Small-sample power of tests for inequality restrictions: The case of quarter-dependent regression errors

Testing for inequality restricted hypotheses has attracted increasing attention in recent years in econometrics. While many tests have been proposed for these testing problems, little is available on the power of these tests. In this paper, we examine the power of two tests in the literature, the locally most mean powerful invariant test and the Kuhn-Tucker test, in the case of testing for quarter-dependent simple AR(4) errors in linear regressions.