Betting on Reality

Everybody’s talking about prediction markets. In the hours before the US struck Iran, a cluster of large anonymous bets landed on exactly the right answer as to when the war would start. 

When Iran’s Supreme Leader Ali Khamenei died, Kalshi (partnered with CNN) froze $77 million in winning bets — declaring it wouldn’t pay out on markets tied to death. Like police prefect Renault in Casablanca, Kalshi was shocked — shocked! — to find that gambling was going on here, while pocketing their winnings.

But almost nobody is explaining what these markets actually are, how they work, or why they keep getting the future right when everyone else gets it wrong. Our guest on this week’s WhoWhatWhy podcast is Robin Hanson, an economist who saw this coming.  

A professor at George Mason University, he started the first internal corporate prediction market in 1990 — decades before any of this was an industry — and invented much of the mathematical machinery these platforms run on today. He’s watched prediction markets get killed by Congress, dismissed by regulators, and ignored by the institutions that needed them most. 

And he’s been asking the same question for 40 years: Why do institutions fight so hard against a mechanism designed to surface truth?

The answer, it turns out, may be the most unsettling thing about prediction markets — not what they get right, but what their cold welcome from establishment institutions reveals. 

Hanson is also the co-author of The Elephant in the Brain, which argues that almost nothing we do is actually about what we say it’s about. Medicine isn’t about health. School isn’t about learning. Politics isn’t about policy. They’re about signaling. 

And prediction markets, which ruthlessly expose the gap between what we perform and what we actually believe, are a direct assault on that entire pretense.

Hanson explains why markets consistently outperform experts and polls, what Kalshi’s freeze on the Khamenei winnings reveals about institutional self-interest — and what happens to a truth-telling mechanism once everyone is watching.

Apple PodcastsGoogle Podcasts RSS

Full Text Transcript:

Jeff Schechtman: Welcome to the WhoWhatWhy podcast. I’m your host, Jeff Schechtman. In the hours before the United States struck Iran, something strange happened on a website most people never heard of. Hundreds of anonymous bets, each over a thousand dollars, landed on a single question. Would America attack Iran before tomorrow? The bets were right. Nobody knew who placed them. Nobody knows to this day whether those bettors had inside information or simply read the geopolitical tea leaves better than the generals being quoted on television. That story lives at the center of one of the most disruptive ideas loose in the world right now. The idea that markets, not experts, not panels, not polls, not pundits, markets with real money on the line might be the most reliable, truth-telling mechanism human beings have ever devised. And that everything we’ve built to tell us what’s true and what’s likely — journalism, polling, peer review, political forecasting — might be elaborate performances, signals of authority and expertise, but have almost nothing to do with actually getting the future right. 

My guest, Robin Hanson, has been arguing this for nearly four decades, long before Kalshi was showing live odds on CNN, long before Polymarket was flashing ‘prediction’ at the Golden Globes, back when he was a researcher at NASA and Lockheed, he started the first internal corporate prediction market in 1990, and then spent the next 30 years watching the world slowly catch up to an idea it kept finding reasons to resist. Hanson is an economist at George Mason University. He’s invented the mathematical machinery that runs many of today’s prediction markets. He’s proposed something called futarchy, a form of government where we vote on our values but let markets decide how to get there. And he’s written a book called The Elephant in the Brain, what may be the most unsettling version of a simple observation that almost nothing we do is actually about what we say it’s about. Medicine is not about health, education not about learning, politics not about policy. They are about signaling, about performing care, performing virtue, performing expertise. And if he’s right about that, then prediction markets aren’t just a forecasting tool. They’re a direct assault on the entire performance. For most of his career, that argument lived in academic papers, blog posts, and conference rooms where people found it fascinating and then went back to doing other things the way they’ve always been done. The idea kept getting rediscovered. It kept not getting adopted. And Hanson kept asking the same question: why do institutions fight so hard against a mechanism designed to surface truth? Then 2024 happened. Suddenly Kalshi was on CNN. Polymarket was flashing predictions. The markets had Trump winning when every poll called it a coin toss and the money followed. Kalshi, now valued at $22 billion, Polymarket at $9, combined trading volume pushing $50 billion. The fringe became the infrastructure almost overnight. The question Hanson has been sitting with for decades has finally arrived. Not as a thought experiment but as a live problem. Does mainstreaming prediction markets make them better, or does it expose them to exactly the same forces of manipulation, performance, and institutional capture that corrupt every other truth-telling mechanism we’ve ever built? It is my pleasure to welcome Professor Robin Hanson here to the WhoWhatWhy podcast. Robin, thanks so much for joining us.

