This transcript was generated using AI automation and may contain minor formatting or transcription errors. Please refer to the original audio to verify specific quotes or context.

Cato Podcast: Prediction Markets, with Robin Hanson

Host: Ryan Bourne, Cato’s R. Evan Schaaf Chair for the Public Understanding of Economics Guest: Robin Hanson, Associate Professor of Economics at George Mason University

Ryan Bourne: How much would you pay for a better forecast of whether inflation will fall, a congressional bill will pass, or a war will begin? Prediction markets produce a real-time probability of an event happening, derived from people staking money on their own beliefs by buying yes-no contracts in market exchanges. Now in recent years, platforms like Kalshi and Polymarket have taken off here in the United States, broadening the range of events we can collectively predict through bets. From how high oil prices will go this year, through to estimating the number of building permits for new homes. Yet the same platforms’ biggest markets offer new ways to bet in-play on sports, politics and celebrity gossip, providing fresh avenues for young people to gamble. With that, almost inevitably comes a policy backlash, and niche opportunities for insider trading are also getting lawmakers hot under the collar. When there’s markets on whether a politician will use a particular word in a speech, there’s plenty of scope for hoovering up dumb money from other punters. So are prediction markets a public information system, or a tilted casino in financial clothing? Will they develop into something genuinely useful for real-world decision-making? And how should they be regulated, like financial products, gambling or something else?

I’m Ryan Bourne, Cato’s R. Evan Schaaf Chair for the Public Understanding of Economics. My guest today is Robin Hanson. He’s Associate Professor of Economics at George Mason University, a pioneer of modern prediction and decision markets. So Robin, welcome to the show.

Robin Hanson: Great to be here.

Ryan Bourne: How does one become an expert in prediction markets? What first got you interested in this topic?

Robin Hanson: Well, I just first wrote about it a lot before other people. But my origin story was that there was a time before there was a World Wide Web, and there were people imagining the Web and predicting it and trying to make it happen. So I got involved with those people who had hopes that the Web would increase and improve public conversation. And I bought into that dream and I worked with them and then I started to have doubts. So their hope was that making criticism easier to find would make us all a little less stupid about believing the various things we do in the public sphere. And I started to think, is that really going to help? How much? And then if not, what else could we do? And because they were kind of rabid libertarians actually involved in this project, the idea was accessible to me who hadn’t been so much of a libertarian before. What if we had betting markets? Wouldn’t that create an honest, solid consensus on important policy, political, public topics? And if we had more betting markets, just have than a better public conversation. That was where I came from. When I had that concept, I started to think, what are the obstacles here? How far could we go? And I started to write about this big potential of betting markets as a way to create a better conversation.

Ryan Bourne: So for those of us who perhaps are less familiar with prediction markets than you, is that how you would essentially define them? Just the application of betting markets to a wider range of topics and applications?

Robin Hanson: Pretty much, yeah. So there are many technologies where we see them different depending on what they’re used for. So for example, post-it notes are basically pieces of paper with bad glue. From the glue industry, post-it note glue is terrible glue. But then somebody thought, oh, what if you had this other use for a glue? Why a weak glue would be great. And that’s what a post-it note is. It’s a really weak glue. And similarly, you can use betting markets for many purposes. Obviously, some people use them just to have fun. Most financial markets are basically betting markets, but they are for the purpose, people say, of hedging and transferring risk. And then as a side effect, we know that they aggregate and distribute information. So each of those is a different purpose or function of the same thing.

Ryan Bourne: So let’s take a simple contract. Something like, will US inflation be below 3% in December? So when somebody’s trading on that, what exactly is being bought and sold? And what is the mechanics of what happens, what occurs, that means the price of those contracts fluctuate?

