Good Disruption: Prediction Markets

UVA Darden.

It's a betting market.

When you put a head to head comparison between this mechanism and other mechanisms for similar questions, similar time, similar resources, you'll get more accurate estimates, that's the main reason to be interested, and that's why I'm interested.

I'd like us to have more accurate estimates on a lot of important questions in the world.

Innovation changes markets.

Disruption changes how we live.

We bring in the experts, challenge the consensus, and decide which innovations are actually driving society forward.

I'm Dean Yael Gushka Kacane, and I'm Professor Mike Lennox.

Welcome to Good Disruption.

Hey Mike, how's it going?

How are you doing today, Yaya?

I am doing great.

I'm excited to be in the studio chatting with you about, um, I think a topic that is near and dear to my heart, and I hope you're gonna get excited about it if you're not already.

I, I know where we're going, and I am very.

Excited to talk about this, uh, this topic.

So it's for our audience, yeah, what are we, what are we talking about today?

Well, um, I wanted to kind of stage it by asking you potentially a side question, and then we'll bring it back to the main topic of the day.

But do you watch, do you watch, uh, America's Got Talent?

I have watched it.

I am not like a, you know, huge fan.

I have to watch every episode, but I do know what you're talking about.

OK, I wonder when you watch it, and maybe you watched the World Cup, um, back in the summer.

Did you, were you watching the World Cup when I was watching the World Cup.

We have established, I'm a big sports fan.

OK, right, so this is gonna be great.

So when you were watching a game, did you immediately, um, think throughout the game, like, oh, what are the odds, as the odds were changing that 11 team was gonna beat.

The other, how do you, how do you think about it as the game evolves?

Does that become suspenseful and you get more popcorn and you kind of calculate the odds?

How do you, how do your family engage?

No, and I, in fact, I love the kind of expos when you see a sports team that had, like, you know, a 5% chance to win with 1 minute to go, and then somehow flips the odds the other way through some, you know, miraculous comeback.

So, uh, you see those graphs about win percentage over time.

Yeah, so, uh, those graphs about win percentage or who's gonna take, uh, take the game, they do change in the run up before the games, let's say during a tournament, over the life cycle of a tournament, you'll see them changing, and then, um, maybe they'll change throughout the game, and then they can flip at the last minute, and my question to you is really like, where do you go to find those curves?

Like, what is it that you think is behind those curves?

That you trust or you don't trust to kind of give you that input.

Well, I think that's an interesting question, because I know in some cases they're kind of proprietary algorithms that are trying to make these predictions on their own, but sometimes when it comes to, and as I know where we're going, but with like betting markets, that's based on the kind of the collective uh choices made by those who are betting on the games.

Yeah, exactly.

And so really what I wanted to ask you about is if, if you ever looked at Kashi or if you looked at, uh, one of those, uh, uh, prediction markets platforms that are out there these days and are gaining momentum and are very popular, uh, uh, for the sake of coming up with these predictions, have you relied on those sources when you tried to come up with an inside information around who's gonna win the US personally, I have not, I've not bet on Koshi or any of these prediction markets, um.

But I recall them going back, I guess at this point, decades, where especially the presidential prediction markets were kind of first established and interesting to note what they were saying versus what maybe what the pollsters were saying, um, that's a different approach, a different set of information about how you assess this.

Yeah, exactly right.

So the history here is, it goes back, you know, over 40 years when you or about that time, when you think about one of the most famous markets today, the prediction markets today, that is still active and still in operation is the IOa um electronics market.

Do you know, did you, have you heard about this one?

I have, I have.

I think this is the one that really got it all, kind of started, at least in the popular imagination.

That's exactly right.

And not maybe not a huge surprise, economists, you know, the, I think, uh, the story goes, 3 economists walk into a bar, very kind of like almost like a joke, but uh really they, they identified the, the idea that you can leverage a lot of people's opinion, and what better way to collect many people's opinion about the odds, or try to learn really what people believe is the case regarding a certain chance of a candidate to win.

What better way to do it, but not to let them kind of trade on it, or actually put some money um at stake, and really let them, uh, as you say, bet on who is going to be the the winning candidate, um, and so it goes back to 1988, when, when that market started, and today you can still kind of track it and follow it, and then other markets popped up in the 19990s, when it came, uh, became apparent that this could be an interesting, uh, mechanism to elicit information.

Then maybe you can apply it for things like box office success and predicting Hollywood success, uh, uh, um, and the, and the kind, and the sort of the of prediction that might interest people, people might have an opinion, and the more people we ask, the better our knowledge would be, so we can really come with a more curated understanding of a prediction of a certain event, um, happening.

Now, of course, like sports betting has been around for probably ever.

Um, how do we Think about this separate from like traditional sports betting, or is it just a continuation and expanding it to other arenas, is what we're talking about here.

In some ways, this is kind of a little bit of a marriage between uh a sports betting in the traditional sense, when you think about horse racing and the like, and other kind of areas where sports betting is pretty popular and common, versus something like trading in in the stock market, right?

Like you have a a an ask, you have a, a somebody willing to sell, somebody willing to buy, and you Have some kind of exchange, you have these contracts where you're gonna spend a certain amount and you're expecting a certain amount in return if certain events happen, and that in a sense, in essence, kind of reveals what you believe the probability is gonna be of that event occurring.

