Modelling conversion options with a mean reversion motion

Commodity prices are generally better modeled by a long-term Mean Reverting Process, than by a Geometric Brownian Motion stochastic diffusion process, which is more generally used to value real options, since it is simpler to use. In this article we model two correlated uncertain variables using a mean reversing process bivariate lattice to value the switch option between outputs available to ethanol and sugar producers, using the same source: sugarcane. The model results show that the switch option adds a signi cant value for the producer income. The article also shows that when modeled by a geometric brownian motion, the switch option yields signi cantly higher values than with a mean reverting model, for the option itself as much as for the base case without exibility. This con rms that the stochastic model chosen can in uence signi cantly the option value.