Robin Hanson: So what I’m going to tell you is just how all financial markets work, but you might not know how financial markets work. So here we create an asset that pays off if an event happens. So if this inflation target, say, is met, this thing is worth $10. So at the end, if you can show that, hey, this target was met, you turn it in and then you just get $10, you won. Now, earlier on, there was a way to create these things. You took, say, cash and you split it into cash if this event happens and cash if the event doesn’t happen. And now you’ve got these two assets, you could trade them separately, and then there’s a market in trading. So let’s say in a market, there’s a bid price that you can buy this for $3.20, an ask price that you can sell it for $3.30. That’s a bid ask on the asset, and that sets a price in that range. And now you ask yourself, what do I think the chance of this event is? And if you think the chance is higher than 3.32%, you think there’s money on the table if you buy this thing. You could buy it now and just wait till the very end, and on average, you will win, or you could buy it now knowing that some information is going to come out soon and the price will go up and then you sell it then. Either way, you made money from this market. If you think the price is lower, you think it’s less than 3.32% chance of this, well, you do the opposite. You sell it first and you buy it later. But with any financial market, with no price, if you know which way the price is going to go in, that’s actually a money pump. And so this has been for centuries something people have known about financial markets. Gee, if only you could guess which way the price is going to go, you can walk away with money.

Ryan Bourne: So I guess one difference here mechanically is that whereas in a sports book, you’re usually betting against the bookmaker who is determining the odds. I mean, obviously those odds are affected by the trades that people are making, but in this you’re directly trading against people taking the other side of the question.

Robin Hanson: Well, not necessarily. You could. But there are often people called market makers who sit between you, and their job is just to make it easier for each of you to trade. So if the people who want to directly trade aren’t that many of them, then I would have to sit and make an offer and wait for you to take my offer, and only then would I have completed a trade. If I’m in a hurry and I just want to go straight and make a trade, a market maker who sits between us will just make sure to offer this bit of mass spread, and then I can trade with the market maker, and then you can trade with the market maker, and now we in effect traded with each other, but we didn’t have to wait.

Ryan Bourne: So make the strongest possible case for the broader social benefit of prediction markets, which you’ve been a proponent of. Why does asking people to put money behind their forecasts produce more socially valuable information than perhaps other sources of information like polling or surveying the public, listening to pundits talk about a particular issue or something like that?

Robin Hanson: Okay, there’s two comparisons to make. So there’s one, the different functions of these markets. We talked about how they can also be used to hedge risk and also be used for fun. And then there’s the other comparison of different mechanisms to achieve each function. So for the information aggregation function, we have other institutions in society who claim their job is to aggregate, summarize, and then distribute information, for example, journalism, for example, academia, many government agencies, many sort of certifying agencies. All of these organizations say that what they do is collect information, summarize it, and then distribute it for a fee, and prediction markets are sitting in that same role. So now which one do better at that role? Well, I’m an economics professor, so I could walk you through lots of theory that you might not understand or even believe because what do we economics professors know? So what I’d rather do is just point you to the data. By now we have many decades, even centuries, of data on financial markets that say, typically financial markets just have so much information that it’s hard to find any information they don’t have. They just do a really good job of sucking up information. Then in the recent decades we have a lot of data specifically on betting markets and prediction markets, comparing them to other mechanisms at the same time, polls, committees, other things, which shows that they typically just are more accurate when you compare them at the same time on the same question with similar resources. They are either about as good or substantially better than these other mechanisms. We’ve just done that experiment over and over again for many decades now. And that’s the main reason I should want you to believe they work is we just have that data.

Ryan Bourne: Now, again, I could walk you through the theory of it, but the data is the most important thing. And the way you’re talking about it, I guess, it’s just an application of the virtues that people like Friedrich Hayek saw of market prices themselves. They embed dispersed information. Not everybody who’s kind of trading on those prices needs to know everything that goes into them. But they’re an aggregation. They’re kind of a wisdom of crowds benefit of dispersed information, supply and demand going into determining a price.

Robin Hanson: In part, but I do want to distinguish different kinds of markets as having different features and virtues and costs. Not all markets are the same. So the Hayekian point is just abstracted from the differences between different kinds of markets. I might say betting markets have unusually specific advantages over many other kinds of markets. And I would say, again, we have data on betting or speculative markets as being especially powerful at aggregating information.

Ryan Bourne: So people can use then the information provided by prediction market prices to hedge, create conditional forecasts, use it for their own applications in terms of creating forecasts, can improve their own decision making. Is there anything else that I’m missing in terms of the advantages to the individual of seeing this socially constructed information?

