| Title: | IMPROVED FINITE-SAMPLE INFERENCE IN OVERIDENTIFIED MODELS WITH WEAK INSTRUMENTS |
| DOI No: | 10.1142/9781860949531_0011 |
| Source: | RECENT ADVANCES IN STATISTICAL METHODS (pp 132-146)
|
| Author(s): | NIKOLAY GOSPODINOV
Department of Economics, Concordia University, 1455 de Maisonneuve Blvd., West Montréal, Québec, H3G 1M8, Canada
|
| Abstract: | This paper investigates the finite-sample properties of the class of generalized empirical likelihood estimators in possibly overidentified models with weakly identified parameters. These nonparametric likelihood estimators satisfy exactly the moment conditions and automatically remove the bias that arises from a lack of centering of the moment conditions. The inference procedure suggested in the paper does not involve any explicit estimation of the variance-covariance matrix. The confidence sets for the parameters of interest are constructed by inverting the χ2 acceptance region of the criterion test. |
| Full Text: | View full text in PDF format (684KB) |
| TOC: | Back to Table of Contents |
|
|