Robin Hanson: Thanks for having me, Jeff.

Jeff: Thank you so much for joining us. You’ve been working on this idea, as I mentioned in the introduction, really since the 1990s, before the internet, before crypto, before the industry of prediction markets existed. What did you see all those years ago that nobody else saw?

Robin: I saw that betting markets do a great job of aggregating information, speculative markets. Now this was a standard result in finance literature for many decades. This isn’t new there. People have known that compared to other mechanisms, financial markets just do a great job of aggregating information. I don’t know why they weren’t grandiose as I was, but I just thought, well, then couldn’t we use this a lot more places? And I guess that was my unique addition there. So it was inspired by being part of a group that was trying to make the World Wide Web. So, there was a time before there was a web, and in fact, I was, there were people who were imagining that. And that was sort of our grandiose ambition, that we could change the world by making a thing that didn’t exist. And we had hopes for it improving the ability to find criticism. And that was our big hope. And then I started to have doubts about that, and I was looking around for something else that could serve this same grandiose vision, a way to change how we think together. And that’s when I came up with, well, why not have more betting markets?

Jeff: And there was at every level, it seems, resistance to this, resistance in the corporate world, even though, as you say, it existed in financial markets for a long time before that. Talk about that resistance.

Robin: Well, the world likes, you know, it gets used to things. And when you want to change things, people are kind of scared because they’ve grown up with something, and their skills are tied to it, and they are worried that it might disrupt things. And in fact, changes do disrupt things. But I was actually initially focused on, say, applying this to public conversation, the sort of things that are in the news or policy debates. And then I realized a few years later that, in fact, the most likely place to get value out of information in prediction markets was in decisions, and in particular, in organizational decisions. And so I started to focus on how could we set up markets in organizations to help their decisions? And that’s when I started to realize just how much politics and, you know, non-transparent strategies and behavior really happens in typical organizations. So most managers like to present themselves as some sort of scientific decision maker with a spreadsheet. They’re trying to fill in the numbers to make the right decisions. But in fact, they’re more typically seen as politicians: they have alliances and they’re trying to set narratives and try to recruit people and undermine their opponents. And all of that can be, you know, messed with if you have markets just telling the truth.

Jeff: You were involved with a DARPA project many years ago, which was set up to predict geopolitical events. And Congress killed the funding for it. There was a real firestorm in opposition to it. Talk about that as an expansion of what you’re saying here.

Robin: So, in the late, I guess, Clinton administration, the Department of Defense said, we’ve heard about these markets, show us the, you know, they can help us do our things. And so they did a call for proposals. And I responded to that call with colleagues from my school where I just got my PhD. And we created a project to explore the concept of combinatorial prediction markets and apply it to geopolitical events in the Middle East. And we were setting up to go live with that. We put up a call for participation and sort of a test system. And then on a Monday morning, two senators had a press conference where they declared that the Department of Defense was about to have betting markets on terrorist attacks, and that was a terrible thing. And less than a day later, in front of Congress, the Secretary of Defense declared the project was dead. Congress didn’t even have to vote on anything. And in that 24 hours, nobody asked us if the accusations were correct. They happened to do their press conference when the DARPA PR person was unavailable. And so, for the next few months, we had a lot of discussion in the press. And I did a lot of interviews about what this project was supposed to do and whether it was a good or bad idea. And the biggest complaints about this market were that people might be enticed to actually cause terrorist attacks in order to win $30 in the market, or that people might trade in the markets to distort the prices and mislead us about terrorist attacks. And those were… But I think the biggest complaint was just the idea that it was bad to bet on death, betting on death was immoral, and therefore, these betting on terrorist attacks shouldn’t happen now. We weren’t actually trying to make a market on terrorist attacks. We were making a market on geopolitical events in the Middle East. But we did have in a sample web page, a sample miscellaneous category of two events that we thought might go in the mix, one of which was, what if Arafat had been assassinated or North Korea did a missile strike? How might those change events in the Middle East? And that was the basis for saying that we were going to make a market on terrorist attacks.

