Evaluating Blue Chip forecasts of the trade-weighted dollar exchange rate

摘要

Existing studies examining exchange rate expectations have used data from surveys which ask participants to provide their forecasts in, for example, 3 months, 6 months, 12 months and so on. This study contributes to the literature by evaluating the Blue Chip quarterly forecasts of trade-weighted dollar exchange rates collected as 3-month averages. As such, the actual rates (against which we evaluate the forecasts) are quarterly averages instead of the end-of-period figures utilized by previous studies. Our findings for 1989-2008 reveal that forecast accuracy improves with a reduction in lead time. The forecasts, however, display a Topically Oriented Trend Adjustment (TOTA) behaviour and thus fail to be forward-looking. Further evidence indicates that Blue Chip forecasts are unbiased but, in general, fail to outperform those of the random walk in terms of predictive information content and directional accuracy. From a more practical perspective, Blue Chip forecasts are generally unable to accurately predict directional change and are thus of no value to a user.