Just what contract you're willing to sign, what agreement you're willing to make, what price you're willing to put on it, is kind of revealing your odds that something happens.

Now, is he here having skin in the game?

Like we can't just ask people's opinions because we don't get the truth.

Truth unless they actually have to put money in.

So it's interesting that you ask.

So, um, I, you know, the reason I'm so enthusiastic, maybe you hear in my voice, is that I've spent a lot of time thinking about wisdom of the crowd and how we can, uh, actually collect and really pull many different opinions together, and there are different mechanisms that work better than others in terms of truth revelation.

I want individuals to share their real beliefs about odds of events happening, and not to try to game it.

And so you could come up with mechanisms like the winner takes all, if you do some kind of competition, a forecasting competition, that could reveal the truth, and you could, uh, see some, you can learn from that type of uh of structure by taking the average uh prediction, by doing certain combinations and certain aggregation techniques, or you can let people bet on money, and you can let them actually exchange these contracts as an alternative mechanism.

And so one of the things that we're seeing recently, and why these um markets have grown, is because uh so many people are taking part, so it becomes more liquid, meaning there is more money involved, more contracts involved, more people are joining, and that activity allows us to be more and more accurate with these predictions, and they really have, you know, surprised uh uh um a lot of individuals.

By how accurate they are and how well they're predicting occurrences that other mechanisms don't.

And just to give you some, you know, we're, we're business school professors, just to give you some numbers here, the US uh prediction market volume in 2025, so last year, exceeded $50 billion across all platforms, so all of these, uh, uh, companies that have been popping up, 4 million unique users, 2 billion trans.

Actions and 100 $126 billion market volume, so millions and millions of dollars in um in action, and Polymarket and Koshi, those are names that maybe people now read and hear about, they're integrated into CNN and other outlets, and so they're collaborating with other news outlets, and so they're really becoming really um ubiquitous and um pretty, pretty reliable, I would say.

And and pretty profitable companies themselves.

Yes, OK, and so there may be, um, I don't know if there's concern or ethical questions around this, like, um, uh, why did, haven't you participated so far in one of these markets, I would say.

Yeah, and so like they're taking, I assume, a percentage of the uh the amount bet here.

Are, are most of these bets, um, dare I say like two sided, there has to be somebody on the other side of the bet to take the contract, um, in these.

Yeah, typically you, there, there, there will be offers and markets that you can participate, you can put it in, and you can see what, if somebody kind of, if the platform is allowing you to to submit that contract, um, these things are evolving.

The real breakthrough actually is regulatory, I would say, the volume has increased because also some of the regulations has softened, and so it used to be the case that you couldn't create these contracts on every type of Events, sport betting wasn't always um allowed or legal in, in different states in the United States.

Some of the election, uh, uh, betting has, has not always been legal, but now some of the regulatory uh landscape has changed, and so we're seeing a lot more activity.

But, um, maybe it's time for us to talk a little bit more in detail on what exactly these markets do, and so we can bring in an expert to share with us their knowledge.

Do you think that this is a good time?

Always a good time for us to.

In the expert, you know, well, I am, I am so excited.

I've heard and been following for years the work of, uh, Professor Robin, uh, Henson, who's joining us today, associate professor of economics at George Mason University, so just up the road from us here in Virginia.

He has contributed dozens, and I mean dozens of publications on prediction markets dating back to the 1990s, so really one of the, um, uh, leaders in the field.

And has been featured on hundreds of podcasts, and now ours too, um, and articles on the in the Financial Times, Forbes, um, ESPN, Atlantic, Wall Street Journal, and many, many more.

He's recently appeared on CNN to talk specifically about this topic, prediction markets, and, um, betting on world events, which we're um seeing more and more of, like, who's gonna win America's Got Talent, um, but Also, um, other, um, events.

Um, Robin is an associate editor of the Journal of Prediction Markets.

There is a journal dedicated to this, uh, topic, and prior to teaching, he was a research scientist at the artificial intelligence at, um, NASA, um, and, uh, Lockheed Artificial Intelligence Center.

Robin, we are so honored that you are joining us here today.

Thank you for your time.

Great to meet you all.

What should we talk about?

Well, prediction markets, obviously.

What's the question?

Why?

What's the issues?

So, so, um, help us understand.

I wanna make sure that we get it right.

Mike was asking, um, some questions around how the swapping and the contracting takes place.

Can you help us make sure that we and our audience understands, like, what is the mechanism that these markets help us understand the real probabilities of an event happening.

Well, it's just a market, it's a betting market.

So there's an asset that pays off if one claim is true, and doesn't pay off, and then there's a complementary asset that pays off if the opposite event is true.

And then you let people buy and sell these assets, and then the price becomes your estimate of the chance of that event.

Or if you make a scaled version, the price could be an expected value for that event.

And, you know, we have lots of theory.

I'm an economics professor, we can walk through the theory, but most people don't believe economists' theories anyway, so the thing to point to is the data.

We've got lots of data on these things for many years, and they just not only are usually well calibrated in the sense of it says 70%, that happens 7 out of 10 times.

But when you put a head to head comparison between this mechanism and other mechanisms, they just beat the other mechanisms.

You just, for similar questions, similar time, similar resources, you'll get more accurate estimates from these mechanisms.

