Alternative Asymptotically Optimal Tests and Their Application to Dynamic Specification.

A method is presented for generating test statistics that share the same first order asymptotic optimality properties of the classical statistics. Generalizing J. Neyman's work (1959), t he linearized classical statistic tests restrictions in implicit func tion form using a parameter estimator that is consistent and symptoti cally normally distributed under the alternative hypothesis. By judic ious choice of estimator and form of restrictions at which to evalua te the statistic, a class of asymptotically optimal statistics is obt ained, among which are numbered some familiar classical statistics. A n application is presented for testing common factor restrictions in a single equation dynamic regression model with moving average distu rbances. Copyright 1987 by The Review of Economic Studies Limited.