Dynamic adjustment cost models with forward-looking behaviour

In this paper we propose a new approach for dynamic decision problems where forward-looking agents choose a set of non-stationary variables subject to quadratic adjustment costs. It is assumed that expectations are computed by a cointegrated Vector Equilibrium Correction Model (VEqCM). The role of feedbacks from the decision to the explanatory variables on solution properties and modelling approach is discussed. We show that once the system of interrelated Euler equations stemming from the agent's optimization problem is embedded within the VEqCM, a switching algorithm based on Generalized Least Squares can be used to estimate and test the model. A labour demand model for two Danish manufacturing industries is investigated empirically. Copyright Royal Economic Society 2006