Control of the LBS Econometric Model via a Control Model

Abstract

As a result of the work that has been devoted to the use of control theory in macroeconomics, it has become clear that demonstrating the basic principles on small models, though useful as a pedagogic exercise, is not likely by itself to achieve the needed transition from theory to actual practice in economic policy-making. When a control engineer has to convince industrialists that he can make a contribution to the solution of their problems he is able to do so, in the final analysis, by showing that the performance of the boiler or the chemical process has been improved. This can be achieved in a relatively short space of time. In the management of the economy, however, the period necessary to demonstrate that stochastic control methods are beneficial may be prohibitively long. There is, therefore, a serious problem of validation since the Government are not going to hand the economy over so that experiments in control can be conducted. One way this obstacle can be avoided is to use a large nonlinear econometric model of the economy, such as that of the London Business School, as an analogue for the real economy, and to control it with a small linear model estimated from economic data. Needless to say, the large model itself is not free from many of the criticisms that can be levelled at the small model. But as long as the model, in terms of its dynamic responses, behaves like the economy, it will provide an adequate test-bed for the use of control methods.