When polls closed in Brazil last week, online bettors were convinced Flávio Bolsonaro would become president, putting his odds near 80 percent. Luiz Inácio Lula da Silva won. It was the latest reminder that election betting odds are not election results, and that reminder comes just before the November 3 midterms in the United States, the biggest test yet for Kalshi and Polymarket.
Both platforms are betting that the midterms will be their moment. More than $221 million has flowed into midterm-related markets, up from $92 million in the 2024 presidential race, according to a report from the Anti-Corruption Data Collective (ACDC), a Washington research group. Wagers on congressional races have more than tripled compared with 2024. Kalshi rented a Times Square billboard showing Senate odds, and Polymarket ran a TV ad urging viewers to “watch the midterms move.” Kalshi says viewers of its midterm markets outnumber traders five to one, so people increasingly treat betting on elections as a news source.
Bu should they? A review of more than a million markets, published in the Proceedings of the National Academy of Sciences, found they were “well calibrated,” and a Vanderbilt study found markets called Donald Trump’s 2024 win well before polls did. The Washington Post looked at 268 primary candidates who had a 70 to 80 percent chance on Kalshi or Polymarket during the two months before their races, and three-quarters won. Kalshi says that since 2024, the candidate its markets favored has won about nine times out of ten.
But accurate on average isn’t the same as right. Prediction markets gave Spencer Pratt, a former reality TV star, a 78 percent chance of advancing in the Los Angeles mayoral primary after more than $47 million in trades, the most of any mayoral race this year. He finished third. Traditional pollsters are skeptica of this new marketl. Election forecaster G. Elliott Morris says people shouldn’t be looking at election betting odds at all, since polls measure public opinion directly.
The bigger concern for many is potential manipulation in the mid-term elections. ACDC analyzed price spikes across Kalshi and Polymarket and found that of 167 congressional markets worth more than $100,000 (38 in total), about 23 percent, saw moves of 30 cents or more. In the previous Los Angeles mayoral race, a single trader spent $1.25 million on “yes” contracts for Pratt on May 20, lifting the price from 29 to 45 cents. Kalshi says the market corrected within nine seconds, and ACDC found a smaller bump lingered for about two hours. In Illinois, a pair of purchases totaling about $8,000 moved a Senate primary contract from 47 cents to 99 cents and back to 52.
And this matters because the influence is real. A study funded by the Brennan Center for Justice, surveying 1,994 registered voters in September, found people were 10 to 20 percentage points more likely to believe a candidate would win if prediction markets favored that candidate, whatever the polls said. Pratt’s supporters used his miss to cast doubt on the election itself, and Los Angeles County’s registrar warned that market projections can create confusion when treated as more authoritative than the actual count.
The platforms say they’re policing it. Kalshi added labels for thinly traded markets and live feeds of individual trades, and says it uses algorithms to flag unusual activity. Polymarket launched an election-specific surveillance system and partnered with blockchain analytics firm Chainalysis, which helped identify a U.S. soldier charged with using classified information to bet on events in Venezuela. He has pleaded not guilty. Both Kalshi and Polymarket U.S. are overseen by the Commodity Futures Trading Commission, and collect customer identity information, though outsiders can’t see who is behind any given trade.
Regulators aren’t convinced. Maryland’s elections administrator, Jared DeMarinis, asked the state prosecutor in July to investigate whether election markets violate state law, and in September barred election judges and board members from trading in them. More than half of states have laws making election betting illegal in some or all circumstances, according to Pew, but the platforms argue the CFTC has sole jurisdiction. New Jersey, backed by dozens of states, has asked the Supreme Court to settle it, and Utah recently charged Kalshi and its co-founders with gambling crimes.
In the coming days after the mid-terms, the story will be which way the odds went. For the platforms, the stakes are whether a good night proves the markets work, or a bad one shows how easily betting on elections can bend what people believe will happen.