Prediction markets face US midterms test as election bets soar

A prediction market is a trading platform for possible future events. Instead of buying company shares, traders buy contracts tied to outcomes, such as whether a candidate will win. The market price becomes an implied probability. This matters because it creates a continuously updated forecast based on participants’ money and information.

A trader who expects an event to happen buys a “yes” contract. Someone who expects it not to happen buys a “no” contract. As traders buy and sell, prices move. A “yes” contract priced near 70 cents is commonly read as roughly a 70% chance of that outcome, although the article focuses on odds rather than explaining contract settlement mechanics.

The process is not guaranteed to be right. Sophisticated traders use polls and other data to correct prices, while large bets can move them. Kalshi and Polymarket say their markets quickly adjust after major trades. Their value is therefore as a live gauge, not as certainty.

US midterm prediction markets have grown into a major financial and information system. Kalshi reported $1.33 billion in trading volume across all midterm bets on Monday. Polymarket had attracted about $145 million to congressional markets, according to the Anti-Corruption Data Collective. These figures show the scale of interest before results were known.

The number of markets was also unusually large. More than 28,000 midterm markets had opened. Only a fraction involved individual contests, and a small group of races generated most of the trading. Senate contests in Texas, Michigan, Maine, and South Carolina were among the biggest contributors.

This scale raises the stakes for the industry. Prediction markets gained attention after signaling Trump’s 2024 victory against traditional polling expectations. If their midterm forecasts perform poorly, the large sums involved could intensify scrutiny of their accuracy, safeguards, and influence.

On Monday, prediction-market prices aggregated by BetInt implied that Democrats had a 95% chance of taking the House of Representatives. They also gave Democrats a 76% chance of taking the Senate. These were market-based probabilities, not guarantees or official forecasts.

The figures came from prices across trading platforms. Traders bought and sold contracts tied to congressional outcomes. Their activity incorporated polling, other data, and their judgments about individual races. Sophisticated profit-seeking traders appear to supply much of the capital and help correct prices, according to research cited in the article.

The odds mattered because the midterm results could validate or damage confidence in prediction markets. Many contests were close, and researchers warned that a market can be well calibrated overall while still missing a particular race. The final results therefore offered a broad test of the platforms’ performance.

Accuracy means whether a market correctly identified the winner in a specific contest. Calibration asks whether probabilities are reliable across many markets. If events priced at 20% happen about one-fifth of the time, the market is well calibrated, even though it misses most individual events.

The article gives a clear example. Kalshi and Polymarket implied only a 20% chance that Luiz Inacio Lula da Silva would defeat Flavio Bolsonaro. Lula won. That result was a miss if the market was treated as a winner-picking forecast, but it was not impossible under a 20% probability. A one-in-five event can happen.

Researchers say markets are generally well calibrated at scale, partly because sophisticated traders use polling and other data. Yet most races are not competitive, and the accuracy of any specific price is impossible to know beforehand. Investors should treat the odds as a useful gauge, not certainty.

Prediction markets rose to prominence after signaling Donald Trump’s 2024 presidential victory while traditional opinion polls pointed elsewhere. Since then, media outlets and investors have followed them globally. Their reputation now depends partly on whether their highly visible midterm probabilities match the results.

The article identifies a clear risk. BetInt prices implied a 95% Democratic chance in the House and 76% chance in the Senate. If the results diverge sharply, analysts say confidence in the industry could suffer. This concern is larger because trading has reached record sums, including $1.33 billion in Kalshi midterm volume.

A poor showing would not prove every market useless. Researchers distinguish overall calibration from accuracy in one race, and unlikely outcomes occur. Still, a sharp miss could prompt more questions about data, market concentration, political manipulation, and controls. The industry says its value is forecasting in an imperfect world, not offering certainty.

Prediction-market prices reflect the activity of many participants, but influence is not evenly distributed. A small proportion of sophisticated, profit-seeking traders appears to control a large share of the capital. They use polling and other data to identify errors and push prices toward their judgments.

Candidates can also affect markets by betting on their own races. This has created misconduct concerns for Kalshi and Polymarket. The Anti-Corruption Data Collective found that political markets can be shifted with relatively small bets. Prices can therefore move because of both informed trading and activity that raises conflicts of interest.

The platforms say they have strengthened controls. Kalshi blocks politicians and campaign workers from political wagers. Both companies say markets quickly correct when large bets move prices. Even so, the article notes that expert traders can be wrong, so price movement is not proof of accurate information.

US midterm elections occur halfway through a president’s four-year term. Voters choose all 435 members of the House of Representatives, while roughly one-third of Senate seats are contested. The article focuses on markets forecasting which party will take each chamber, but it does not explain these election mechanics.

Control matters because the House and Senate pass legislation together with the president. The chamber majorities organize proceedings, control committees, and decide which proposals receive consideration. The Senate also handles confirmations of many presidential nominees. A president generally finds governing easier when the president’s party controls Congress.

When the opposition controls one or both chambers, legislation can face greater resistance, investigations can expand, and spending or appointments may be contested. A president can still use executive powers and veto legislation, subject to constitutional limits. These implications are established background knowledge beyond the article’s detailed discussion of prediction markets.