That's the main reason to be interested, uh, and that's why I'm interested.

I'd like us to have more accurate estimates on a lot of important questions in the world.

So Robin, maybe we can, you know, explore that a little deeper, right?

So if you think about alternatives like polling when it comes to political races and the like, what is it about the betting markets that gives us more accurate predictions, you know, is, is it, is it something about the skin in the game, as I was saying, so there could be a truthfulness, you know, revealed here that wouldn't be true when someone just asked me a question and I have no consequences for not expressing my true opinion.

Um, is it about who participates, you know, is it not the wisdom of crowds as much as it is the smart money is playing and that gives us a better accurate, uh, prediction.

What, what's the actual underlying mechanism that makes these better than other ways we might forecast?

Well, before we discuss mechanisms, just to distinguish three different kinds of things.

There's information sources, like a poll is a source of information, or a news article is a source of information.

Then there's different methods of aggregating sources, different ways to use statistics or, you know, calculations to put together an estimate from all the different sources you have.

And then there are different forums in which these different methods compete.

So, we could either, you know, put a crown on somebody's head and say, you're the czar of forecasting and we'll just trust your method.

Or we can have a forum in which people have different methods, and they compete, and then the met forum itself gives us the weighted average of the winners, and we can just trust the forum.

So, I'll recommend these prediction markets as forums.

They aren't a method, the participants will bring their methods, and they will try lots of methods, and the forum doesn't need to care what their methods are, the forum just needs to judge among the participants who's doing better.

So, First of all, the one key thing is just to have a form, to have lots of different people who can compete with different approaches, and have a clear incentive for whoever is winning, doing better in that, to be rewarded.

So that's the first power of this method, is, it's giving people nice, clear, strong incentives that if you are more accurate than other people, you will in fact profit and win.

Now, but that's not the only thing, we also have a powerful incentive to self-select.

So, if you're a reporter and you call somebody up, they'll usually like blather at you, even if they don't actually know what they're talking about because they're honored to talk.

They don't have a good incentive to just shut up and say, I don't know, and hang up.

But with the markets, you have a strong incentive to only talk to the topics you actually think you're world-class at.

And you have pretty strong incentives for that.

So, you know, there's the old saying in poker, you sit down at the table, you look for the fool, uh, you're all gonna be making money off the fool.

If you don't see the fool, he's you, walk away.

A good attitude in these markets, you should think you're world-class, and you need to have a pretty good reason to think you're world-class, or you should just shut up about that topic and let somebody else speak.

So there's a strong incentive to self-select in the markets and only speak up on the topics you actually have a decent reason to think you're best.

I, I get that from a sense of like who wins, but I'm thinking about like sports betting, right?

And um I'm a Philadelphia Eagles fan, big Philly sports fan, I grew up in that area.

Um, pretty irrational fan base, right?

And so now we can be self-deprecating, but let's just go with the idea that you have a group of people who are overly optimistic about some outcome and are willing to put money into it because of their passion, their blind passion, irrational passion for something.

Now the smart money is like, I'm gonna take advantage of that, right?

The line's gonna shift, and now I can take advantage of that.

But tell us how the mechanism then equates to a better prediction rather than just being like, yeah, I'm the smart money and I want a bunch of dumb money out there in the market.

It looks crazy, but I'm the one who's winning because I'm the smart market money in the market there.

And I actually don't want much smart money in the market because I'm taking advantage of this situation.

These market prices are dominated by who you call the smart money.

They not necessarily the majority of traders, they are often a minority of the ones who are actually paying the most attention and actually working the longest, hardest at getting it right, but they set the prices, and that's the key point.

All the other people are dragged along, and they're losing money, and the smart money is setting the prices and getting it right.

So, there's a key um selection effect I just talked about, and there's also just the look for biases effect.

So, Everybody's invited to look at the prices and ask, where is it wrong?

Right.

And if you can find any way there's a bias, anywhere, you can make money on that bias.

So you don't necessarily have to look at the ground truth and the core, like, you know, this player's left leg or whatever.

You can look at these overall patterns in the market and see.

Any pattern you can find that represents some sort of bias you can make money on, that's how much of the financial markets work.

So everything we're talking about is basically about how financial markets work.

These markets aren't that different from financial markets in these basic mechanisms and the basic way in which they produce accurate prices.

And what's interesting to me, and um, correct me, Robin, if this is the wrong way to think about it, but going back to your question, Mike, about how it compares to other environments, um, Phil Tetlock, who I know you know well, Robin, um, recommend, uh, talks about super forecasts, right?

Like some of the mechanisms of other environments have shown us over time that if you're gonna rely on a group of not just one person, one expert, as Robin mentioned before, but if you You're gonna rely on a group of people to make predictions, you'll, you, you really need those super forecasters in there because they, they at the end of the day are gonna carry the day and help you be more accurate.

In the markets, those super forecasters naturally emerge by what they're willing to put down in terms of money and, and how much they're willing to pay for it, and how much sophistication they have in their in their sense of, of the prediction.

Is that, is that fair to say?

Markets are an information institution, and I think you should compare them to the other information institutions in society, like journalism or academics, or, you know, gossip or government, you know, agencies.

If you think about those other organizations, if you see some mistakes in the newspaper, you don't have a way to make money by fixing that.

You see mistakes in academic journals, you also don't have a straightforward way to make money by fixing that.

