On forecasting UK consumption

This paper is concerned with the search for the appropriate methodology for constructing an econometric model for forecasting aggregate consumption. Forecasts of the various national income categories such as consumption, investment, exports, etc. is an important pre-requisite of macroeconomic policy. It is such forecasts welded within the framework of a macroeconomic model which provide the policy makers with the information on which to base economic policy. Ideally a good forecasting model should have the properties of: 1. being parsimonious in the use of information, especially current information on endogenous variables, 2. be simple to manipulate, and 3. should ‘work well’ in the econometric sense. A fourth criterion, namely, being strongly founded in economic theory is more open to dispute. For example, the time-series forecasting models developed by Box and Jenkins need not rely on economic theory at all. Without pushing the point too far, one could argue that, at least to economists, economic theoretical coutent is, ceteris paribus, a desirable property of a forecasting model. In an important study, Hall (1978) 1 All future references to Hall are to the same paper. has provided a model for forecasting US consumption which satisfies all the criteria above (including the last) to a large extent. The main purpose of this study is to examine the extent to which Hall' method can be successfully used in the UK. In section I below, Hall' method is outlined and compared with some currently available UK models. In section II, results for the UK are presented. Section III offers some concluding remarks and observations.