Robin Hanson: Well, first, let’s just notice that with these other institutions like academia or the news media, often individuals need to pay for this information and the information won’t exist unless the suppliers believe there are enough customers willing to pay. So in journalism, people write stories on the topics people would like to read about, not just random topics in the world. With speculative markets, typically the people are paying to participate in the markets to give you information, but the people who get the information don’t pay much. So mostly speculative markets, information is a freebie. It’s something everybody gets to free ride on and to get for free. And so the topics on which they’re creating information are whatever the people want to trade on, not so much what other people want to learn about. And that’s been the way most financial markets have been for, again, centuries. The thing I started to think about decades ago when we were first building the World Wide Web was what if the people who wanted information could be the customers? And then the topics on which they wanted information could be the ones the speculators focused on because the people who wanted the information were willing to pay for it. And that’s my key concept of information markets or prediction markets. Now we’re not there yet. So these big providers, for example, at the moment, Kalshi and Polymarket, who are booming because legal changes allowed them to create lots of markets, they’re being paid mostly by the traders who participate, having fun or hedging risk, and not by the people wanting the information. So I think we’re a long way from where I’m hoping to go. They’re going to help. What they’re doing is helping us get there. But still my vision is can we get to the point where if you want to know something, you pay for the information by subsidizing some markets and then traders turn their attention to your topic, aggregate information on it, and give you answers to your questions. And now we can go into more detail about lots of scenarios, how that could work.

Ryan Bourne: Yeah, I was going to say that from the information that I’ve seen, 80 to 90% of the trades on those two platforms at the moment tend to be in sports, politics, or crypto. Now, some might say that’s a distraction from the serious information that other markets could provide. But I guess there’s advantages too. Those sorts of paying customers help finance the technology. They provide liquidity onto the platforms. So they might provide some of the funds for the infrastructure to be developed for the types of applications you’re talking about.

Robin Hanson: As economists, we know that if somebody buys a hot dog, they’re not preventing you from buying a ham sandwich, right? You shouldn’t be jealous that somebody else is buying another product that you’re not buying. If you notice that many kinds of products have synergies, that is if other people buy hot dogs, that can lower the cost of bread, which can make your ham sandwiches cheaper. And you could appreciate that they are helping you, really, even if they’re not directly buying the thing you want. Sure, if they bought the ham sandwich instead of the hot dog, then they would lower the cost of ham sandwiches even more than they do by buying a hot dog. But you shouldn’t complain that somebody else buys a different product than what you want. To the extent that it’s similar, they’re still lowering your cost somewhat.

Ryan Bourne: I guess one consequence, though, of those applications dominating on these platforms is that it’s likely to create a regulatory pushback. So at the moment, there’s a lot of young people using these platforms, as I say, the vast majority of money in sports and entertainment. And that’s creating what we might describe in economics as a kind of bootleggers and Baptists coalition that are really driving concern and pushing for heavy regulation of these platforms. So on the Baptists side, you have those who just don’t want new platforms for gambling, just think gambling is immoral or corruptive to the youth, and so we need restrictions on it. But at the same time, you’ve got the bootleggers, in this case the state governments, worried about losing lots of local gambling tax revenue, and their gambling licenses being undermined by these platforms doing things that traditional bookmakers have done. So I guess my question to you is, if these end up being regulated like historic traditional gambling, what damage could be done to the development of the industry?

