Senate Watch: September 15
Fewer key states were polled for Senate candidates over the past two weeks than we had last time. But we did get some movement in a few races.
Here is the current map:
Now for the details.
As always, “Continuing” refers to the seats in Senate Classes 2 and 3, which are not up for election this cycle.
Here the highlights of the past two weeks, walking from the Pacific to the Atlantic:
- Arizona: One new poll, from Public Policy Polling, shows Representative Jeff Flake (R-Mesa) ahead of Democratic former Surgeon General Richard Carmona by one unadjusted point (which adjusts to four). Polling has been light and infrequent, but Carmona has yet to lead in a single poll. Last time I said I was torn between keeping Arizona at “Leans Republican” and moving it to “Likely Republican”. I’m not torn anymore; Arizona is now “Likely Republican”. Intraders seem to agree; Flake is trading at 87 percent, and Carmona at 22 percent.
- Montana: One new poll from Public Policy Polling gives Democratic incumbent Senator Jon Tester a two-point unadjusted lead over Representative Denny Rehberg (R). This adjusts to a tie. Coupled with Rasmussen’s poll from last time, I still don’t have sufficient evidence to consider Montana anything but a “Tossup”. Intraders are more inclined to believe Rehberg will win here; he is given a 65 percent likelihood, with Tester at 44 percent. It’s a lightly-traded market, though, so I don’t give much credence to it.
- New Mexico: Another pair of new polls. One was from the Albuquerque Journal, who found Representative Martin Heinrich (D-Albuquerque) ahead of his predecessor, former Representative Heather Wilson (R-Albuquerque) by seven points. The other, from Public Policy Polling, indicates a nine-point unadjusted Heinrich lead (which adjusts to seven). The numbers are consistent with every poll since April, and lines up perfectly with Rasmussen’s poll from last time. Wilson has yet to even tie a single poll, let alone show a lead. Moreover, when looking at the same polling firms over time, the trend is a consistently increasing gap between the two candidates. For this reason, I’m moving New Mexico from “Leans Democrat” to “Likely Democrat” a couple of weeks earlier than I had originally expected to.
- Missouri: Since last time, we’ve had one poll in Missouri, from Rasmussen. He saw Democratic Incumbent Claire McCaskill ahead of Representative Todd Akin (R-Wildwood) by six unadjusted points (which adjusts to seven). Ordinarily, this would be cause for moving Missouri from “Tossup” to “Leans Democrat”, since no polls after Akin’s rape comment have shown the challenger with a lead. However, Rasmussen saw McCaskill ahead by four more points two weeks ago, which could suggest a narrowing lead. Until the polls show me more, Missouri stays a “Tossup”. Intraders disagree, giving McCaskill a 63 percent chance to Akin’s 40 percent.
- Michigan: One poll in the past two weeks, from EPIC-MRA, gives Democratic incumbent Senator Debbie Stabenow an 11-point lead over former Representative Pete Hoekstra. Last time, I noted that both Mitchell Research and Baydoun/Foster looked like outliers in showing Hoekstra ahead, given the historical trend, though it’s uncommon to have two outliers in the same time period. It now appears my caution was justified, so I’m leaving Michigan as “Likely Democratic”. Intraders agree, giving Stabenow an 85 percent chance to Hoekstra’s 23 percent.
- Ohio: Three new polls this time, from NBC/Wall Street Journal/Marist, Public Policy Polling, and Gravis Marketing. Marist gives incumbent Senator Sherrod Brown a seven-point unadjusted lead (which adjusts to five points) over Treasurer Josh Mandel. PPP came away with Brown up by eight unadjusted points (five adjusted). Gravis sees Brown up by six unadjusted (eight adjusted). Last time, I thought we might be seeing evidence of a tightening race. This time, it looks like Brown has a comfortable lead. Given how noisy polls have been in Ohio this year, I’m leaving Ohio in the “Leans Democrat” column until we have a longer period of time without the close numbers of two weeks ago. If we see the same thing in the next installment, I’ll move Ohio to “Likely Democrat”. Intraders are as bullish for Brown as they are for McCaskill, giving him a 65 percent chance to Mandel’s 25 percent. It’s a rather lightly-traded market, though.
- Florida: Four polls this time, from Public Policy Polling, Gravis Marketing, SurveyUSA, and NBC/Wall Street Journal/Marist. If you look over the two-week period, incumbent Senator Bill Nelson has gained a little, but his challenger, Connie Mack, IV (R-Fort Myers), has rapidly dropped about four points over that time. Nelson’s lead is approaching double digits. I moved Florida into “Leans Democrat” a month ago, but I’m nearing the point where I’ll push the Sunshine State into “Likely Democrat”. At this point, I’m waiting to see if this is a temporary bump or a new plateau. This market is too lightly traded on Intrade to be useful.
- West Virginia: This is a rarely-polled state, and only Public Policy Polling has bothered to look anyway. Regardless, with margins well above 20 points each time, there’s no doubt that Democratic incumbent Senator Joe Manchin will handily defeat ultraconservative Republican John Raese. West Virginia is “Likely Democrat”.
- Virginia: Two new polls this time. One from Gravis Marketing gives former Governor George Allen a five-point unadjusted lead (which adjusts to two) over former Governor Tim Kaine. NBC/Wall Street Journal/Marist sees an unadjusted tie, which also adjusts to two points in Allen’s favor. Two points is within the margin of error of the historical trend of tossup. Until I see data to the contrary, I will continue to believe Virginia will be a “Tossup” right up to election night. Intraders are more confident in a Kaine victory, giving him a 59 percent chance of winning, to Allen’s 30 percent.
California, Minnesota, New Jersey, and New York were also polled, but their outcomes have never really been in question.
The number of seats left up for grabs is diminishing, with the only remaining leaners of Ohio and Florida on the left (both rapidly moving to “likely” status), and possible flippers North Dakota and Wisconsin on the right. Our six tossups remain so: Connecticut, Indiana, Massachusetts, Missouri, Montana, and Virginia. Republicans’ best chance is Indiana, while Democrats’ best is probably Connecticut. The other four are currently coin-flips. That makes it credible for Republicans to win at least three of the six, but winning three would not hand them the majority unless Representative Paul Ryan (R-Janesville, WI) becomes Vice President, a decreasingly likely event. Moreover, North Dakota, which has been lightly polled for some time, may not still be “Leans Republican”; the absence of data leaves us effectively blinded there. The Intrade markets moved a decent amount: they now give Republicans about a 45 percent chance of taking the Senate (down nine points from two weeks ago), with a 33 percent chance of Democrats holding at least 51 seats (up seven points from last time), and now a 21 percent chance of exactly 50 seats (down four from last time). The sum is 99 percent, which suggests that these markets are being traded heavily enough to carry some weight.
How credible do you think those market numbers are? Do you agree or disagree with my state analyses above?
- Senate Watch: September 1 (logarchism.com)
- Poll: Akin surging behind McCaskill (voices.kansascity.com)