A SIMULATION APPROACH TO THE PROBLEM OF COMPUTING COXS STATISTIC FOR TESTING NONNESTED MODELS

This paper proposes a new procedure for computing the Cox statistic fo r tests of nonnested hypotheses using the method of stochastic simulat ion. The procedure is applicable to a wide class of probability distri butions, is relatively simple to implement, and does not require an an alytic derivation of the pseudo-true estimators that enter the Cox tes t statistic. The paper also contains an application of the proposed me thod to the test of probit versus logistic formulations of the univari ate binary choice models. The empirical results show that the simulati on method works reasonably well even for a moderate number of replicat ions.