Welcome to Futures Industry
Michael Gorham
Published 1/1/2004

This is the story of a new family of markets: event markets. Event markets generally allow the participant to profit from the occurrence of a specific event. These markets go by a variety of names: prediction markets, decision markets, proposition markets, opinion markets, idea markets, claim markets, information aggregation markets, unconventional markets, and non-traditional markets.

Defining an Event

            What is an event? It is something that happens or doesn't happen. It is a binary, digital, zero-one sort of thing. Either Howard Dean will be the Democratic nominee for president in 2004 or he won't. Either a Palestinian state will be established by 2005 or it won't. Either American Airlines will go into Chapter 11 by next summer or it won't. And there are currently markets that allow you to buy options on each of these occurrences or events. If the event occurs, you get a fixed pay-off. If the event doesn't transpire, neither does your pay-off.

            Of course much of the world is not binary. It’s continuous. Rainfall, crop yields, temperature, and prices are all measures that can take on lots of different values. But the event market makers have an easy solution for this. They simply carve up these continuous things into a small number of discrete outcomes. The event becomes whether the price of soybeans, for example, exceeds $4 a bushel by July 1 or whether the Dow closes under 10,000 at the end of the year. One marketplace, for example, has taken the Nymex February crude oil closing price on a specific date and chopped it up into eight possible outcomes: below $16, above $40 or within one of six $4 ranges in between, i.e. 16 to 20, 20 to 24, etc. A market participant can then take a position on any one of these outcomes.

The Early Pioneer

            The Iowa Electronic Markets (IEM) was one of the earliest ventures, starting in 1988 as a research project of the University of Iowa Tippie College of Business. The Iowa project is not-for-profit in every sense of the word. While there is a one-time fee of $5 to set up an account, every penny lost by one trader is gained by another trader. There are no transaction fees. Administration costs are covered by grants and donated time. Traders can open an account with as little as $5 up to a maximum of $500 and the typical trade is for less than $50. While the political markets are open to anyone worldwide, the other markets are restricted to academics. It is truly a teaching and research market.

            It’s worth explaining IEM’s products because much of the event industry has used them as models. IEM’s election contracts come in two basic forms—“winner takes all” and “vote share.” Let’s look at the market for the Democratic National Convention, which uses the “winner takes all” contracts. These contracts are really options—binary, European call options to be specific. They are binary or digital because there are only two outcomes for each one. If you own the Lieberman option, you will receive $1 if Lieberman wins the nomination and zero otherwise. They are European, because you can’t exercise the options until expiration, i.e. right after the convention.

            Each of the major candidates (Dean, Gephardt, Kerry, Clark, and Lieberman) and one non-candidate (Hillary Clinton) have their own options, while all other candidates (Kucinich, Edwards, Sharpton, Moseley-Braun, and anyone else who might declare) are relegated to a single “rest-of-field” option. These seven options represent all possible outcomes—either one of the major candidates will win or someone else (rest of field) will win. If you held the full bundle until expiration, at the end of the convention, one of the options would pay $1 and all of the rest would pay nothing. At any time, IEM will sell you a bundle of all seven options for $1 (and IEM always stands ready to buy back the bundle for $1 as well). If you believe Clark was the most likely candidate, you could buy the seven-option bundle for $1 and then sell off everyone but Clark, or you could just buy Clark directly in the open market. And you would do whichever gave you the cheapest access to Clark. As of mid-December, a Clark option was going for 8.6 cents, so if you bought him at that price and he won the nomination, you’d be able to sell the contract back to IEM for $1 and earn 91.4 cents per contract. Incidentally, the market was pricing a Dean option (which pays $1 if Dean wins) at about 75 cents. Everyone else was going for a dime or less.

            In over a decade of experience, what have the Iowa academics learned from their markets? This past summer a group of Tippie researchers compared the forecasts inherent in presidential election markets to the forecasting ability of polls. The authors note that the polls and the markets are really asking two very different questions. The poll asks for whom you would vote. The election market essentially asks for whom you think the country will vote. For a poll to provide good predictions, the sample polled should be representative of the kinds of people who will vote in the election. For a market to provide good predictions, you just need good, informed traders. (Interested readers may want to read their paper, “Accuracy and Forecast Standard Error of Prediction Markets” by Joyce Berg, Forrest Nelson, and Thomas Rietz of Tippie College of Business Administration at the University of Iowa.)

            Even though the Iowa Electronic Markets admit their traders are not at all representative of the voting population (they are too white, too male, and make too much money), their study found that IEM markets do a better job of predicting the actual election results than the polls do. The researchers found that the markets will stay considerably above or below the polls for considerable periods of time and the markets are generally closer to the actual outcomes, even months in advance. Also, the markets don’t show the convention bounce generally found in the polls—i.e. the polls generally rise for a party during its national convention and then settle back down after the convention is over. This is the result of the steely-eyed trader predicting what will actually happen, compared to regular mortals who report to the pollster their enthusiasm for a particular candidate after just having watched his very moving acceptance speech at the convention.

