Abstract

Snowfall derivatives are important complements to other weather derivatives such as the most popular temperature derivatives. However, non-arbitrage models could not be used to price snowfall derivatives because the snowfall index is not traded on the market. Also, utility maximization methods are normally too complex to use and the results are sensitive to departures from the models' assumptions. Therefore, I use statistical models to price snowfall derivatives, by modeling the index and the daily snowfall. I use numerical simulations to test the validity of all statistic models that I used. The explanatory power of historical index and daily snowfall values and the prediction accuracy of snowfall derivative prices are used to estimate the models' efficiency. The best model should well explain the past historical pattern and well predict the derivative prices.