Tests of seasonal integration and cointegration in multivariate unobserved component models

The paper considers tests of seasonal integration and cointegration for multivariate time series. The locally best invariant (LBI) test of the null hypothesis of a deterministic seasonal pattern against the alternative of seasonal integration is derived for a model with Gaussian i.i.d. disturbances and deterministic trend. A test of seasonal cointegration is then proposed, which parallels the common trend test of Nyblom and Harvey (2000). The tests are subsequently generalized to account for stochastic trends, weakly dependent errors and unattended unit roots. Asymptotic representations and critical values of the tests are provided, while the finite sample performance is evaluated by Monte Carlo simulation experiments. We apply the tests to the indices of industrial production of the four largest countries of the European Monetary Union. We find evidence that Germany does not cointegrate with the other countries, while there seems to exist a common nonstationary seasonal component between France, Italy and Spain.