We provide some evidence of Empirical Likelihood's (EL) practical value in econometrics. We present EL as an alternative to GMM estimation and assess the finite-sample properties of their overidentification tests (size and power) through Monte Carlo simulations. We address the issue of the importance of the results to applied workers and use as laboratories to our experiments two settings with potential empirical applications: the Mean-Variance and Three-Moment CAPM and a dynamic panel model with individual effects. In cases in which we found important size distortions we introduced efficient bootstrap critical values. Prior research applied this bootstrapping technique to the GMM (GMM-bootstrap) and we present results for the EL (EL-bootstrap). We also include an empirical example on a United States panel cash-flow model. Even if our findings do not uniformly support the conclusion that one estimator dominates the other, we found evidence that EL and EL-bootstrap are good alternatives to GMM and GMM-bootstrap in some econometric applications.
Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:428456 |
Date | January 2005 |
Creators | Gonzalez Solano, Flor Angelica |
Publisher | University of York |
Source Sets | Ethos UK |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Source | http://etheses.whiterose.ac.uk/14074/ |
Page generated in 0.0136 seconds