Before Kalshi and Polymarket there was the Iowa Electronic Markets

Hey guys, we have an episode for you, we always do, but first, we have a question. A lot has happened in the past five or so years, and we want to know how you all are holding up. Did you make any big decisions that are maybe backfiring now? Did you move to Montana during the pandemic, and now your work wants you back in the office in Seattle? What's your plan? Or is your money in the stock market right now, and you are just like, "Watch it, grow, grow, grow, grow! What are you spending it on?" Yeah, we want to check in with our whole planet money community. Does it feel like things cost a lot more these days? Groceries going on dates, doing something fun for your family, have you found any great life hacks to help you get by, or maybe even thrive in this economy? Yeah, we want to hear from you, and maybe we'll call you up for a future show about how things feel right now. Send us an email about how the economy is affecting you and your life at Planet Money, at npr.org, and maybe we'll call you up to chat. That email, [email protected]. Alright, Erica, you have a show to start. This is Planet Money from NPR. Recently I heard a story that changed how I think about prediction markets. One moment in the story in particular, it was part of the history of where prediction markets came from, which as you may have heard, goes back hundreds of years in some form, but also in another form, prediction markets, as we know them today were cooked up by economists trying to test some theories on markets. And in this story of the evolution of prediction markets, there's a moment when an economic historian studying the long ago early roots of prediction markets realizes there's already a well-functioning design of a prediction market at the racetrack. Now of course, many of you will say, yeah, obviously gambling is gambling, of course, but the part I didn't quite know is that prediction markets for elections were popular and robust until they mysteriously faded away. And one theory is, of course, racing took over, you know, because instead of an election every year, whatever, you can have a dozen races a week. This all comes from this fascinating history of the economic origins of prediction markets. The economists who cooked up the proto-prediction markets, the political machines that basically had to financially bet on their candidates hundreds of years ago and even farther back than that. So today we're going to share an excerpt of this story for you. It comes from our friends at ThruLine, NPR's excellent history podcast. They take one big question every episode and bring you the history and context behind it beautifully sound designed and deeply researched. Today we pass it off to host Rund Abdul-Futta and ThruLine for a history of prediction markets. Part one, three professors walk into a bar. On a spring afternoon in 1988, Robert Forsythe and two of his colleagues were sitting around a table at a local sports bar called the airliner in downtown Iowa City. Unlike the often did, the three economics professors got to talking about the news. And I would say the conversation was helped because it was a three-year lunch, which led to this creativity, I think, some days. Three beers each or each of them. Each, each. It was an election year. After Super Tuesday, Michael Ducakis was the presumptive frontrunner for the Democratic nomination. That is, until a huge surprise upset. The polls had predicted a Ducakis victory, but it didn't happen. It was the day after the Michigan caucus, which you may recall, Jesse Jackson won. Jesse Jackson. He scored a stunning win yesterday in Michigan's Democratic caucuses, defeating Michael Ducakis by a margin of nearly two to one in the popular vote. And that was a big surprise. The polls missed it all together. Knowing in, political polls had shown Jackson and Ducakis running neck and neck. But what happened was a blowout. Jackson won the Michigan caucus with 53 percent of the vote. And Ducakis only got 29 percent. And that's where it came from. He said, "Well, gee, you know, as economists, what would we do if we were going to try to predict the outcome of something?" And what's natural for a bunch of economists has said, "Well, let's run a market on it." Like a stock or commodities market. So Robert and his colleagues got together a couple hundred people, students and faculty at the University of Iowa. They set up an online interface where those people could buy stock in political candidates. And just like on Wall Street, these political stocks could be traded. So for example, when George H.W. Bush and Michael Ducakis ultimately went up against each other in the general election. If you didn't like Ducakis' chances, you could sell your shares in him. To be overarching idea here, the market might show what people were actually thinking, even better than polls could. The wisdom of the crowd. And lo and behold, at midnight, the night before election day, we ended up predicting the outcome of the popular vote within two tens of one percent. That result outperformed major polls, including Gallup, Paris and CBS New York Times. It was so effective that Robert and his colleagues thought, "We'd like to do this again in next election, this time with a