The Interfaces between Policy Makers, Markets and Modellers in the Design of Economic Policy: An Intermodel Comparison

Abstract

Economic policy makers wish to bring about the 'best' set of economic circumstances within the constraints set out by the economic system. A natural way to express this problem formally is in terms of a policy maker who maximises a given objective function (or equivalently, specifies a set of policy priorities or trade offs) in the light of a set of beliefs regarding the workings of the economy. As detailed models representing views of how the economy works have been available for over twenty five years, it is natural to formalise this notion within the context of the large macroeconometric models and to use optimal control techniques. The basic technology for carrying out such exercises has been available for many years (Ball Committee on Policy Optimisation I978). Optimal policy exercises published ex ante and described by Bray (I 994) on the Treasury model and by Westaway (I995) on the National Institute of Economic and Social Research models yielded policies which would have avoided some of the major economic policy errors that have been made in recent years. This in itself does not validate such policies, but it justifies their further development. Comparisons have been made between optimal policies derived from different models by Whitley et al. (I992) and Bray et al. (I993). This paper simulates current UK economic policy making by policy optimisation exercises on the Treasury, Oxford Economic Forecasting (OEF), London Business School (LBS) and National Institute of Economic and Social Research (NIESR) models, earlier versions of which were used in the above comparisons. In this exercise, recommendations for instrument trajectories and expected outturns for the economy differ substantially. This is because of differences not merely in the picture the different models give of the objective behaviour of the economy, and the assumptions they make about policy makers' objectives and priorities. It is also because of the assumptions made about the way economic agents and the authorities learn, and form expectations, and the nature of the strategic games they are playing.