Policy Trade-Offs in the NIESR Model: Exercises Using Optimal Control

Abstract

This article investigates the consequences of alternative fiscal packages by applying optimal control techniques to the National Institute econometric model. Britton (1983) provides an assessment of alternative packages by using ready reckoners derived from simulations of the NIESR model reported in Brooks and Henry (1983). The present investigation may be seen as a more systematic analysis, using the techniques of optimal control. Such an assessment inevitably concerns the measurement of trade-offs between the major objectives of government policy, an issue which has generated much controversy over the last decade. Early interpretations of the Phillips Curve for example tended to regard the empirical finding of an inverse relation between inflation and unemployment as providing a menu for policy choice. According to this interpretation, aggregate demand policy could be used to select the best combination of inflation and unemployment rates available subject to the constraint set by the Phillips Curve. Since that time, parallel developments in the theory of wage inflation, expectations formation, and macroeconomic policy have produced considerable scepticism about the existence of this trade-off, questioning whether the implied choice between alternative combinations of inflation and unemployment is open to governments using aggregate demand policy. Much of this scepticism was coincident with growing dissatisfaction with the Phillips Curve as a statistical representation of wage inflation, but we will not be directly concerned with this aspect of the argument here. Two points however do arise from this debate which are of direct relevance to the subject matter of this article. The first is that a trade-off between inflation and unemployment may exist even if the labour market is not characterised by a Phillips Curve. The second is the central role played by natural rate concepts in the debate on the effectiveness of aggregate demand policy.