ON SCORING ASYMMETRIC PERIODIC PROBABILITY-MODELS OF TURNING-POINT FORECASTS

To forecast the turnaround of an economy we do not usually take season al effects into account. Recently, the author showed that business cyc le turning points as well as durations do not appear to be uniformly d istributed throughout the year (see Ghysels, 1991a). In this paper we suggest improving the forecasting performance of turning-point predict ions by adopting periodic hazard models. Following Diebold and Rudebus ch (1989), we construct several formal probability models and score th eir prediction performance. The results indicate that for sequential f orecasting rules significant gains can be made by exploiting periodici ties in turning-point probabilities.