Soothsaying or Science?: Falsification,Uncertainty and Social Change in Macroeconomic Modelling

Abstract

Macroeconomic models play an increasingly important role in economic policy-making and evaluation. They cost large sums of money, yet appear unable to forecast the ups and downs of economic activity with any accuracy. Drawing on material gathered during interviews with leading UK economists, I draw a distinction between `strong' and `weak' claims in econometrics, and discuss how these relate to the falsification of economic models and theories. Empirical work in econometrics is shown to be chronically ambiguous, and it is argued that the stochastic properties of macroeconomic models swamp any trends identified. Furthermore, although macroeconomic models can support economic forecasts, they cannot produce them unaided — forecasters must use their skill and judgement to anticipate future events and render forecasts plausible. The conclusion drawn is that although their pedagogic value is undisputed, the case for using macroeconomic models to guide, frame and evaluate economic policy has not yet been made.