How the Races Are Breaking at the End

The wave has put Republicans within striking distance of 54 Senate seats, but the gubernatorial map is a lot less clear.

I t is Election Day at long last. Barring a runoff in Georgia, there will be no more Election 2022 polling. And while we won’t know all the results by this time tomorrow, we’ll know a lot more than we do now.

Throughout the fall campaign, I have adhered to the same view of elections for the Senate and governorships — particularly in midterm cycles — that I’ve held since 2014. In short:

• Mid September is really the start, not the end, of serious campaigning; we should not expect most races to remain static from then until Election Day.
• The national environment — the president’s approval rating, the generic-ballot polling, and to a lesser extent voters’ views on the direction of the country — tends to exert a powerful gravitational pull that drags most races in the same direction. This gravitational pull tends to have a stronger effect on Senate races than on gubernatorial races.
• A crucial factor is not just the size of a candidate’s polling lead heading into Election Day but how many voters remain undecided, as we can typically expect undecideds to break in the direction of the gravitational pull — i.e., the wave.
• Even the final polls are not the last word; there are likely to be some races that break hard at the end, as there were in 2018 and many other recent cycles. There have been 33 upsets since 2010, races in which the final polling leader lost, and several of those losers had a lead of three points or more on Election Day. We can expect, if history holds, two or three such upsets this year.

In mid September 2014, I concluded from projecting the historic trends that these dynamics pointed in the direction of an eight- to ten-seat pickup for Republicans in the Senate. That outcome was widely scoffed at by many of the leading poll analysts at the time, but Republicans gained nine seats.

I have continued to update the progress of the polls across all of these races here, here, here, here, here, and here. One of my frameworks is the distinction between Republican “Wave Surfers,” who could bring their races into a statistical tie or lead in the polls by the end, and the “Beached Boats,” who were going to fall short in the polls but be close enough to have hope of pulling off one of the upsets.

So, how have things broken at the end of the cycle?

Leaving aside Alaska, which at this point is utterly impossible to project, it appears that every Republican incumbent should be safe, and four of the five Republican-held open seats (plus the Oklahoma special Senate election) are effectively locked down, with only Pennsylvania still seriously in doubt. Democrats have likewise locked down their one open seat, in Vermont.

Republicans now look solid in Nevada and Georgia (although in the latter Herschel Walker will still need a late break in his favor to avoid a runoff), have narrow leads in Arizona and Pennsylvania (both races where they have long trailed), and are just 1.4 points back in New Hampshire. In none of these are races would it be much of a surprise to see the Republican candidate win. If all of them were Republican wins, that would yield 54 GOP Senate seats, which is what Henry Olsen at the Washington Post is predicting. At this point, I’d say that the odds of his being right are about even, although I think as a betting man, I’d probably err on the side of caution and predict Republicans will end up at 52 or 53.

At the outer limit of Republican hopes for the Senate, Washington looks like the more plausible upset than Colorado. Tiffany Smiley and Joe O’Dea have both run good campaigns against entrenched incumbents. I had higher hopes for O’Dea than for Smiley: Michael Bennet has never cracked 50 percent of the vote, Colorado isn’t as blue as Washington, and Republicans won a Senate seat there as recently as 2014. But Bennet seems to have proven a tougher nut to crack than expected, perhaps helped by the popularity of Governor Jared Polis, while Murray goes into today still below 50 percent in the polls, and she is alone at the top of her ticket. Democrats have acted as if Murray is in trouble and Bennet is not. Either state would be a gift for Republicans, but neither seems particularly likely to yield an upset.

It’s when we turn to the shifts since mid September that we see the full power of the wave. Of the 20 races for which we have polling both before and since mid September, 14 have shifted at least three points towards the Republican candidate, and 17 have shifted at least two points in that direction; the only non-trivial shift towards the Democrat was in Vermont, where the hopes Republicans were given by a single Trafalgar poll in early September proved illusory. And most of the net movement has come from added support to the Republican column rather than declines in the Democrats’ support, which is entirely consistent with the historical pattern of wave elections breaking towards the out-of-power party.

