Towards Empirical Assessments of Controlled Cointegrated Models

ABSTRACT This paper explores control theory and stabilisation policy within the framework of a cointegrated vector autoregressive (VAR) model from the perspective of an econometrician concerned with inference and identification. We demonstrate that a new process derived from control theory should be treated as a series of observables rather than as a latent series. This process can be viewed as being driven by a vector autoregressive moving‐average (VARMA) model, which can, in turn, be interpreted within the framework of structural vector equilibrium correction. We also introduce a data‐driven procedure for classifying intermediate and final policy targets within the model. The practicality and effectiveness of this procedure are demonstrated through a counterfactual policy analysis based on observations of the new process simulated from actual New Zealand monetary policy data.