bigger sample size." So we went through the commodity futures trading commission and asked them whether they would give us permission to operate nationwide. And just so I understand, what was the law that you needed to get around? Well, anytime you have people exchanging real money, or the outcome of something, the question is, is this gambling? Yeah, even though there's some form of regulation that's going to oversee you. Back then, and now, the body that oversees that regulation in the US is called the commodity futures trading commission, or the CFTC. It's been in charge of commodity futures like markets that set the future price of grain since the 1970s. They issued us, but they called it no action letter. That letter was basically like a permission slip. As long as the Iowa electronic markets followed a few set rules, the CFTC would allow it to do its thing. And the rules were, we stayed small. We couldn't accept the cons over $500, and we didn't engage in paid advertising. So they could only run markets on presidential elections. Sports were off the table. And the Iowa electronic markets did not make a profit. Between 1988 and 2004, the Iowa electronic markets grew from a one-off experiment to one of the most reliable predictors of American presidential elections. During that period, its predictions beat traditional polls 74% of the time. And as the market continued to grow, so did the public's interest in it. I'm Bob Edwards, and this is NPR's Morning Edition. If you think you know who the next president of the United States will be, there's a place in Iowa where you can back up your convictions with a buck or two. At what point did you realize that people outside of the University of Iowa, outside of your orbit, were starting to pay attention to what you all were doing? Well, we started to get a lot of national publicity. This was different enough that before we knew it, the Wall Street Journal and the financial times of London and NPR. NPR and a lot of the television media started picking us up because it was just so different, right? I mean, who would have ever made a market on an election? Oh, so they thought this was the first time something like this was happening. That's right. That's right. It was, of course, not the first election betting market, not by a mile. After the break, we go back farther and meet an economic historian trying to find the origins of election betting more through the line after the break. Let's call this part two, the racetrack. We pick it up with run from through line and Coleman Strumpth. I am an economics professor at Wake Forest University. They've been teaching, of course, on prediction markets for almost 20 years. Coleman may be an economist, but his driving interest is in human behavior. On top of prediction markets, he's written papers about illegal file sharing, tax evasion, and the economics of addiction. And he says, if you want to understand where prediction markets came from, just head down to your local horse track. My uncle used to take me to the racetrack and have me bet on horses. I guess I was already a social scientist in training at 10 years old because, first of all, yes, back many, many years ago, you could be 10 years old and go to a racetrack and batting. I couldn't even reach up to the counter, and nobody seemed to care very much. But it was as interesting as much to me to watch what everybody else was doing. And a lot of things that I see when I look at markets today, I could sort of first see then. Watching people read the odds and make bets taught Coleman something essential about gambling. Winning makes people feel good. And the longer the odds, the better the winners feel. I remember when I was younger, and I was watching people do it, if somebody would bet on a long shot horse, and that horse would actually win. The person would not only win a lot of money, but they would. So we're going around to all their friends saying how smart they were because they managed to figure out this very, very unlikely thing from happening. And I sort of realize that part of the sort of psychology of these markets is people like to be smart. People like to use this as a way of sort of showing off their smarts of how they figured these things out. I should say here, there's been a lot of back and forth over the years about whether or not prediction markets are really just betting dressed up as something fancier. Or if you're looking at legal definitions, whether they might count as gambling, we're not here to litigate that. Coleman's point is as long as people have been putting money on things, whether it's on the stock market, in insurance, or at the race track, social scientists and economists have been able to learn something about us from those bets. So flash forward to the 2000 presidential election, George W. Bush versus Al Gore, one of the closest presidential races in a century. Coleman was doing research about election forecasting at the University of North Carolina. And he was keeping a close eye on the Iowa electronic market. I was writing a paper about the Iowa electronic market when my then colleague is at the University of North Carolina at the time Paul Rodi came by my office, asked me what I was doing, and I said I'm looking at the first political prediction market