Diagnostic Checks for Single-Equation Error-Correction and Autoregressive Distributed Lag Models

Abstract

Single equation error correction models (ECMs) are widely used in the analysis of cointegrated variables. It is important to check the specification of ECMs using diagnostic tests. Monte Carlo evidence is reported that shows that the finite sample significance levels of such tests can be sensitive to the method used to estimate long-run coefficients that yield the error-correction term of the ECM. Estimates of long-run coefficients based upon autoregressive-distributed lag (ADL) models are recommended. The applicability of diagnostic checks to ADL models for integrated variables is examined. An indirect approach to obtaining asymptotically valid checks is proposed. Copyright 1998 by Blackwell Publishers Ltd and The Victoria University of Manchester