compared to just 41 percent of Democrats.
It’s anybody’s guess how much of an effect this enthusiasm gap will have on turnout, but it almost surely will have an impact of some sort. To be on the conservative side, it seems safe to estimate that (at the least) it will increase Republican turnout by a point and decrease Democratic turnout by a point.
Under these assumptions, turnout would be quite different from what the polls are projecting. However, most polls provide a breakdown of what percentages of Republicans, Democrats, and independents, respectively, plan to back a given candidate in a given race. These party breakdowns allow us to apply the polls’ tallies to the turnout projections summarized herein.
Doing so yields the following list of the ten closest Senate races:
The 10 Closest Senate Races:
1. California: tie (Real Clear Politics average as of 12:00 A.M. on Election Day: Boxer (D) by 5)
2. West Virginia: Raese (R) by 1.5 (RCP average: Manchin (D) by 4.5)
3. Connecticut: Blumenthal (D) by 5 (RCP: Blumenthal by 9)
4. (tie) Nevada: Angle (R) by 6 (RCP: Angle by 3)
4. Pennsylvania: Toomey (R) by 6 (RCP: Toomey by 4.5)
4. Washington, Rossi (R) by 6 (RCP: tie)
7. Illinois: Kirk (R) by 6.5 (RCP: Kirk by 3)
8. (tie) Colorado: Buck (R) by 10 (RCP: Buck by 3)