Multinomial Approximating Models for Options

Many problems in valuing complex derivatives are solved by using discrete multinomial approximations. In this article, we suggest two modifications for Kamrad and Ritchken (1991) multinomial approximating model. First, we propose the inclusion of an omitted second order term to reduce errors. Second, we ensure non-negative probabilities by bounding the stretch parameter, which parameterizes the size of the up-and down-jumps in the lattice. From a standpoint of assessing the computational effort, we derive mathematical expressions to determine the number of nodes generated by the approximation process for a k asset model. Numerical examples are presented to illustrate gain in accuracy of the proposed model on pricing options and computational efficiency.