Thomas Peterffy has a problem with prediction markets. Not with the concept, he loves the concept. His problem is with what they’ve become. And I agree with him.
Peterffy, the founder of Interactive Brokers, recently joined the Odd Lots podcast to talk about his new venture, Forecast Trader, IBKR’s dedicated prediction market venue. His pitch is straightforward: prediction markets are one of the most powerful tools humanity has ever invented for aggregating dispersed knowledge into a single, honest number. And we’re filling them with garbage.
“Will Jesus return by 2026?” That’s a real contract trading right now. So is “How many times will Trump tweet today?” This is the problem. When you build a market around questions nobody serious would bet on, you attract nobody serious. You get a casino dressed up as an oracle.
Peterffy’s solution is to focus exclusively on questions that matter: recession probabilities, climate timelines, AI adoption curves. The goal isn’t entertainment. It’s expert consensus, expressed in dollars.
Insider trading isn’t the enemy of prediction markets. It’s the whole point. The reason a prediction market produces useful information is precisely because it rewards people who know things. The monetary incentive is the alignment mechanism: it pulls the informed off the sidelines and into the price. An “insider,” in this context, isn’t a bad actor. They’re what make the product tick.
The real insight is that the insider problem is actually a question-design problem. There is, by construction, no such thing as an insider on the contract “There will be a recession in twelve months”. Nobody has that information. The market works by aggregating thousands of people who’ve each done a piece of the homework.
Now consider the other kind of question, “Will Trump tweet twenty times today?”. Here, the insider problem is real and obvious. Someone close to the principal (or the principal himself) has a genuine information edge that no amount of public research can close. But notice what that question actually is: pure gambling. It produces no useful signal for society (but there is some social value in separating stupid people from their money). It tells us nothing about the future worth knowing. The solution isn’t regulation. It’s curation. Ask better questions and the insider problem largely disappears.
Peterffy’s vision for the future is one where we consult markets instead of economists. He envisions a world where “good” economists participate in prediction markets to prove their theories with capital, and where prediction prices become as standard as credit spreads for reference data.
Peterffy’s vision is compelling enough to prompt a practical question: if this world actually arrives, how should investors position for it?
Before getting to that, though, it’s worth pausing for a detour through financial archaeology. Because this isn’t the first time Wall Street has tried to turn economic uncertainty into a tradable asset.
The Rise and Fall of NFP Options
Long before Kalshi and Polymarket, Wall Street tried to professionalize economic betting…and failed. In 2002, Goldman Sachs and Deutsche Bank launched what they called “Economic Derivatives”: binary options on the Non-Farm Payroll report, sold through high-tech auctions. The premise was elegant. Hedge funds could hedge the specific risk of a surprise jobs number without routing the trade through Treasuries or S&P futures. By 2005, the CME had taken the concept mainstream. By 2008, NFP futures were trading on Globex. By December 2011…they were dead: delisted for lack of interest.
The failure revealed something important: institutional traders didn’t need a specialized market when massive, liquid ones already existed. When the jobs number surprised, Treasuries moved instantly. Why bother with the sideshow?
Peterffy calls this the “chicken and egg” problem. No liquidity means no serious participants. No serious participants means no liquidity.
I have a personal footnote here. Around 2006 (2007?), I was a (very) junior trader on the FX desk of a large Italian bank. On one of those NFP Fridays, I pulled up the CME Auction Markets, read the implied probabilities, compared them to the Bloomberg consensus, and suggested to my colleagues that we take a position in EUR/USD accordingly. To my genuine surprise, they listened. And we made money. We ran the strategy for a few months before I left trading for the corporate world, and the whole thing faded into the background noise of a career.
But the memory never fully left. Not because of the profit, the amounts were modest and I got nothing personally out of it, but because of the feeling. That rare sensation of thinking: maybe I’m onto something here.
How to use Prediction Markets
Assume, for a moment, that deep and liquid prediction markets on recession risk, inflation, and growth become real. What do you do with them?
Three ideas (worth taking seriously?).
First: use them as better economic cycle signals. On our podcast website, we run a model developed by Verdad. It takes a single input, the high-yield spread, and uses it to identify where the economy is in its cycle, then recommends which asset classes to hold accordingly. It’s elegant in its simplicity. It’s also fragile for exactly the same reason.
Here’s the structural problem: high-yield spreads are a derivative signal. The people moving that market aren’t all primarily trying to forecast the economy (or make money). A CFO pricing a bond isn’t trying to send a macroeconomic signal. They’re trying to fund their business. That noise pollutes the indicator. It works until it doesn’t, and there’s no fundamental reason to believe it will keep working forever. Maybe one day companies will start funding themselves using alternative channels, and bye-bye liquidity.
A direct prediction market on recession probability or inflation would be cleaner. The participants have one job: be right. No conflicting incentives, no institutional constraints. Just information, priced.
And I do not think markets will arbitrage away these signals.
The efficient market instinct says that once everyone can see the recession probability, it gets priced in immediately and the edge disappears. But look around. Markets are full of participants who buy short-duration bonds because they like the “surety” of getting their nominal principal back, or chase high-dividend stocks because yield feels like safety. If that level of persistent irrationality coexists with today’s markets, there’s no reason to think prediction market signals will be instantly consumed either.
I didn’t invent the theory that markets are micro-efficient while macro-inefficient, but this is basically a bet on the same.
Second: use them as a portfolio hedge. Can a recession prediction market replace a tail-risk strategy? Probably not cleanly. The payoff is capped: you don’t get the convexity that a well-structured options position provides. You still can get leverage if you buy cheap, when the recession probability is low, but the overall payoff structure is worse.
And you introduce basis risk. But there’s a scenario where prediction markets are better than tail hedges: when the downturn is real but mild. A modest recession might barely move a volatility strategy into the money. A prediction market contract that you bought at 15 cents and closes at a dollar? That pays.
Maybe it is a better hedge if you consider your portfolio and career combined? There are some instances where your portfolio might not need protection, but this would be a…better, capital-efficient emergency fund?
I would like to read more expert ideas on this.
Third: harvest the risk premium on the other side. If prediction markets mature into genuine hedging tools, someone has to take the other side of those hedges. Historically, that’s a good business. What makes it even more attractive right now is a behavioral quirk: the current prediction market crowd skews heavily toward longshot buyers, people buying “Yes, there will be a recession” not to hedge, but for the thrill of the big payout. That’s the same bias that makes lottery tickets persistently mispriced. If you’re systematically the house against longshot bettors, you should, over time, collect a meaningful premium.
What I am reading now:
Follow me on Bluesky @nprotasoni.bsky.social
0 Comments