Robin Hanson: It’s definitely true that this industry could face a backlash, wherein it became much more heavily regulated. And in that case, it would have to reduce its scope and therefore the rate of development. Then we would less get the gains that we’re getting now from lowering infrastructure costs, legal precedents and customer familiarity. That would be the cost of a backlash. Now, a few years ago, say 10 or 20 years ago, you might well have thought that the main backlash would come from people who objected to gambling. But I don’t think that’s what we’re seeing actually at the moment. So we are seeing the objections from people who do state gambling, saying you’re taking the gamblers away from us. But I don’t actually hear that much broader complaint about gambling. And I think to notice is, centuries ago we were pretty prudish about lots of kinds of fun. And over the last few centuries, our world has become much more open to people having fun, even when it seems wasteful or even ill-considered. So for example, we let lots of people try to get careers in acting or music or athletics when their chances are really crazy low. We let young people decide who to date when their adults don’t think those partners are very wise. We let people waste lots of money on lots of hobbies that seem not very productive. So our world is full of ways that we let people waste lots of time and money and take risks for fun. So on that reference point, I don’t know that sports betting is especially harmful compared to other ways people waste themselves on fun. Again, you might have thought that we were just retaining some prudishness about sports betting compared to other kinds of prudishness, but I’m not so sure we are. So it’s important to keep clear, all the gambling that you can do in betting markets on sports, you can do exactly the same level of gambling in terms of speed and risk in ordinary financial markets. So why don’t people do it there? Well, it’s not as fun. The reason they do more of it in sports is because they have more fun betting there than they do in the other places. So again, the main complaints we’ve seen, I think, lately is more on this insider trading, which isn’t really about fun or gambling. It’s a whole separate set of complaints, and that’s, I think, interesting. People have decided that’s their best place to go to object to these things. Even the state people who are trying to protect their revenue of state gambling — what they’re complaining more about is the insider trading and less about the gambling.

Ryan Bourne: Let’s tackle the insider trading question, because this is really an issue in which libertarians’ position is often considered pretty provocative, because to the extent that we see prices in prediction markets as an aggregation of information about the reality of the world, of course, insider information becomes a feature rather than a bug, because it actually makes the market price more correct than it otherwise would be. It gets us closer to accurate probabilities. So do you have a framework to think about when insider trading is helpful or when it might be more problematic?

Robin Hanson: Yes, but let’s first set aside the old framework that doesn’t really apply here. So a century ago or so, U.S. regulators declared that in stocks, they thought it was more important that more people buy stocks than that the stock prices be accurate. And that’s why they banned corporate officials from trading in stock markets. That was what they called insider trading a century ago. And for a century, that’s what insider trading meant is when corporate officials traded on stocks and we had laws against that. And they actually aren’t very effective. So that on average, when a company makes an announcement, the stock price moves in response to that announcement. On average, half of that move happens before the announcement. On average, half of that is insider trading. So as a matter of fact, insider trading is in fact rampant in ordinary stocks. But nevertheless, we’ve had these rules against it for a century. And that’s just not relevant for these prediction markets, unless you think the public policy priority should be to get more trades in these markets rather than more accurate prices.

Ryan Bourne: Can you just explain, I think I understand, but just explain for the listeners why that type of regulation would lead to less aggregate trade, isn’t it? Because people wouldn’t trade because they’re worried they’re going to be duped by people who have insider information.

Robin Hanson: Right. So a standard rule in poker is when you sit down at the poker table, look around and find the fool. The rest of you will be making money off the fool at the table. If you don’t see the fool at the table, it’s you. You should walk away. That’s a good way to think about financial markets in general and prediction markets in particular in terms of speculation. If you’re there to hedge risk or to transfer risk, okay, you know that it’ll be a cost that you might trade against people who know more than you, but you’ve got this bigger reason to be there. But if you’re just mainly there because you think you know something other people don’t know, then you should consider this poker rule. And so in fact, these markets are a contest between people who think they know more than other people and the question is who’s right. Now, if there’s somebody who really does know a lot more than the rest of you and you know they’re there, that’s a reason for you to walk away from the table because they’re going to win over you. Again, as you say, if the purpose of these markets is to aggregate information, that’s fine. We’ll just have fewer people who play at the table and we’ll get more accurate prices. Now, the markets themselves like Kalshi and Polymarket, they probably just want more trading. They don’t care so much about the accuracy of the prices. So they’re not going to be so eager to suppress insider trading. But the rest of us maybe shouldn’t be so eager either.

Ryan Bourne: So let’s test that against an especially difficult, prominent case recently. So federal prosecutors allege that an army master sergeant used classified information about the overthrow by the US government of Nicolas Maduro, former Venezuelan president, to make more than $400,000 on Polymarket. Now, his trades may have made the market’s forecast more accurate, but he was allegedly cashing in on a secret that he had a duty to protect. So is there a crime there? Should the crime be the misuse of the classified information, or the trade itself? How do you think about an issue like that?

