Why Prediction Markets Matter for Business Intelligence

Steve Odland: Joining me today is PJ Tabit, Principal Economic Policy Analyst at The CEO Center of The Conference Board. PJ, welcome.

PJ Tabit: Thanks for having me, Steve.

Steve Odland: So you've written this great new paper on prediction markets. First of all, where can our listeners find your paper?

PJ Tabit: It's of course on our website under The CEO Center.

Steve Odland: Yeah, so tcb.org, and then look under Centers, and it's The CEO Center, and you'll find this paper. Now, you can't turn on the television today without, "You should bet on this. You should bet on this." What's the range of what you're referring to when you talk about the prediction markets?

PJ Tabit: Simply, a prediction market's an online platform where people make bets about what may happen in the future. And when you go on these markets, you'll see pretty quickly that the kinds of things you can bet on really run the gamut, from things with actual economic consequences-- maybe it's the price of oil at the end of the year, to what the GDP numbers may be in the next quarter-- to things that are a little bit more exotic, like whether the US government will confirm the existence of aliens in the next couple of years. You can bet on pretty much anything that you're interested in betting on.

Steve Odland: Now that you've said that, what's the prediction on that?

PJ Tabit: That's a good question. I haven't looked in a while. I'm guessing the odds are probably pretty low but if you have some insight into that maybe you can place a winning bet.

Steve Odland: I don't know. It's an alien concept to me. But every sports show, all these betting sites. Is that a prediction market?

PJ Tabit: It's definitely part of it, and that's one of the ongoing debates in the policy questions around prediction markets.

If you go to the major prediction markets, sports is a huge part of their business. Of course, sports betting has been around for a while. There are sports books, both physical and online, and there's this ongoing debate about what qualifies as sports betting-- much of which is regulated at the state level-- and then what's considered a prediction market, which is regulated at the federal level. It's an ongoing policy debate.

Steve Odland: That's an interesting distinction. And it used to all be illegal, or it was for a while, and then the regulations changed, which allowed this. And so anyone who's watching football or baseball or any of these, any of these sports, you see these things and you're able to get it on your app and you do it.

But what people may not realize is it's not just these betting sites on sports. You're talking about predictions of almost any event, right? On any timetable. And it can be really wacky kinds of stuff but it also can be really important economic data that they're predicting.

PJ Tabit: Absolutely. And there's a quote from one of the CEOs of a major prediction market where he said, "The goal is to financialize everything." Every aspect, both economic and, as we talked about, some of the more quirky predictions, is to have a market where you can bet on whether or not that thing will happen.

Steve Odland: If you're betting on something, do they all have a payout?

PJ Tabit: They do, yeah. All of the bets are things that can be quantified and part of the terms is saying this is how this will be decided we'll use this data source that will tell us at this point in time whether or not this thing happened.

Steve Odland: When you're talking about bets, you win or lose, right? In all of these prediction markets, is there the possibility of earning money by being right?

PJ Tabit: Absolutely.

Steve Odland: Okay. So that's ultimately the goal here. Where's the money come from?

PJ Tabit: From users. Yeah. Users online who think they have an insight into whether or not something will happen.

Steve Odland: And they put money down and then, obviously, whoever wins gets the pot. Essentially. We're simplifying it. There are algorithms that allocate the winnings. But it's basically, it's for entertainment but it's also for monetary gain on the individual's part through the betting, right?

PJ Tabit: Absolutely. And in some cases getting to where businesses may be interested in this. There was a story I read recently about an ice cream stand that was buying contracts on prediction markets about what the temperature may be in the future because, of course, sales of ice cream are highly correlated to what the temperature is. And so they hedged their risk on temperatures going down by buying contracts on the prediction market.

Steve Odland: Yeah, it's really interesting. Your paper doesn't really focus on sports betting. It focuses on prediction markets as they relate to broader economic stuff, which again, most people don't think about.

But there is this market out there. It can be on the broader economic themes, governments, whether one country is going to bomb another. So there's geopolitical stuff, military stuff. But there's also betting on companies and what's going to happen to their earnings and that kind of stuff too, right?

