Sampling the Variance-Covariance Matrixudin the Bayesian Multivariate Probit Model

摘要

This paper is concerned with the Bayesian estimation of a Multivariate Probitudmodel. In particular, this paper provides a method to sample the restricted variancecovarianceudmatrix directly from its conditional posterior density. The method allowsudthe application of a standard Gibbs sampling algorithm to sample from the posterioruddensity of the parameters, and hence it avoids the use of a Metropolis step. The methoduduses a decomposition of the Inverted Wishart density and alternative identificationudrestrictions.ud