As you know, I take a pretty dim view of prediction markets, especially for capturing anything more than conventional wisdom when it comes to election outcomes. There’s also a whole controversy about how much prediction markets are vulnerable to insider trading. But the truth is that insider trading is actually an intentional feature of prediction markets. Perhaps not literal insider trading — as in an officer of a public company placing bets on non-public information — but people trading on non-public information. Indeed, many don’t remember this, but one of the seemingly harebrained ideas to come out of George W. Bush’s first term Pentagon was the idea of creating a terrorism prediction market. I’m not kidding about this. It would have basically been a Kalshi or Polymarket about future terror attacks, with the idea being that you’d get people with knowledge of future of attacks to place wagers, and thus, give U.S. policy planners and counterterrorism officers leads on future attacks.
In any case, I note this because again, the real logic of prediction markets is you get the people with the inside dope to place money making bets on future outcomes. You thus use the incentive of winnings to work to pull information out of the realm of secrecy into the open.
I thought of this today because Israeli polls closed at 10:00 p.m. Israel time today and the first exit polls were set to be released immediately afterwards. As we got close to 10 p.m., Gadi Eizenkot’s odds were on the upswing. This seemed more encouraging to me than the vibes and shards of information I was seeing. But I was happy to see it.
The second the exit polls were released his numbers collapsed.
They’re now going back up a bit and people are trying to make sense of what these results tell us. But the point is that I really would have expected people to be doing some trading on the information in those exit polls or among political operatives who sort of do their own version of exit polling. But that sharp drop-off strongly suggests that no one did that.
Now, you can say, ‘great, people acted with integrity and didn’t place trades on the basis of non-public information.’ And I 100% agree. But the promise of these prediction markets is that they will do that, that they’re a way to surface non-public or little known information. Otherwise they’re just opinions and vibes (which I think they mostly are.) I’m not saying this never happens. But this was a good test case. And it pretty clearly didn’t.
