In: Metron, 2004, vol. 62, no. 2, p. 161-184
We propose robust counterparts to tests of equal forecast accuracy such as those proposed by Diebold and Mariano (1995) and West (1996). We illustrate the robustness problem and evaluate the size and the power properties of the classical and robust tests under various types of deviations from model assumptions. The new robust test has a correct size and larger power across a wide spectrum of...
|