Cointegration Rank Tests In Vector ARMA Models <Articles>

This paper proposes several parametric hypothesis tests to determine the cointegration rank in multivariate time series expressed as vector ARMA models. The tests are constructed based on the instrumental variables (Ⅳ) method. It is established that adopting critical values for the standard Johansen likelihood ratio (LR) test is valid in that the same limiting distribution reached. Some Monte Carlo experiments also show that the proposed tests exhibit sufficiently desirable finite sample performances.