You might have seen the videos lately of dudes sitting in front of a TV, waiting for a politician to say the words “Good morning” at the beginning of the press conference and either cheering or yelling expletives, depending on what is said. This dude is likely betting from either Kalshi or Polymarket, the most popular prediction market apps of today. These apps let you bet on a particular event, which comes in the form of a yes or no question: “yes” or “no”, will the fed chair say “Good morning” at the beginning of his speech? There is a price associated with each voting option that corresponds to the likelihood of that option being true at the current moment; the price of the “yes” and “no” votes will add up to $1. So a “yes” vote priced at 20 cents means the market currently believes there to be a 20% chance of “Good morning” being said and an 80% chance of it not being said. Each bet, or “market”, has a cutoff date when the result is determined and the winning bettors are paid out. When the date arrives, those who chose correctly are paid out 1 dollar for each vote they purchased. Those who were not correct lose all of the money they bet. This means that if I bought 1 “yes” vote at 20 cents and I am correct, I will make 80 cents profit, but if I am not correct, I lose 20 cents. This is simplifying things a bit, but that’s the gist of it.

Turns out out prediction markets are not a new concept. Political betting dates back to 1503 to predict papal and royal successions. In the 1980s, they were used at the University of Iowa to forecast US presidential election outcomes with surprising accuracy. Out of all polls, the prediction market was the closest estimate of each candidate’s share of the popular vote. Why are these things so accurate? Well, one answer is that prediction markets force bettors to be more thoughtful about their answers, as wrong bets mean financial loss. Furthermore, when many bettors make more thoughtful choices, a collective wisdom can emerge. Meanwhile, any noise such as uninformed bets or biases are filtered out, making the outcome more accurate than any single expert. Another reason has to do with the information itself. Informed bettors will likely bet more money on an outcome than will uninformed bettors. Due to the immense speed that information travels today, the message will soon trickle out to wider and wider groups, causing more informed bets and tipping the price/odds towards the correct answer.

Tech CEOs of companies like Kalshi are harping on the accuracy of prediction markets to convince the public that their apps are the news outlets of the future. I think there is some truth to that, especially because crowdsourced information is generally less biased than information obtained from any single news outlet (think Wikipedia vs Fox News). However, what they don’t tell you is that people with control over what’s being bet on can very easily exploit prediction markets for financial gain by deliberately changing outcomes. Brian Armstrong, the CEO of Coinbase (which has partnerships with both Kalshi and Polymarket) rattled off a list of words from a Kalshi market (“Bitcoin”, “Ethereum”, “Web3”, etc.) at the end of an earnings call. With this type of manipulation, the prediction market ceases to be the source of truth for future outcomes and instead becomes an opportunity for powerful people to make lots of money.

Believe it or not, Kalshi is considered a U.S. Commodity Futures Trading Commission (CFTC) regulated exchange, meaning it is a trading venue that is registered with and overseen by the CFTC and complies with the Commodity Exchange Act (CEA). The CEA specifically states in it’s Core principles that contracts must not be readily susceptible to manipulation. To their credit, the CFTC has prohibited congressional decision contracts, to prevent members of congress from having their decisions swayed by markets. This is totally a step in the right direction, but I still don’t think there is enough regulation of these markets as they exist today. I’m not sure how they would be able to successfully mitigate even a small amount of the exploitation happening today without prohibiting certain types of bets altogether. In order to avoid the possibility of manipulation, prediction market events would need to be limited only to categories that a regulatory agency can reasonably oversee. For example, keeping bets on US elections seems fine as they are already heavily regulated for fraud, making them more difficult to manipulate for personal gain. Whereas betting on what a a public figure says in their speech (AKA Mention markets) can be easily manipulated and therefore shouldn’t be allowed. It’s similar to the new proposal for removing prop bets in sports betting – allowing people to bet money on easily manipulatable events is a recipe for insider trading.

Prediction markets can be a useful tool for predicting outcomes, but they are vulnerable to abuse as they exist today. The classification of these markets as legitimate financial exchanges gives people the wrong idea about what they’re in for when signing up. Even if you are making an “informed bet” on an market, you could still get the rug pulled out from under you if people with closer connections to that market can change the outcome. So in short, I wouldn’t put your hard earned money into a prediction market unless 1. you have high signal information about the outcome and 2. the outcome can not easily be manipulated.