Split Ticket

Our forecast this week sees Kamala Harris’ odds tick down a notch, from 55% to 53%. Readers may note that this marks the fourth straight week of decline from her peak of 62% when the model launched, which happened fresh off a debate that Harris was widely seen as having won.

This begs the question: has there been tightening in this race?

Other models certainly think the answer is “yes”; for the first time in weeks, both Silver Bulletin and FiveThirtyEight now have Trump as a nominal favorite. Ours has moved around less than theirs, but it does think that there has been a dash of movement against Harris.

This is driven mostly by a minor amount of attrition in the Blue Wall states of Michigan, Wisconsin, and Pennsylvania; in the former two states, her lead has actually shrunk by two points. That, along with Trump slightly widening his lead in Georgia and especially Arizona, has dragged the race back to a pure tossup.

It’s important to note that polling shifts can be fleeting, and that they can be driven as much by coverage and response bias as they are by actual change in voter preferences. Sometimes, it can be impossible to disentangle the two. It’s also worth pointing out that we’ve had a surprising paucity of high-quality polling of late, especially relative to the time of year. But our model accounts for that, and the few high-quality surveys we do have mostly show some degree of state-level movement against Harris in Wisconsin and Michigan.

These shifts are still fairly minor on the whole, and we remain very open to the possibility that this is noise, or that it’s a coverage or differential response-related bounce that could reverse as the election nears. In fact, depending on your priors, you may credibly argue that this is the likely outcome. But models like ours are about making inferences from polling. And at this point, the data suggests that there has been a minor amount of tightening in core battlegrounds from the post-debate high that Harris enjoyed.

If you forced us to pick a winner from this incredibly uncertain picture, we’d still probably pick Harris, simply because she continues to narrowly lead polling in our averages, while also maintaining a commanding edge in head-to-head favorables. But it’s not as clear of a choice as the one we would have made a month ago (though the data still doesn’t quite agree with the full-blown Democratic panic of late).

We want to stress that the choice of words we used was uncertain rather than close for a reason. As we’ve mentioned before, you should expect polls to be wrong — polling overestimated Republicans by 3 points in 2012 before overshooting Democrats by 3 and 5 points in 2016 and 2020, respectively. The issue is just that we don’t know which way things will break ahead of time.

But before you jump to perhaps the most obvious conclusion, we’d caution that it’s not as simple as saying polls will continue to undercount Trump supporters. There are a variety of indicators that suggest this time might be different.

We’ll talk more about those next week.

House

As of last week, Democrats had a 52% chance of winning the House, with their shrinking lead on the generic ballot accompanying the tightening in odds. In today’s update, these probabilities are similar: Democrats now have a 53% chance of retaking the House from the GOP.

The race for control is truly anyone’s game; by our estimates, Democrats are favored in 212 seats, Republicans are favored in 209, and 14 seats are tossups. The House and the presidency are extremely likely to track closely with each other, as our forecasts have now converged, and whichever party wins the presidency is also exceptionally likely to win the majority. A polling miss in either direction is likely to carry over to downballot races, dragging the president-elect’s party over 218 House seats.

Following the second to last pre-election quarterly reporting deadline, we incorporated the latest fundraising filing reports into our model. This week’s temperature check reviews some of the closest races with the biggest total spending deltas between the parties. These are the seats where a last-minute injection of funds by one party or the other could make a difference both in terms of the district winner and the party that ends up controlling the House as a whole.

Let’s start with Washington’s 3rd district, a Trump-won district where freshman Democrat Marie Gluesenkamp Perez is facing a tough rematch against Republican Joe Kent (who finished ahead of incumbent Republican Jaime Herrera Beutler in the 2022 jungle primary, following her vote in favor of President Trump’s second impeachment).

Gluesenkamp Perez won the Trump +4 seat by less than one point in 2022, and our model currently gives her a 52% chance of winning reelection with a predicted margin of 0.3%. While Perez outperformed expectations by roughly 5 points in 2022, the slim advantage she currently has actually stems from the huge Democratic spending advantage in the seat. In total, Democrats have outspent Republicans in WA-03 by over $8 million, which is the biggest delta out of all the tossup seats. We certainly do not expect the GOP to close the gap here, but narrowing it could make the difference between a win and a loss for Kent.

The story is similar in another Trump-won district: Pennsylvania’s 8th. Democrat Matt Cartwright has held different versions of this seat in presidential years before and our model expects him to do so again this year, giving him a 56% chance of winning reelection with a predicted margin of 1%. Still, the seat is clearly within striking distance for Republican Rob Bresnahan, who simply lags due to an immense cash deficit of nearly $5 million. Trimming the massive spending gap could make a meaningful difference here, especially if Trump exceeds expectations in Pennsylvania.

