Reading Tea Leaves or Science? Forecasting for Beginners
The apparently low reliability of economic forecasts is at center stage of the current public debate. This note uses a stylized version of the problem of predicting future economic growth rates to introduce the basic forecasting principles and the unavoidable problem of forecasting errors. Furthermore, this note clarifies what a best forecast should be, why this best forecast should be corrected or updated regularly and why such a correction does not render the approach inappropriate. Moreover, we discuss identification assumptions as the formal prerequisite which are necessary to extract sensible forecasts from historical data.
Fertig, M. and C. Schmidt (2002), Reading Tea Leaves or Science? Forecasting for Beginners. AStA Allgemeines Statistisches Archiv, 86, 4, 459-467