Some of us wish we could.

We indeed we did.

Right.

Exactly.

So.

You'll notice that with most of our other information institutions, you could be a smart super forecaster, you could see errors and biases in the, you know, what everybody else is seeing, and you can't make money on that.

You do not, you're not invited to come fix it and make money.

That's the miracle of the financial markets.

There you have the opening.

Any bias you can see, any error you can see, you are invited and paid to fix those errors, and that's why they can beat out these other institutions.

Because in the other institutions, we have lots of people, your colleagues, probably yourselves sometimes, you go, that's not right.

You read a journal article, you read a newspaper article, you can see how that's in error, and you can't fix it.

I'm curious, not in, you know, comparison to other methods, I, I'm, I'm right with you.

I, I'm thinking about like even in the stock market, um, there, there are, at least to me, seems to be absolutely bubbles, there seems to be manias, there seems to be bandwagon effects.

Some of these behavioral issues, and we still hope the smart money is there, but history seems to prove that the smart money sometimes gets, uh, out over their skis, and we can argue about the long-term efficiency of capital markets, but the short term.

Potential for there being bubbles and the like.

Do we have the same phenomena, let's say in a more, um, endpoint prediction market, whether it's, you know, some event or the like where people get overly excited about a certain political candidate and their, their value starts trading at a huge value, and then, yes, maybe the day before the election, it collapses again, but there are these moments where the mania takes over.

Look, these are all humans participating in these markets.

There are no like, you know, wise aliens or angels, OK.

They're all humans and they all suffer all the usual things you've seen humans suffer from all your life, OK?

They aren't special, they aren't perfect, they fail in all these ways you're talking about.

But the question is, how do we empower somebody to find these mistakes and fix them.

So like think about the 2008 financial crisis, right?

A big example would to say, look, prices were wrong.

And then there was a big correction to fix them.

And we now ask, where did we first learn the prices were wrong?

Where was the first place in society where we started to get the message, these prices are wrong.

And then that message spread and then the correction happened.

Well, it was in financial markets.

Right?

Read the film, see the film The Big Short.

The first people who were telling the message, so.

If it was somewhere else in society like academia, where we're first learning the message, then you'd have a basis to say, oh, well, markets have this problem that academia can fix, and we need to like empower academia to oversee the markets to make sure we fix these problems, but it's actually typically in these financial markets where the first news is revealed that the markets have a problem.

But to the point, that was a very contrarian position.

And if you think about credit, credit default swaps, um, there's a belief in financial markets, right, that the creation of new derivatives always creates efficiencies ultimately in capital markets.

But with credit default swaps, there was a belief that, well, that might not have been quite right, that it actually created confusion and actually obfuscated information.

Uh, and if you write, read, you know, if you watch and read The Big Short, of course, they talk about this, right?

Like they we had these bundles of assets that were actually Um, basically, removing information from the market, not adding information to the market.

And I know I'm, I'm pushing you here, but why aren't prediction markets then creating similar types of, um, almost information obfuscation, um, when you have You know, it, it just proliferation of these out there.

And let me, as, as you're thinking about your response, Robin, to that, I just wanna, um, mention a few like, um, examples from this week, what's trading just to get us thinking about, about the opportunities here and the challenges.

For instance, South Carolina Republican Senate primary is being, uh, there are contracts being traded on that.

How high will Bitcoin get in August?

Who will Trump's next press secretary be?

Uh, will open AI or anthropic, um, IPO first?

So if you have some insight into that, please go ahead and, and, and put, put your money on that.

Uh, Big Brother Elimination week seven, if you're following Big Brother, uh, Time Person of the Year for 2026, uh, Mike and I are available for that.

Um, Iran nuclear deal, The Odyssey total domestic gross, I paid my $13 to see it, um, and, and whether on any given day in any given city.

That's the kind of things that are being traded now, OK?

Um, Robin, back to you.

OK, centuries ago, journalism got started.

And at the time, there were the existing authorities at the time, the church, the state, the aristocrats, they didn't like journalism, and they said it was this upstart, disrespectful, undermining proper authority, it was sensationalist, and they thought the proper authority should be controlling journalism, making sure that its problems didn't infect the public spirit and world.

So, we long just have these competing information institutions, each of which says, Those other ones are good, but they have problems, you need to trust me to fix them.

And so we have this meta-roblem of who do we trust as the as the institution to to oversee and fix the other ones, cause they all have problems.

So that's why we have to do this comparative analysis and say, OK, in a head to head comparison, who's actually doing better?

That's we're gonna say, who would we empower to fix or oversee the other.

And I'd say, whatever problems you Seeing with the financial markets, you're not seeing some other party who does better that you should empower to oversee and fix them.

Fair enough.

Like, it's in that market are who are the best people to fix it.

That's when we saw the mistakes, it was people in the markets who found the mistakes, who profited from telling us about the mistakes, and that's how the mistakes got fixed in that case.

Uh, when we think about disruption, we've often talked on this show about why now.

And so part of what I mentioned earlier on, and I'm curious to your thoughts.

On this, part of why we're seeing, seeing this growth in, um, popularity of these prediction markets or the valuation that goes along with the companies that are involved, part of it is because the regulatory um landscape has changed and so the, um, the change that we saw, the CFTC, the, uh, Commodity Futures Trading Commission, has really allowed these types of trades to occur and to take place and most states, I think it's by state still, and, and correct me if I'm wrong, that there's a, a.

