Kalshi has reported more than $2.5 billion in trading volume on its 2026 U.S. midterm election markets, nearly twice the $1.3 billion recorded during the 2024 election cycle.
Summary
- Midterm election trading surpassed the 2024 cycle’s total by more than $1.2 billion.
- Kalshi said favored candidates won 479 of 487 races resolved by early Nov. 4.
- Its research put winner-prediction accuracy at 98.2% when polls closed, up from 93.8% three months earlier.
- Election-night updates showed traders backing Democratic victories in Texas and South Carolina as results arrived.
According to Kalshi’s Nov. 4 recap, traders anticipated Democratic control of both chambers of Congress before Election Day, although they underestimated the size of the party’s gains.
Using the company’s rounded figures, the reported volume represents growth of more than 92% from the previous election cycle. The comparison spans two different contests: the 2024 cycle included a presidential election, while this year’s voting centered on congressional, state and local races.
Kalshi reports a 98% hit rate across resolved races
In its initial tally, Kalshi said market favorites won 479 of 487 races resolved by 6 a.m. ET on Nov. 4. The count covered House, Senate and gubernatorial contests alongside races for attorney general, lieutenant governor and secretary of state.
The company listed 356 winning favorites across 360 House races, 32 across 33 Senate contests and 33 across 34 gubernatorial elections. Eight favorites lost across the full tally, giving the platform a combined success rate of approximately 98.4%, rounded to 98% in its announcement.
Several contests remained unresolved when the recap was published, including the Alaska and Mississippi Senate races. The reported hit rate therefore covered the results available that morning.
Kalshi said its House control market reached a 99% probability of a Democratic majority at 8:10 p.m. ET on election night. Senate control crossed the same threshold at 9 p.m. ET.
Election-night odds changed as results came in
The platform’s Nov. 3 live updates documented how individual markets moved during the count.
At 6:55 p.m. ET, its feed put the probability of Democratic House control at 97.2%. By 8:06 p.m., Republican incumbent Greg Abbott was no longer favored in the Texas governor’s race.
Later entries placed Democrat James Talarico at a 99% probability of winning the Texas Senate contest at 9:35 p.m., followed by Annie Andrews in South Carolina at 10:10 p.m. The feed reached the same threshold for Texas gubernatorial challenger Gina Hinojosa at 11:35 p.m.
Kalshi identified those entries as market projections, separate from official race calls or contract settlement.
Reuters reported that prediction markets anticipated Democratic control of Congress but missed the strength of some candidates. Analytics firm BETINT, cited in its coverage, counted 14 Democratic winners who had been market underdogs a week earlier and found that markets and polls performed roughly equally well overall.
Kalshi research measures forecasts before polls closed
A separate Kalshi Research analysis examined races called by 6:26 a.m. ET on Nov. 4, tracking forecasts from three months before voting through poll closure.
The researchers reported that directional accuracy rose from 93.8% to 98.2% over that period, a gain of 4.4 percentage points. Brier scores, which measure the gap between forecast probabilities and actual outcomes, fell from 0.040 to 0.016; lower scores indicate better performance.
The accompanying working paper, written by Nicole Kagan and Rubens Baiocchi, explains that simply counting winning favorites does not measure confidence. A successful forecast priced at 51% receives the same directional score as one priced at 95%.
For its comparisons, the study used prices available before each measurement time and matched election-model snapshots at 9 a.m. ET on Election Day. Its methodology anchored forecasts to voting times because election contracts can remain open after the outcome becomes clear.
U.S. election markets drew scrutiny before voting
Before the election, crypto.news covered the CFTC’s reported election-market surveillance, including Chairman Michael Selig’s comments about monitoring possible manipulation.
The Oct. 24 coverage cited Dune research showing $250 million in combined midterm open interest on Kalshi and Polymarket as of Oct. 18. Kalshi accounted for $210 million and Polymarket for $40 million.
Candidate participation prompted a separate election trading proposal from Rep. Don Davis. Introduced in October, the No Betting on Your Own Race Act would restrict federal candidates, their spouses, dependent children, and authorized campaign committees from trading contracts tied to their elections. Proposed penalties would be $10,000 or three times the transaction’s net financial gain, whichever is higher.
