With all the attention being placed on the battle for the House in 2022, realistic analysis of the battle for the Senate has been lacking, so I tried to provide some at New York:
Political handicappers looking to the 2022 midterms have focused on House races because the very predictable pattern of midterm House losses by the president’s party makes continuation of a Democratic House a real long shot (and probably a prohibitive long shot unless Joe Biden’s job-approval rating shows significant improvement soon). The loss of either chamber, of course, means the governing trifecta that has made enactment of part of Biden’s legislative agenda possible will be gone, probably for a good while (at least until 2026, by my reckoning). But there is some independent value in continued Democratic control of the Senate thanks to that chamber’s role in confirming Biden’s executive branch and judicial nominees along with the ability to control committee and floor action in a way that gives Democrats significant leverage and opportunities for conveying their message.
Because only one-third of the Senate is up for reelection every two years, there is not the sort of predictable relationship between Senate outcomes and the general political climate. In other words, a bad year for either party in presidential, House, or gubernatorial contests doesn’t mean a bad year in Senate races if the landscape is positive. We saw that most recently in 2018, when Republicans lost 41 net House seats and seven net governorships yet picked up two net Senate seats because the landscape (with 26 Democratic Senate seats and only nine Republican Senate seats at stake) was very positive for the GOP.
The Senate landscape is modestly positive in 2022 for Democrats, who have to defend only 14 seats as compared with 20 seats for Republicans. Moreover, as Amy Walter points out, none of the 14 Democratic seats are in a state carried by Donald Trump in 2020. Meanwhile, Republicans are defending two seats in states carried by Biden in 2020, Pennsylvania and Wisconsin.
But at the same time, Democrats are defending three Senate seats (in Arizona, Georgia, and Nevada) in states Biden carried very narrowly (he won by 0.30 percent in Arizona, 0.24 percent in Georgia, and a relatively luxurious 2.39 percent in Nevada). Republicans in the two nominally blue states whose Senate seats they control don’t have much ground to make up, either (Biden won Pennsylvania by 1.17 percent and Wisconsin by 0.63 percent). They also control an open seat in North Carolina, a state Trump won by only 1.3 percent.
To give you an idea of how much “swing” Republicans might rationally expect in a midterm, consider that Republicans won the national House popular vote by 1.1 percent in 2016 and Democrats won it by 8.6 percent in 2018. That’s a lot of movement against the party controlling the White House. Anything remotely like that in 2022 — again, controlling for state aberrations despite the trend toward straight-ticket voting in recent years — and Republicans could pretty easily sweep the six contests mentioned above, all rated as toss-ups by the Cook Political Report, and take control of the Senate by a 53-to-47 margin, assuming neither party breaks serve by winning in a less competitive state.
What may give Democrats better Senate odds is the current nature of Republican intrastate and intraparty dynamics. There are potentially fractious GOP Senate primaries in Arizona, North Carolina, and Pennsylvania that could produce nominees with real weaknesses. Moreover, in these states and others (notably Ohio, a “red” state that recently reelected a progressive Democratic senator), Trump’s insistence on turning GOP primaries into referenda on loyalty to his ludicrous 2020 election claims could interfere with the expected pro-Republican midterm trend.
Potential Trump-generated problems affecting Senate races aren’t limited to his involvement in just those races. Georgia is a classic example. Freshman Democratic Senator Raphael Warnock, who along with Jon Ossoff won by an eyelash in 2021’s unique dual general-election Senate runoff in what has become the ultimate battleground state, ought to be a sitting duck in 2022 with even a minimal midterm swing. But Trump enormously complicated Georgia politics by pushing the man Ossoff beat a year ago, David Perdue, into a primary challenge to the incumbent governor, Brian Kemp, as part of a purge effort aimed at those who didn’t support the 45th president’s efforts to overturn the 2020 election results. The Perdue-Kemp primary is sure to be an extremely expensive and divisive affair. It could weaken the ultimate winner in a general election against Stacey Abrams and might spill over into the Senate race, where Republican front-runner and Trump favorite Herschel Walker hasn’t shaken questions about his background and temperament (or rid himself of primary opposition).
Divisive Republican gubernatorial primaries seem likely in Arizona and Pennsylvania, as well, and could extend to Wisconsin, where incumbent GOP Senator Ron Johnson is struggling with low favorability numbers.
Republicans should be considered the slight favorites to flip the Senate (and much stronger favorites to flip the House) in 2022, assuming Biden’s popularity doesn’t seriously improve by November. But Mitch McConnell should not be making big plans for 2023. His party’s lord and master, Trump, could screw things up yet, and you never know entirely what will happen in a wide array of competitive Senate races.
The interesting thing about the ABC News poll is that it shows union household figures for Bush and Kerry, and something else I’ve not seen before: “Contact by the campaign” numbers.