Robin Hanson: We have some information institutions like prediction markets and journalism and academia, and their primary function in this view is to collect and generate and share information. Now, sometimes information isn’t always good. For example, people argue that on election day it’s bad to tell voters how close the election is because if the election is not close, they won’t go vote. And we should then hide information about who’s winning on election day. That’s a reason then to suppress information institutions for some topics of information. But in general, we have these information institutions because we think it’s good for people to become informed about things, and that’s their purpose. But we also think that organizations have some legitimate interest in keeping some secrets. Now, in journalism, a competitor for prediction markets, many of the most famous celebrated cases of journalism in the last few decades were cases where journalists extracted some secrets from organizations that the organizations didn’t really want to have leaked. And now we say, well, there’s a conflict of interest here, right? Organizations have legitimate interests in keeping their secrets, and then often the rest of us have legitimate interests in finding things out. So I think we want policy to sit in the middle there, letting organizations have some tools to keep secrets, i.e. contract or even crime of theft for things that they have stolen on the one hand, but also let journalists and others probe to find things out that might be in the rest of our interest to know. That’s where I would sit on this trade-off, and I would say, hold the same policy for prediction markets as you would for journalism as a test case. So for example, some people recently had a bill before Congress proposing that all government officials should be forbidden from trading in these markets because they might thereby reveal secrets. I’d say, okay, forbid them from talking to reporters too, because they might reveal secrets that way. If you’re willing to be consistent, I’m more okay, but I’d say, be consistent across these different information institutions between your policies of when it’s okay for people to probe and reveal secrets versus when you want to make sure organizations are allowed to keep their secrets.

Ryan Bourne: I guess there’s a question of precision, though. So we had Alvin Roth on the podcast relatively recently. We touched briefly on prediction markets and some of the concerns that people have, and he kind of indicated an instance where one could imagine in the Maduro case, for example, Venezuela might actually see that price movement and react accordingly. So there can be harms to the national security interests of the United States in that case, if their information is revealed in a more precise way than a leak to a newspaper that suggests that this might happen in the near future.

Robin Hanson: If you’re willing to shut down an information institution by saying this information is not good to get out, then you should shut down journalism on the same topics in the same way. Many government officials and military officials have definitely sought to prevent journalists from speaking on particular topics at particular times under the same rationale that that information was better kept secret. So that’s a key trade-off you have in our First Amendment public journalism institutions. We have typically allowed journalists to say what they want, but we have many concrete specific exceptions where government has been able to prohibit journalists from talking about things because the government thought that was not in the public interest. So again, I would mainly argue for uniform treatment here. I’m not going to say all information should always be revealed, but I’m going to say people are inclined to dump more on the new institution rather than the old one for interesting reasons. So when journalism first got started several centuries ago, it was complained about and insulted by the usual authorities at the time. The Crown, the government, aristocrats, the church, they said this new institution was disruptive, disrespectful, was making people focus on sensational things and not being serious enough and that these people didn’t have enough skin in the game to have good incentives. These were all complaints made about journalism at the time, and many governments did in fact suppress journalism to a great degree several centuries ago when it first got started. Now today, journalism is among our most respected institutions for information.

Ryan Bourne: Indeed. It is, and it’s seen as the proper authority, and this new upstart of prediction markets is threatening the prestige and revenue of journalism.

Robin Hanson: For example, there was an op-ed in the New York Times a couple months ago saying it’s just bad when people go to prediction markets as their first place to get news on something. They should be going to newspapers because they are the proper authority. And I remember we had this big blowup in 2003 on something called the policy analysis market where we were going to set up markets on geopolitical events in the Mideast, and one of the main complaints people had is, don’t we pay prestigious intelligence researchers a lot of money to do this for us? Those are the proper authorities on this topic. Having markets on a topic where intelligence authorities speak is just improper, and they are just not the right authority. So I think many people find this distasteful — that ordinary people, low-status people, ugly people, sweaty people are out there making prices that then compete with prestigious, Pulitzer Prize-winning, Harvard-trained professionals. This is just not right. Alvin Roth calls it repugnant markets.