The Infamous Event Market

            There is another event market that for one week last fall was the best-known market in America. In a single week it went from obscurity to fame and was then dissolved even before it began trading. It was called the Policy Analysis Market, but you will remember it better as the “terrorist market,” because some of its contracts involved whether certain political leaders would be assassinated, or certain governments overthrown. It was a partnership of Net Exchange (of California), the Economist Intelligence Unit (of London), and our own taxpayer-supported Defense Advanced Research Projects Agency (more affectionately known as DARPA). DARPA was really just trying to apply the success of the Iowa experiment to the problem of predicting important economic and political events in the Middle East.

            PAM would have listed contracts on the familiar March quarterly cycle, up to one year out, on various indices of civil stability and economic performance of, and U.S. military or other involvement in, various Middle Eastern countries. One specific contract was on whether the U.S. would recognize a Palestinian state by the end of the first quarter of 2005. The market would have traded 24 hours a day, seven days a week, same as Iowa. Despite its defeat last fall, the Policy Analysis Market has announced that it will resurrect and recreate itself as a kinder, gentler, assassination-free market in March 2004, with no involvement of government money or agencies.

            Even though the Policy Analysis Market was shut down because it would have been a taxpayer-supported market listing contracts considered inappropriate, distasteful, and offensive by U.S. legislators, Americans are currently taking positions in very similar markets based offshore. TradeSports, for example, listed contracts on whether Saddam Hussein and Osama bin Laden would be captured or “neutralized” on or before a specific date and about 14,000 of these contracts were traded. In mid-December, the Irish market was showing a 20 percent chance of something happening to Saddam, but only an 11 percent chance of something happening to Osama. The market paid off on Saddam following his capture.

            TradeSports is one of at least 14 offshore markets offering event contracts. The market lists about 1,300 different contracts, with sports obviously being the largest category. Despite its Dublin home, TradeSports has a distinct American flavor. Contracts are dollar-denominated and it appears that most of the sporting events covered take place in the U.S. A recent glance at its 10 most active markets revealed five options for the U.S. presidential election, three options on the upcoming Super Bowl, and two on whether the Dow would close over 10,000 or under 7,000 on New Year’s Eve. TradeSports, like many other offshore markets, uses real money and seems designed primarily to be a gambling enterprise. In fact, a third of the offshore sites use the term “bet” somewhere in their names, as in Betfair, InventaBet, BETonSPORTS, BetOnGames, and BETDAQ. None of the U.S. markets, except for Iowa, use real money.

            There is a rich variety of event markets available. There are the natural events, things that the earth does, usually to us. There are economic, financial and business events like inflation, unemployment and FOMC actions. There are political events ranging from elections, to FDA approval of drugs, to armed conflict between states. There is a large group of entertainment events, which includes sporting event outcomes and celebrity behavior, such as will Ben and J. Lo marry and divorce by the end of 2004. And finally, there is a huge set of science event markets covering when theorems will be proven and humans will be cloned. My personal favorite is the market in whether the universe will eventually collapse. The marketplace notes that the universe doesn’t have to actually collapse for the trader to collect, it just has to show signs that it has stopped expanding.

Commercial Success and Regulatory Oversight

            Although these markets can amuse, entertain, and, as we learned from the Iowa markets, generate useful information and predictions, the role that they will play in the U.S. marketplace is still relatively murky. These markets are all Internet-based, non-intermediated, retail markets, offering very small contract sizes. This is a very different model from the traditional U.S. futures exchanges, which are highly intermediated and traditionally offer contract sizes appropriate to commercial users. These markets also raise some very interesting business and regulatory questions. Will they be economically viable for the organizers with contract sizes of $1 to $10?

            While some of the offshore markets appear to be quite active and possibly profitable, the U.S. markets are still operating largely with play money. Part of the reason for the lack of vibrant commercial growth in the U.S. could be regulatory uncertainty. While the real-money Iowa market received a no-action letter from CFTC staff in 1992, all the other U.S. markets have avoided using real money, most likely to keep from running afoul of state gambling laws. The CFTC has not yet made any public statement or taken any action to indicate which, if any, of these markets fall under its jurisdiction. Even the no-action letter given to the Iowa market did not clearly assert jurisdiction over these markets. In fact, CFTC staff, in response to a subsequent request for no-action relief by an individual seeking to operate a similar market, stated that the agency had no reason to assert jurisdiction over the market. One way of thinking about this is whether these markets are trading futures or options that serve a significant economic purpose and should be regulated by the CFTC. Or are they merely vehicles for gambling? Or are they something in between?

            A CFTC statement on the matter could have substantial regulatory repercussions. If such markets did involve futures and options contracts, then they would be subject to exclusive CFTC regulation and would have a degree of immunity from other federal laws and state gambling and gaming laws. However, they would also face the non-trivial cost of applying to become designated contract markets. On the other hand, if event markets were outside of the CFTC's jurisdiction, then they would have to deal with 50 different state regulatory schemes. CFTC staff has been approached by several entities interested in becoming designated contract markets and listing event-type contracts, so resolving this issue has become a high priority. 

None of the views expressed in this article in any way reflect those of the Commodity Futures Trading Commission or the Commission staff and should not be interpreted as in any way foreshadowing any future Commission action or lack thereof.




Michael Gorham is director of the Division of Market Oversight at the Commodity Futures Trading Commission.
Change text size
Print this article
Email a link