This paper examines the evaluation of models claimed to be relevant for policy making purposes. A number of tests are proposed to determine the usefulness of such models in the policy making process. These tests are applied to three empirical examples.