Jeff: And really, it seemed that the whole business of prediction markets faded away, or at least from public consciousness for a while after that, really until 2024 and the Trump election. Talk about that. What went on during that intervening period and why all of a sudden it is so front and center again? What was the inflection point?

Robin: So in the aughts, there were a number of prediction market projects, and one of them was something called the Hollywood Stock Exchange, and it was focused on movies, and it was a play money betting market on movies. And it did quite well, aggregating information about movies and forecasting various movie events. And the people who did it thought this was so good, they wanted to create real money versions. And so they spent the millions of dollars required to go through the official regulatory approval process at the CFTC to get their movie markets approved, which they did. And then a few months before they were going to go live, Hollywood executives got wind of this and didn’t like the idea, and they lobbied Congress to pass a law to make it illegal to have movie markets. And a similar thing happened in the ’50s with onion futures. Apparently, people who grew onions did not like the idea of onion futures, and they also then got Congress to pass a law to make onion futures illegal. So, that was a big blow to the possibility of prediction markets in the aughts because having gone through all the official process to get approved, they were still shut down by Congress changing the law. Now, in the last few years, basically, regulators changed their interpretation of the law to allow Kelsey, first, and then a number of others have recently got approval, and that’s the major thing that happened in the last few years was a change in regulatory interpretation.

Jeff: And one of the other things that gave credibility to these markets contemporaneously here is that in 2024, the polls all said the election, that the presidential election was a toss up. And yet the prediction markets had a different story, which turned out to be right.

Robin: So, betting markets on elections have been a big part of the story for many decades now. And we have seen in head-to-head comparisons that betting markets just do better than polls in forecasting. And we can go into the details of that. And sometimes that’s involved with, you know, dramatic successes, like in the Trump winning, the regular sources were much more skeptical about that happening than the betting markets, and the betting markets turned out to be right. Although in Trump’s first election, the markets had said he was only at a 15 percent chance of winning and he won. Then people complained that the markets were wrong then, I guess. A similar thing happened in Brexit around that period. Brexit was also thought to be unlikely, even though it happened. But we have a large data set on all these things and we can just say, overall, the betting markets are more accurate than the polls in forecasting elections.

Jeff: And how much of that has to do with the success of the betting markets or the failures of polling, which has been broken for a long time in terms of response rates, structural undercounting, there have been problems with polling. To what extent is the failure of polling really driving prediction markets?

Robin: Well, it’s more that, you know, it takes some intelligence and finesse to try to make mechanisms work well. So a story of the Trump recent election was that the usual polls weren’t that good. And a French bettor, somebody with money in France, was trying to bet in the prediction markets, commission their own polls, using some better methodology to get more accurate estimates, and then they bet on the basis of those more accurate estimates. So, you know, prediction markets are like a meta institution in that they give people incentives to try to find all the other institutions that can work and make them work well. So polling or even journalism doesn’t necessarily work well by itself. They need some sort of incentive or push. Somebody has to figure out how to make them work well.

Jeff: So what happens, though, when the mechanism for prediction markets really becomes mainstream? Does it make the prediction markets better because it’s a wider audience, or does it make them more arguably corrupted in some way?

Robin: We’ve had financial markets for a long time now. Well over a century we’ve had large financial markets. So all the issues you’re seeing in prediction markets are also issues that have been in financial markets for a long time. So I don’t think you necessarily have to imagine a strange new world. You can just look at the familiar world that we’ve seen. Financial markets have given people incentive to uncover information about topics relative to financial markets. They sometimes give people incentives to go change the world in order to get advantages in the markets. They have given people incentives to try to manipulate markets, to change people’s impressions. And notice that these are all just generic issues of information institutions. They also appear for journalism, gossip, academic journals, even polling. That is, we just have a set of information institutions in the world and they all have a set of issues or concerns you could have about them in common. And I think people are somewhat unfairly complaining about these issues showing up in prediction markets when they also show up in all the other institutions, including all the other financial markets we’ve had for a long time.