A state level regulation that also goes hand in hand with that.

Like, what, what happened?

Like how, how is it that this has changed and what kind of impact has it, has it had on these prediction markets?

Well, you're right that it was primarily a regulatory change, although it was initiated by lawsuit.

Calci sued the CFTC saying, hey, you can't stop us the way you're trying to do, and the courts agreed with CalSci, and that created this regulatory change that then has allowed the burst of Kalsi Polymarket and many new competitors.

And that's the news, that's the thing that's happened in the last few years, is this change of regulation.

We're now having a lot of activity, and the reg the new regulation basically said that overrules state regulation.

And the state regulators are not happy because we have a lot of gambling industries in states who were making money under the state regulations, and now their customers are being taken away by these new federally regulated competitors who are offering cheaper prices for similar products.

And there's a lawsuit still going on saying, no, no, the state regulation should be the one that rules, but so far, the status is that federal regulations rule, and now this is new and different, and that's, you know, bothered the people who are making money in the state.

But a lot of other people are bothered, you know, a lot of people for a long time have thought we shouldn't have so many markets like this.

They call them gambling, and they see problems with gambling, and so there's a lot of backlash, a lot of uh complaining.

And there's this key question, will there be some sort of regulatory backlash to uh rein this in?

And obviously some people wish for that.

I'm not seeing it soon.

You could imagine Congress, but you know, as long as Trump's president, that's probably not gonna happen.

Trump might flip, because he does flip sometimes, he's not that reliable an ally, but that still seems unlikely.

The courts could flip and say no, no, and make a surprising ruling, but that seems unlikely.

So I gotta say the the best chance is the next Democratic president, which is a ways off.

And then the key question is, will people still be mad?

And in the mood to do something because as you know, the public is kind of fickle in terms of what it's mad or upset about, you know, these just things go in cycles or they go away, they kind of switch to, you know, actually data centers.

Lately, I see less complaining about prediction markets, more about data centers.

So, you know, maybe people, maybe that that's.

The question, well, I wanted to, I wanted to ask actually and pivot just for a second, we can come back to that because it's an interesting dimension on the future of those specific companies, right?

Cashi and Polymarket, they make their market from the public and that the regulatory environment that supports that.

But you know, going back from when you started to do this work, um, I've been following work, um, um, by, by other academics also like.

Like us to do like wisdom of the crowd stuff, you know, companies have been trying to run these internal prediction markets for, for years now.

I know Google is, is one of the kind of more celebrated cases.

I think HP did it when HP was still, um, uh, uh, you know, active largely in that domain.

So they tried to run prediction markets internally to gather information from the employee.

Employees about certain things that happen internally, like what are the chances that we'll get our products or that there'll be a supply chain delay or that we'll launch a new product on time, but I think they struggle to get these markets off the ground.

I'm curious if you have um a prediction, um, no, no pun intended here, around the adoption of these prediction markets in-house.

This has long been my hope that we could use this powerful mechanism to aggregate information in lots of other social contexts.

And so, many, I did push and many people did try to get companies and other organizations to try this, and there's a lot of organizational obstacles there, and I think we learned about what some of them were in that process.

So, I like to give the analogy of putting an artist in the C-suite table.

Imagine, you know, we've got all the top executives of a company and.

There's one person at the table who understands the company really well, but they have no sense of anybody's agenda or narratives or things they're trying to push, and whenever a subject comes up at the table, they just blather whatever they happen to know about that subject.

That person just typically won't be allowed to sit at that table, because they'll be too disruptive to the other people who are trying to coordinate to like make alliances and undermine their rivals and push narratives.

So, that's what a prediction market is though, in an organization.

It's a knowledgeable source that has no savvy or sense of how to play ball and play politics.

It just blathers whatever it thinks, regardless of who's gonna be upset by that, who it'll undermine, and that's a key obstacle.

So we need to find ways to feel these things so that they can be less disruptive and therefore more acceptable for uh management to to use as part of the process.

So, Most, as you probably know, as business schools, you know that most managers, at least who come from an MBA program, present themselves as scientific decision makers, right?

What they've got a spreadsheet, they're trying to fill in the numbers, then they're gonna calculate the best thing for their division to do.

And in fact, most managers are politicians, OK?

They represent a part of some coalition.

Coalition has an agenda, has a narrative, it's trying to push.

They're playing ball with their coalition to try to, like, you know, get their stuff happening, right?

That's what managers really are.

And their pretense of being the spreadsheet filler says not our Darden MBAs.

OK, fine.

You guys must be an exception.

I'm happy to accept that for, for, for the purpose of argument here, but the point is.

You know, the standard image they project says, of course they want more information.

More information will help them fill in the spreadsheet, get more accurate estimates, but in fact, more accurate information that's out of control is very disruptive to coalitions trying to push agendas through managing narratives.

Yeah, I'll give you an example.

I do a lot on, uh, the interface between business strategy and public policy, and I'm always amazed by companies who do incredible rigorous analysis on market demand and market features, and then they turn to policy risk and they just throw up their hands like, I have no idea if the Strait of.

It's gonna open or not.

But I'm like, but you must have some expectation and prediction about what is going to happen here.