It is also consistent with two other dynamics we have seen in 2022.

One dynamic is that the Democrats consolidated their base vote in the summer after the Dobbs decision was issued, Biden got the Inflation Reduction Act passed, and gas prices declined somewhat from their high in early July. Most of those voters were likely to come home to Democrats anyway, but proof that Biden wasn’t entirely inert helped lift morale, and the abortion issue fired up the party’s base, which had become depressed and disenchanted. Unfortunately for Democrats, they confused this with a big counter-wave of swing voters in their direction, and many of their candidates have hit the polling ceiling well south of 50 percent.

The other dynamic is that Republicans had some divisive primaries that yielded unimpressive, uninspiring, and often inexperienced candidates. Candidates such as Walker, Smiley, O’Dea, J. D. Vance, Mehmet Oz, Blake Masters, and Don Bolduc needed to introduce themselves to the public (or, in the case of celebrities such as Oz and Walker, reintroduce themselves as politicians), mend fences with their primary opponents, and get past rookie mistakes on the campaign trail. Eric Schmitt, who already holds statewide office in Missouri, got to 50 percent in the polls much earlier than many of these newcomers, despite having survived a nasty primary contest. Adam Laxalt, who holds statewide office in Nevada, also started off in better shape than some of the rookie candidates but has not gained as much ground since. As we went into the stretch run, voters simply either warmed more to these candidates or just decided to hold their noses and throw the current bums out in favor of new bums. That’s how wave elections work.

The races for governor are less inherently national, which means that Republicans have ended up short at the end in more of them. I expect some upsets in these races, but with (depending how you count them) between seven and nine contested races in which Democrats lead heading into Election Day — all of them in states with Democrat incumbents or an open seat held by a Democrat — it will take a major surprise for Republicans to flip more than a couple of them.

Democrats are now certain to gain two open-seat governorships currently held by Republicans, in Maryland and Massachusetts. They could still surprise with a third if Kevin Stitt falters in Oklahoma, although late polling has made him look safer than he once did, and Oklahoma is not Maryland. In Arizona, Kari Lake seems to have crossed into safe territory in the other closely contested race for a Republican-held open seat. Republicans are certain to retain the open governorship in Arkansas. Despite the absence of incumbents, nobody has bothered to poll either Nebraska or Hawaii, so we must assume those will go the way of the state’s partisan tilt, as well.

If Republicans are down two governorships, then the question becomes: How many can they pick up? Nevada and Wisconsin feature the most imperiled Democrat incumbents. Michigan, Oregon, Kansas, New Mexico, and (maybe) Minnesota look like wild cards. Kansas has been lightly polled and the final Michigan polling puts Tudor Dixon closer to striking distance. By contrast, Christine Drazan has led for months in Oregon, but has sagged in the last two polls as supporters of third-party candidate Betsy Johnson seem to be abandoning her in droves for Democrat Tina Kotek. Lee Zeldin, for all of his momentum and all of the visible Democratic panic, lethargy, and disorganization in New York, must still hope for a stunning upset that the polls just don’t see coming.

The overall picture of the trend lines in the governor’s races since mid September is not uniform. Many powerful Republican incumbents have pulled away in governor’s races: Ron DeSantis, Mike DeWine, Greg Abbott, and Brian Kemp have all assured themselves of comfortable reelections. Elsewhere, little-known and often-neophyte Republican candidates have introduced themselves in Arizona and tapped into dissatisfaction with the incumbents in Minnesota, Michigan, New York, New Mexico, Wisconsin, and Nevada. In Pennsylvania, I expect Doug Mastriano to do better on Election Day than his polling would suggest, but he clearly has failed to capitalize on the same dynamics that have buoyed other Republican gubernatorial candidates.

If I had to place a bet, I’d say Republicans net a gain of two governorships. But substantially more than that could be in the cards if the final numbers are just a few points to the right of the final polls across the board. I can easily imagine a scenario in which Republicans lose eight or ten races for governor by two to three points, but it is not that big a stretch to picture something that looks more like a massacre of Democrat incumbents, either.

All that’s left now is to vote (if you haven’t already) and wait.