Key facts
- Research indicates U.S. political prediction markets are vulnerable to manipulation.
- A single bet of less than $1,000 could shift the probability by 10% in 94% of congressional race markets studied.
- Market shifts sometimes lasted for 24 hours or up to four days.
- Operators like Kalshi and Polymarket argue that market forces naturally correct distorted prices.
Research published by the Anti-Corruption Data Collective suggests that prediction markets for U.S. elections are highly susceptible to manipulation, with even small bets capable of significantly altering probabilities. The study found that in over 11,000 markets tracking congressional races, 94% would experience a 10% shift in probability from a single bet of less than $1,000. In hundreds of instances on Polymarket, these new price levels persisted for 24 hours and typically lasted around four days.
ACDC researcher Michelle Kendler-Kretsch stated that the large dollar volumes in political prediction markets are not solely motivated by short-term profits, which undermines their accuracy claims. She noted that the conditions for accuracy are not being met in markets concerning the November midterms and primaries.
Representatives for Kalshi and Polymarket disputed these findings. Kalshi cited a case study where a market corrected within nine seconds after over $1 million was invested. A Polymarket spokesperson argued that mispriced odds present opportunities for traders, incentivizing them to restore accurate pricing.
Discussion