Jeff: Except that there are no insider trading prohibitions with respect to these platforms that we have today, these prediction market platforms.

Robin: That’s not true. The CFTC has declared that insider trading is illegal, and in fact, the platforms have declared that they are not going to allow it. And that’s been, those laws were even strengthened in the last two decades, compared to what the legal precedent was before that. I’m not so sure these are a good idea, to go so strong with these rules, but those are the rules.

Jeff: Talk about that and why you think that it may not be such a good idea.

Robin: Well, so a century ago we had insider trading laws for stocks. Those were created by regulators reinterpreting their legislative authority. And their rationale was that it was better to have less accurate stock prices if we could have more people buying stocks. The quantity of stock purchases was the priority, and thereby reducing insider trading, they would therefore reduce the thing that would push people away from buying stocks, is that they would be trading at a disadvantage. That rationale just doesn’t really apply to these other prediction markets. There’s no particular priority we have in more people trading in prediction markets and having a higher quantity of trades. But people have now substituted a rationale of just, organizations should be allowed to keep their secrets. That is, many organizations have secrets, and they get their employees and associates to promise to keep their secrets. And then sometimes those people don’t keep their secrets. And many famous news stories over the last decades have been based on people failing to keep organization’s secrets. I’d say that overall, as a society, we don’t want to go to either extreme. We don’t want to just make everybody reveal all their secrets. That’s too far, on the one hand, but we don’t want to also just require everybody at the criminal penalty to help all organizations keep all their secrets. Many good things have come by journalists uncovering secrets that organizations had tried to keep. So I would rather let us sit in the middle saying, okay, as an organization, you can use contract and even laws of theft to help people keep your secrets. But then journalists and others can try to find out your secrets, if the larger world wants to know.

Jeff: What happens, though, when the line between uncovering the secrets and these markets themselves basically disappears. I mean, CNN and CNBC have financial relationships now with Kalshi. And so the market moves and these networks cover the move as news, and then the coverage moves the markets, and there’s this feedback loop going on there. And there’s the danger one would think for manipulation to take place. Talk about that.

Robin: As I said before, all the issues you could complain about with prediction markets already exist with all the other information institutions, including journalism. Set aside prediction markets. Journalism has for a long time had people trying to manipulate the world and manipulate the news in order to get more favorable coverage, and there’s lots of money that goes on in journalism, and money has influenced what journalism does for a long time. So all of these issues are already issues with journalism. There are also issues with gossip. There are issues with academia. These are just general issues in having information institutions. And we should just talk about for each issue, what’s the best way to deal with it in all the institutions.

Jeff: Then there’s the issue — and you brought this up before in terms of some of the early work with respect to terrorism — of betting on death, I guess. Kalshi froze, you know, what was it, $70 or $80 million in winning bets, based on Khamenei’s death? Talk about that.

Robin: When journalism first arose centuries ago, the usual authorities of the time, the church, the state, aristocrats didn’t like it. It was out of control. It was sensationalist. It wasn’t giving them proper respect. It was degrading their authority and prestige. And they crushed it, to many degrees. They regulated and crushed and pushed down on journalism until it could conform to give them proper respect. And I think today, prediction markets are now seen as an upstart, and that journalism is seen as the proper authorities. And many people are upset that their proper authority and deference is being questioned and displaced by these markets that people are turning to instead of to journalists. And they’re holding the markets to standards that they’re not willing to hold journalism to. That is, journalists write about death, you know, and journalists make money writing about death all the time. That’s one of their big topics to write about. And so complaining that people might make money in the markets telling people about death seems a bit hypocritical to me, if you also make money in journalism writing about death. If you want to just make a whole nonprofit death journalism thing, that would be kind of hard to set up.