PJ Tabit: Yes. And that is one of the areas I think that should get attention from executives, that even if you're not actively participating in the prediction market your company may show up on one of the prediction market sites, from everything from a product launch to an executive change to your stock price.

There's an example recently of a CEO who ended an earnings call just by reading a list of words because there was a contract being traded on whether or not they would say those words during the earnings call, and they thought that it would be fun just to hit all of those at the end of the call. So companies are engaging it in different ways but a lot of it is really outside their control.

Steve Odland: And that raises the specter of real legal challenges because you have to keep material non-public information non-public. There are laws around how that's managed and yet there are people betting on this stuff. It's fraught with risk for companies but it's also subject to a lot of misdeed, yeah?

PJ Tabit: Certainly a lot of risk for companies. And it's a compliance concern around equities markets. There are well-established processes for insider trading, monitoring that, prosecuting violations for that.

Prediction markets, they're less well monitored. It's hard to predict what kinds of things may show up on prediction markets. It's not necessarily just simply the price of the stock of the company, as I said, executive changes. And there have been examples of employees of companies being indicted because they traded on insider information on prediction markets.

So it's a whole new frontier of compliance risk for firms.

Steve Odland: Yeah. And we're going to get into this some more, we'll come back to it. The risk is for the companies themselves but it's also for the individuals who are engaged in it, if that individual is misusing private information. We'll come back to that for our listeners.

One of the things in your paper that you point out is that data are now starting to show that these prediction markets are kind of accurate.

PJ Tabit: Yeah. It's really a surprising thing. One of the studies that I cite in the report is a Federal Reserve staff paper that compared traditional economic data around job numbers, GDP, other economic variables to bets on those things on the predictions markets, and they actually compare quite favorably.

And so I think that's particularly in this era where there are concerns about response rates for surveys, for job numbers for example, and interest from both businesses and policymakers to have more timely and granular data. This is an area where maybe we could actually get an edge by relying on data from prediction markets.

Steve Odland: Yeah, and what's the old saying about the crowd?

PJ Tabit: It's the wisdom of the crowd. Going back as early as Aristotle. For those that are proponents of prediction markets, one of the things they point to is that you're getting data from all kinds of sources, maybe people on the ground. If you're predicting weather events, it's maybe people who live in those areas. You're not relying on quarterly surveys or something like that. So it really gives you access to timely and granular kind of data that you may not have access to otherwise.

Steve Odland: Let's just do a quick example of that, which can bring it to life. If you have 360 million people in the country and you were able to ask them the question, get all of them to engage in a prediction market of whether they're going to buy a car or not this year. In other words, are car sales going to go up or down? But you ask the question in such a way, are they going to buy a car or not?

Essentially, the breadth of the crowd participation has more data points and that are accurate. Yes or no, I'm going to buy a car, I'm not going to buy a car. And therefore, the prediction markets can actually be more accurate because you're aggregating those data or those answers or those predictions from people who are actually engaged in the action, right?

PJ Tabit: That's absolutely right. And an added layer is the idea that they have a financial incentive to give you accurate information.

Steve Odland: Not to lie.

PJ Tabit: Exactly because maybe they bought a contract on the prediction market.

Steve Odland: And they want to win that, right? Unless you get some contrarian who's trying to-

PJ Tabit: Game the system.

Steve Odland: Yeah, game it. But it doesn't really work with prediction markets. You really can't game it when the participation is wide open like that, unless you are able to corner the market on everybody who is participating, like the silver market. But that can't happen either because of the way the regulations are. It has to be this broad participation.

PJ Tabit: That's the idea. But as I mentioned in the report, there are examples of people trading on contracts where they do actually have the ability to influence the outcome.

There was a story recently of a contract that traded on a prediction of the temperature at the Charles de Gaulle Airport in Paris, and somebody who was trading that contract went out and manipulated the thermostat that was measuring the temperature. I think they had a hairdryer and they warmed it up. There are other examples like that, where people are trying to manipulate the actual measure that decides the outcome of the contract.

Steve Odland: I think as long as there are people involved, there are people who are going to try to game it. Yeah. You didn't say in the paper it was 100% accurate. No, of course. You just said there's a higher degree of accuracy sometimes. And it's hard to forecast, as you said, especially when it relates to the future.

