Close this Window
Mattias Villani
Abstract
Key words: Cointegration,
Bayesian inference, Estimation, Rank.
A Bayesian reference analysis
of the cointegrated vector autoregression is presented based on a new prior
distribution. Among other properties, it is shown that this prior distribution
distributes its probability mass uniformly over all cointegration spaces
for a given cointegration rank. A simple procedure based on the Gibbs
sampler is used to obtain the posterior distribution of both the number
of cointegrating relations and the form of those relations, together with
accompanying error correcting dynamics. Simulated data are used for illustration
and for discussing the well-known issue of local non-identification.