Prediction markets have an elections problem
Prediction markets for elections often misprice long-shot outcomes and near-certain events, not just because traders are irrational but due to structural issues like high fees, small bet limits, illiquidity, and the time value and platform risk of locked-up funds. Commenters argue that these frictions cap how accurate markets can be, especially for very low-probability or already-decided outcomes, and allow “dumb money” driven by partisanship or signaling to persistently distort prices. There is broad interest in better-designed, better-regulated markets—potentially with interest-bearing deposits, higher limits, and lower fees—but also concern about ethical risks, regulatory barriers, and whether such markets can scale without incentivizing manipulation of real-world events.
Market inefficiencies and structural issues
- Many comments argue that election prediction markets, especially PredictIt, are structurally distorted: high fees, low bet limits, illiquidity, and slow settlement all prevent “smart money” from fully correcting prices.
- Events with probabilities near 0 or 1 are seen as systematically mispriced because potential arbitrage (e.g., 1–2% return on a near-certainty months out) is dominated by safer, higher-yield alternatives.
- Bet-size caps (e.g., ~$800–$850 per market) mean a single informed trader cannot neutralize a large long-shot order; multiple rational traders are needed to move the book.
Fees, opportunity cost, and platform risk
- Platform fees (e.g., 10% on profits, ~5% withdrawal) plus the time value of locked capital set an effective lower bound on exploitable mispricing; small edges are not worth it.
- Counterparty risk (platform shutdown or biased resolution) further reduces willingness to tie up money in “sure thing” bets.
- Some argue these frictions alone can explain 5–10% price distortions, especially on long-dated, high-certainty markets.
Behavioral biases and “dumb money”
- Long-shot/favorite biases are widely observed: low-probability “YES” contracts, including conspiratorial or partisan outcomes, tend to be overpriced because they are “fun” lottery tickets or expressions of identity.
- In 2020, many bettors seemingly spent money to signal allegiance (e.g., betting on Trump after loss), turning markets into emotional or tribal outlets rather than profit-seeking venues.
- Others counter that some “impossible” scenarios (e.g., post-election procedural maneuvers) were at least non-zero probability.
Usefulness vs limitations as predictors
- Some commenters still find markets better-calibrated than many pundits and roughly comparable to polling aggregators, especially pre-election.
- Others emphasize that markets are not pure probability estimators but “weight of money” systems, easily skewed when informed capital is constrained.
Regulation, ethics, and scale
- Legal restrictions limit who can run markets and cap stakes, which keeps out institutional expertise but also avoids extreme cases like de facto assassination or death markets.
- There is concern that large election markets could create incentives to influence or rig outcomes, not just predict them.
Trader experiences and improvement ideas
- Several report profitable but unscalable arbitrage and edge-exploiting strategies.
- Proposed fixes include interest-bearing deposits, lower fees, higher limits, interval-based probability displays, and better visualization of uncertainty.