PJ Tabit: Sure.

Steve Odland: I mean, Absolutely, it's an old saw. But this is proving to be an interesting tool. It's been used mostly for entertainment until recently, when people are starting to see the benefits here as a true forecasting tool because of the number of data points and everything that we talked about. You've got economists that are working for governments or for organizations, businesses, and typically people listen to the economists and so forth. But are the prediction markets more accurate than the experts?

PJ Tabit: There's another paper that I cite in the report that found exactly that, that in a lot of cases, the prediction markets are actually more accurate. I start the piece by talking about the 2024 election, where, you know, all the experts, the polls showed a very tight race, but when you actually looked at contracts trading on the prediction markets, it didn't show a very close race.

And that's more accurate to what we ended up getting in the end. And I think that was really a breakout moment for prediction markets where people took notice and said, "Hey, maybe there's actually something here."

Steve Odland: It also goes to the state of polling in this country. The private polls are for their own use. The public polls are to try to sometimes manipulate the opinion. If people want to vote for the winner, then if you're predicting, even if the candidate's losing and your poll shows something, all of that gets manipulated. Which to your point, if you're doing in the prediction markets, they're less able to move the whole prediction market. You tend to get a more accurate view of that.

Our listeners are mostly, you know, senior executives in companies. Should executives start to look to the prediction markets as it relates to their own business?

PJ Tabit: I think it certainly depends on their business. It's probably worth looking at but, as with any single data source, you probably don't want to put all your eggs in one basket. Certainly for deeply traded contracts, where there's a lot of action going on, that's probably going to give you a more accurate prediction of what may actually happen.

But in a lot of cases on these markets, there isn't a lot of trading going on. It's maybe a handful of people who are actually involved and that may be less accurate. Executives are really going to have to use a lot of discretion and judgment in determining whether it's useful to them.

Steve Odland: So your point is, if it's more of a general subject that people are more interested in, it probably has more participants and therefore you've got more data points, right? Fewer data points, the less potential significance there is.

And so going with some real discrete thing, what's this margin going to be on this product and this whatever this week, you're not going to have a lot of people betting on it. And so you get a couple people and so who the heck knows, right?

It's the broader subject and therefore the number of data points that really help to drive the level of reliability of the data.

PJ Tabit: Absolutely.

Steve Odland: But isn't this a little like AI, where it's like the frontier, and trying things and experimenting. Is that kind of where we are with prediction markets?

PJ Tabit: I think it's certainly early days. One of the stats that I mentioned in the report, as of late last year, about $24 billion in action was happening on the major prediction markets. Certainly $24 billion is a large number but when you compare to larger equity markets, for example, it's a drop in the bucket.

So they're very new relative to some of the more established data sources out there. May not be as trustworthy but there's emerging research and data that show they can be quite accurate, so it's worth taking notice.

Steve Odland: So people should start playing with them a little bit. No, don't just jump in and rely on it, but start learning, right?

PJ Tabit: Absolutely.

Steve Odland: Yeah. One of the things that you talked about in the paper was how markets could be a better way to hedge risk. Talk about that.

PJ Tabit: Yeah. I said earlier in our conversation the very small example of an ice cream stand that's trading contracts, right? Based on what the temperature may be. If you broaden that out to a company that maybe has exposure to a particular geographic region traditionally, companies may hedge that risk. For example, if they have exposure to the Middle East, maybe trading contracts related to the price of oil. But it's an indirect and maybe imprecise measure of what may be going on in that region because the price of oil moves for all kinds of reasons.

And so for those that are proponents of prediction markets, they say you can trade a contract that's specific to whether a government of a particular country will be overthrown, whether the Strait of Hormuz will be open or closed by a particular date. So it gives you a more precise way of measuring what your risk may be to a particular event.

Steve Odland: Yeah and you talked about tariffs and interest rates as being subject to these prediction markets as well, which is an interesting look.

PJ Tabit: Absolutely. Yeah. And you may have hedged those risks by trading equities or currencies or other kinds of derivatives. But in this case it gives you exposure to the exact decision that you're hoping to hedge risk against.

