Will Abdul El-Sayed win the 2026 US Senate election in Michigan?

TL;DR

I estimate a 52.0% chance that Abdul El-Sayed wins. The race is effectively even: current polling gives Mike Rogers a small edge, while Michigan’s Senate history and the anti-Republican national environment pull back toward El-Sayed. The decisive question is whether El-Sayed can close his candidate-specific gap with other Michigan Democrats without losing the turnout energy that won him the primary.

Context

The Associated Press called the August 4 Democratic primary for El-Sayed after he led Haley Stevens by 14,893 votes—just under one percentage point—among more than 1.5 million votes counted. Stevens and withdrawn candidate Mallory McMorrow subsequently backed him. Rogers won the Republican nomination without opposition.

This is an open-seat election for the term beginning in January 2027. Rogers is already a proven statewide competitor: he lost the 2024 Senate election to Elissa Slotkin by only 19,006 votes, while Donald Trump carried Michigan on the same ballot (Michigan Department of State).

Evidence

The historical prior favors Democrats, although the recent margins are far more relevant than the older landslides. In the complete certified Michigan Senate history from November 7, 2000 through November 5, 2024, Democrats won all nine elections. Margins below are percentage points calculated over every valid candidate vote; the vintage is the state’s current certified archive (Michigan Department of State).

The strongest analogue is 2024. Slotkin beat Rogers by 0.34 points even as Trump won Michigan by 1.41 points, meaning Rogers ran about 1.75 points behind the top of his ticket (certified 2024 results). That is evidence against treating Rogers as an unusually strong crossover candidate. It is not enough to make El-Sayed the favorite by itself, because Slotkin was a more conventional nominee.

Here is the full set of 2026 El-Sayed–Rogers ballot tests I could verify, covering fieldwork from January 2 through August 10. Margins are El-Sayed minus Rogers; LV means likely voters and RV means registered voters.

I converted each poll to a Democratic margin, weighted by the square root of sample size, used a 30-day recency half-life, doubled the weight of post-primary polling, reduced weights for lower-transparency and sponsored polls, applied a two-point sponsor correction, and capped any adjusted poll at six points to prevent the Glengariff outlier from dominating. This produces a current polling estimate of Rogers +1.1. Reasonable variants range from roughly Rogers +0.6 to Rogers +1.4; using only non-sponsored surveys gives Rogers about a 2.8-point lead.

The two true post-primary polls explain most of the uncertainty. They used identical August 6–10 field dates but differed by ten points. Fox’s registered-voter sample had Rogers ahead by four, while the Democratic-sponsored GBAO likely-voter poll had El-Sayed ahead by six. Fox also found El-Sayed ahead by four among definite voters, seven among the extremely motivated and eleven among the extremely interested; Democrats were more engaged than Republicans on all three measures (Fox News). I treat that as evidence for a turnout adjustment, not as a substitute for an independent likely-voter poll.

Fox found 91% of Democrats supporting El-Sayed against 96% Republican unity for Rogers, while El-Sayed’s net favorability was -13 and Rogers’ was -6. In the same sample, Democratic gubernatorial nominee Jocelyn Benson led John James by five points while El-Sayed trailed Rogers by four—a nine-point ticket gap (Fox News). Other same-sample surveys also showed El-Sayed running behind Benson, though by an unstable two to fifteen points (Michigan governor polling). I read this as real candidate-specific drag, but not as a permanent nine-point penalty.

The national environment pulls the other way. Silver Bulletin’s August 18 generic-ballot estimate was 48.4% Democratic to 41.6% Republican, D+6.7, and had stayed near six to seven points over its May 21–August 18 coverage window (Silver Bulletin generic ballot). Its voter-only Trump approval estimate on August 17 was 38.1% approve and 58.0% disapprove, net -19.9 (Silver Bulletin approval tracker). Republicans won the national 2024 House vote by about 2.6 points, so the national environment has moved roughly nine points toward Democrats since Rogers’ last race (Office of the Clerk). Only part of that swing should transfer to one state race, but it is too large to ignore.

El-Sayed’s primary contains both positive and negative signals. He survived a near ten-to-one television and radio spending disadvantage, $52 million to $5.2 million, and won 48.5%–47.5% (Axios). But he lost Wayne, Oakland and Macomb counties, while Stevens exceeded 60% in Detroit; his strengths were university areas and Election Day voters (Bridge Michigan). Democratic primary turnout exceeded 1.5 million and set a state midterm record, but the primary was unusually competitive and expensive, so I give this only a modest general-election weight (Associated Press).

Money favors Rogers, but less decisively than the headline suggests. The Senate Leadership Fund raised its announced Michigan commitment to $51 million, including an early $6 million post-primary campaign (Axios). Senate Majority PAC had already announced an initial $20 million Michigan television reservation and additional digital spending (Senate Majority PAC). El-Sayed’s primary result is direct evidence that advertising has diminishing returns, but Rogers still holds an early-definition advantage.

My independent synthesis gives the election-day polling path to Rogers by about 0.5 points after allowing for Democratic consolidation and turnout intensity, while the fundamentals model gives El-Sayed about a one-point lead. Weighting polls at 60% and fundamentals at 40% produces an expected margin close to El-Sayed +0.1. With a 3.8-point election-day standard deviation, the independent forecast is 51.0%.

I checked markets only after completing that estimate. Around the cutoff, Democratic contracts were about 56%–57% on Kalshi, 56%–57% on PredictIt, and 58%–59% on Polymarket. I gave the roughly 57% market consensus a 15% weight in log-odds space because the venues share the same polling and national-environment inputs. That yields the final 52.0% estimate.

What's non-obvious

Michigan’s Democratic Senate streak is real, but it is not the central reason to favor El-Sayed. The stronger clue is the gap between the Senate and governor races in the same surveys. Michigan voters currently look willing to vote Democratic while withholding support from El-Sayed. His problem is candidate-specific, not primarily partisan.

The polling disagreement is also less mysterious than it looks. Fox measured all registered voters immediately after a bitter primary; GBAO used a Democratic sponsor’s likely-voter model. The right answer is not to choose one. It is to infer a race where Rogers leads the broad electorate, El-Sayed may lead the most engaged electorate, and turnout modeling decides which picture is closer to November. Markets appear to put more weight than I do on the second picture.

Uncertainties

• There is no independent post-primary likely-voter poll. Two or three such polls would sharply reduce the current ten-point methodological spread.
• El-Sayed’s eventual support among Black voters, older Democrats, Jewish Democrats and Stevens primary voters is not yet clear. Wayne County and Detroit data will be especially valuable.
• His negative image may reflect temporary primary attacks, or it may persist after tens of millions of dollars in Republican advertising. September favorability polling will distinguish those cases.
• The national environment is strongly Democratic now, but a one-to-two-point movement in the generic ballot would matter in a race centered within a few tenths of a point.
• Announced advertising reservations are not the same as final spending. The resource adjustment would shrink if Democratic commitments approach Republican levels.