Posterior analysis of restricted seemingly unrelated regression equation models: a recursive analytical approach
提出一种递归分析方法，对系数受线性约束的似不相关回归（SURE）模型进行精确贝叶斯后验推断，并应用于英国货币与利率的两方程模型，证明该方法能提高复杂模型分析的可靠性和效率。
The scope of exact analytical results in Bayesian econometrics is known to be quite limited. It is, however, shown here to be broader than the simple natural-conjugare framework. Restricting the coefficients of a SURE model in a recursive linear way can not be accommodated in a natural-conjugate analysis,but still allows for analytical ingerence, exploiting the recursive characteristics over equations. These finding are used to obtain analytical posterior results in a two-equation model for money and interest rate in the UK. Subsequent research shows that such methods can substantially increase both reliability and efficiency in the analysis of more complicated models than the ine under scrutiny here.