文档介绍：Computational Economics 17: 265–284, 2001. 265
© 2001 Kluwer Academic Publishers. Printed in herlands.
Bayesian Analysis of the Stochastic
Switching Regression Model Using
Markov Chain Monte Carlo Methods
MARIA ANA E. ODEJAR 1 and MARK S. McNULTY 2
1Kansas State University, .; E-mail: leilannie@; 2Los Alamos National
Laboratory, .
Abstract. This study develops Bayesian methods for estimating the parameters of a stochastic
switching regression model. Markov Chain Monte Carlo methods, data augmentation, and Gibbs
sampling are used to facilitate estimation of the posterior means. The main feature of these methods
is that the posterior means are estimated by the ergodic averages of samples drawn from conditional
distributions, which are relatively simple in form and more feasible to sample from than plex
joint posterior distribution. A simulation study is paring model estimates obtained
using data augmentation, Gibbs sampling, and the maximum likelihood EM algorithm and determ-
ining the effects of the accuracy of and bias of the researcher’s prior distributions on the parameter
estimates.
Key words: stochastic switching regression model, conjugate priora, posterior mean, Markov Chain
Monte Carlo method, data augmentation, Gibbs sampling, EM algorithm
1. Introduction
Economic systems are intrinsically dynamic. These dynamics are characterized
by changing economic relationships. Shocks to the economy may be due to an
economic crisis, a change in the society’s economic behavior, improvement in
technology, economic policy revisions, or a major revamp in the political system.
Whatever the source of the change, econometric models must be able to incorporate
the alterated economic relationships.
An econometric model that allows for change is the switching regression model,
where it is assumed that an observation yj may be generated by one of s alternative
regression models or regimes, .,
= + = =
yj xij βi εij ,yj ε Regime i, i