Bootstrap Inference in a Linear Equation Estimated by Instrumental Variables
Bootstrap Inference in a Linear Equation Estimated by Instrumental Variables
复制标题
由工具变量估计的线性方程中的自举推理
DOI：
10.1111/j.1368-423x.2008.00247.x
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
J. MacKinnon
中科院分区：
文献类型：
--
作者：
R. Davidson;J. MacKinnon
我们研究了工具变量估计的线性方程中单个右侧内生变量系数的几种检验方法。我们表明，将所有的检验统计量--学生t，Anderson-Rubin，Kleibergen和Moreira的LM统计量(K)，以及似然比(LR)--写成六个随机量的函数，可以得到一些关于weakinent渐近下检验性质的有趣的结果。然后，我们提出了几种新的方法来自举这三个非精确检验统计量，并且还提出了一个新的条件自举版本的LR检验。与现有的方法相比，这些方法使用了更有效的简化形式方程的参数估计。当使用这些新过程中的最好的时候，K和条件自举LR测试在零下都有很好的性能。然而，电力方面的考虑表明，后者可能是选择的方法。
We study several tests for the coefficient of the single right-hand-side endogenous variable in a linear equation estimated by instrumental variables. We show that writing all the test statistics—Student's t, Anderson-Rubin, the LM statistic of Kleibergen and Moreira (K), and likelihood ratio (LR)—as functions of six random quantities leads to a number of interesting results about the properties of the tests under weakinstrument asymptotics. We then propose several new procedures for bootstrapping the three non-exact test statistics and also a new conditional bootstrap version of the LR test. These use more efficient estimates of the parameters of the reduced-form equation than existing procedures. When the best of these new procedures is used, both the K and conditional bootstrap LR tests have excellent performance under the null. However, power considerations suggest that the latter is probably the method of choice.