I'm always surprised how wildly the market swings with like every pronouncement the president makes.

It's like, haven't you baked that into your estimate yet, that that's gonna be part of the uh of the game.

So I, I think a question related to that is when you think about these markets, one of the things that I've heard said as a feature, but I'd be curious, is the ability to bring inside information.

Now, we have reasons in our financial markets that we don't want to create a playing field with those who are trading on inside information.

What what are your thoughts on this?

Like, you know, is it good that there are people playing in these markets who actually might know things, uh, that others don't, or is that a, is that a problem here?

So, when you set up an information institution, you're basically creating incentives for people to find information any way they can and include it in the process.

Now, that can conflict with organizations legitimate interest to keep some secrets.

And in journalism, we've long seen that conflict.

Many of the most famous, uh, you know, journalist cases of decades have been cases where journalists revealed secrets to some organizations were trying to keep.

So, uh, I think as the policy for an information institution shouldn't go to either extreme, it shouldn't try to just always make sure all organizational secrets are kept on the one hand, on the other hand, it shouldn't just always try to reveal all secrets, it should be setting some intermediate, you know, position where try to let people legitimately keep secrets, but also let them reveal things.

Now, I gotta say that ordinary uh stock markets do in fact quite consistently reveal secrets.

Insider trading is rampant there.

I think we just should admit that the status quo in financial markets.

Like, when a company makes an announcement, OK, on average, the price moves in response to that announcement, but half of that move happens before the announcement, and about half of that's insider trading.

So insider trading is in fact rampant in ordinary stock markets.

And the world you've seen is a result of that.

So I gotta say, maybe we'd be better if we cut down on that, but it is really hard to cut down on that, and the world we're in seems to be somewhat OK, so I gotta say, I guess it's somewhat OK if a lot of insider information comes out, cause that's where we are.

And I think both, both Koshi and polymarkets, I think, have debated or have, they all say they're stopping it.

They, they all have rules against it.

Government says there's rules, but that's also true in the stock market, you know.

Everybody says they have rules against that.

The SEC has strong rules, and what we see is that in fact it's rampant.

And with anonymity, like if you can have anonymous trades, then it makes it harder to kind of really kind of scrutinize some of this, and at the volumes that we're seeing, it's also gonna be really hard.

But um uh and I like your point from before when we talked about in in-house in-company trading, or in in-company prediction markets that It's good to crowdsource information, but the naivete of different individuals across the organization about various agenda or their vantage point might be disruptive, right, from a business perspective, right?

If I am a leader of the organization, uh, we can call it political, but I'm also in a position of, uh, I'm put in that position because I have some responsibility.

To make certain decisions that not everybody else could really understand or kind of think strategically about.

Of course, that's why people didn't like journalism a long time ago, you know, the, you know, the state and the aristocrats in the church, they had some agendas and they coordinated on the official story and then journalisms were undermining that, and we've decided over time that it's more OK for the journalists to undermine those stories.

And maybe we should be a little more OK with official corporate stories being undermined too, but That's, you know, where we have to figure out where we're gonna go.

Beyond the, the information piece, um, there's also the potential in some of these markets for, um, affecting the outcomes, right?

Like direct, uh, like I called endogeneity here in these, um, you know, sports betting markets, it's been interesting, you know, maybe hard to like fix a, uh, outcome of a game, as we know, for a variety of reasons, you, you know, the coordination costs and the like.

But now we have bets where you can say, you know, well Yael get 4 rebounds in the 3rd quarter of a game, and Yael's sitting there going like, I actually can control that outcome.

I'm not getting any rebounds.

I'm 5'2.

Don't, I'm just saying, but you can control that outcome to some extent, uh, and we have cases of like athletes betting on themselves, betting in, um, um, you know, insider betting, I guess.

You will, but not just on the information advantage, but in the actual outcome with policies you could imagine like if we're gonna start a war, let's say, uh, with Venezuela.

I know there was a lot of debates of whether we actually started a war, we just had an insurrection and whether that counted for what the polymarkets were, um, you know, uh, betting on there.

How do we think about that, that endogeneity?

So this is again where I just compare this to other information institutions.

It's long been true that people have changed the world to get favorable news coverage.

News doesn't just report on the world, news induces the world to change in order to get the news to, you know, treat something well.

And academics are the same, and gossip is the same.

Information institutions induce changes to the world, sometimes they're good, sometimes they're bad.

It's also true that financial markets have long had this potential, and many people have feared that, for example, having markets on stock prices means that somebody would go sabotage a company in order to make money uh on the in the stock market on that price.

So, what we've seen when we've looked at financial markets is that surprisingly little sabotage happens of corporations.

We've found it very hard to find any examples where that actually happens.

Uh, what we do see maybe is people buying life insurance and other people and maybe, you know, hurting them.

So, the way to think about it is, you need a combination of an amount of money at stake and an ease of one person influencing, and a combination of that could be worrisome as in the case of my buying life insurance on you, a stranger, for example.

So, that's the place to set the boundary, but most stock markets aren't in that combination space.

The events are too hard to influence for individual groups to get together to, to, to who can put the amount of money that would be required to make money on that.

And that's a surprising fact about most financial markets is there's relatively little sabotage there.

And that sort of sets a reference point for these prediction markets too.

