Republicans seem to have the momentum heading into next Tuesday’s midterm elections. My read of the evidence is that the Republicans’ fundamental advantage as the out-of-power party in a midterm year will be partially offset by Democratic backlash to the Dobbs decision overturning Roe v. Wade. My best guess is that Republicans will retake the House, but Democrats will—just barely—hold on to the Senate.
Before I went on vacation in June, I argued Republicans were very likely retake the House in next Tuesday’s midterm elections. The president’s party almost always does badly in midterm election years. Inflation—which may be the thing voters hate most—is running at a 40 year high. Republicans were, not surprisingly, leading FiveThirtyEight’s generic congressional ballot average by more than two points. Because congressional maps broadly favor Republicans, Democrats would probably have to win the national popular vote by something like two points just to hold on to the House. A Republican wave looked more likely.
That changed with the Supreme Court’s decision in Dobbs v. Jackson Women’s Health Organization to overturn the 50-year-old precedent of Roe v. Wade guaranteeing a woman’s right to terminate her pregnancy. I knew the Supreme Court’s decision in Dobbs would galvanize Democratic turnout, but I thought it was unlikely to be one of the rare events that fundamentally shift the dynamic of an election. In the months that followed the decision, however, Democrats—buoyed by falling gas prices and a string of legislative successes—took a two point lead of their own on the generic congressional ballot. There was a surge in new voter registration in battleground states—particularly among young women—and Democratic candidates substantially overperformed in a series of special elections. It began to look like Democrats would turn out in unusually high numbers for an election they’d normally be expected to lose.
Inevitably, the surge in Democratic enthusiasm has faded over the last six weeks as the Dobbs decision gets further away and the election gets closer. Inflation is still high. Senate races that looked like they wouldn’t be competitive now look like dead heats. Republicans are now back to being up 1.2 points in the latest generic congressional ballot average. The two parties are dead even in Split Ticket’s average of nonpartisan pollster generic congressional ballot polls. With four days left before Election Day, the polls suggest we are somewhere between where we were in June and where we were in September.
Polls are at best imprecise measures of voting intentions, simply because it’s hard to accurately sample the voting population. Surveying voters is getting even harder as people become less likely to answer their phone. There are two major sources of uncertainty around this year’s polls. On one hand, the polls may not be accurately counting Republican voters, after having underestimated the strength of the Republican vote in 2016 and 2020. It’s possible that certain Republican voters are simply hard for pollsters to reach. But pollsters work hard to address this kind of sampling problem, so historically the polls don’t tend to miss the same way over and over again—and anyway the polls were fairly accurate in the last midterm election. On the other hand, polls may be underestimating the surge in enthusiasm among new, young Democratic voters, who may be similarly hard to reach and hard to model. The current Republican lead on the generic congressional ballot is less than the amount the polling average is usually off—the final average missed the popular vote margin by 4.2 points in 2020. Another error like that in favor of Republicans would mean a decisive Republican victory, while the same error in the Democrats favor would be enough for them to hold on to the House. Based on polling alone, the election could go either way.
My best guess is—while a range of outcomes is plausible—that the election will be close. Republicans are clear, but not overwhelming favorites. Their modest lead in the generic congressional ballot is consistent with the structural advantage they have as the party out of power being somewhat offset by the backlash to the Dobbs decision. In other words, Republicans are likely to do less well than their fundamental advantages would suggest because Democrats are likely to turn out in larger numbers than they usually do for midterm elections when their party holds the presidency. The Republican advantage is also likely to be somewhat offset—particularly in a few key senate races—by the fact that Republicans are running exceptionally poor quality candidates in competitive states. The main question is whether Republicans will underperform the fundamentals by a little—and still manage a decisive victory—or a lot—and fail to take control of one or both chambers of Congress.
According to 270toWin’s consensus model of election forecasters, there are 222 House races that lean toward Republicans, 200 races that lean toward Democrats, and 13 tossups. If the parties win the races that are expected to go their way and split the tossups, Republicans will pick up 16-17 seats. That would be a relatively modest gain by historical standards, but it would be enough to win the House. 270toWin’s consensus model likewise suggests the median outcome is the Senate remaining evenly split and, thanks to Vice President Harris’ tie-breaking vote, in Democratic hands. If the polls are off in either direction, one party could win almost all the tossups as well as a number of races the other party was favored to win. That means that something like eight senate races could go to either party. Republicans could plausibly win over 240 House seats, but Democrats could also plausibly retain control of the House.
Political betting and prediction markets (Betfair, Good Judgment Open, Insight, Manifold, Metaculus, PredictIt) think Republicans are somewhat more likely to win one or both chambers of Congress than forecasting models (Decision Desk, Economist, FiveThirtyEight) do, although both agree Republicans are favorites. On average markets give Republicans an 87% chance of winning the House and a 65% chance of winning the Senate. Forecasting models give Republicans an 81% chance of winning the House and 55% chance of winning the Senate. The political markets seem to reflect the conventional wisdom that the polls are biased against Republicans, but I don’t think there’s a compelling reason to discount the polls much. Nevertheless, I think the political markets are right that Democrats are unlikely to hold on to the House because doing so would require almost everything to go Democrats’ way. But I also think that because Democratic have the better candidates, they’re slightly more likely than not to hold on to control of the Senate. We’ll know in a few days.
My Forecast
Republicans have a 90% chance of winning a majority in the House
Republicans have a 18% chance of winning at least 240 seats in the House
Republicans have a 48% chance of gaining control of the Senate
Please help me out! In a few weeks, I’m submitting a report to the Centre for Effective Altruism on the progress of my newsletter grant project. If you’ve enjoyed Telling the Future, I’d very much appreciate it if you would send me a message or leave a comment letting me know why, so I can share your thoughts with them and use them to continue to improve my newsletter. Thanks for reading and supporting my work.
I'm a fan of Telling the Future!
I read every issue of Telling the Future because doing so makes me a smarter human being. Please keep it coming!