Jeff: So did Kalshi do the wrong thing in freezing this money?

Robin: Kalshi is a for profit organization, and it fears regulatory disapproval and public backlash. And so I expect it’s doing what it thinks wise to avoid that sort of backlash. So people trying to prevent a public backlash is quite an understandable thing to do, but it might not be for the best in terms of public information.

Jeff: I want to talk about something that’s a subset of this, something fascinating that you’ve written about in your book, The Elephant in the Brain. The idea that nothing is really about what we say it’s about. Talk about that and as a concept, and then we can talk about how it may apply to this whole idea of prediction markets.

Robin: Sure. So, I was an economist — I am an economist — and over the decades I have learned many areas where economists and other social scientists try to understand the world and build theories about human behavior. And repeatedly I’ve come across areas where our standard stories about what we are doing are just puzzling, in the sense that they don’t make sense of a lot of the particulars of what we do. And over time, I realized that many of these puzzles could be explained if you were just willing to offer a different theory about what our motives were. So if we look at politics and charity and medicine and consumption and school — in a wide range of areas, it just looks weird what we’re doing when you look at the details. But if you just say maybe we’re just doing something else, a lot of it makes a lot more sense. And so that’s the basis of saying we’re just often wrong about why we do things. The reasons we give for why we do things make us look good, or at least avoid looking bad. But they aren’t the real reasons. So, for example, we don’t actually go to school to learn. We don’t actually learn much. Whatever we learn, we don’t remember much, but still we get paid more for going to school. And plausibly that’s because we go to school to show off. We show that we’re a smart, conscientious, conformist, and employers value that, even if they don’t much value what we’ve learned because we hardly earn anything, we hardly remember anything. So that’s an example of hidden motives. 

So that’s plausible all through our lives. And this framing helped me understand how corporations are often not that interested in prediction markets that might help them better understand their business. So here’s a picture I like to give for this. Imagine a C-suite conference room, a big table with all the top people in a company sitting around the table. Imagine we put at this table an autist, that is somebody who knows the company really well, and any topic comes up, they can say relevant things about that topic, but they have no sense of what the agendas of people around the table are, what people want to hear, what sort of narratives they’re trying to push. This person just speaks up whatever they know that’s relevant for the topic at the table. This person won’t be allowed to sit at this table very long. Most of those tables are quite political, and people have agendas and narratives they’re pushing, and the people around the table know that, and they either try to sort of cooperate with the dominant coalition, or maybe try to undermine it, but either way, they’re not just oblivious to the agenda. But a prediction market is in fact such an oblivious autist. They just give their current estimate, they update it immediately as soon as somebody knows something, and they have no sense of what anybody wants to hear. And so even though the management might claim, yes, we want to know everything relevant to our business, as soon as that information is available, we are eager to get more accurate information about our business, that’s just not true. That is, they might privately want to know things, but in terms of what gets said publicly, they have strong considerations other than that. And that’s an example of hidden motives in business. And I could walk you through more detailed examples of how this plays out with a particular market, say, on a deadline.

Jeff: Please do. Yes.

Robin: Okay. So one of the most popular applications initially, at least, for prediction markets and organizations, was deadlines. Many companies have projects, projects have deadlines, and there can be hell to pay if you don’t make your deadline. So a key question is often: will we make our deadline? And one way to find this out, and usually done, is you have project meetings. People sit around every month or so and say, how are we doing? What’s it look like? Are we going to make our deadline? And typically people say, yes, we’re going to make the deadline, it looks like we’re on track. And then quite often, they just don’t make the deadline. And so these project meeting forecasts are famously inaccurate at predicting whether the deadlines will be met. Now, a very simple solution is just to set up a betting market on will we make the deadline and let people anonymously trade in these markets. And what typically happens is the betting market odds of making the deadline are just much lower than what the committees have said, and typically more accurate. 

