Macro-Economic Forecasting and Modelling

Confidence in macro-economic forecasting has periodically been punctured by episodes of economic turbulence and concomitant predictive failure. The poor performance in predicting the consumer boom in the late 1980’s, and the depth and duration of the recession in the 1990’s, can be viewed as merely the latest examples in a catalogue of failures, with noteworthy antecedents including the under-prediction of post-war consumption, and the 1974-5 and 1979-81 recessions.1 Periods of economic turbulence may be informative in highlighting model inadequacy, but they also bring into question the usefulness of economic forecasting.2 Should we be surprised that such failures have occurred? Econometric theory comprises a body of tools and techniques for analyzing the properties of prospective methods under hypothetical states of nature. In the forecasting context, the methods are forecasting models and procedures, and the states of nature relate to the properties of the variables to be forecast. For an econometric theory of forecasting to be useful in terms of delivering relevant conclusions about empirical forecasting, those states must adequately capture the appropriate aspects of the real world to be forecast. However, the traditional, text-book theory of economic forecasting fails to allow for a number of vital features of the economy. In particular, analyses of forecasting have usually been based on assumptions that implicitly rule out structural change, or regime shifts, in the economy, namely: