Market watchers have reason to be disappointed by last week’s news that Intrade, a popular online betting exchange, is to close the accounts of all its US-based customers. The announcement follows a decision by the Commodity Futures Trading Commission to file a complaint against the company which runs the site, alleging that it has illegally been operating an unregulated options market. Non-US citizens can however continue to use the site, which offers punters the chance to place bets on a variety of political and economic outcomes.
Prediction markets such as Intrade have become increasingly popular in recent years, and not just with gamblers tempted to take a view on the outcome of such events as the Presidential election, the Eurozone crisis or the situation in Syria. The sites are also regularly used by investors and analysts who value it not for the opportunity it provides to place bets but as a way to keep track on the implied probabilities of potentially important events happening.
On Friday last week, for example, Intrade’s prices implied a better-than-evens (59%) chance that President Assad of Syria will have lost power by the end of June 2013, and a 43% chance that either the United States or Israel will launch an overt air strike against Iran before the end of next year. The odds on any country leaving the Eurozone before the end of next year have meanwhile fallen sharply from over 50% to around 34%.
Intrade has had a run-in with the CFTC before, and it does appear that the regulators have a decent case in arguing that the site, which is based in Ireland, has been operating illegally as an unregulated derivatives exchange according to US law. Its main beef appears to be that in addition to its various political and non-financial bets, Intrade also offers prices on financial markets such as gold and the Dow Jones index, which does take it clearly into regulation territory. The company says it has now closed all its US clients accounts but intends to come back shortly with a new predictions market service that it is confident can be legally used by both US and non-US citizens.
Whether the CFTC is right to take action in this particular case is a moot point. Intrade is a small company whose turnover, its defenders say, barely registers statistically when measured against the annual trading that takes place on the major regulated futures, options and commodity exchanges. Should the regulator be taking such a heavy-handed approach to a business that is clearly providing a service that its 100,000 or so clients find valuable?
The reason for thinking not stems from the fact that prediction markets such as the one offered by Intrade are increasingly recognised as a valuable source of market information that may serve a wider purpose than merely indulging the habits of gamblers. In 2008 a group of prominent economists in the United States, including Kenneth Arrow and Robert Shiller, wrote an article in Science magazine in which they called for prediction markets to be encouraged on public interest grounds, rather than hounded out of business by over-zealous regulators.
Their argument was that prediction markets help “to produce forecasts of event outcomes with a lower prediction error than conventional forecasting methods”, and as such as are widely used today by governments, businesses and others to improve decision-making. It urged the CFTC to explore ways of creating “safe harbour” prediction markets that could fulfil this important social purpose without fear of regulatory clampdown. The Iowa Electronic Market, a prediction market that is run by a group of academics as a research experiment, and only allows small bets to be placed on a narrow range of outcomes, was offered as a precedent. It continues to operate with a guarantee from the CFTC that it will not be pursued by the regulators.
The evidence that prediction markets do outperform traditional forecasting methods is clear and strikingly robust. It has been most prominently observed in the political arena, where the betting exchanges have repeatedly demonstrated that their implied prices give a better guide to the outcome of elections than any number of opinion polls. This was borne out again in the latest US Presidential election, when the odds on an Obama victory on Intrade never fell below 50%, despite the media excitement that greeted the late Romney surge following the first televised debate.
Nate Silver, the nerdy pundit who shot to fame by correctly calling the outcome of the 2008 Presidential and Congressional elections in almost every detail, using the same careful statistical techniques as those which have transformed the analysis of baseball, says that even he expects to struggle to do better than Intrade in calling the result of elections. (He lays out the evidence for the superiority of market-based forecasts in convincing detail in his recently published book The Signal and The Noise).
But he also introduces one important caveat. For prediction markets to work, they have to be big enough and deep enough (a) to be free from manipulation and (b) for their odds to be statistically robust. Only that way can their implied odds be usefully relied on for a wider social purpose. A market on the US Presidential election that excludes millions of US voters from voting is self-evidently less reliable than one which does not. And that is why the hounding of Intrade, which has proved itself to be the most commercially successful of the prediction markets, is potentially a cause for regret.
Jonathan Davis 