The red wave builds in the Senate races, but so far, not a red tsunami.
The red wave builds in the Senate races, but so far, not a red tsunami.
T his year’s midterm elections will be held in wave-election conditions, at least by the broadest definition of the term: It’s a first presidential midterm, the president’s party holds both houses of Congress, the president’s approval rating has been double-digits underwater all year, and the party out of power regained the lead in the generic ballot in late September and has been widening it since. That bodes well for Republican Senate candidates down the campaign’s home stretch, and the direction of the polls over the past month tends to confirm this.
The Red Tide
Look at the RealClearPolitics poll averages, as I will throughout this article. One advantage of the RCP averages is that we have them going back several years, so it is feasible to compare their predictive power over time. At this writing, Joe Biden’s approval rating is eleven points underwater (43.1 percent approval, 54.3 percent disapproval) — improved from his midsummer nadir, but the worst it has been since early September. Presidential approval is typically the strongest predictor of how a party’s fortunes will go in a midterm election, especially in Senate and House races.
The Republican lead in the generic congressional ballot is 2.8 points (47.8 percent Republican, 45 percent Democrat), the biggest Republican lead since late June, but also not matching the GOP high-water marks of the first half of 2022. The Republican lead in the generic ballot is encouraging, but it trails where it stood at this point in 2014 (a 3.4 point lead) or 2010 (a 6.9 point lead). Notably, Republicans underperformed the generic ballot in the 2018 Democratic wave by 1.1 points and overperformed it by 3.3 points in the 2014 Republican wave, although they underperformed by 2.6 points in the 2010 Republican wave. In the last three presidential elections, by comparison, Republicans overperformed the generic ballot by 3.7 points in 2020 and 1.7 points in 2016, but underperformed it by 1.4 points in 2012, in each case reflecting presidential-year turnout factors that are absent in the midterms. In short, the generic ballot is likely to be off a few points this year — we can’t predict in which direction — but it is likelier to underestimate Republicans both because of a tendency to favor Democrats and because wave turnout tends to run ahead of the general poll.
Historical Trends, Recent Deviations
This points to the broader trend: Historically, midterm wave conditions typically mean that we should expect most of the contested races to break after mid September in the direction of the out-party: the Republicans. That was dramatically illustrated in 2014, when Republicans won nine senate seats, an outcome few of the professional election analysts saw coming around Labor Day.
That seemingly iron rule also didn’t work as consistently in Trump-era elections, suggesting either that deepening partisanship is narrowing the playing field or that polls are less reliable in capturing a segment of the electorate that trends Republican. In 2016, every single Senate race ended up going the same way as the presidential race in that state — the first time that had ever happened in the era of popular Senate elections — including races where Republicans trailed in the fall. True, 2016 wasn’t really a wave year: Barack Obama’s approval rating was positive by the fall, but his party at the presidential level faced the powerful undertow of the historical trend against parties following the reelection of an incumbent.
Then, there was 2018 — a Democratic midterm wave election by any definition. And yet, Republicans did something unprecedented: They unseated four incumbent Democratic senators into the teeth of the Democratic wave. While three of the four had already trailed in mid September, in three of the four races (Missouri, North Dakota, and Indiana), the Republican finished in a better place than where the mid-September polls stood. The same was true of an open-seat race in Tennessee and a closely fought race won by a Democratic incumbent in Michigan. By contrast, the normal wave dynamics held in Arizona.
Polling errors severely underestimating the Republicans were doubtless a factor in some of these races: Republicans vastly outperformed the final polls in Tennessee, Missouri, Michigan, Indiana, Florida, and West Virginia. Parallel polling errors in the governors’ races suggest that the pollsters just got Florida, Tennessee, and Ohio wrong. But the late, Senate-focused Brett Kavanaugh hearings also appear to have driven the trend.
Then, we had 2020, which, again, is hard to put precisely in a midterm-wave framework given that Donald Trump’s approval rating was mediocre, and he lost narrowly in a very-high-turnout election. The polls, again, were off across the board, mainly by underestimating Republicans in Maine, North Carolina, South Carolina, Kentucky, Kansas, Oklahoma, Arizona, Michigan, Iowa, Montana, Alabama, and Texas.
So, on the one hand, we have a historic trend that favors Republicans; on the other hand, we have a recent trend of polling underestimations that has also favored Republicans. None of this is especially encouraging news for Democrats, although some of them hold out hope that this year’s late-June abortion decision will have the same late effect as the late-September Kavanaugh hearings in 2018. Even then, the “Kavanaugh effect” mattered because crucial Senate races were being held in very-deep-red territory. That is true in this year’s governor’s races, but mostly not in the Senate races.
Republicans are targeting some Democratic senators in very-blue states (such as Patty Murray in Washington), but those are reaches. The big races are in 2020 swing states such as Georgia, Pennsylvania, Arizona, Nevada, Wisconsin, and North Carolina; light-blue states such as New Hampshire and Colorado; and red-leaning states such as Ohio, Iowa, Missouri, Utah, and Florida. It is, however, true that Republicans are defending 21 Senate seats while Democrats are defending only 14, more than half of which are far into blue territory. That places a hard upper limit on the possibility for Republican gains even if things break very strongly Republican at the end.