You have some world event, then it's hard for one or a few people to influence them.

But if we're betting on what one guy will say in one meeting, then yeah, of course that guy can influence it.

But I do think that in those cases, it's pretty obvious.

So I think if a bunch of people decide they wanna bet on what's one guy's gonna say or how the dance he's gonna do when he makes a touchdown or something, they all know they wanna do that, they know this guy can influence it, they think it's fun.

My general attitude for a lot of this is, I don't mind people having fun, if sort of they know what they're doing.

Well, that's an interesting point because, you know, we used to have uh more stringent qualifications for those who actually participated in our financial markets, right, qualified investors and the like with the fear that the average, um, you know, citizen.

Uh, doesn't have the sophistication, you know, it goes back to your point about playing, you know, um, I will not go online and play poker because I know I am, you know, the one they're all training on there.

Uh, but some people don't, right?

Like some people actually, you know, either have decision bias or maybe think they're better than they are, and they're out there, and now we're giving them the mechanism, you know, the forum, as you said before, to be able to trade in ways that they never would have been able to do before.

So, you mentioned the super forecasters before.

Uh, an interesting thing they showed is that if you have some sort of process like a market that everybody can participate in, and then you get some sort of accuracy out of that, and then you take from that market only the best traders, and you have a whole new market where only the best traders can participate.

That new market will be more accurate.

Right.

And that's the thing we're not doing that we could do with all familiar financial markets.

It's because we let everybody play in them, they in fact are less accurate probably than if we had versions of them that only the best traders could trade in.

So we are giving up the option to have more accurate financial markets by not doing this thing of having the specialist superfirecaster only markets.

That's maybe something to think about policy-wise for our financial markets and our, you know, financial economy.

We could in fact have more accurate prices on most financial markets.

Interesting, um, I know, um, you're obviously, as we said up front, you're an expert in this field, um, just by your voice, you're an enthusiast about, uh, prediction markets, but still there are ethical concerns that people have, like, uh, with many things these days, especially when it's online and there's money involved.

Um, what, what to you is like a valid concern that we should be thinking about or that Koshi and Poly Marets and their, their peers should be, uh, trying to resolve?

The valid concerns, I think, are the ones that are in common across the information institutions.

I think I'm suspicious of people complaining about the prediction markets for issues that are also in all of the other information institutions, but having a different standard for them.

So for example, right from the basics, all information institutions add more information.

There are plausibly topics where more information is not good.

For example, most people, many people say that on election day it's bad if people can find out who's winning toward the end of the day, because then they might not vote because they think they know the answer's already set.

That's an argument for information being bad.

Therefore, you might want to limit all information institutions, prediction markets, journalists, everybody, and saying, OK, this information's not good, let's shut it down, let's not have it.

OK.

Similarly, we might say sabotage, uh, you might do illegal things in order to make money or reward in any of these information institutions.

The main thing to do there is to just have records be subpoenaed by people who suspect suspect foul play.

That's how we do it in many of these sorts of areas where there's a possible crime.

We try to make sure that if somebody suspects some foul play, that there's a way to go dig into the information and check on various suspicions by having information that can be subpoenaed.

That doesn't, that's different than having just shutting it all down.

Uh, that's way too, you know, extreme a response, but yes, you should worry about people, you know, having incentives to do foul play.

So again, There's a number of those sort of things you could worry about, with many of these institutions and other worries, just manipulation.

Like people often, you might not realize this, try to distort news articles so that they can get favorable coverage out of the other end.

They try to manipulate the inputs into the process.

Now it turns out betting markets and speculative markets are unusually resistant to that.

So, compared to other institutions, they are very good at resisting manipulation.

So good in fact that when people try harder to manipulate, the prices get more accurate as a result.

So, that's the one case where I can point to and say, this institution just does much better at dealing with that problem.

That's super helpful.

I appreciate the way that you think about it, Robin, and everything that you've taught us today.

Um, Mike, do you feel more informed about prediction markets?

I feel much more informed.

I really appreciate it, Robin.

We do appreciate your time on this.

This is a, it is such a, a nuanced topic, I feel, and there is so much discussion of it out there, but maybe not.

Um, the depth of the understanding that you're bringing to us.

So thank you, thank you, yeah, thank you for your time, Robin.

We learned so much and it's wonderful to hear from an expert on the topic, and, um, and good luck.

Um, I hope you, you have some, some contracts, uh, unfolding on the, on, on the prediction markets today.

Uh, uh, great to meet you all, and I'm very honored because I just gave my first lecture in IO last night, where I said, look, this is basically a business strategy, people.

And you know, in business school, they're talking about these same subjects and it matters there, and I try to inspire my students to pay attention to the subject there by comparing them to you guys.

Well, that's great.

Amazing.

Well, it was nice to meet you and thank you so much.

Um, Mike, um, disruption, are you, do you think that this is gonna change the way that we think about forecasting?

Yeah, I think, I think it's fair to say this is a disruption that's upon us.

Um, it's, I think, as we've talked many times before about the, uh, the market structure implications here.

I mean, clearly there's an emerging market sector, uh, in these markets, uh, that seems to be very viable and, and growing there.

Um, how this impacts other kind of secondary markets or derivative markets, um, again, I mentioned like, you know, the prediction for policymakers and how that then affects corporate action.

