ABSTRACT
This chapter brings out the principles involved in short-term economic forecasting. Given the detachment of economic theory, it is not surprising that forecasting should have developed autonomously, according to the practical needs it was designed to meet and the factual data made available for it. Since the quantities are aggregates, an econometric model has a minute number of equations compared with that, scarcely conceivable, required for the system of a Leon Walras or a Vilfredo Pareto. An econometric model is a bad predictor if and only if the ex post substitution of true values for the exogeneous variables yields a wrong "prediction". In the method of leading indicators, the relations between the presages and the phenomena are not supposed to be necessary. Studies of consumer intentions are now made in great number. In George Katona's opinion, an "index of consumer attitudes", derived from answers to a series of questions, is of greater use in forecasting than are announcements of intentions.
