Information Criteria, Model Selection Uncertainty and the Determination of Cointegration Rank

The determination of the cointegration rank of a multivariate cointegrated system is usually carried out assuming that the lag order of the vector autoregressive process representing the multivariate system is known. Since such an assumption is unlikely to hold in reality, the question of how the lag order selection may affect the determination of the cointegration rank arises. This paper aims to address this question in a Monte Carlo framework. Additionally, we provide a formal justification for the application of model selection criteria in selecting the cointegration rank.