Ryan Bourne: Indeed. But there are a few other concerns people have. One is what you’ve described in the past as sabotage. That is, markets offer trades where people can bet on things that they perhaps can influence, like a player’s action in sports, or candidates withdrawing from political races. Just last week, there was a story about a White House teleprompter operator who made around $100,000 betting on the contents of President Trump’s speeches because he’d seen them in advance. But one could imagine another scenario where he might try and slide a word into the speech and then profit from trading that that was going to be said. Recently, I saw that there might even be a market in flights being delayed. If you’re a passenger who’s willing to put a bet on that, there’s probably things you can do on the margin to slow down the boarding process. So how worried are you about sabotage? How do you think of it from a policy perspective?

Robin Hanson: So notice that all the other information institutions in principle have exactly the same problem. People do, in fact, do things different in the world in order to produce favorable news coverage. People, in fact, do things different in order to just directly cause news coverage because actually most revenue of news comes from sponsors, not the actual news readers. So academics often affect the world in order to have interesting things to write about. So the nature of an information institution is that as soon as you reward people for having information about the world, that also rewards them for creating information about the world by changing the world. That’s just a general problem with all information institutions. So why do you see it more about prediction markets? Well, one again is the fact that you’re dumping on the new thing, but the other is actually showing this is an unusually powerful information institution. This can actually induce more specific effort because it’s actually more effective at translating resources and rewards into actions and information. So just for the same reason that these markets beat other competitors in terms of being accurate for similar resources in question, they also beat other institutions for inducing effort to be accurate because they are just more powerful.

Ryan Bourne: So another concern is manipulation, different to sabotage. Just a hypothetical: imagine a billionaire or well-funded political group wants to defeat a bill in Congress. So it places enormous bets that the bill will fail, not because it genuinely believes it will fail, but because it hopes lawmakers internalize the market price and conclude that the bill is doomed and then abandon it. Now that group may lose money on the bets — it’s basically paying for the political outcome that it wants. Now you’ve argued that attempts like this are often self-correcting, and that this isn’t something that exercises you a lot. Can you explain why you think that?

Robin Hanson: Well, first let’s just notice that academia and news media are enormously susceptible to manipulation. Again, news media are actually paid more by sponsors trying to influence what positions they take and articles they write than they are by news readers, and much of academia is as well. Many people pay for academic research exactly to induce research results that favor their positions. So manipulation is an enormous fact of the other competing institutions. That makes the following fact all the more striking. In speculative markets, the more people try to manipulate the market, and then the more other people expect that to happen, the more accurate prices get. So this institution is not only better able to resist manipulation, it feeds on manipulation. And that’s a fact we’ve seen in theory, in the lab and in the field. And we have a nice theoretical way to understand why this is true. But the basic fact is when people try to manipulate financial markets, they become more accurate as a result, because they create profit opportunities that induce entry from more people being willing to take the opposite trade.

Right. So one way to think about this is that the simplest model of ordinary financial markets is that there are two kinds of traders. I’ll call them sheep and wolves. Sheep are people who have a generic random reason to trade other than the reason wolves have. Wolves are trading because they know something. So a wolf says, I have some information, where can I go to trade on that? And they want to trade in the market hoping that because the price is wrong, they’re going to make money on their trade. Now, if a wolf trades against a wolf, they don’t actually profit. One has a clue that the price should go up, the other has a clue the price should go down, their clues cancel, they don’t make money. So wolves want to trade against sheep. Wolves want to trade against people who are there for some other reason than trading. So for example, in the stock market, you might be saving for retirement. So every paycheck you put some money and you buy some stock, that’s a trade that’s not based on any information you have. Then later when you retire, you need to sell stock every paycheck period so that you have some more money to spend. Again, those are trades not based on information. So you’re a sheep in those trades. You don’t have information, but you need to make a trade. The wolves want to trade against you because you don’t have information. And so in ordinary financial markets, what we see is that the more sheep there are in a market, i.e. the more money trading in a stock market, the more accurate the prices are because more wolves show up to trade against the sheep. They not only bring their information with them, they try harder to get information. And so a general fact about financial markets is the more sheep, the more wolves, the more accurate prices get. The last thing you just need to understand is that a manipulator is a sheep. They are someone there for a reason other than the information they have. They want you to believe something about the price, but they don’t actually have information to back that up. So they are trading like a sheep. And they are someone that a wolf wants to trade against. So when a wolf believes there are more manipulators, they basically believe there are more sheep here and they are more eager to trade against the sheep, which makes the prices more accurate.

