83rd Annual

Thursday, 13 February 2003: 8:45 AM
Blending Climatology and Forecasts to Compete in the Weather Market
Jeff Shorter, WSI, Billerica, MA; and R. J. Boucher
The weather trading market in North America and Europe started in the energy industry and was principally an "Over the Counter", OTC, market. This made it inaccessible to a broad spectrum of investors. In the start of 2002, a shift in the weather market from OTC to financial exchanges occurred in the US largely due to credit concerns of the big energy companies. With the shift to the financial exchanges, principally the Chicago Mercantile Exchange (CME), came the ability of all financial traders, both companies and individuals, to trade weather futures. In the presentation, three topics will be discussed: 1) Computing potential gains, 2) Evaluating Risk, and 3) Effectiveness of Forecasts in the Weather Market.

Computing the potential gains from a weather future contract is easy and straightforward. Two pieces of information are needed, the prices and the forecast. The prices can be obtained from the CME or from venders. The forecasts can either be made personally or acquired from a weather company, such as WSI Energycast (the authors employer). With both of these pieces, the payout can be calculated simply by subtracting the forecast from the price (if going long) and multiplying the result by the payout per tick. If the payout is positive, then you have an expectation of making money.

Evaluating the Value at Risk, VAR, is a little more difficult and can vary from company to company. The VAR is the amount that a trade could lose if everything went against the trade. For example, if you thought it would be warm in Chicago and it turned out real cold, instead of making some money, you would lose a lot of money. The approach described in the presentation will leverage 40 years of climatology to produce a worst-case scenario for computing VAR.

Finally after seeing the potential to make money in the weather futures market and computing the risks involved in potentially making money, results from simulated trading are presented. The simulated trading began in July 2002 and results to date will be presented. The weather futures prices were from the CME and forecasts from WSI's Energycast WeatherTrader product. In the month of July 2002, the simulated trading make 14 trades. Thirteen trades made money and one trade lost money. The total simulated payout was $22k.

Supplementary URL: