Simulation Based Finite- and Large-Sample Inference Methods in Simultaneous Equations
In the context of multivariate regression (MLR) and simultaneous equations (SE), it is well known that commonly employed asymptotic test criteria are seriously biased towards overrejection. In this paper, we propose finite and large sample likelihood based test procedures for possibly nonlinear hypotheses on the coefficients of SE systems. We discuss a number of bounds tests and Monte Carlo simulations based tests. The latter involves maximizing a randomized p -value function over the relevant nuisance parameter space. This is done numerically by using a simulated annealing algorithm. Illustrative Monte Carlo experiments show that (i) bootstrapping standard instrumental variable (IV) based criteria fails to achieve size control, especially (but not exclusively) under near non-identification conditions, and (ii) the tests based on IV estimates do not appear to be boundedly pivotal and so no size-correction may be feasible. By contrast, likelihood ration based tests work well in the experiments performed.
Year of publication: |
1999-03-01
|
---|---|
Authors: | Dufour, Jean-Marie ; Khalaf, Lynda |
Institutions: | Society for Computational Economics - SCE |
Saved in:
Saved in favorites
Similar items by person
-
Structural Estimation and Evaluation of Calvo-Style Inflation Models
Dufour, Jean-Marie, (2006)
-
Are New Keynesian Phillips Curves Identified ?
Kichian, Maral, (2004)
-
Testing Financial Integration: Finite Sample Motivated Mothods
Beaulieu, Marie-Claude, (2006)
- More ...