|
| |
| Wenjie Wang and Qingfeng Liu |
| |
| ''Bootstrap-based Selection for Instrumental Variables Model'' |
| ( 2015, Vol. 35 No.3 ) |
| |
| |
| This paper develops bootstrap-based method for addressing the “many instruments” problem in the context of instrumental variable estimation. We propose a plug-in restricted-efficient-residual-based (plug-in RE) bootstrap for choosing optimal number of instruments used for two-stage least squares (TSLS) and limited information maximum likelihood (LIML) estimator. In Monte Carlo experiments, we find that the instrument choice based on our plug-in RE bootstrap generally yields an improvement in finite sample performance. |
| |
| |
| Keywords: Instrument selection, Many instruments, Bootstrap, TSLS, LIML |
JEL: C1 - Econometric and Statistical Methods: General
C2 - Single Equation Models; Single Variables: General |
| |
| Manuscript Received : Jun 14 2015 | | Manuscript Accepted : Sep 07 2015 |
|
