Guest contribution by James Annan of FRCGC/JAMSTEC.
“The more unpredictable the world, the more we rely on predictions” (Steve Rivkin). The uncertainty of an unknown future imposes costs and risks on us in many areas of life. A cereal-growing farmer risks a big financial loss if the price of grain is low at harvest time, and a livestock farmer may not be able to afford to feed his herd if the price of grain goes up. One way to reduce the risk is to hedge against it in a futures market. The two farmers can enter a forward contract, for one to deliver a set quantity of grain to the other for a fixed price at a future date. And indeed farmers do routinely use futures contracts to reduce their risks.
Weather is a major uncertainty affecting our futures (it is one of the main sources of risk for those farmers), and weather futures markets can be used to hedge on the monthly/seasonal time scale. In the longer term, changes in climate will also bring a range of costs and benefits, and a market in climate futures would allow anyone who is vulnerable to hedge against these risks too. But how can we assign fair prices to the contracts? One obvious starting point would be to look at model predictions and historical data. This is essentially what the IPCC does, eg with its estimate of 0.3+-0.1C
/decade for anthropogenically-forced warming over the next 20 years in the absence of substantial mitigation of emissions (at the “likely” level, ie 66%-90% probability). If we want to work out the probability of global mean temperature being warmer 20 years from now, we could take this 0.1C in uncertainty of anthropogenic forcing (which I will assume to be 1 standard deviation of a Gaussian deviate), and add another 0.15C of independent natural internal variability, which gives a combined estimate of 0.3+-0.18C warmer overall or about a 5% chance of cooling. To this, we can add perhaps another 5% due to the possibility of a large volcanic eruption at the right time, making 10% in total. Now before you all write in telling me my assumptions are wrong, the real answer should be 20%, or 4%…that’s precisely my point. Of course my simplistic assumptions can be questioned, and I could have performed a more accurate calculation, but however carefully it is approached, this sort of forecasting inevitably involves subjective judgement and assumptions. The IPCC estimate depended on expert judgement, so someone who believed that they markedly overstated the anthropogenic influence might deduce that the chances of cooling were closer to 50%, and an advocate of the more extreme solar forcing theories might even confidently predict significant global cooling. So ultimately it seems like we have no really firm, provably correct and objective basis for setting a market price.
Fortunately, there is a solution to this problem. Efficient market theory states that the price in a free market should accurately reflect the aggregated information that is available, and so in 1990 economist Robin Hanson noted that we can turn the problem around: use the market as an aggregation mechanism to tell us what the odds are on any particular event. This, he claims, could be much more effective than relying on panels of government- (or self-)appointed experts and media pundits to predict the future, since the financial penalty would keep the incompetent away from the marketplace, and the real experts would be automatically rewarded for correcting any errors in the market prices. There is now a growing body of evidence in support of these theories, including for example the analysis of existing markets and results from a market which was set up specifically to investigate the idea further. Of course the market price does not exist in a vacuum: in practice, traders would perform calculations as I did above (but rather more carefully), and the market price would reflect their consensus. The market is a highly inclusive mechanism: rather than having to debate to the death, anyone who has an opinion can invest as much as they want (which will relate to their confidence), and in the long run the winners will drive out the losers. If I offer (and accept) bets on global cooling at 10:1 based on my rough calculation, someone who does a better job of estimation will probably be able to take money off me, earning a reward for their skill and effort.
The Foresight Exchange market is an internet-based game which runs an ideas futures market, covering a wide range of claims including a number of environmental issues. The basic betting mechanism is based on the concept of a pair of coupons labelled “Yes, claim X is true” and “No, claim X is false” (where X is a proposition whose truth is not yet known). These coupons can be bought and sold individually on the open market, and a pair of them can always be bought or redeemed for $1 from the bank. When the claim’s truth is determined, the correct coupon is redeemed for $1 and the false coupon becomes worthless. The market price of the Y coupon is the market’s estimate of the probability of the claim being ultimately proved true. Here are a couple of examples – indicating how confidently (or otherwise) the market predicted the re-elections of Clinton and Bush in 1996 and 2004 respectively.
The Foresight Exchange is a small market with a limited number of players, and is certainly not perfect – I presented a simple analysis in a poster at the EGU in April, and I have also discussed a couple of claims in more detail here and here. (These are a slightly different type of claim, where the payout of the Y coupon depends continuously on the value of a variable at a specified future time – such as the global mean temperature in 2030. The N coupon pays $1-Y as before.) As Hanson’s theory predicts, by participating in the game, I have both improved the price on these claims (increasing market skill) and simultaneously increased my wealth (received a reward for my contribution). Although this small and non-serious market has noticeable inefficiency, a real money market could certainly hope to do better. So what are the drawbacks? It seems that there are are surprisingly few. Fears that rich vested interests could rig the market are generally misplaced: it would cost them a lot of money to do so, potentially much more than the price of influencing a few high-ranking experts. Robin Hanson’s pages have a great deal of discussion about the various criticisms of the idea that have been offered, and also how the basic idea could be extended to cover a wide range of applications.
I’ve recently been trying to establish consensus on the subject of global temperature rise, by arranging bets with sceptics who claim that the IPCC TAR is overly alarmist. Richard Lindzen was the first I noted who forecast here that over the next 20 years, the climate is as likely to cool as warm, and said he would be prepared to bet on it. However, when challenged to a bet, it turns out that he expects odds of 50:1 in his favour, ie he will only bet on the chances of cooling being at the 2% level or higher, far short of his 50% claim. My quick and dirty estimate above based on the IPCC TAR suggests that they would put the probability at more like 10%, so his offer actually appears to affirm the IPCC position. He also suggested an alternative bet (see here for my comments on this article) based on the amount of warming: >0.4C warmer and I win $5,000, <0.2c and he wins $10,000. Again, no-one who believes the IPCC summary would find his offer attractive, since it has negative expected value. The chances of winning and losing are roughly equal, so there does not appear to be any possible justification for his expectation of a 2:1 ratio (in his favour) in the stakes. In both cases, in contrast to his words, his position seems to be more alarmist than the IPCC!
The list of sceptics who have refused to bet against the IPCC position has grown steadily since then, and now also includes Michaels, Jaworowski, Corbyn, Ebell, Kininmonth, Mashnich and Idso (all my blog posts and related comments are linked from here). While I would be happy to take money off any or all of them, there is more to this than sceptic-bashing and a few high-profile bets – it could also perhaps result in a working market that would generate a true consensus and, furthermore, provide the socially and economically valuable function of allowing the vulnerable to hedge against risks.
[2005/06/24: comments on this post are now closed. This post generated a record number of comments: thanks to all those who commented. For any further follow up, you are recommended to the discussion on the newsgroup sci.environment - William]