2.4
The Miami Hurricane Event Market (MAHEM): A Hurricane Futures Market
David Letson, Univ. of Miami/RSMAS, Miami, FL; and D. S. Nolan, D. Kelly, and F. Nelson
Abstract: Can a hurricane futures market efficiently aggregate a large body of expert opinion in tropical meteorology and in so doing advance understanding of hurricane track forecasts? We have created an electronic futures market to harness the expertise of tropical meteorologists as to where a given hurricane will land (http://mahem.miami.edu/). Each time a new hurricane or tropical storm is named, MAHEM opens trading in a new set of futures contracts, each of which is associated with a specific segment of U.S. coastline, Atlantic or Gulf Coast. After the hurricane lands, one dollar is then paid for each contract with the geographic coordinate matching where the hurricane made landfall. Prices at which contracts trade while the market is open indicate beliefs regarding landfall probabilities.
MAHEM is an experimental research tool whose primary purpose is not to predict but to show what people know. If, for example, MAHEM shows that people are using bad or misleading information, that itself can be an important finding. Numerous hurricane forecasting models are already competing, in a sense, and MAHEM simply enlarges and aggregates, in real time, expert opinion on which models are most credible.
We will test various economic models of how new information is incorporated into prices of financial instruments. For example, we will contrast the change in contract prices to the change in probabilities that Bayes' Rule would predict. If the market prices of contracts change dramatically upon the issuance of an official forecast, then the market is placing a high value on the National Hurricane Center information.
We believe the study of a futures market tied to hurricanes will lead to improved understanding about the hurricane warning process and to advances in the study of efficiency of forecast markets in economics, design of catastrophic insurance, hurricane modeling in meteorology, risk communication and weather forecasting.
.Session 2, Economics and Weather: Methods and Applications
Thursday, 2 February 2006, 8:30 AM-12:00 PM, A307
Previous paper Next paper