we have. And Paul, who is an economic historian, said nope, that's totally not correct. He was MIT and I'm Stanford and a lot of times creative frictions happen from people with different perspectives. That's Paul Rodi. He's an economic historian at the University of Michigan, but back in the early 2000s, he worked down the hall from Coleman at the University of North Carolina. I happened to be reading some microfilm at the library and came across news stories in October, November, 1924 about election market. So I go up to Coleman's front to his office and say, hey, did you know this happened? And he wasn't aware that they happened, so it was an interesting surprise for both of us. Paul and Coleman started digging into the archives in earnest. And it turns out pretty much as long as there have been elections, people have been betting on them. Back to the 16th century, there's betting on who would become the Pope. Elections in city states and Venice in Genoa. And in the US, you can find markets going back to George Washington. How long the Stamp Act would be in place before the American Revolution? Paul and Coleman say, for centuries, these markets stayed relatively small. There wasn't a ton of money on the line. Sometimes the bets didn't involve money at all. If the candidate I support doesn't win, I'm going to cut my beard. Or I'm going to walk from New York to Boston. Or you have to eat a crow if I'm right. You have to push me in a wheelbarrow down Main Street. But around the turn of the 20th century, election markets in the United States really started to gain steam. The main place that people would trade on, elections was something called the Kerb Exchange, which literally has the name suggests was the Kerb outside the New York Stock Exchange. They'd be sitting on the Kerb, and they'd be trading stocks and signaling the people in the offices about them. Stocks to be clear in political candidates, people running for office. Reporters for the Wall Street Journal of New York Times could go down to this pit and not only tell us what the overall price was, but the names of the people trading. And if you look, the people were trading these were like the elite of the city. These were people from Tammany Hall, these were bankers, Wall Street folks, people on hotels. People would be doing this very publicly. So if you were like the head of the Republican Party or you're the head of the New York Tammany Hall, you'd be expected to go to the batting commissioner and be willing to offer money for your candidate. So be part of the publicity about you're standing behind this person, you think they have a good shot of winning. Were you expected to bat on your own candidate or could you bat against your candidate if you thought your candidate wasn't going to win? So we knew that like William Jennings Bryant, the populace from Nebraska, was not popular with the workers in New York. The Democratic machine in New York has to like bet for him, but then the stories are behind the scenes. They're placing bets the other way to so that they cancel out because they don't want to lose money. Like publicly, they'll say, we love him. We endorse him, but they don't want to lose their money in the market. Yeah. So they're going to be like hedging or going against their candidate. All of this was happening in a sort of legal gray area, making a friendly bet on the outcome of an election wasn't against the law. But if you did that, you weren't supposed to vote. But I don't think anybody ever said we're turning away from the polls because we know that you bet on elections. Were these prediction markets from as far as you can tell, good at predicting the outcomes of elections and whatever else was being bet on? So this was like the second thing that I was wrong about. My sense was, okay, the markets will do as well as they can, but there's not much information to be had. These markets won't really tell us anything. Well, that I was definitely wrong on. They were always right, basically. They would tell us who would win. They could give you a sense of whether they're the landslide. It was really pretty remarkable. By the 1980s, the Iowa electronics market would put that theory into practice. But why did it take them so long? That's after the break. And now, part three, the other golden age of prediction markets. Back to run from through line with economist Coleman Strumpf. You kind of pointed out that late 19th century to world or to era as this kind of golden age almost of prediction markets. Why did they fall off after that for a while? It's a combination of factors. One of the things that both drove the popularity of the markets as well as the discussion of the markets was the press coverage. Through this period, all the newspapers covered it, but they were never very comfortable doing this. Then in the 1930s, that's when the scientific polls galloped and some other folks came around. And so the polls were doing something that was kind of the same thing. And newspapers were much more comfortable writing about polls than they were with markets. The other thing, in some sense, has to do with interest to the people trading. So if you're somebody who really likes this for whatever reason you like to be known as a good forecast or you like to make money, you