| Author/Presenter |
Gubhinder Kundhi (Carleton University) |
| Co-author |
Paul Rilstone (York University) |
| Title |
Approximate Empirical Saddlepoint Approximations for Nonlinear Estimators |
| Abstract |
Empirical Saddlepoint Approximations to the density of Nonlinear Estimators are derived. The results are shown to apply to most of the common extremum estimators used in applied work including Generalized Method of Moments, Maximum Likelihood Estimators and Generalized Empirical Likelihood Estimators in an i.i.d sampling context. The approximations are illustrated for a number of popular nonlinear estimators. Saddlepoint Approximations generally provide accurate approximations to density and distribution functions, particularly in the tails. A Monte Carlo Experiment is conducted for the exponential regression model. In this experiment the performance of the Empirical Saddlepoint approximation for the MLE is compared with other methods commonly used in finite sample inferences such as the Edgeworth approximation and the Bootstrap. |