Prof. Dr. Dietmar Bauer
Project Duration: 01.12.2015 - 31.07.2020
While the estimation and inference theory for integrated and cointegrated processes is well researched in the vector autoregressive (VAR) setting, the analogous results do not exist for the vector autoregressive moving average (VARMA) setting.
The same holds true for the - in a certain sense- equivalent linear state space framework for the empirically most relevant case of integrated processes of order one and two (I(1) and I(2)) as well as seasonally integrated processes.
The main goal of the process therefore is to develop such estimation and inference theory which (i) allows to include relevant restrictions on the dynamical properties of multivariate processes (in the form of knowledge on the integration and cointegration properties) and (ii) uses the state space framework.