Selecting the Best Instrumental Variables Estimator
考虑在联立方程模型中常用的有限信息估计量类中找出“最佳”估计量的问题，基于Edgeworth展开的集中度比较给出了足够简单的选择规则，对应用计量研究有用。
This paper considers the problem of finding the "best" estimator among a class that includes most of the commonly used limited information estimators in simultaneous equation systems. Concentration comparisons, based on the Edgeworth expansion of the distribution of these estimators, lead to selection rules of sufficient simplicity to be useful in applied econometric research.