Estimation of Cointegrated Multivariate Continuous-Time Autoregressive Moving Average Processes

In this thesis we consider cointegrated MCARMA processes. A canonical representation is derived and the probabilistic properties are investigated. A step-wise estimation method of the model parameters from discrete-time observations is derived. Super-consistency for long-run and consistency for short-run parameter estimator are established. The limiting distributions of the estimators are deduced. Lastly, a simulation study to demonstrate the applicability of the estimation method is presented.