UCLA-Anderson's Ely Dahan on Prediction Markets, Obama and McCain

Last Updated Oct 29, 2008 1:39 PM EDT

ely_dahan_photo.JPG Ely Dahan is an assistant professor of marketing at UCLA-Anderson School of Management and is currently serving as a visiting professor at MIT's Sloan School. Much of Dahan's research focuses on the power of predictive markets to harness the wisdom of crowds. Indeed, James Surowiecki discusses Dahan's work in his seminal book, "The Wisdom of Crowds." I asked Dahan how markets could serve as a crystal ball, especially in this election season.

Dann: How can you figure out how fickle consumers -- and voters for that matter -- will really behave?

Dahan: Imagine we can pull away from just using markets for stock pricing and instead use them to predict presidential elections, future box office success of movies, the results of a sports match or -- in the case of my work -- which of 10 product ideas people will like the best. With predictive markets as a marketing tool, you can actually fairly quickly test the predictions against the actual results.

Dann: Well, people would be very interested in what these predictive markets are saying about our presidential election--

Dahan: If you look at the Iowa Electronic Markets, the leading market for presidential and really all elections in the U.S., the "stock" where you get one dollar if Obama wins is actually worth a lot more than the "stock" where you get one dollar if McCain wins. This is the main market in the country where there's real money on the table -â€" you do have to pay and you do collect real money in the end. They got an exemption from the SEC early on. All of the other markets that predict things like movie box office receipts or whether a certain company's software will ship on time are not permitted to use real money. IEM has an amazingly good track record over many, many years, outperforming the most prominent pollsters. IEM is much more accurate.

The key thing to note is not everyone would agree with these numbers. This is the consensus number. This is where two people are willing to exchange a share of stock together at a certain price, what we call the "marginal traders." These are not the "extremists;" a very liberal democrat would probably put a much higher probability on the likelihood of Obama to win and the right-winger would go the opposite direction. The people who trade are really those who can see both sides of the coin. These people are very influential in real markets; they really determine the price of a stock.

Dann: So, is the marginal trader like the "swing voter" the campaigns try to win over?

Dahan: That's like a swing voter-- but not exactly. Remember, you're not betting on "who I want to win" or even "who I think will win." This is "who I think everybody else will say will win." [He clicks on the IEM site] Before we looked at this page, I thought Obama would be 65 and McCain 35. These people are saying it's more like 80-20-- almost a lock for Obama.

[This interview took place on 16 October, the day after the final presidential debate. Obama's stock rose from 84.5 cents the day before the debate to 86.8 cents the day after. Obama's stock started the month of October priced at 69 cents and McCain's at 31 cents.]

There are high volumes on the days with new information [the day of the debate represented the highest trading day of the month for Obama's stock]. People get nervous or get excited. If things are boring and stable, people are not going to trade as much. But the prices are what's important in this market; looking at the volumes isn't going to tell you nearly as much.

Dann: How difficult is it to establish a predictive market like this?

A market this elaborate might seem like it would be difficult to implement, but they are trivially easy to set up, but I am doing them in every single class I teach. I apply the same methodologies to judging 15 student projects, and it works amazingly well. You don't need lots of trading; you just need a survey instrument and a reward system.

Next week, we'll discuss just how Dahan is using predictive markets as a tool for teaching product development and how he would overall the teaching approach at top business schools so U.S. MBAs can compete in the new global economy.
To see the latest stock prices of America's presidential candidates, please visit the Iowa Electronic Markets' 2008 Presidential Election Vote Share Market page.
Editor's Note: Predictify is another Web-based prediction tool. The Predictify widget below lets you enter the prediction market for the 2008 presidential election, and create your own on-the-fly prediction markets. Who will win? Obama or McCain?

  • Jeremy Dann

    Jeremy Dann is a Lecturer in Marketing at UCLA's Anderson School of Management and an innovation consultant and writer. He has been a contributor to several business and technology publications and is the founding editor of "Strategy & Innovation."