Moment based regression algorithms for drift and volatility estimation in continuous-time Markov switching models
Date
2008
Authors
Elliott, R.J.
Krishnamurthy, V.
Sass, J.
Editors
Advisors
Journal Title
Journal ISSN
Volume Title
Type:
Journal article
Citation
Econometrics Journal, 2008; 11(2):244-270
Statement of Responsibility
Conference Name
Abstract
We consider a continuous time Markov switching model (MSM) which is widely used in mathematical finance. The aim is to estimate the parameters given observations in discrete time. Since there is no finite dimensional filter for estimating the underlying state of the MSM, it is not possible to compute numerically the maximum likelihood parameter estimate via the well known expectation maximization (EM) algorithm. Therefore in this paper, we propose a method of moments based parameter estimator. The moments of the observed process are computed explicitly as a function of the time discretization interval of the discrete time observation process. We then propose two algorithms for parameter estimation of the MSM. The first algorithm is based on a least-squares fit to the exact moments over different time lags, while the second algorithm is based on estimating the coefficients of the expansion (with respect to time) of the moments. Extensive numerical results comparing the algorithm with the EM algorithm for the discretized model are presented.
School/Discipline
Dissertation Note
Provenance
Description
Access Status
Rights
Copyright 2008 Royal Economic Society