Ryan Bourne: Well, something in sheep’s clothing, but not another wolf in sheep’s clothing — they don’t want to trade against another wolf. So let’s move from prediction markets to decision markets, which you’ve described as your best idea. Could you just define for us what a decision market is? How does it differ from or relate to a prediction market? And perhaps offer us a concrete example.

Robin Hanson: So the history here was that when I first thought about prediction markets, I thought about it in the way that most people have thought about it when they first hear about it over the decades, which is to say, let’s have markets on the usual public policy topics — the things in the news or politicians’ speeches or pundits, think tank reports — let’s have markets on those things because that’s what we see when we hear people talking. And then after a few years, I realized, oh, standard decision theory says that information is valuable when it affects decisions. What this implies is that the most valuable markets should be markets that sit right next to a decision, advising that decision. So here’s an example. For many decades we have had markets in US presidential elections for candidates in terms of the chance they will become president and the chance they will be nominated by their party. The ratio of those two chances is the chance of winning if nominated. That’s advice to the party about who to nominate if you want to win. So that’s what I call a decision market — a set of prices that sits right next to a decision and tells you here are the consequences of your decision. Something that we haven’t seen yet that I hope to see more of is a market for stocks saying, should you keep the CEO for the firm? So this is another way to see a decision market. A firm needs to make this choice. The board needs to decide, do we keep the CEO or do we get a new one? That’s a key decision, one of the biggest decisions firms make. And what they care about is the stock price being higher. They want to make the decision that makes the company worth more. So in order to achieve this, what we can do is we can have conditional stock prices. So a usual stock price is how much cash a particular stock is worth. And when you try to trade in that market, you have to ask yourself: think of all the scenarios the company can be involved with over the coming years, and in each scenario, how much is this company worth? That is, how many customers does it have? What are its costs? How much net profit does it make? And you average over all those scenarios and you come up with your estimate of how much this company is worth. And if the price looks lower than that, you should buy. If the price is higher than that, you should sell. That’s what you do in an ordinary stock market. Now we’re going to make conditional stock markets. These are markets where you trade stock for cash, except the trades will be called off if a condition isn’t met. So now we’ll have the stock price if the CEO stays, stock price if the CEO leaves. And when you trade in those markets, you should ask yourself, if the CEO stays, what are all the scenarios a company could be involved with? How much is it worth in each of those scenarios? Average up that, but you’re only averaging the scenarios that are consistent with this condition — the CEO stays. The other market, you’re only averaging over the conditions consistent with the CEO leaves. That means you’ll get two different stock prices now: a stock price if the CEO stays, a stock price if the CEO leaves. If the first price is higher than the second, you keep the CEO, otherwise you don’t. That’s a decision market. You see, it’s prices that sit right next to a decision, telling the decision maker — in this case the board — what is the consequence you care about, i.e. the stock price, conditional on the particular option you might choose, in this case, keep or dump the CEO. That’s another example of a decision market. But I hope you can see the structure by now. This can be used to advise anybody’s decision on anything. As long as we can see after the fact the consequences you cared about, and we can see which specific decision you might pick, we can advise a college on which students to admit. We could advise students on which college to go to, which major to take. We could advise you on who to date. We could advise firms on when to raise capital, when to issue new products, when to do mergers. We could advise governments on key choices about which bills to pass — does nationalizing the health industry make medical prices lower, make life spans longer. Any set of decisions we have in the world where you can see the consequences later and you can see which choice you picked, we could advise that. I will claim that the latent demand for better decisions in the world is really much larger than the other sorts of demands in financial markets. The demand for risk hedging, the demand for fun — those demands are really just actually smaller than the demand for the world to know more about what to do. Just a note in case my wife is listening: my dating days are done, so I don’t need a dating market.