The ABC News poll discusses the degree of partisanship in Wisconsin itself, as opposed to general arguments based on the 39D-35R numbers usually discussed on this site. The ABC News memo (see the section titled “Party” on p. 6) discusses how partisanship on election day has varied widely in Wisconsin — from +6 for Dems in 2000 to only +1 for Dems in 1996. Average it out (I wish ABC News had provided partisan numbers for 1992 and 1988 also, but oh well…) and that’s only +3.5% for the Dems.
Not much of a cushion. Look at the numbers: It makes Bush’s lead in the highly Republican Milwaukee suburbs +25, and Bush’s lead in the Northeast (Green Bay) +22. It’s heartening to see Kerry polling extremely strongy in the Southwest (Racine over to Walworth county). But it’s shocking to see Kerry polling LOWER in Milwaukee proper than Kerry is polling in the Southwest.
The ABC News poll says Kerry is up four points over Bush among union households. I’m hearing numbers that say that in south and west Milwaukee, as much as a quarter of all union households remain “undecided.” In some heavily union precincts, as many as 30%-35% of union households remained “undecided”! In an election with a virulently anti-union president (as Bush is), it is astonishing that union households would be anywhere this ambivalent about John Kerry as an alternative to George W. Bush.
Worse: Union turnout for Al Gore in 2000 proved decisive in pushing Wisconsin into the Democratic category. The ABC News poll points out that Gore won union voters by 16 points. John Kerry’s numbers are nowhere near that.
What makes me worried about Kerry’s campaign in Wisconsin is the following: “Registered voters are six points more apt to have been contacted by his campaign than by John Kerry’s, 25 to 19 percent. And six in 10 of those reached by Bush’s campaign support him, while Kerry’s supported by fewer than half of the Wisconsin voters his campaign has personally contacted.”
I don’t know what “personally contacted” means. Telephone? Face-to-face? In on Labor Day, many voters were saying that they had already burned out on telephone contacts, and had received almost no face-to-face contact (canvassing). Indeed, only progressive groups (League of Conservation votes, unions, etc.) had done face-to-face canvassing.
In one way, this result can be taken to mean that the Bush campaign is contacting random votes and finding massive support for Bush, while the Kerry campaign does the same and finds little support.
But, another interpretation could be that the Bush campaign is contacting its base. Meanwhile, the Kerry campaign is contacting its base but also attacking the Bush people. We’re given no analysis of whether the Kerry contacts are changing minds.
But the idea that the Bush campaign is more active in Wisconsin than Kerry’s makes me very nervous. (Will the Dems blame former Kerry campaign co-chair Matt Flynn for abandoning the Kerry ship for a Quixotic run at Congress, if Kerry loses?)
Smooth Jazz-
You do not give the source for that long article you posted. I gather it’s from someone on the right given (a) pointing out Democratic-related firms, but not Republican-related firms; and (b) various snarky comments about Democrats.
Makes me wonder about the validity of the analysis overall.
I will note one thing that stood out to me. There’s a 15% response rate for the NBC survey. I’d guess that’s fairly typical. If so, it really makes me wonder about these surveys. Some of the 85% non-response is doubtless because of random numbers that are not in fact residences. But with such a high non-response rate, small variations in people choosing not to respond would seem able to produce large variations in poll results.
I read some negative comment on RDD in an academic piece sometime in the last week or two, which also makes me wonder whether that is, in fact, a sufficient methodology for getting a random sample. It’s 9 pm here on a Friday night, so I don’t think I’m going to go searching right now. Perhaps over the weekend.
And who’s cherry-picking? I’d note that I used all the polls I could find about Wisconsin. It’s the sort of argumentation that you engage in that continually reminds me why I can’t vote Republican. And I came pretty close in ’00, when I thought about going for McCain, till Bush decided to trash him.
Loooking over the post above from Smooth Jazz:
Interesting that so many polls conducted with such similar methodologies should be coming up with such wildy different results. Although, I do note that a lot of the apparent discrepencies between polls that we have seen lately tend to be more in the “LV” subsamples than the unadulterated RV segments and the article really doesn’t get into LV screens to any meaningful extent.
So I would say this article tends to back up Mr. Tiexera’s assertion that we should give more credence to the RV data wherever available and poke the LV data with a long stick before going anywhere near it.
ABC and Gallup put Bush ahead in Wisconsin and West Virginia? That means it’s a dead heat.
Part of the problem in Wisconsin results from the poor governance by Democrats as Mayor of Milwaukee and Milwaukee County Executive, alienating goo-goos and giving suburban Republicans stories to frighten their children. We’re not seeing the movement toward Democrats in the Milwaukee suburbs that shows up very clearly in the Chicago, Philadelphia, and New York City suburbs over the past 8 years.
I’ve lived in Madison for 27 years, and I would not be surprised to see Bush win here if the national election is very close. But if Kerry winds up winning nationally by 3 or 4 points, Wisconsin will very likely remain Democratic.
The last 2 times Zogby polled Wisconsin (Sep 17th and Sep 20th), they found Kerry ahead by 2.
