Gas storage valuation : a comparative simulation study

The purpose of this paper is the comparative analysis of four natural gas storage valuation approaches. In competitive natural gas markets the optimal valuation and operation of natural gas storages is a key task for natural gas companies operating storages. Within this paper, four spot based valuation approaches are analyzed regarding computational time and accuracy. In particular, explicit and implicit finite differences, multinomial recombining trees, and Least Squares Monte Carlo Simulation are compared. These approaches are applied to the valuation of a gas storage facility considering three different underlying price processes. Major characteristics of historical natural gas prices are: seasonality, mean reversion and jumps. Therefore, we consider a mean reversion process as underlying price process. In a first step, we extend this mean reversion process to a mean reversion jump diffusion process, to account for jumps, occurring in historical gas spot price time series. Moreover, we consider a more general price process accounting for mean reversion as well as seasonal patterns as observed in the historical time series. Besides the analysis of the numerical results, the benefits and drawbacks of the methodologies are discussed.