This paper reports results on building transfer function models with linear combinations of quick indicators as inputs for very short-term prediction of the monthly time series of the volume of industrial production in Finland. The number of input variables in the transfer function models is reduced in two alternative ways: by replacing the original indicators by their two first principal components and by omitting certain indicators. The prediction accuracy of the transfer function models is checked outside the sample and found superior to that of corresponding ARIMA models. Neither of the two ways of reducing the number of input variables leads to consistently more accurate forecasts than the other. It is also found that the prediction accuracy of the transfer function models compares rather favourably with the preliminary values of the volume of industrial production published by the Central Statistical Office during the periods of rapid growth. KEY WORDS Estimation-principal components Time series-ARMAX APPL-MACRO: Scandinavia APPL-MACRO: Output Business indicators-methodology Comparative methods-estimation regression This paper reports results o n constructing a quantitative prediction model for quick forecasting of the present and the very near future of the monthly volume of industrial production in Finland using time series with short publication lags. The first preliminary value of the monthly volume of industrial production is available from the Central Statistical Office (C.S.O.) only after a lag of 2 months and, for a number of reasons, shortening this lag would be desirable.
The nature of the problem rules out structural models for the volume of industrial production, as at least some of the necessary time series are published no faster than industrial production itself. The potentially useful time series naturally have to be related directly o r indirectly to the industrial production, but they have to be published quickly. Models based on these variables, which will be called indicators, will unavoidably be non-structural, as the criteria for choosing the time series emphasize quick publication and by-pass most economic theory.
The simplest models in this work are merely based on the past values of the volume of industrial production itself. They are used to check the prediction performance of more complicated models