Efficient Posterior Simulation for Cointegrated Models with Priors on the Cointegration Space
针对协整模型中协整向量空间先验导致计算复杂的问题，提出了两种高效吉布斯抽样算法（折叠吉布斯和参数增广吉布斯），显著提升后验模拟速度。
A message coming out of the recent Bayesian literature on cointegration is that it is important to elicit a prior on the space spanned by the cointegrating vectors (as opposed to a particular identified choice for these vectors). In previous work, such priors have been found to greatly complicate computation. In this article, we develop algorithms to carry out efficient posterior simulation in cointegration models. In particular, we develop a collapsed Gibbs sampling algorithm which can be used with just-identifed models and demonstrate that it has very large computational advantages relative to existing approaches. For over-identifed models, we develop a parameter-augmented Gibbs sampling algorithm and demonstrate that it also has attractive computational properties.