like the adrenaline rush, any of these factors, the problem with elections, or at least I'll just say US elections is we don't have enough of them. This is around the time when thoroughbred racing started to really take off. And so instead of a couple events a year, you could have 12 races a night. And I think for people who were interested in that, I think the horse track was more attractive as a thing to do. And I think some of the interest among traders kind of dissipated at that time. As far as Coleman and Paul can tell, presidential election markets went dark sometime in the 1940s. And then for the next four decades, radio silence. I have never found one person who could have been around during that period, or even books that talked about elections during that period, that mentioned these markets. This was not some small, tiny thing. I don't know how they managed to slip through the cracks of what's known about that time, but they seemingly did. Until 1988, when three professors in Iowa went for a three-beer lunch, and the Iowa electronic markets catapulted prediction markets back into the public consciousness as an alternative to political polls. Mr. Robert Forsyth again, one of the founders of the Iowa electronic markets. Once we went nationally, two things happened. We would get phone calls from traders around the country, got to realize that we had a $500 limit, and $22, to send us a check for several hundred thousand dollars to invest in the market. And we'd have to say, well gee, that'd be great, but we only can take $500 of your money. But we've got to realize that many of these people weren't speculators. These are people that were involved, and we're trying to hedge some political risks that would affect their company or their operations. Say you're worried a certain candidate might win and pass laws that will hurt your business. You can hedge your bets. Put some money on the person you do not want to win. If they win, you cover your losses. So that was the one kind of phone call, and then we would occasionally get a phone call saying, this is great, but you know, you have to stay small, we stay in the United States once you come with us and come over to the Cayman Islands and you operate it there, and you can run the car restriction. And maybe we should have done that. I don't know. At that point in time, we were a bunch of academics who were mainly concerned with our teaching and our research, and so we turn those opportunities down. Looking back on it now, why do you say, well, maybe we should have done that. Well, you know, if there are days, and I'm sort of jealous of a call sheet in a polymarket. I mean, they've really taken the same idea. I mean, they're running the basic same prediction markets we did when I'm on a much bigger scale. In fact, they basically use the same rules for trading and issue contracts that we used back to the day. in '88. But they started with the expanded basket, until you can just about trade on anything there. The Iowa electronic market established something important. It demonstrated to modern economists that prediction markets worked, and that sparked a chain reaction. It says economist Coleman Strump. Once we see, oh look, these markets kind of work in the modern period, could we use these markets to forecast other things? So could we look at other elections? Could we look at current events? And I think it kind of showed to the world that these things could work, but maybe it was only in this one specific domain. And we didn't yet know, would that same kind of forecasting savvy translate into other situations? So that was for other markets to to help point out. That was economist Coleman Strump speaking with Rund Abdelfatta and through line and PR's history podcast. Each episode they take today's big questions and find the context in history. Deeply research stories beautifully sound designed. This is just a part of their episode on prediction markets. They have much more on terrorism future markets and military applications. Check it out in their podcast feed. We will put a link in our show notes. Thanks as always to our NPR+ supporters. Reminder that we have a live virtual book tour event for you this Thursday, June 25th at 3 p.m. Eastern. Check out the start of our most recent bonus episode for details on joining. We're hoping to do more virtual events like this. And if you are a supporter and want to become one, yes please go to plus.npr.org. This episode was produced for Planet Money by Sam Yellowhorse Kessler and edited by Alex Goldmark. America Bears and I'll let through line and run give you the credits in their classic through line style. I'm Rund Abdelfatta. Through line was created by me and Romtean Adabliwi. This episode was produced by me and Sarah Wyman. Julia Redpath. Casey Meiner. Cristina Kim. Devin Caudillama. Kiana Mojave. Irene Muguchi. Julie Kayne. Thank you to David Bieri. Brett Neely. Johannes Dergi. Dylan Kurtz. Rebecca Farar. Yolanda Sengweni. And Tommy Evans. And shout out to Michael Strong. That checking for this episode was done by Kevin Vocal. This episode was mixed by Maggie Luthar. Music for this episode was composed by Romtean and his band Drop Electric, which includes. Navid Marvi. Show Fujiwara. Anja Mizani. This is NPR. Thanks for listening.