A Monte Carlo Comparison Of Non-nested Encompassing Tests
Author: Fakhr E Alam

Encompassing tests are used to select between non-nested models. There are number of encompassing tests available in literature. We have chosen three tests used for encompassing named Cox test, Sargan test and Ericsson test. This study compares size and power of non-nested encompassing test. Size distortion was negligible which implies the powers are comparable and the power curves are obtained using powers on different alternatives. The analysis reveals that Ericsson test is the most powerful test among the three tests and this test should be preferred when using encompassing tests. The analysis also reveals that Sargan test should be ignored because of its very low power. Supervisor:- Dr. Atiq-ur-Rehman

Meta Data

Keywords : Encompassing Tests, Monte Carlo Comparison
Supervisor: Atiq Ur Rahman

Related Thesis​