Steve Odland: You're not saying that the company should throw out their heat maps and their entire enterprise risk management (ERM) process. But you are saying, look, on certain dimensions here, there are some of these subject areas which may be worthy of adding to your risk profile assessment and learn about it, right?

PJ Tabit: Yeah. It's one more tool in the toolbox that a risk officer may have access to.

Steve Odland: All right. We were talking before the break also about the potential for misuse, misdeeds. There are regulations around this, as you said, at the federal level for the prediction markets, state level for betting markets.

When you're thinking about this, there's risk of insider trading, there's risk of market manipulation, and you know, you've given some of those examples. If you're an executive and you're thinking about starting to look at this, where do you think the biggest risks lie in the market itself and what are the safeguards that people should take?

PJ Tabit: I think certainly employee trading on proprietary information is a major risk. Because prediction markets are so new, employees may not think of it as the same kind of behavior as if they were trading stock related to the company on insider information. Some employee education about what is allowed, what is not allowed would certainly be in order.

There are also risks to firms, as I said earlier, about contracts being traded about the company without the company having any knowledge or involvement in that. That could create reputational risk for the company. There are also examples of prediction markets using proprietary data from companies in ways that violate the terms of service of that company. So that's another area where companies will need to be vigilant to make sure their data is being used in ways that they're okay with.

Steve Odland: Let's talk about a couple of those things. If you've got now this whole brave new world, you could have employees out there doing things. Maybe it's nefarious but maybe it's just they think they're having fun and they don't know. Doesn't that say that human capital executives probably need to put some guidelines out there and some rules around this for their employees? Not only for the safety of the company but also for the safety of the individuals themselves, right?

PJ Tabit: Absolutely. Employee education I think is really critical here because in a lot of cases, the employees may not even have knowledge that what they're doing is illegal or improper.

Steve Odland: Yeah. I think most companies have employee handbooks. There should be a little section added here on what the appropriate engagement is. The safest thing is don't bet on anything related to your own company, right? Isn't that the safe harbor?

PJ Tabit: It is. And even at a broader policy level, there are concerns certain around national security. There's an example of a US soldier being indicted because they made trades related to US military operations. Both at a corporate level and a policy level, it's a major concern.

Steve Odland: Yeah, you're right. We're talking mostly private sector but this is a big deal in the public sector.

PJ Tabit: It is, absolutely.

Steve Odland: Yeah. Both at the national level but also state and local levels too, right?

PJ Tabit: Yes.

Steve Odland: You mentioned earlier privacy concerns on that. Just can you just tell us more about that?

PJ Tabit: Sure. One of the pieces that I mentioned in the report is that an analysis found that trades related to military operations had a higher-than-expected hit rate, indicating...

Steve Odland: Hit rate meaning?

PJ Tabit: Successful bets.

Steve Odland: Interesting. That's all top secret. How would anyone know?

PJ Tabit: That's the concern. That is why this particular soldier was indicted by the Department of Justice, because there's evidence that he traded on his insider knowledge of what that military operation would be.

People who are maybe looking for an edge have looked at military contracts trading related to military operations and said, "Hey, I may have some insider knowledge based on the prediction markets about what a country may decide to do."

Steve Odland: Because who else would be in there, participating in that? It's like, will XYZ happen? If you're just a normal citizen, you wouldn't have an interest or knowledge or even awareness that would be an issue. So the assumption is that it's people with knowledge that are in there moving it, and therefore, the reliability of the prediction is higher.

PJ Tabit: And it obviously raises national security concerns. Also market integrity concerns. It's a major issue.

Steve Odland: You talked on this earlier, but if you're betting on succession in companies or you're betting there's succession in public sector, which are elections. The company's earnings announcements or regulatory approval of a drug, all of this stuff has to start with some knowledge. People don't just make up and start guessing some of this stuff. Maybe you can guess on whether the Green Bay Packers are going to do a passing play or a running play.

But this requires some sort of internal stuff and that's why you're saying it's risky to the company and needs to be watched.

