OPTION PRICING UNDER THE NORMAL INVERSE GAUSSIAN DISTRIBUTIONS

摘要

This paper discusses European style option pricing for bothrnpath dependent and nonpath dependent cases where thernlog returns of the underlying asset follow the normal inversernGaussian (NIG) distributions. The moment matchingrnmethod is used in estimating model parameters. ThernMonte Carlo method and the Sobol' sequence based quasi-rnMonte Carlo method combined with some variance reductionrnmethods are used in simulating option prices. Ourrntest results show that the (randomized) quasi-Monte Carlornmethod is more efficient than the Monte Carlo method ifrnboth with the same variance reduction method.