Two Republican incumbents who are currently projected to lose have also been convincingly outspent. One, Ken Calvert, is in as close to a pure tossup race as one could be. Calvert had a closer than expected victory over Will Rollins in 2022 and is now in a dogfight rematch. Democrats have amassed a considerable spending advantage in California’s 41st district, but Rollins still only has a 51% chance of winning, so it’s easy to see how this contest could tip back into the GOP column if the gap tightens.

California’s 45th presents the opposite case. There, Republican Michelle Steel is facing an extremely close reelection against Democrat Derek Tran, and she currently has just a 51% chance of winning. Steel would be in considerably worse shape if Tran was at financial parity, but Republicans maintain a substantial spending advantage of over $5 million in this seat, making it extremely difficult for Democrats to take the edge.

Another district in which Democrats face a serious cash disadvantage is New Jersey’s 7th, where freshman representative Tom Kean, Jr. is locked in a tight race against Sue Altman. Both candidates have matched each other blow for blow in fundraising, and a high-quality poll from Monmouth University recently showed a tie in the seat.

Yet despite virtually all signs pointing to a very close election in this highly-educated, Biden-won district, national Democrats have chosen to triage this seat altogether, sending virtually no money to Altman despite Kean receiving a heavy helping of national GOP assistance. This has allowed the incumbent to remain a 63% favorite in our forecast.

On the face of it, this decision seems inexplicable. On paper, Kean is a more vulnerable candidate than many of the other representatives that Democrats have chosen to target (such as Monica de la Cruz in Texas’ 15th and Jen Kiggans in Virginia’s 2nd). There is no logical explanation for their decision to abandon Altman, but if the House ends up being as close as it currently looks, Democrats could come to rue the decision. And with only 14 tossups remaining, any future development to tip each seat in either direction could be crucial to building a House majority for the 119th Congress.

Senate

The race for Senate control remains mostly unchanged; though Democrats have outside chances in Nebraska, Texas, and Florida, Republicans remain firm favorites in 51 seats and are extremely likely to win back the majority this cycle. Though Sherrod Brown has held up remarkably well in Ohio against Bernie Moreno, data points to the Senate seat in Montana returning to the Republican column this cycle.

If Democrats want to play for a majority, they would be wise to invest in Texas (or figure out how to boost Independent Dan Osborn in Nebraska without directly associating with him). However, a look at the expenditures of the Democratic Senatorial Campaign Committee and the Chuck Schumer-affiliated Senate Majority PAC suggests national Democrats view Jon Tester’s odds more favorably than our model does, because while they have sent $60M to Montana, they have not even bothered to invest in Texas.

We’ll see if that bet turns out to be correct.

Governors

This year’s gubernatorial races have remained somewhat sleepy affairs. There is no evidence anyone believes embattled Republican Mark Robinson can win in North Carolina. The polls back this up, with Robinson consistently trailing by double-digits. New Hampshire, the only truly competitive race in the country, still looks to be a nail-biter. 

However, we do have one ratings change to make: Indiana is moving from Safe Republican to Likely Republican. Senator Mike Braun remains a strong favorite to keep the governorship in GOP hands, but his path to victory has been somewhat complicated by Donald Rainwater, a Libertarian who is polling in the upper single-digits.

In 2020, Rainwater won 11% of the vote, but incumbent Gov. Eric Holcomb easily won 57–32% anyway, thanks to substantial crossover appeal among Democrats. In rural counties, Libertarian support was even higher, with Rainwater coming in second place in 33 counties. Rainwater’s unusually strong performance was attributed to right-wing frustration with Holcomb’s COVID restrictions.

This year, with the pandemic long in the rear-view mirror, Rainwater doesn’t appear to be as strong as he was four years ago. Nevertheless, he remains a potential issue for Braun, who lacks the crossover appeal that Holcomb had. The Democratic nominee, Jennifer McCormick, was elected statewide as Superintendent of Public Instruction in 2016, albeit as a Republican. She switched parties in 2021, and, sensing an opportunity, national Democrats have invested in the race.

Third party candidates usually poll higher than their eventual result, as partisans come home in the end. Additionally, Rainwater’s voters are not entirely Republican (third party voters are often idiosyncratic in unusual ways), although he is likely pulling from a universe of voters who would otherwise be inclined toward Braun. The likeliest effect of Rainwater’s candidacy is that it lowers the threshold of victory for McCormick. Instead of needing 50%, she may be able to win with 46% or 47%.

We remain skeptical that a Democrat can get to that level of support in Indiana, but the possibility cannot be foreclosed entirely. Likely Republican.