Optimal stabilization policy with delayed controls and imperfect state measurements

The standard optimal control solutions of the macroeconomic stabilization problem - i.e. essentially: the open- and closed-loop solution - are not necessarily implementable or optimal in real-life situations. This is because they do not take into account the time necessary to measure the economy's state and to realize the policy measures physically. In this paper, Dynamic Programming is used to derive, the "best implementable" solution to the optimisation of a quadratic welfare loss-functional subject to a linear econometric model when there are such delays. Two cases are considered: a) Perfect, but delayed state measurements are possible; b) Only imperfect, delayed measurements are available. In both cases, the analytical characterization of the solution immediately suggests practical schemes for the numerical computation of the optimal policy sequence.