October 2003
Volume8/No7

Geopolitical futures

The politics of betting

Using markets to forecast political events may not be as strange an idea as it seemed in July, when a terrorism futures scheme collapsed. But there is still scepticism as to whether such an approach would be ethical or effective. By Maria Kielmas

Ephraim Clarke is no longer a lonely man. Professor of finance at Middlesex University Business School in the UK and head of risk consultancy Countrymetrics, Clarke has spent years working on a methodology to develop derivatives securities that could be used to manage country risk.

But in late July the Pentagon’s Defense Advanced Research Projects Agency (Darpa) launched a pilot project called the Policy Analysis Market (Pam), which would forecast political events through a speculative financial market (see box, page 16). Although the US quickly abandoned Pam, amid strong opposition on moral grounds, many academic economists emerged to defend the idea of political betting.

Clarke says: “I thought I was the only guy working on it.” His approach is to create a country risk implied volatility swap and use it to hedge country risk. Clarke does not assume the political risk to any foreign direct investment to be a single loss-making event, but multiple sources of mutually dependent events. For example, he says, a tax increase may cause a strike that harms the investor company and in turn causes the government to grant wage increases that also hurt the company.

Underlying security
But to create derivative products, you need an underlying security that gives the option its value, Clarke says. This factor also bothers Hal Varian, a professor at the School of Information Management Systems, University of California at Berkeley. “Can you have a market in a thing like political stability where there isn’t an underlying commodity?” he asks.

Such a market might be possible if it were to be subsidised by government, says Varian, but it is not viable on a stand-alone basis. “You have got to have people interested and informed enough to participate as traders,” he adds.

The University of Middlesex’s Clarke suggests in a recent paper* that the underlying asset could be an indicator similar to the degree-days used in weather derivatives. The indicator should reflect the volatility of the country’s economic value compared with the rest of the world.

Clarke has called this indicator a country’s “international market value” and says it is related to the country’s cashflow discounted back to the present. He models expropriation risk – the cost to the investor when a host government expropriates his asset – as an American-style call option whose value is equivalent to the insurance compensation for the expropriation.

Clarke also uses an American-style call option to model sovereign debt default risk. He assumes a government will choose to default when the benefits of defaulting are big enough to offset the cost of doing so. So the higher the value of the option to default, the less willing the government is to continue servicing its debt. The arrival process – that is, the occurrence frequency – of these political events is modelled as a learning, Poisson process, which indicates that each political event can have a different arrival rate, its value can change through time and the events can be inter-dependent.

Using these principles, Clarke forecasts that unless the Brazilian economy receives a large amount of new money, we can expect a major rescheduling of foreign debt, a default or a bailout in the country in the next year. He also believes his model is suitable for creating a derivative on the risk of a French public-sector strike. France has a very high level of political risk associated with public-sector employees. “They can close down the whole country, and that kills a lot of small companies,” says Clarke.

Quantitative modelling
Meanwhile, the quantitative modelling of terrorism risk has gathered pace since the September 11 World Trade Center attacks. Consultants such as California-based Risk Management Solutions (RMS), Boston-based Air-Worldwide and Guy Carpenter, a division of New York insurance broker Marsh Inc, all offer models, which use game theory as their main tool. Game theory attempts to combine different disciplines, mathematics, economics and social sciences to study human behaviour.

Gordon Woo, a mathematician at RMS in London, who has developed mathematical models for natural catastrophes and is now pioneering terrorism modelling, says game theory is superior to any market or futures mechanism. In RMS’s case, the methodology receives substantial input from geopolitics experts, who explain how terrorist targets are prioritised.

The quantitative modelling approach presents a contrast with the Darpa project, which was dubbed “terrorism futures” by the international media, amid expressions of moral outrage from politicians and various commentators. Robin Hanson, associate professor at George Mason University in Fairfax, Virginia, was the brains behind Darpa. “We weren’t intending to create a risk hedging market,” he says. “The idea was to pioneer a new concept for a speculative market in information.” But he acknowledges that a hedging market would have been a long-term hope.

