Prediction markets have surged in popularity in 2026, making companies like Kalshi and Polymarket more prominent than ever during the November midterm election.
Kalshi and Polymarket saw their monthly global trading volume more than double in only two months earlier this year, rising from $26 million in May to $53 billion in July, according to a Pew Research Center analysis of data from The Block, a digital assets media and information firm. Much of that increase was a result of sports trading, but with election season looming, it was only the beginning.
Financial services company Fidelity defines prediction markets as “marketplaces where participants buy and sell contracts tied to the outcome of future events.” A specific event is clearly defined, often with a yes-or-no outcome, each outcome is given a tradable event, participants trade contracts with one another based on their expectations, the event resolves and the event contracts settle at their stated value, typically at $1 per contract.
Meanwhile, many of the major polling organizations — including NY Times/Siena, Marist and YouGov — conducted their final rounds of polling in October. On average, a majority of the polls favored the Democratic candidates in Ohio’s gubernatorial and U.S. Senate races leading up to Election Day.
Decision Desk HQ provides independent results and analysis across every level of U.S. elections. Its forecast is based on a calibrated combination of past election results, race-level polling, markets and the national environment, according to the organization’s website.
“Election forecasting models tend to work very well for presidential elections,” said Dr. Justin Buchler, associate professor with Case Western Reserve University’s Department of Political Science. “It's much harder to predict seat swings, and it's particularly difficult to predict seat swings for Senate elections because in any given year, we only have about a third of the Senate up for election, and that means that Senate elections, the total number seats that can swing, will be determined very much by idiosyncratic factors.”
Among those factors, Buchler said, was the case of U.S. Sen. Lindsey Graham passing away coupled with his sister, Republican candidate Darline Graham, underperforming.
Comparing forecast analytics from DDHQ with Kalshi’s midterm forecast, the prediction market showed earlier levels of confidence across three of Ohio’s major toss-up races, though both entities accurately predicted the victors.
Ohio’s hotly contested U.S. Senate race between Democratic candidate Sherrod Brown and Republican Sen. Jon Husted didn’t show much separation until the later stages of the campaigns.
Brown’s uptick on Kalshi began on Sept. 10, when he had a 51.1% chance of victory. His chances of winning, according to Kalshi, began steadily increasing before ultimately reaching 70.3% on Nov. 1, an increase of almost 20 points. In that same span, DDHQ never had Brown’s chances of winning higher than 57%. Polling data compiled by RealClearPolling, a poll aggregator site, had Brown as the favorite in eight of the final 10 polls conducted leading up to Election Day.
In the race for Ohio’s 7th Congressional District between incumbent Republican Rep. Max Miller and Democrat Brian Poindexter, Kalshi favored Miller in late July with a 63.5% chance to win reelection. His ex-wife, Emily Moreno, went public with allegations of domestic violence in May, but widespread media coverage didn’t spike until the end of July after a Mother Jones report detailed those allegations.
While not stating it as a concrete fact, Bowling Green State University political science professor Dr. Robert Alexander acknowledged that the increased attention to those allegations could have been a factor in Kalshi’s odds shifting in Poindexter’s favor.
"I don't want to pretend to say that I know exactly why that occurred, but that would seem to make some sense,” Alexander said. “There's definitely some vibe checks that go on when you're looking at prediction markets. And if you're the incumbent and there's more media coverage of you, that's generally not a good thing.”
On July 29, Kalshi had Poindexter as the favorite in the race for the first time. He never trailed again, with his chances of winning reaching as high as 89% on Oct. 23.
Inversely, DDHQ had Miller as the favorite in the race until Oct. 21, at which point it had the odds at an even 50-50 split. DDHQ didn’t list Poindexter as the favorite until Oct. 26, hinting at Buchler’s earlier statement regarding the difficulty in predicting seat swings. By that point, Kalshi gave Poindexter an 86.1% chance of winning, while also having him defeating the incumbent by eight points.
Ohio’s gubernatorial race was one of the most widely discussed in the country, and DDHQ and Kalshi mirrored one another closely.
Kalshi had Democratic candidate Dr. Amy Acton ahead of Republican opponent Vivek Ramaswamy for the first time on Aug. 13. Following a batch of September polls, DDHQ put Acton’s likelihood of winning the election ahead of Ramaswamy’s on Sept. 17. She held those leads in the respective probability forecasts through Election Day.
Ultimately, the polls, election forecasts and prediction markets acted in a relatively similar fashion, though prediction markets locked down their eventual winners sooner.
“It would appear that they were pretty much in tandem with one another in most cases,” Alexander said. “Where they were successful, both were successful, polling and the prediction markets. And where they failed, they both failed.”
With the 2028 presidential election on the distant horizon, prediction markets have emerged as a player in political prognostication. Alexander sees the Kalshis of the world as having the possibility to stick.
“It doesn't appear as though that is something that will fade,” Alexander said. “Again, because they pretty much moved in line with the polls and the forecast, there didn't appear to be any big misses, per se. I think that there will continue to be interest in these prediction markets.”
