Optimal control of linear econometric models with intermittent controls

Dynamic Programming is used to derive the optimal feedback solution to the minimization of a quadratic welfare loss-functional subject to a linear econometric model, when the value of some instrument variables can not be optimized in every model period, but only in single ones. In this way, the relative inertia of fiscal policy-making, as compared to monetary policymaking, can e.g. be taken into account. Analytical expressions are derived for the optimal feedback rules and for the minimum expected losses, and iterative schemes are proposed for their numerical computation. It is suggested that a numerical analysis of the economic gain to be realized by making more frequent adjustment of fiscal policy variables than is actually the case could yield valuable information for policy-makers. r. INTRODUCTION In recent yea r s , several authors have addressed problems in macroeconomic policy-making with the help of l i nea r -quadra t i c control methods, see e .g . Pindyck [1973], Chow [1975]. Spec i f i ca l ly , these authors consider the l i n e a r or l inear ized econometric model in s t a t e v a r i a b l e form (la) x = A x + B u + D z + £ , t = 1 , 2 , . . .