Approaching rainfall-based weather derivatives pricing and operational challenges

Abstract

This article approaches some of the current rainfall derivatives pricing and operational challenges through an empirical application to Comunidad Valenciana, Spain. Regarding the former, two different issues are addressed. First, we examine the rightness of suggesting the Gamma distribution to price rainfall contracts, which is the alternative chosen by previous authors applying the Index Value Simulation technique. This is done for the purpose of determining whether the consideration and comparison of other alternatives may lead to more accurate valuation results. Concretely, two different distributions, in addition to the Gamma, are proposed: the exponential and the mixed exponential, whose fits are assessed through the Kolmogorov–Smirnov/Lilliefors test and graphical analyses. The outcomes attained indicate that this selection process leads indeed to a precise generation of the rainfall index’s moments. Next, we examine the viability of using a unique distribution to model the rainfall risk of regions located nearby, since this would considerably decrease valuation complexity. Our analysis shows that the most convenient choice depends on the period and location considered, although the mixed exponential appears as a reasonable option in most cases. Finally, a relevant operational challenge related to geographical basis risk is approached. Concretely, an evaluation of this type of risk among the locations studied is conducted. The results attained indicate that, given the insufficient degree of correlation between nearby locations, rainfall risk hedging measures may rely on compound derivatives referred to several neighbor stations.