Appendix: Recursive Marginal Quantization of Higher-Order Schemes

Abstract

This document serves as the online appendix to McWalter et al. [2017]. It provides efficient matrix implementations for both vector quantization and the generalized recursive marginal quantization, which allows for the implementation of the higher-order discretization schemes presented in the main paper. It includes graphical examples of how the higher-order discretization schemes reduce the error in the implied marginal distributions for both the geometric Brownian motion process (GBM) and its generalization, the constant elasticity of variance (CEV) process. It concludes with an example of the effect of modelling the CEV process with an absorbing or reflecting boundary at zero. The full-text version of this paper can be found at http://ssrn.com/abstract=2894753.