Small bets sway US congressional prediction markets

Prediction markets have drawn intense scrutiny in Washington, and new research suggests the congressional side of the business is especially easy to nudge. The Anti-Corruption Data Collective said a $3,500 bet, equal to the maximum personal campaign contribution allowed under federal law, could move 97% of midterm-related markets by five cents or more.

Across more than 11,000 markets on Kalshi, Polymarket and Polymarket’s U.S. platform, ACDC found that 94% would shift by about 10 percentage points in probability after a single bet of less than $1,000. It also said 806 of 1,094 markets priced at five cents or less could be moved five cents for less than $100, and that most markets would move 25 cents after a $25,000 bet.

The report said those moves often lasted. On Polymarket, it identified 353 cases in which one or two wallets moved a market by five cents or more. In 211 cases the new price held, in 62 it kept moving and in 80 it snapped back, while hundreds of markets retained the new level for 24 hours and typically for about four days.

ACDC warned that a manipulated price can ripple through the information ecosystem as if it reflected real momentum. Michelle Kendler-Kretsch, one of the researchers, said the conditions for accuracy were not being met in markets about the November midterms and primaries.

Kalshi and Polymarket pushed back. Polymarket said traders are incentivized to profit from an artificial boost and push the price back to where it should be. Kalshi called the methodology flawed and argued that, in liquid markets, mispricing quickly corrects itself, pointing to a Spencer Pratt case study in which more than $1 million moved a market before prices corrected in nine seconds.

ACDC’s Michael Hornsby said the Pratt example involved a heavily traded market, while thin congressional markets often lack the active trading needed for self-correction. He said the mechanism for manipulation matters more than the motive.