Automatic Diagnostic Checking for Vector Autoregressions

• Activity Number: 173
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Ignacio Lobato*+ and Juan Carlos Escanciano and Lin Zhu
• Companies: ITAM and Indiana University and Indiana University
• Keywords: Autocorrelation ; consistency ; power ; Akaike's AIC ; Schwarz's BIC

Abstract

This article introduces automatic tests for the correct specification of the lag order of a Vector Autoregression (VAR) model. The proposed tests extend the automatic test for serial correlation of raw data introduced in Escanciano and Lobato (2009) to the multivariate and to the residual case. A main characteristic of the proposed tests is their simplicity: the researcher does not specify the order of the autocorrelation tested and the asymptotic null distribution of the proposed test statistics is chi-square, so there is no need of using bootstrap procedures to estimate critical values. In addition, one of the considered tests is robust to the presence of conditional heteroskedasticity of unknown form. Simulations show that the proposed test presents higher power than the existing ones for models commonly employed in empirical macroeconomics and empirical finance.