Well Rick, if you assume that Gallup and ABC oversample Republicans in their state polls (as they did in their national polls) then these Bush leads are probably not as large as they seem, or even extant at all.
Penn and Wisconsin are two battleground states with among the worst-maintained lists of registered voters, with many erroneously purged, according to a recent study by Scripps Howard News Service.
Some other items pulled as well from Joshua Kurlantzick’s eye-popping piece, “2000, The Sequel”, in the current (Oct) issue of The American Prospect, on how the Help America Vote Act (HAVA), passed by Congress in 2002, may make things even worse than 2000:
(Link is: http://www.prospect.org/web/page.ww?section=root&name=ViewPrint&articleId=8544)
*Rep. Rush Holt (D-NJ) introduced an amendment to HAVA in May ’03 that would require touchscreen machines to have a paper record. Bob Ney, Chair of the House Administration Committee, has not allowed Holt’s legislation out of committee. House and Senate Republicans have offered similar “smokescreen” legislation–that would take effect by the 2006 election.
*There has been major WH footdragging on getting up and running the Election Assistance Commission (EAC), authorized under HAVA to disburse money to states to upgrade voting systems, issue guidelines, and hold hearings to help make voting as fair as possible. EAC was supposed to be set up within 120 days of HAVA’s passage but did not even have office space until April of this year. Congress gave the EAC $2 million out of a proposed $10 million budget for ’04 and Bush asked for less than half of the $1 billion proposed in HAVA for overall election-reform efforts.
*Under HAVA all first-time voters have to have valid ID. NY PIRG says it is illegal for local election boards to tell poll workers not to accept a student ID as proof. U Wisc and Penn State Students for Kerry, do you copy? One study showed that in NY election officials in only 18 of 45 counties even understood voter-ID requirements.
*Some states simply do not count “provisional ballots”, intended to give those wrongly denied the opportunity to cast a vote on Election Day via the usual process the chance to have their votes tallied if their wrongful exclusion is established. HAVA offers no national guidelines for counting these provisional votes. This raises just a few questions for state and local election officials and newspaper editorial boards, among others.
*After 2000, Jeb Bush rejected the recommendation of a bipartisan task force that he make state and county election supervisors nonpartisan. Shortly afterward, the Republican-controlled Florida legislature made all top election officials appointees of the governor.
I have not seen these specific items, several strongly suggesting the GOP does not want the sorts of problems that turned up in 2000 and since to be fixed, reported by any of the media bigfoots. Another shocker.
A bit OT, but I thought I’d post a version of a comment I made on the Mystery Pollster site, regarding his argument that Party ID was quite unstable, and therefore should not be used as a weighting factor. In fact, as I point out, the results he quotes from studies deal ONLY with increases in the number of INDEPENDENTS:
——————————————————–
One significant weakness in your [i.e., the Mystery Pollster’s] logic, as you present it, is that you don’t get at the nerve of Ruy’s argument, which assumes that the DIFFERENCE between party identification for Republicans and that for Democrats has been stable.
The particular research you’ve quoted demonstrates something VERY different, namely that the number of INDEPENDENTS has varied a great deal. This is completely consistent with the idea that the difference between Republican PID and Democratic PID are stable. And, indeed, it stands to reason that a convention might polarize independents, drive them equally in both directions into the hands of the ideologically nearest party.
Indeed, one of the very studies you quote says that the largest amount of variation in Party ID comes from movement in and out of independence. It seems very plausible to believe that politically potent events or circumstances might in general simply polarize independents, leaving the difference essentially intact.
If you have any evidence that the DIFFERENCE between Rep and Dem Party ID varies as wildly as we’ve seen across these polls, that would be very good.
Otherwise, you’ve proved exactly zero.
Another note on Wisconsin. In the ABC poll, the breakdown was 35% R, 29% D, 36% I/Other. In the Moore (GOP) poll, it was 39% R, 39% D, 22% I/Other. In 2000 it was 32% R, 37% D, 31% I/Other.
So…it’s back to that issue we keep debating. Has there really been a sizeable shift to the Republicans in registration, or is something else going on. I lean toward the latter, but I don’t think we’ll know till election day.
Rick-
I don’t know what to make of Wisconsin. As with national polls, I’d point out variability. As best I can tell, there are 10 polls of Wisconsin released 9/12 or later:
ABC, Bush +6
Moore (GOP), Bush +3
Badger, Bush +14
TNS, Bush +10
Zogby, Kerry +2
Mason-Dixon, Bush +2
ARG, tied
Strategic vision (GOP), Bush +6
Gallup, Bush +8
Rasmussen, Bush +2
The median for those is Bush up by 4.5%. That’s probably not a terribly bad guess. That’s quite a recoverable lead, and then there’s turnout.
Just keep saying the mantra…it’s a close race. Let’s work hard to get Kerry the victory. And say it if the polls show us ahead, tied, or behind.
What do we make of Wisconsin? This looks like a point of concern.