Okay, so that sounds great, right? Well, people running projects with deadlines do not want these markets. Let me explain why. So if you have a project with a deadline, you care about whether you’re going to make it, but you care even more about having a good excuse if you don’t. And most everyone’s favorite excuse is the following: we were going along just fine until the last minute, when something weird came out of left field and knocked us down, and it’s such a weird, unusual thing. It’s never going to happen again. So let’s just forget about it and go on. And everybody likes that excuse. But if you have a betting market that’s been saying for six months, you’re not going to make the deadline, you don’t have this excuse anymore. You can’t say we were all going along fine until the last minute because you weren’t. You were not going along fine. You were not going to make it six months ago. So people would rather have a good excuse, you see, than know, for example, what they’re going to do. Also, if you, say, have a group of people who are working toward a deadline, imagine either they’re pretty sure to make the deadline, pretty sure not to make the deadline, or right on the borderline of maybe we’ll make it, or maybe we don’t. It’s that last state that has people put in the most work, so managers are often trying to put their employees in that state of thinking, they’re just on the borderline of maybe making the deadline, so that they’ll get more work out of them. And that’s also a reason why you don’t want an accurate forecast of whether you make the deadline. You want people to believe you are just on the verge of maybe, or maybe not.

Jeff: So talk about it as it relates to this broader idea of futarchy, and I hope I’m pronouncing that right. The idea…

Robin: “Futarchy” is the way I say it.

Jeff: Right. The idea of it really being institutionalized in government.

Robin: So, at the moment when a CEO and the board say to the world, we want to raise more capital for this company, if that approval can just go straight without shareholder approval, then on average stock prices drop 2 percent. But if the shareholder approval is required, stock prices go up 2 percent. That is, investors think that shareholder approval being required for such a choice is good for them. It’s going to help them avoid situations where CEOs are trying to raise money and it’s not so good for them. Maybe the company is in trouble and they should really just cut back. So that’s an example of how investors don’t always just want to have CEOs be in charge of everything and do whatever they want. It can be valuable to have oversight where a larger pool of investors can check the CEO’s choices. And that’s a motivation for wanting other ways that it might be cheaper to do that sort of checking. And so futarchy is another name for decision markets or a way to use markets, speculative markets, to do that sort of checking. Let me give an example of whether to keep the CEO as the sort of decision. Obviously, you wouldn’t want to leave up to the CEO or necessarily even to the board, when a public company, you have a stock price, i.e., you have a stock and it has some price, and when you trade that stock, you trade it for cash. And you ask yourself, let me average over all the situations the company could be in and in each one say, how much is the company worth? And when I get this average, compare it to the price. And if the price is lower than that, I should buy. And if the price is higher, I should sell. And that’s how you do stock trading. Now what we can do is have conditional stock prices. We could have trades of cash for stock, but we call off those trades if a condition isn’t met. So we can then get conditional prices for the stock. And in particular we could get the price of the stock if the CEO stays past the end of the quarter and the price of the stock if the CEO is gone by the end of the quarter. And if those two prices are different, that difference is what the speculators are telling you about the value the CEO is giving to the company. If that value is positive, you should keep them. But if the value is negative, then you should swap them out. And that would be a way in which the market could directly give you advice about that key question of keeping the CEO. It could also give you advice about whether to approve the the CEO’s proposal to raise more money or to approve a merger or acquisition or restructuring. Any of these major decisions of a company, we could have the markets evaluate those proposals from the point of, will the company be worth more or less if we did that? And that’s a general way we could do governance for companies.

Jeff: Talk about the way it might apply to governance in a political context.

Robin: So for many decades, we’ve had markets in presidential elections, we’ve had, for each candidate, the chance that they will be nominated by the party, and then the chance that they will win in the general election. Now, the ratio of those two probabilities is the chance of winning if nominated. And that number is advice to the political party about who to nominate if you want to win. The numbers aren’t all the same. So a decision a political party has to make is who to nominate. And one of their major goals is to win in the election from their nominee. And these prices just directly tell them who to nominate. That’s what I call decision=market decision-advice to the parties. In addition, in the last presidential election in the US, we had two websites, Manifold and Metaculus, that gave advice to the voters about who to vote for. That is, they estimated different outcomes, conditional on whether Trump or Harris became president. And voters could then look at these different outcomes and ask which outcomes they wanted and use that to decide who to vote for. So that’s, again, the markets giving advice to voters about who to vote for. That’s decision-advice that markets can give.

