Most people probably don’t realize it, but the prediction/betting market like Kalshi and Polymarket have begun taking bets on things like the success of Phase III drug trials. The New York Times highlighted this recently, and I wanted to weigh in on this idea. Not least because it’s something that I had though about before!
That goes back to the early 1990s, when I discussed the idea (with a few colleagues) of what I envisioned as more of a futures market in drug discovery. Like other futures markets, a primary purpose of it would have been a place for financial hedging of risk (that’s what happens with commodities futures in things like grain or fuel oil). Running a Phase III is indeed a substantial risk, and my initial thought was that if there were a way to distribute this risk more widely or evenly then it might lower the barrier to running the trials.
But the more I thought about it, the more problems appeared. A market like this only really works if there are a substantial number of active participants, enough to take both sides of the contracts and thus to set what should be a reasonably accurate price for them. But the prices of commodities are in the end a matter of supply and demand - for futures contracts, what traders think will be the future course of both. Information about the likely weather, international affairs, inflation and interest rates, supply fluctuations around the world for the same commodities - all of those factor into the price of a contract. And that’s because these are largely fungible commodities - a barrel of oil of defined grade or a container load of wheat is useful around the world and can be bought and sold on that basis.
Drug trials, though, are largely a yes-or-no bet on a single entity. Each clinical candidate is a different story - it’s much closer to betting at the beginning of a season on which team you think will win a championship (except in that case, you know that one of them at least will succeed!) But those tend to be “yes or no” bets and the odds that obtain when you placed the bet are the odds that you get if it pays out for you. If I’m enough of a wild man to put money down today on the Cleveland Browns winning the next Super Bowl, I deserve the reward if I’m right!
A grievously high percentage of Phase III trials fail, of course, so the only rational “yes or no” betting strategy would always be on failure. The way betting markets handle this sort of thing is by (for example) having a particular contract be worth anywhere from 1 to 99 units, with that price fluctuating as more information comes in. I could get that Cleveland Browns contract at 1 right now (no thanks), and if they went on a mighty tear through the regular season and the playoffs, that price would of course go up. I might actually choose to sell my position before the actual Super Bowl game and take my winnings rather than face the maximum win/maximum loss resolution waiting for me in the game itself.
But that position’s value will change on the basis of real information coming along (just as the price of futures contracts on oil or grain do). And in the quotes from a Kalshi executive in the Times story, you can see that’s how they’re picturing it:
He argued that Kalshi was enabling the wisdom of the crowds to cut through public relations spin and hype, providing a clear signal that could direct investment and resources toward the most promising experimental drugs.
“This is an incredibly important, incredibly sensitive, incredibly impactful subject matter area,” Mr. Such said. “This is why the people making decisions need to have the best possible information available, and the avenue for the best possible information is traders on Kalshi putting money where their mouth is.”
Here’s a big problem, though: very little information is being revealed in a drug trial! There are very, very few updates about how a particular trial is going - at the most, there might be a predefined point where a committee looks over the results and decides whether to keep going, halt for futility, or (rarely) halt for overwhelming efficacy. And that’s it! Everyone just gets to grit their teeth and wait until the actual results are announced, as stock price movements vividly reveal. Now, it would be nice to have an implement to cut through the spin and hype, but as it turns out, prediction markets of this kind ain’t it. They are, in actuality, another playground for more of both.
My drug-trial-futures-market idea came to nothing of course, not least because the regulatory environment of the time wouldn’t have allowed for any such thing, any more than it would have allowed for the existence of Kalshi or Polymarket, etc. It would have been very unlikely to serve its intended financial purpose, and on top of that I was not sufficiently cynical about human nature, which is something that you can learn about by looking at those markets at present. There is currently a lot of agitation about the way that some prediction markets are allowing bets on events like wildfires, and the obvious problem here is that there are people who will take it upon themselves to provide the wildfires if the payout is good enough. The corrosive effect of the betting markets on professional sports is becoming apparent, and the same thing can apply anywhere. It’s not like an outside bettor can do much to make a trial successful, but there might well be ways to mess one up if the money is right. Perhaps a bet on how many people drop out of the trial or experience serious side effects? Kalshi has already filed documents saying that it’s thinking about offering those.
No, the effort that would be involved to block off all the avenues for trouble does not seem to be something that the betting/prediction companies would be willing to make, assurances on their part to the contrary. And what does this add that the stock and options markets don’t? OK, there are a few nonpublic companies running trials, but not very many at all. And the market on a specific trial does break that out from the rest of the company’s activities and prospects, fine - but the sorts of bets that will be offered will be the ones where the trial results are big enough news that the existing financial markets are already reflecting what outside observers think. As for that, and the wisdom of crowds - I think in cases like this, where the size of the crowd that actually understands what they’re talking about is relatively small, and where the amount of real information they have is similarly constrained, well. . .how much wisdom is really available?
I’ve saved one last consideration. Kalshi and Polymarket say that they do not offer life-and-death bets (such as whether some public figure will be assassinated by a certain date), and that’s for what should be blindingly obvious reasons. But as the Times article mentions, a bet on the success of an investigational cancer drug is in fact usually a bet on how many people are going to die while taking it. Cancer patients themselves believe that have enough tension in their lives without feeling as if they’re horses running around a track while money changes hands around them. Taking positions in the company’s stock does not feel quite as offensive, somehow.