1. Midterm prediction markets enter their first full election cycle at scale
On November 3, midterm prediction markets will face their first large-scale test against traditional polling, with Kalshi and Polymarket trading far more volume than they did in 2024. Two weeks before the vote, veteran political journalist Chuck Todd is hosting the NEXTPredict summit in New York on October 22 and 23, where pollsters, analysts and exchange executives will debate which forecasting method deserves more trust.
Todd, a former network political director who also moderated Meet the Press, has said he wants to learn how much pollsters and prediction markets depend on one another, and whether the growth of trading on elections is healthy for politics. He has not committed to either camp as a general matter, describing himself as open-minded about what market exposure will do to the political system.
The timing gives the debate unusual weight. In 2024, the platforms operated in a very different environment, and the questions now being asked about accuracy, liquidity and influence are being asked of markets that have grown considerably since then.
2. How much has the market grown since 2024?
Election contracts became legal in the United States only about a month before the 2024 presidential vote. Benjamin Freeman, Kalshi's head of politics growth, has said the industry now has a much longer runway to plan, which means exchanges had months rather than weeks to build products, recruit traders and publish tools for the midterm cycle.
The scale difference is visible in the volume figures. An analysis by the Pew Research Center found that the combined monthly volume of Kalshi and Polymarket rose from $26 billion in May to $53 billion in July, with most of the growth coming from sports contracts. Kalshi and Polymarket together topped $50 billion in monthly volume in July, more than double the May level.
Politics remains a small slice of that activity. According to Pew, Polymarket has seen roughly $2 billion a month in political trading this year, while Kalshi has seen under $300 million. By July, more than $30 million had been wagered on which party will win control of the House and the Senate. For a market whose headline numbers are measured in tens of billions, those political totals show that election trading is still a niche inside a much larger business.
3. What is the regulatory position on election contracts?
The federal view of election contracts has shifted. The Commodity Futures Trading Commission once barred Kalshi from listing contracts on which party controls Congress, describing them as gaming. In 2024 the agency proposed classifying election contracts as gaming, then withdrew that proposal in February. On June 10 it published a proposed framework under which elections would not be treated as gaming, according to an analysis by the Congressional Research Service.
The commission's proposal rests on an argument about information. It holds that event contracts can aggregate the judgment of many participants into a signal that is more reliable than other forecasting techniques, the research service's analysis found. That reasoning is the intellectual basis for treating the products as financial instruments rather than wagers, and it is exactly the claim the midterms will put under scrutiny.
The proposal is still a proposal, so the legal footing of election trading continues to depend on how the rulemaking proceeds. For traders and exchanges, the practical effect is that products already listed operate under a framework that regulators have signaled they intend to formalize.
4. How are the exchanges presenting odds next to polls?
Kalshi introduced a Midterms Hub in July that places traders' odds for individual races beside polling averages. The layout invites direct comparison: a reader can see whether the crowd of traders agrees with the pollsters or diverges from them, and by how much.
That comparison is the premise of the NEXTPredict event. Todd, who is the official host, will be joined by working journalists, including Contessa Brewer and Marshall Cohen, so the exchanges' executives will share a program with reporters who cover the same elections.
In a September 18 interview with NEXTPredict co-founder Pierre Lindh, Todd summarized the theme as pollsters versus prediction markets. His central question was how much each needs and feeds the other. Markets often cite polling averages when traders form views, while pollsters and commentators increasingly cite market odds, so independence between the two signals cannot be assumed.
5. Hedging, entertainment and the effect on politics
Todd also wants to understand how much election trading is genuine hedging and how much is entertainment. The distinction affects how much weight a price deserves: a market dominated by people with real exposure to an outcome may carry different information from one driven by enthusiasts who enjoy a wager.
He has said he is open-minded about whether exposure to markets is good for politics or whether it could produce a new class of politician, one whose decisions are shaped by contracts on the outcomes. He raised the question without answering it.
The question has practical stakes for the exchanges as well. Platforms that want to be treated as serious sources of information, rather than entertainment venues, need to show that their prices reflect considered judgment and that participants cannot easily distort them. Their ability to make that case will shape how regulators, banks and the public view event contracts after the election.
On the Texas Senate race between Democrat James Talarico and Republican Ken Paxton, Todd said he sides with the markets and would be more surprised if Talarico lost. He added that Trump voters tend to engage later than Democrats and that September optimism often fades by October, a pattern he said makes late-season movement worth watching.
6. Are congressional markets too thin to trust?
Critics focus on liquidity. The Anti-Corruption Data Collective found that a $3,500 bet, the federal maximum for a personal campaign contribution, could move 97% of midterm-related markets by at least five cents. Co-director Michael Hornsby said most congressional markets are too thin for the active trading needed to correct prices that have been pushed away from fair value.
Kalshi spokesperson Dani Lever offered the opposing view. She argues that in a liquid market, a mispriced contract is effectively free money, because other traders will take the other side and restore the price. The disagreement is therefore about whether enough capital and attention reach the smaller contests to make that correction reliable.
The dispute matters because individual House and Senate races attract far less trading than headline national outcomes. A price that looks authoritative on a screen may reflect only a handful of orders, and users comparing it with a polling average have to judge how deep the market behind it actually is.
7. What does the institutional research show?
Evidence on accuracy is mixed in emphasis but notable in scale. Morgan Stanley's Counterpoint Global studied more than 72 million Kalshi trades and found that prices tracked outcomes closely. Lindh says research desks at banks want to know whether market odds are a better read than polling, which is part of why institutional capital features in the summit's agenda.
Morgan Stanley will lead a panel on institutional capital on the second day. About 2,500 attendees are expected, a figure that indicates how far the industry's audience has moved beyond retail traders.
The study speaks to trades across many contracts rather than to any single election, so it does not settle whether markets will outperform polls in November. It does show that large pools of data are available for that kind of evaluation, and the midterms will add to it.
8. Why the comparison with polls is harder than it looks
Judging markets against polls is not as simple as counting winners. A poll reports a snapshot of stated preferences among a sample, usually with a stated margin of error, while a market price expresses the collective probability traders assign to an outcome. A contract trading at 70 cents implies a 70% chance, not a prediction that the favorite will win, so a single upset does not by itself show that the price was wrong.
Fair evaluation therefore needs many races and many price points. A forecasting method is considered well calibrated when events priced at 70% happen roughly seven times in ten. The enormous number of Kalshi trades examined by Morgan Stanley's Counterpoint Global is relevant for exactly this reason, and the midterms, with hundreds of House contests, dozens of Senate and governor races and several ballot questions, offer a broader sample than a presidential election does.
The two signals also differ in timing. Polls arrive in batches, with fieldwork finished days before publication, whereas prices move continuously and can react within minutes to news, candidate scandals or turnout reports. That speed is an advantage for markets, but it also makes them sensitive to rumor and to the behavior of a few large traders, a risk that is sharpest in the thin congressional contracts described above.
9. What happens next?
The summit's second day includes a panel titled Who Calls the Midterms: Polls, Models or the Markets? Todd will moderate with Rajiv Sethi, a professor at Columbia University, Geoffrey Skelley, chief elections analyst at Decision Desk HQ, and John Phillips, chief executive of PredictIt. They are expected to examine where polls, models and markets fail and which source to trust as results arrive on November 3.
After that date, the comparison becomes concrete. Observers will be able to check final market prices and polling averages against actual results across the Senate, House and governor races, and to see whether thin markets were moved by small bets. The outcome will inform both regulators weighing the commission's framework and institutions deciding how seriously to treat market odds.