Jeff: Does it matter? And if so, how big the pool is of people participating in these prediction markets.

Robin: So you can think of these markets as an information production process. There’s a supply and there’s demand, or there’s producers and there’s consumers. So the price itself is the outcome. And people who want to see that price in order to inform their decisions, they’re the consumers, and people who trade in the markets, they’re the people who have information. They are producing the information in the market. And what we want is for good price that matches the consumer and the producer. Now, in most financial markets, consumers pay nothing for the information. The information is a byproduct of the various traders trading against each other. And so in a sense, the traders are so eager to trade that they trade without being subsidized at all. And then the people who look at the prices get their information for free. However, I think it makes more sense when we have information that only you would care about, say only this company with the CEO would care about knowing whether they should keep the CEO, that that consumer of the information should pay for it by subsidizing the market, and then the traders in the market would be more eager to trade because of the subsidy. And now there would be a price of information. The consumer would set up a subsidy, which would set a price for how much they’re willing to pay for information. Then the traders respond to that price by supplying information whose cost is less than this price offered. And that’s what I would like to see as a general structure for these markets in the future. At the moment, financial markets are pretty much all. The customer is the trader who just pays a tax in order to be allowed to trade, and all the information that it’s produced is a side effect that everybody can get for free. I would rather people who get the information pay for it.

Jeff: At this point in time, where you have some people that are participating in these markets and others that are either not aware of them or are not participating, is that influencing how the markets play out? And as it grows, as it becomes seemingly more inevitable, how will that change the markets or will it?

Robin: Well, again, we’ve had financial markets for a long time. These are a new kind of financial market, but all the basic behavior and structure is what we’ve seen in financial markets. I don’t know if you know that roughly a century ago there was more money in the US bet in the US presidential betting than there was in the stock market at the time. Stock markets are very small, but over time the stock market has grown. The betting markets receded for a while, and we now have large financial markets where people do a lot of things. So I think you want to ask, what does the world look like when there are large financial markets nearby? Look at ordinary financial markets. The companies that have stock prices, etc., that’s what that world looks like.

Jeff: We’ve seen these markets be very wrong from time to time. We’ve had crashes in markets.

Robin: Of course. The claim is that they are better than your other alternatives for getting these estimates. Not the claim as they’re always perfect or that they could never make mistakes. But we’ve had a large literature of head-to-head comparisons between speculative markets and other institutions estimating the same thing at the same time, with similar resources, and the markets either consistently or about the same or substantially better — almost never substantially worse. And even older literature in financial markets showed that it was just hard to find any information that wasn’t embodied in market prices. That’s how powerful these institutions are. But being the best institution available for aggregating information does not mean you know everything. It’s not omnipotence. It’s just the best you can do with the available sources.

Jeff: What, if anything, is the danger that could come from these markets?

Robin: Well, they are information institutions. Sometimes information is dangerous. Information isn’t always useful or valuable. So, for example, many people think it’s bad to have information about who’s likely to win an election on Election Day, because that will discourage people from voting late in the day. They would rather we forbid information about that. When somebody tries to extort you through, say, a threat of when they steal some… I’m blanking on the word. If somebody grabs one of your associates and demands money for them. You’d actually rather credibly not be able to believe them because then they won’t do this, and then you won’t have to pay. But there are many cases in the world where information isn’t always the best. We talked about organizations having secrets. And for example, if the military has secrets and we want our military to keep secrets, then information institutions risk those secrets being exposed. So the most basic thing to worry about a competent information institution is the fact information isn’t always good.

Jeff: Are we going to get to a point, do you think, in 10 years where these prediction markets, the data from these prediction markets becomes as routine as a stock ticker or even a sports book kind of situation?