The Breakers
In past years, I have run what I call the “breakers report” across the major statewide polls. This is not, by any stretch, a scientific method; my goal is simply to present the state of the polling, how far it has moved since earlier in the race, and how far the Republicans need to go from the current poll averages in order to win. A healthy dose of skepticism based on experience, rather than some sort of formulaic adjustment, is recommended.
The basic concept is to look at the percentage of voters not committed to one of the two major-party candidates and ask how many of them would need to break for the Republican candidate to get to 50 percent. Of course, a bunch of cautions apply: Some of these are voters committed to third-party candidates, some will stay home, and people win Senate races all the time without getting to 50 percent of the vote. As I said, it’s a rule-of-thumb metric, not Science™.
The second thing I put on the chart is how many polls make up the current poll average, and when they completed their survey. For example, the Nevada and Pennsylvania poll averages both include six polls, all six of which were completed after September 15. By contrast, the poll “average” in Vermont is a single poll that was last in the field on September 7. With my customary apologies to color-blind readers, the chart is color-coded.
Also, bear in mind that a few states have distinctive rules. Alaska’s polling is complicated by the state’s new ranked-choice voting and jungle-primary system, under which there is a Democrat on the ballot, but the final choice is likely to be between incumbent liberal Republican Lisa Murkowski and conservative, Trump-backed Republican Kelly Tshibaka. The one poll we have, which is over a month old, shows that they would be tied. Georgia, of course, has a runoff system, so Herschel Walker and Raphael Warnock could easily go to a second round. Louisiana also has a jungle-primary system, although incumbent John Kennedy’s lead is prohibitive even in the single available poll. Then there are races nobody has bothered to poll, including an open-seat race in Alabama that is likely to be won by Republican Katie Boyd Britt. Seven of the eight races with no public polling are in red states, and other than Britt, they all feature Republican incumbents.
Walking through the list, there are twelve races that appear to be locked down: four Republicans and eight Democrats. If you’re an optimist about late momentum, you can tell a story where Republicans pull out Vermont, Washington, and Connecticut — Vermont hasn’t been polled since Trafalgar’s early-September poll, and in Connecticut, Leora Levy was down just five to Richard Blumenthal in the most recent poll — but those are still long shots.
That leaves 14 races that are seriously contested. Republicans plainly have the upper hand in five of those: Wisconsin, Missouri, Iowa, Kansas, and Florida. Also, Kansas and Utah each have very sparse polling to work with, and I’d be hesitant to describe Mike Lee as being in serious danger without more evidence. In Wisconsin, it is remarkable to see Ron Johnson ahead. In 2010, he pulled ahead in the polls in mid July, but in 2016, he trailed in the poll average all the way to the end and led in only a single poll (by Gravis, in early October), yet won by 3.4 points. Ted Budd in North Carolina and J. D. Vance in Ohio are still a ways from locking their races up, but each now has enough of a lead that they can win without banking on a big break in their direction.
Democrats, by contrast, have what looks like really solid leads in just two of the contested races, behind incumbents Michael Bennet in Colorado and Maggie Hassan in New Hampshire. Tightening of those races over the next round of polls would be a serious sign of a major wave.
That leaves us with the four races that are likeliest to decide control of the Senate. The good news for Republicans for some time has been Nevada, in which Adam Laxalt has led consistently over incumbent Democrat Catherine Cortez Masto. In the past, the state has tended to go more Democratic than the polls showed, but that was due in good part to Democratic strength with working-class Hispanic voters in Las Vegas who are hard to poll but easy to organize. Those are precisely the voters who are most open to Republican appeals this year in Nevada.
If Laxalt wins and all the other races go as currently polled, Republicans need only to go one for three in Pennsylvania, Georgia, and Arizona, all of which have been intensively and recently polled, and all of which Joe Biden won by a hair in 2020 after Donald Trump carried them in 2016. Republicans trail in all three with dubious candidates, and trail by enough to require around two-thirds of the undecided voters to break their way. That is where the real test of wave conditions comes in. It is also where it pays to look at the trend lines.
Red October
Looking at the movement in the polls over the past month — which limits us to the 18 races that had even been polled by mid September, then polled again — we see a picture that is mildly optimistic for Republicans, but not greatly so. The big moves at the edges are mostly in races that have not been very competitive or highly polled, such as in New York, Connecticut, and Oklahoma. Those aside, Ron Johnson has made the biggest move, with a net 4.5 point swing in his favor. Among the most hotly competitive races, there has been movement in the Republican direction of at least 0.7 points in North Carolina, Arizona, Pennsylvania, and Nevada, with Democrats really only gaining that kind of ground in a seriously close race in Georgia.
Also, not all poll moves are created equal. In Florida and Pennsylvania, both candidates have declined in the polls, which creates more room for late breaks. That might be better news for Val Demings if not for the history of poll underestimation of Republicans in Florida and the headwinds from the governor’s race, in which Republicans hold a fundraising edge of staggering proportions. Under those circumstances, Demings’s dropping below 42 percent is probably fatal. But in Pennsylvania, John Fetterman was closing in on the magic 50 percent line a month ago, and voters seem to be more hesitant now — a bad sign going into his late debate with Dr. Oz, which will contrast a veteran TV performer against a man struggling to speak and to process what he hears after a stroke.
The Senate races aren’t over. Whether Republicans still have a chance of a real wave — one that delivers more than a one- or two-seat net gain — depends on how many grains of salt you apply to the current polls. But the odds of a minimum net gain of one seat to take control of the chamber look fairly good, if you expect historical trends to hold.