In my role as a political commentator, I enjoy making predictions. I possess no special talent for this and am often hilariously wrong: Unlike most other pundits, I famously clung to my stupid belief that former President Joe Biden would never step down as the Democrats' 2024 candidate, for instance. I called the 2020 presidential election correctly (and got most states right, only flipping Georgia and Florida), but I messed up in 2024. (Nevertheless, I still demolished my co-host, who didn't appear to take this challenge very seriously.)
The hardest part about making predictions is separating what you want to happen from what is actually likely to happen. Commentators often feel inclined to rally for their preferred outcome, as if shouting about it on television is going to make it come true. In some ways, this is an easier temptation for us libertarians to avoid, since all outcomes generally seem suboptimal. Once again, that is the case: See, for example, the upcoming 2026 midterms.
It seems inarguable that President Donald Trump has squandered much of his political capital by failing to focus narrowly on the economy and instead launching another unpopular war in the Middle East. It also seems obvious that the Democratic Party, once returned in power, will do everything it can to thwart the administration's shrewdest policies—on tech and AI in particular.
But that is my prediction: Democrats will retake the House of Representatives and the Senate.
I don't have any specific map in mind for the House (it's too complicated), but here are my Senate predictions:
Here is my REVISED prediction for the Senate. For the last several weeks, everything has moved in Dems' direction, and I now expect them to take it.
I still say Collins holds on. Michigan and Ohio seem lost for GOP, though. And I'm not super confident in the Iowa call. https://t.co/oph3Swtgzw pic.twitter.com/FtsrGxmukJ
— Robby Soave (@robbysoave) October 20, 2026
When I looked at the map back in August, I still thought Ohio was a long shot for the Democrats and that Michigan was in play for the GOP due to Abdul El-Sayed being so far left. Now it looks to me like Republicans will lose both seats, along with Texas and Alaska. I'm still betting that the formidable incumbent in Maine, Sen. Susan Collins (R), will hold on—the polls have misjudged her in the past—and that Iowa is a bridge too far for Dems. If Republicans manage to keep those two seats, then it's a 51–49 map in Democrats' favor.
But wait…all is not lost for Trump.
In the event of a 51–49 Dem advantage, it seems somewhat likely to me that Sen. John Fetterman (D–Pa.) could switch parties, which would restore Republican control because of Vice President J.D. Vance's tiebreaking vote. With his conservative rhetoric on Israel and other issues, Fetterman has no future as an elected Democrat in his state—he could never ever ever win a Democratic primary—and will have to strike some sort deal with Trump before his reelection in 2028. Fetterman should agree to become an independent who caucuses with Republicans in exchange for Trump ensuring that no Republican dares run against him in the future. Absent such a deal, Fetterman is virtually guaranteed to be a one-term senator, so if he harbors any desire to keep his job—perhaps that's an open question—he will need to make a move. A 51–49 Senate will give him the exact leverage he needs.
So those are my predictions! What are yours?
This Week on Free Media
I'm joined by Amber Duke to discuss the riots in France, Vivek Ramaswamy's terrible campaign ad, and more.
Worth Watching
Here's another shameless plug for my upcoming debate with Katherine Mangu-Ward, Mehdi Hasan, and Jason Stanley! (Katherine is on my side.) If you are in Washington, D.C., please get tickets here.
The post I Predict the GOP Will Lose Its Senate Majority (and Magically Regain It) appeared first on Reason Magazine.