Robin: I actually hope they get to be a lot more routine than that. That is, I’d say there’s a huge demand in the world for better informed decisions. Companies make decisions, nonprofits make decisions, governments and even people make decisions all the time, and the ability to make better decisions is worth a lot more than all the value that financial markets have been acquiring in the last century. That is, there’s just a lot more potential value to be achieved by having more informed decisions. And so I’m hoping we can have a lot more markets than we have today because, for example, every time you go on a date, we could have a market about how long will that relationship last, if you date a new person. Or a student choosing what school to go to or what major to take, we could have markets and how will that go for you? Will you graduate? Will you get a job? Will you like it? We could have markets in most major decisions people make, or that organizations make and get information about those. That promises to have just a much better informed world that’s making better decisions.

Jeff: How does AI play into all of this now, which has the ability to aggregate huge amounts of data to make predictions, essentially.

Robin: So I’m an economist who studies institutions, and we have long sought to make institutions that are robust to varieties of situations and varieties of participants. And I think we have, in fact, made many institutions that are robust. So, different kinds of people, different times of day, different topics — our institutions cover a wide range of scenarios, and yet they still seem to function. I think swapping humans for AIs is also a change that these institutions should be robust to, and it seems that they are so far. There’s a question of whether there are any ways you might want to change these institutions to accommodate particularities of AI. But so far I don’t know that we see any.

Jeff: And what about backlash at this point? What does that look like to you? And what impact do you think that it can have on the markets?

Robin: So all the familiar financial markets you know of were once somewhere illegal as gambling. And over time, what happened is we carved out exceptions and we no longer called them gambling because we saw they had useful social functions, even though in some literal sense they are all still gambling. So the hope here is that we could carve out some new exceptions where even if people say, “No, no, there’s too much gambling here that has to cut back,” those exceptions could remain because we no longer see those as gambling, because we see those as having a social function. Now, if you notice that centuries ago, many societies were not that into letting people have fun. There’s lots of kind of fun they didn’t let people have. It was important that people were serious and worked hard and didn’t waste their time and money having fun. Our society is more open to people having fun, and I think that gives us a possibility that, in fact, people will be more tolerant to people having fun through the betting markets. All the gambling you could want to do with any of these markets, you can already do with ordinary financial markets. But many people think that’s just not as fun. They think it’s more fun to bet on sports than it is to bet on stocks. And maybe it is. But then, if you shut that down, what you’re really trying to shut down is the fun. I’d say, in our world, we let people have fun by many risky things. We let them try to be actors or musicians or journalists, or even professors, when the odds are long and most people fail. We let young people date all sorts of people adults don’t think are very appropriate for them, and in the past adults wouldn’t have allowed that. But now we let people have their fun. And so the question here is how much fun will we let people have with these various betting markets? If we decide sports betting is too much fun and we’re going to cut back on that, that could be a backlash. And the risk is that they shut down other more useful stuff with it. So if I was going to draw a line, I’d say, let’s let people bet on questions that matter, topics that matter, topics where, if we have a consensus estimate, that might plausibly inform people’s decisions, and most sports markets don’t seem like they matter that much other than to the people involved in the game itself, or their fans. But most of these other markets that are on Kalshi and Polymarket, they are bigger questions that matter. And I’d rather we kept those because of the rationale that they have some other purpose besides fun. They are helping the rest of us learn about our world and make decisions.

Jeff: And finally, to come back to where we started. You’ve been at this for 40-something years. Are you surprised at where we are now in this process?

Robin: Well, I was always uncertain about when this would happen and how fast. And we’re in a recent burst of activity. But I’m in this for the long run, so I’m not so sure how long this burst will last, but I’m happy to let it go as long as it can. And even if there’s a backlash, I’ll still be in this for the long run, trying to help us do this, eventually, later. But for now, let it ride.

Jeff: Professor Robin Hanson, I thank you so very much for spending time with us.

Robin: Nice to meet you, Jeff.

Jeff: Thank you. And thank you for listening and joining us here on the WhoWhatWhy podcast. I hope you join us next week for another WhoWhatWhy podcast. I’m Jeff Schechtman. If you like this podcast, please feel free to share and help others find it by rating and reviewing it on iTunes. You can also support this podcast and all the work we do by going to WhoWhatWhy.org/donate.