The Michigan senate race is Abdul El-Sayed’s to lose

Our model shows Abdul El-Sayed can survive a "squad"-level underperformance

Abdul El-Sayed won a bruising primary. Now comes the hard part. (Photo by Finn Gomez/Getty Images)

The Split Ticket midterms model is updated daily at 5 p.m. Eastern with the latest polls on our website.

---

We’ve seen a lot of cases where swing-state Republicans nominate ardently ideological, base-aligned candidates over more “establishment” options. Abdul El-Sayed’s primary victory may be the first instance in modern memory where Democrats have done the same in a critical statewide race.

Primary voters are not reflective of the general electorate, and many of those Republicans lost winnable races as a result. But there’s reason to believe El-Sayed’s fate could be different.

There are real risks associated with his candidacy: Our Split Ticket midterms model sees him as a slightly weaker candidate than his primary opponent Rep. Haley Stevens would have been. Stevens had an 80% chance of winning the general election, according to the model, while El-Sayed’s odds are currently at 70%. While we estimate that Stevens would have won by 5.4 percentage points, we think El-Sayed will win by just 3.6 points, thanks to his substantially weaker polling.

Why is El-Sayed still forecast to win, despite his relative weakness? The single biggest reason is a national and state environment that’s uniquely unfriendly to Republicans.

In 2024, when Donald Trump won the popular vote by 1.5 percentage points nationally, Michigan was an R+1 state. Today, that number is closer to D+9 by our estimation (or D+8, if you look at Nate Silver’s modeling). In that sort of environment, even if El-Sayed’s candidate quality is similar to that of the average hardline progressive from “The Squad” (meaning he would perform 5 points worse than a replacement-level Democrat), he would still win.

Are national polls simply wrong about how blue 2026 might be? I don’t think so. Every other indication we’re seeing, ranging from special elections to the Washington state jungle primary, suggests that this really is a blue tsunami of a cycle.¹ And in a situation like that, El-Sayed doesn’t need to overperform or even match the baseline performance of generic Democrats. He simply needs to avoid being a disaster.

Can he do that? It’s early, but I’d say the initial signs are promising. He has distanced himself from controversial remarks about 9/11 by Hasan Piker, who campaigned with him. He appears to be embracing party unity and has disavowed his own prior left-wing comments on defunding the police.

In doing so, El-Sayed appears to be avoiding many of the same missteps that plagued Republican firebrands like 2022 Arizona gubernatorial candidate Kari Lake. Also working in El-Sayed’s favor is the fact that Republican nominee Rep. Mike Rogers has made absolutely no effort to distance himself from Trump, flirting with election denial despite running in a state where the president is supremely unpopular.²

That’s why, if you forced me to pick today, I’d say that El-Sayed still wins this race.

Our model is more bullish on El-Sayed than other quantitative forecasts,³ in part because we don’t pay as much attention to pre-primary polling,⁴ and in part because our generic ballot modeling of Michigan suggests El-Sayed would have to underperform at the level of Elizabeth Warren (who underperformed by 11 points in her last Senate race) in order to lose.

This means he’d have to be nearly twice as weak as the average member of “The Squad,” for whom the typical underperformance is 5 points. While there’s plenty of evidence to suggest El-Sayed is weaker than Stevens would have been, there’s simply no evidence to support the hypothesis he’s that much weaker than a replacement-level Democrat.

This has implications for 2028 as well. Because while election nerds like me will look at overperformance and evaluate whether a win was by “more” or “less” than expected, that’s just not how the median voter looks at things. For a party whose left flank is ascendant, any win from El-Sayed is likely to be held up as proof that their candidates can actually win swing states — and probably gives fuel to the left in the 2028 presidential primaries.

For a long time, candidates like El-Sayed were simply boxed out of Democratic primaries — for anything other than safely blue races — due to concerns about electability as well as struggles with nonwhite segments of the party base. It was this phenomenon that tanked both Warren and Bernie Sanders in the 2020 presidential primary. And if the primary between Stevens and El-Sayed had been held in 2020, it is virtually certain that Stevens — who lost by just 1 percentage point this time — would have won.

But amid the historic unpopularity of both Trump and the Democratic establishment — and in the context of the never-ending conflict in Gaza, which dominated this primary — it is clear that 2026 is a different landscape altogether. Rank-and-file Democrats may well see El-Sayed’s nomination as a relevant test of the left’s ability to win swing states.

Whether that’s true or not, it could have seismic implications for the direction Democrats choose to go in 2028.

Recommended reading:

1. These types of measures struggle to pick up low-engagement voters, but this is a much bigger problem in a presidential year than in a midterm, where they generally just don’t vote. This is also a big reason polling doesn’t overestimate Democrats in non-presidential elections as much as in presidential years.

2. Our modeling finds that more than 60% of Michigan voters disapprove of Trump right now.

3. We knew one of Stevens and El-Sayed was likely to be the nominee, so our model accounted for their polling even before the primary election happened.

4. Pre-primary polling is much less predictive, historically, and candidate vote shares tend to be artificially depressed after a bruising primary before voters “come home.”