Ryan Bourne: And the market could confirm that for you. But you’re talking about business applications — I imagine there’s non-profit applications, but of course there are applications for public policy as well. So you’ve described a vision of futarchy. What does that mean in context, and how does that affect the way public policy decisions might be made?

Robin Hanson: So I showed you examples of decision markets, and to motivate people to see how far we could go with them, I’ve asked them to think about how we use this for governance. So I don’t propose that we immediately use this for governance. I propose that we test this mechanism on small scales and work our way up to a larger scale test as smaller scale tests work out. But to see how far we could go, I said let’s imagine how we could run a government this way. So now if we had a for-profit government, then we’d just do the stock thing. And every time we have a bill to decide what to do, we would ask, does it increase the stock of the government? Now most people today don’t want stock-based governments. So what we’d have to do is create a measure of national welfare, some measurement of what we want. So say start with GDP, add leisure, add respect, add environmental conditions. The point is you just take some weighted average of all the things we care about, and we’d make a number that represented all those things so that we measure how much we’re getting of what we want. And then we create an asset that tracks this, that pays off more according to this number. And now we’ll do the same thing as with the stock example I showed you. Whenever somebody has a proposal for a new change to government, we will ask the markets to say, well, how much of what we get will we get if we adopt this policy versus if we reject it? And then we would only approve it if the first was higher than the second. Now we could use this mechanism also for more general corporate governance, and that’s where you’d want to experiment with it. And there are now some crypto firms doing these experiments. They are actually using this mechanism in the last few years to make their basic governance decisions in their firms. And the key thing is, when there’s a high level decision about some change of who’s in charge, or allocating a budget to a project, they basically ask the market: well, are we worth more if we do this versus if we don’t?

Ryan Bourne: If the information case is so powerful, and you’ve persuaded me that it is, why do governments and companies use these markets so rarely? Is it technology? Is it liquidity? Is it regulation? Or that it’s difficult to design the contracts? Or is it something about the politics of these institutions?

Robin Hanson: Innovation is hard. So our world is much more innovative than worlds were in the past, but we are far less innovative than it’s possible to be. Most people with ideas for change get shot down. Most ideas for change don’t get tried out. And ideas for change — it’s easier to try them when they don’t threaten so many incumbents. So governance mechanisms are actually harder to innovate on, because typically you’re asking some people in charge of something to threaten the stability of their position by allowing some other mechanism to challenge them. So in the long run, I expect innovation to work out, but the fact of the matter is innovation is just slow in our world, because typically the gains that innovations produce are much larger for the world than for the people who introduce them. And that’s the major reason we expect our world just has too little innovation overall on average.

Ryan Bourne: And a lot of these probabilities imply things that are kind of uncomfortable truths for many people. Jonathan Haidt’s work has actually suggested that libertarians are less likely than others to adopt disgust or discomfort as a sufficient reason to prohibit something. I think that might help explain why many of us are willing to look at unsettling probabilities in prediction markets and ask first what information they actually reveal. Because, of course, a market probability is not an endorsement of an outcome — it’s an estimate produced by people combining different information and putting something at stake. And so we really recognize that these prices are socially valuable information even when the event being predicted makes us feel uncomfortable. Now, none of that means every contract should be permitted — you’ve outlined some of the more difficult cases. But I’ve really been surprised at how reflexive some of the calls for regulation of this industry have been. So this is a debate we’re going to be following closely in the months ahead, to ensure the pro-liberty perspective is heard. So, Robin, thank you very much for joining me today.

Thank you for listening to today’s episode of the Cato Podcast. Once again, I’m Ryan Bourne, joined by economist Robin Hanson. If you enjoyed today’s discussion, please subscribe and leave a review wherever you get your podcasts. To learn more about the ideas and research discussed in this episode, visit Cato​.org. The Cato Podcast is a production of the Cato Institute, dedicated to advancing individual liberty, limited government, free markets and peace. Join us next time for more insights and conversations on the issues shaping our world.