What happens when election betting goes mainstream

So are these midterms the prediction market election? That's what the CEO of Kaoshi, the largest prediction market, claims. Users can place bets not just on who will win in the midterms, but on the turnout numbers, endorsements, if candidates will debate, and even what will be said at rallies. So what exactly does that mean for their effect? This is the biggest experiment that we ever had in American politics when it comes to prediction markets that are open to the general public, that don't have caps on betting. That's because in 2024, Kalshi won a landmark battle that opened the door for Americans to bet on elections in prediction markets, something that's illegal in Canada. And since then, they have become extremely popular. This will be the first year that the prediction markets are really mainstream for a major election cycle. According to CNN, more than $1.4 billion has been traded on midterm elections in markets like Kalshi and Polymarket. Not to mention, Kalshi has data-sharing partnerships with CNN, CNBC, and Fox, while Polymarket has them with Dow Jones and Substack. According to the Kalshi prediction market, look at this. It's a 64% chance that Democrats do in fact win back control of the Senate. Now, if this is the prediction market election, there's been a lot of debate over whether they might be more accurate than polling. But is this really the case? Valerie Cross is a journalist who covers prediction markets. She says that polls provide more of a snapshot in time, while prediction markets fill that gap. They're also more of a real-time barometer of the implied probability of who's going to win the election. Back in 1988, when the first prediction market was created, it was designed to do just that. Civil rights leader Jesse Jackson had just won the Michigan presidential primary, despite polls predicting Michael Dukakis would win. And three economists walked into a bar, quite literally. Robert Forsythe at the University of Iowa was one of those three economists. As economists, what would we do if we were going to try to predict an election? Well, let's run a market on it. The three economists built the market at the university as an academic endeavor, And only faculty, staff, and students could bet a maximum of $500 on a given bet. Lo and behold, by the time the election came around, our market predicted the outcome within two-tenths of one percent. That formed the foundation of what is now the Iowa Electronic Markets, still running today. They are used for academic purposes to better understand business and technology. And Forsythe says he sees Kalshi and Polymarket as a large-scale version of what he and his colleagues created. Now, although there have been many headlines about their accuracy, this economist warns that they can be volatile. He analysed the 2024 election and found that while prediction markets performed well when it came to predicting a presidential winner, they weren't as accurate when it came to down-ballot races. It's a mixed bag and that's not really surprising because there's a long literature in finance about how markets react very, very quickly. But they're also subject sometimes to what's referred to in economics as excess volatility. Sometimes they'll overreact to news. And while they react faster to unexpected events or news compared to polls, Sethi says pitting prediction markets against them can be a bit misleading. Polls are really the principal ingredient for making forecasts, even for traders on prediction markets. When the polls move, the markets move. It's just that the markets can also move based on other factors and they can move in the absence of polls. So how else might they affect these midterms? Despite the uptick in their use, bets on elections still represent a small fraction of prediction markets, but that doesn't mean they haven't had some effect. According to Kalshi, more and more users are going to the site for information over trading. And when it comes to midterm markets, viewers outnumber traders five to one. And this might be a concern for how these markets could impact the midterm elections. One study suggests that using these sites for information over trading is influencing US voters. This study showed the higher a candidate's odds of winning on Kalshi, the more US voters were likely to think the candidate was more electable, by either 10 or 20%. That same poll found that while US voters might be influenced by these markets, they don't trust them, with 43% saying they did not think they represented the public's opinion. And 80% of US voters said they wanted prediction markets to be more tightly regulated or banned outright. The vulnerability there is that many of these markets, there actually aren't that many people that are betting. David Zucconi is the co-founder of the Anti-Corruption Data Collective. His team found that the majority of bets are concentrated on a small number of races. He found that specifically on Polymarket, the top 1% of wallets That refers to the small group of traders making the most profit, were responsible for 70% of all betting on congressional and midterm markets. And that means that basically anybody can spend $100 or $1,000 and shift the odds on a public-facing website and create the impression that a candidate is doing better or worse than they are in reality. Plus, there's a category of questions in election markets like, is this candidate gonna enter the race? Or is there gonna be a public appearance from this candidate? That, Zucconi describes as a knowable market, which makes them vulnerable to insider trading. All of these things reflect candidate strategy or internal discussions of which there could be an insider that could exploit that information and then make money for themselves. So does that mean that this is really the prediction market election? The experts I spoke to all say we'll have to wait and see how effective they are in predicting the outcomes and how much influence they really have. We're kind of living this experiment in real time. No one knows whether or not the last couple of months of prediction market information saturating the environment has been a good thing or a bad thing, but they've already kind of ingratiated themselves in American society. It would be pretty hard to, you know, put the genie back in the bottle. Certainly election trading in some form is here to stay.