Forward-looking variables in deterministic control

In this paper, the authors adapt the Fair and Taylor method for forward-looking variables in simulation models. In particular, they develop a procedure for solving quadratic linear control models when there are forward-looking variables in the system equations. The simplest way to do this for deterministic problems would be to stack up the variables for all time periods using Theil’s procedure, as suggested by Hughes-Hallet and Rees for simulation models and done by Becker and Rustem for perfect foresight problems. However, the authors plan to continue from the current paper and develop similar procedures for passive and active learning control problems, and the stacking procedure does not seem as natural for those problems. Therefore, we will use the Fair-Taylor approach here and adapt it for deterministic quadratic linear problems.