But Leighton Williams, professor of economics and director of the Betting Unit at Nottingham Trent University in the UK, says: “There is a blurred relationship between what is betting and what is hedging.” Williams says this is explained by the fact that bets can be hedged.

Users of online betting markets, such as Dublin-based TradeSports.com, Costa Rica-based BetonSports.com and the Iowa Electronic markets can speculate on political questions such as the outcome of elections. Speculators in Europe have been able to bet on the outcome of local and national elections, the trades forming the basis of academic research topics at the Max Planck Institute in Jena, Germany and the Humboldt University in Berlin.

Williams suggest that a speculator could bet on the outcome of an election – say, Bush versus Gore in the US – and also on the outcome of US federal litigation against Microsoft, and hedge one against the other. “The speculator could arbitrage it and get a profit,” Williams says.

The proponents of trading or betting on political events base their confidence on the notion of market efficiency. They see betting as the best way of aggregating information and opinion. The online markets have given more accurate forecasts of election results than have opinion polls on either side of the Atlantic.

Vernon Smith, Professor of Economics and Law at George Mason University and the 2002 winner of the Nobel Prize in economics, says: “The horse most people are betting on is the one most likely to win.

“99% of people don’t have a clue how markets work,” he adds. “They don’t know that the orange juice market in Florida is a remarkable predictor of the weather.”

The problem is that we do not really know why markets work as they do, says Smith, and for this reason the Darpa exercise would have had some merit. The power markets would gain from a similar exercise, where energy demand responses could be simulated , he adds. This involves setting up an exercise where participants simulate a wholesale power market in which a number of constraints, such as interruptibility, are introduced. The generators in this simulated market then bid to supply power to end-users. Such an exercise will illustrate what is required to keep demand in balance with supply, Smith says.

Potential for manipulation
But David Rothkopf, president and chief executive of Washington, DC-based Intellibridge Corp, a consultancy providing open-source intelligence and analysis to government and the military, has doubts about political markets. While there is a real need to analyse large pools of knowledge in a systematic way, he says, a political market probably does not provide the highest value.

The sovereign debt markets and stock index futures already aggregate huge amounts of information of a political nature, says Rothkopf. But he feels a political risk market on its own is open to manipulation.

Nottingham Trent University’s Williams and George Mason University’s Hanson acknowledge the potential danger of market manipulation. Supporters of one candidate in a US presidential election campaign tried to create a bandwagon effect by putting money on their principal runner. Similar manipulation has also happened in betting on municipal elections in North America and Europe, Williams says. But it is always difficult to identify market anomalies correctly. “There is a fine line between conspiracy theories and reality,” Williams says.

Intellibridge’s Rothkopf takes a more prosaic attitude. “Most of the things that move markets most violently are known only to a small bunch of people who are not in the market.” This can distort the picture the market provides. Rothkopf says: do stock markets give an adequate picture of the underlying economic reality in a country or of the value of a stock? Ultimately, he says, markets become historical tools for economists rather than accurate predictors of where the market is going.

Rothkopf is also sceptical about quantitative modelling of political risk. “What you end up with is a bad imitation of how a human mind works,” he says. And while game theory helps us understand certain behaviour, he adds, it does not work well for predicting the behaviour of terrorist groups, as the theory involves an infinite number of targets at an infinite number of levels.

What’s more, quantitative models do not take account of terrorism as a business, where each attack raises the profile and the fund-raising capabilities of the perpetrator, say security experts.

Intellibridge’s Rothkopf thinks that when it comes to assessing political risk, no alternative has been found to taking all the best qualitative knowledge and data and putting it in front of experts with a long professional experience of say, 30 years, and who can use their intuition to analyse it. It seems there is still a lot of work ahead before a geopolitical risk futures market can take off.