This paper introduces a semi-parametric bootstrapping approach to Bayesian analysis of structural parameters in simultaneous equation systems that extends the single and multivariate regression approaches of Heckelei and Mittelhammer (1996, 2002) to models with endogenous regressors. Monte Carlo evidence demonstrated the considerable accuracy of the procedure in approximating posterior distributions, even for small sample sizes.

本文介绍了联立方程系统结构参数贝叶斯分析的半参数自举方法，将Heckelei和Mittelhammer(1996, 2002)的单变量和多变量回归方法扩展到内源回归模型。蒙特卡洛证据表明，相当准确的程序在接近后验分布，甚至小样本大小。