Study: Small Bets Can Sway US Election Prediction Markets
A new study says small wagers can shift US election prediction market odds, raising integrity concerns; platforms push back.
Prediction markets that track US congressional races are highly susceptible to manipulation, according to a study released Wednesday by the non-profit Anti-Corruption Data Collective (ACDC). The research found that in 94% of the more than 11,000 markets examined on Kalshi and Polymarket, a single bet of under $1,000 could shift the implied probability by 10 percentage points. In some cases on Polymarket, such shifts persisted for a day or more, often lasting around four days.
The findings add to concerns voiced by Democrats and consumer advocates that political betting could undermine public trust in elections by creating doubts about results or enabling insider trading. Proponents of prediction markets argue that the aggregated bets provide valuable real-time forecasts, and several media outlets cite these odds in their coverage.
ACDC researcher Michelle Kendler-Kretsch said the markets for the November midterms and primaries do not meet the conditions needed for accurate pricing, noting that large volumes are often driven by motives other than short-term profit.
Representatives for Kalshi and Polymarket rejected the study's conclusions. Kalshi pointed to a case where a $1 million wager on a market involving former Los Angeles mayoral candidate Spencer Pratt was corrected within nine seconds. A Polymarket spokesperson said mispriced odds create opportunities for traders to profit by pushing prices back to expected levels.