Prediction markets notch an important win with Biden's drop out
Add Axios as your preferred source to
see more of our stories on Google.
The prediction markets were right — and Joe Biden himself was wrong.
Why it matters: Accurate polling is increasingly difficult — and many questions, like whether Biden would drop out, can't be answered by polls at all. Enter prediction markets, which provide a minute-by-minute gauge of electoral probabilities.
The big picture: Prediction markets in general, and Polymarket in particular, look pretty smart right now on two different levels.
- Polymarket started making a market on the probability of Biden dropping out of the race on Sept. 21, 2023. Even then, its traders said there was a 22.5% chance he would drop out — much higher than the conventional wisdom inside the Beltway, and roughly three times the likelihood that Trump would drop out.
- More recently, Polymarket's traders flatly didn't believe Biden when he explicitly said he was staying in the race. On the day he said, "as clearly as I can" that "I'm staying in the race," for instance, they still priced in a 66% chance that he would drop out.
How it works: Prediction markets are based on contracts that pay out $1 if the event in question ends up happening, or $0 otherwise. If the market is efficient, then the price in cents corresponds to the perceived probability the event will happen.
- Volumes are much higher now than they were just a couple of years ago, with more than $400 million changing hands on Polymarket alone in 2024.
- A lot of that volume is a function of crypto traders doing short-term arbitrage between different markets, however, rather than representing informed investors getting real skin in the game.
Reality check: It's dangerous to infer or extrapolate too much from a single successful prediction. Prediction markets got the 2022 midterms wrong, for example, and especially when it comes to political events, they have some profound weaknesses.
- Traders tend to rely quite heavily on polls, creating a GIGO (garbage in, garbage out) problem. If the polls aren't reliable, then the prediction markets won't be either.
- While financial markets are dominated by highly-informed investors, political prediction markets are often dominated by so-called noise traders, also known as "degen gamblers" in the crypto world.
- Nate Silver, who founded polling analysis firm FiveThirtyEight and recently joined Polymarket as an adviser, writes in his forthcoming book that when it comes to big elections, "there's so much dumb money out there... that there isn't necessarily enough smart money to offset it."
- Silver points in particular to the millions of dollars spent in 2020 betting that Donald Trump would get re-elected after Biden had already been declared the winner.
The bottom line: Prediction markets are surely more accurate now than they were a few years ago. But elections by their nature tend to be noisy and unpredictable, and no amount of gambling on their outcomes is going to change that.