I think that's gonna be interesting to think through.

Um, there is a marketplace for whether it's Bloomberg or S&P and these companies who provide information to financial investors.

This is another mechanism they're partnering with some of these, uh, prediction markets indeed to kind of establish that, that, uh.

That insight, yeah, and as Robin talked about, we, we, we've done this in financial markets forever, but now we're starting to be able to do this, like I say, in these more, um, general, uh, population things like whether it's a a world event or the likelihood of some, I, I imagine you could predict like hurricanes and things like that, um, and there's some, some value there.

What, what, what are your thinking, you're thinking, yeah overall in terms of like the, the societal worth of, of this.

Well, for me, um, I've been for a few years now, I've been fascinated by, um, the chance that, uh, we're finally in a time and age where more people will be paying attention to probabilistic thinking, meaning thinking about the future in probability terms, and understanding what that means, and even if I say that something is 90% chance of happening, or 85% chance, what exactly do I imply?

If it didn't happen, does that mean I'm wrong?

And how do I kind of assess that calibration that Robin talked about?

So for me, if more people are participating and it raises awareness to thinking probabilistically, the net outcome of that is good.

Um, you know, we saw in the pandemic that we were using prediction curves and probability curves.

You mentioned hurricane forecasting that.

It's a huge factor in how you think about evacuation and and dealing with the public when you understand the probability, and we still don't really understand probability that all that well, so in that sense, from uh my academic mission, I would say this is a good, a good disruption, that it's just a fun thing to take part in, people make some money, people lose some money, but suddenly they're paying more attention to information and what it means.

Yeah, I, I, I actually might be falling on the other side of this.

OK, good, disagreement rare for me.

I, I am totally sold on the value of aggregation and the value of more accurate predictions.

That is useful for society in innumerable ways.

So that I'm fine.

We're creating those accurate predictions on the back of individual, um, for lack of a better term, betters in these markets.

And as we talked about, there are sophisticated betters, there are less sophisticated betters.

Um, I, I worry about the societal implications when we, we're basically opening up gambling in all its facets to all possible things you could ever think about, uh, and monetizing that, you know, there's been epochs in the US going back to like the 1920s where we, we pulled back from gambling because of the fears, and it's really only been over the last 30 years or so that we've really expanded gambling back out to society, um, more broadly.

Sports betting is a great example where's huge.

Used to be very restricted and very regulated, and now it's just opened up where you can just go on the app and do that and, and there's been a lot written about, you know, especially young men who are kind of getting addicted to this and they're, they're losing money, uh, they're not the sophisticated investors.

And then there are those secondary impacts in terms of incentivizing, um, bad behavior, we'll call it, you know, like we're saying, you know, the, um, insider trading is one issue, then the actual manipulation of markets by manipulating outcomes.

Um, clearly agree with Robin.

There's certain events that are not going to be manipulable.

Hurricanes, good example.

Um, sporting events, for example, absolutely, and we're getting more and more cases of people manipulating those markets for financial gain.

Um, I think one of the things we saw in previous epochs where gambling was rampant, uh, was organized crime that kind of goes along with this, and it was one of the reasons as a society we pulled back.

So, we just seem to be in an epic right now where we're allowing more and more of this type of, um, Uh dare I call it gambling, and I, I'm not sure where this ultimately ends.

So, like many things in life, I think there's gonna need to be a balance between the advantages that prediction markets give us in that aggregate sense, but then how do we think about the individuals who participate, the implications for them, and then the secondary impacts on, like I said, manipulation of markets, um, that could be, that could be rampant.

And I think that repeatedly in our, um, disrupt in our good disruption show, even when we Both are on the good disruption side or or or or somewhere in the middle we recognize the importance of guardrails and policy, um, both guardrails for the platforms themselves and the market makers themselves and how they think about trying to minimize some of the impact of those, uh, kind of negative, um, uh, both externalities and, and factors, uh, within, um, and so some of their maturity and evolution will take care of some of your concerns, but maybe we're left with some residual that is harder to, to tackle.

Around just how accepting we are even of sports betting or of people's willingness to go into the casino and play, uh, um, uh, play poker, um, I think that that sometimes is a matter of taste, um, and, and how do we put the right limitation.

So, um, in conclusion, Mike, I guess I'm saying good good disruption, and you're saying no.

I'm gonna go say bad.

I'm gonna go just because we never do it, because we almost never do it, and so I gotta do, I gotta do it every once in a while.

Well, you'll, you'll come back to us and maybe you'll.

Share with us when you, uh, actually go ahead with your first contract on one of these platforms, you'll come back and say maybe you've changed your mind.

OK, so maybe that's right, we'll give you the opening for that.

OK, good.

Um, all right, well, thank you, uh, you know, Robin, of course, thank you to Gary, our producer.

Thank you to our content producer Becky, our, our new title for Becky, um, you know, extraordinaire and all things good disruption, uh, Yael, as always a pleasure.

Thank you, Mike.

I enjoyed talking to you about prediction markets today.

All right, take care, everyone.

Good Disruption is a podcast of the University of Virginia Darden School of Business, produced by Darden Video Production in cooperation with the Batton Institute for Entrepreneurship, Innovation and Technology.

Technical producer, Gary Peters.

Content producer, Rebecca Duff.

Look for new episodes the first Monday of each month.

Darden