PJ Tabit: Certainly, because of the precision of the kinds of bets you could make, the real granularity, the surface area on which a company or department of an agency at the federal level would need to be vigilant has expanded greatly because of the expansion of prediction markets.

Steve Odland: Yeah. And because you have so many regulatory agencies in the country, that's another whole kettle of fish. But they're dealing with a lot of stuff which can have big economic consequences. A pharmaceutical approval is just one example of that but there are hundreds and hundreds of these things sitting there every day that may be approved, or mergers and acquisitions, products. Or the other way, where people are doing enforcement actions of some variety which may in fact hit the economics of a company. All of these things, right?

PJ Tabit: Absolutely. And as I mentioned in the piece, the US Senate, for example, has chosen to prohibit senators and staff from trading in prediction markets. And there are debates about whether there should be wider prohibitions on public servants having access to this kind of information.

Steve Odland: I mentioned some of the regulatory issues. You talk about the regulatory framework being pretty fragmented. What are the key policy areas or questions that still need to be resolved?

PJ Tabit: I think there are a couple. One is just the debate about which things are regulated at a state level and which things are regulated at a federal level.

We talked about gaming, sports betting earlier. You can trade on sporting events on prediction markets but when it happens on a prediction market, it's regulated at the federal level through the Commodity Futures Trading Commission (CFTC). When you do it through a sportsbook, it's regulated at the state level.

Steve Odland: But why is that? Why wouldn't you do everything at the federal level?

PJ Tabit: That is an ongoing policy debate, that states and the federal government are...

Steve Odland: Goes back to federalism.

PJ Tabit: Absolutely. And the rights of the states and there's ongoing litigation about that. So we'll see how that shakes out both at a judicial level and maybe at the legislative level how Congress decides to respond.

Steve Odland: But your point is it is fragmented. We didn't choose it, but it is.

PJ Tabit: Yes.

Steve Odland: And therefore you have to navigate it.

PJ Tabit: Absolutely. We'll see where that goes at the federalism debate. Even at the federal level, there's debate about which things are regulated the CFTC and which things are regulated by the Securities and Exchange Commission (SEC) because you can actually trade contracts related to the price of a stock. And so that seems to fit more squarely into securities regulation, which would go to the SEC compared to the CFTC, and they're working out coordinating their regulatory jurisdictions.

And then a third bucket of major debates in prediction markets is just what I would categorize as things that may seem distasteful or not in the public interest, bets related to possible assassinations or natural disasters.

Steve Odland: Does that happen? People are betting on that?

PJ Tabit: Assassinations are now prohibited through federal legislation. But there's a whole host of things that you can imagine that would not seem like something people should be able to bet on and that's an ongoing debate.

Steve Odland: Yeah, that kind of crime or violent crime because then, if you want to, you know, if it's leaning against it you could, you can impact that.

PJ Tabit: Absolutely. Yes.

Steve Odland: Yeah, which would really be awful. So you have in your paper a whole list of recommendations for businesses. Just talk through some of those.

PJ Tabit: As we said, employee education about what kinds of things are illegal or improper to trade on is one step. I think on the other hand companies should examine whether prediction markets offer an opportunity to hedge risk in a more accurate and more direct way than some of their existing mechanisms.

They also want to think about whether proprietary information about that company is being used in a way that they would not be okay with. If it's being used to decide contracts on a prediction market, that could be an issue. And also reputational risk for a company when they choose to hedge risk in ways that maybe if they are seen to be making money off of something like a natural disaster or a military event, that may reflect poorly on the company even if from a financial or economic perspective it's the rational thing to do.

Steve Odland: Yeah. Lots of things here, lots of great advice. Any final thoughts that we didn't cover?

PJ Tabit: As I said, if you look back just a couple years ago, there was about $1 billion in volume on these platforms. It's $24 billion as of late last year. Who knows what it'll be in a year or two years from now.

So as these markets grow, as they become more complex, maybe as the regulatory landscape becomes a little bit more clear, companies will really keep an eye on it and see how it impacts their business.

Steve Odland: Great paper, PJ Tabit. Thanks for being with us today.

PJ Tabit: Thank you, Steve.