Improving Business Cycle Forecasts’ Accuracy
17
Table 4 cont.
Test for Information Efficiency Based on a Ranked Sign Test1 1991-2004_____________________________________________________________________________ | |||||||
Indicator |
RWI-7 |
RWI-4 |
RWI-3 |
GD-6 |
GD-4 |
GD-2 | |
GDP |
26* |
21** | |||||
PC |
82,5* | ||||||
Real short term interest rate-1 (RSR-1) |
GC IEQ IS | ||||||
EX | |||||||
IM | |||||||
GDP |
24* |
18,5** | |||||
PC |
81* | ||||||
Real effective exchange |
GC IEQ IS |
16** | |||||
EX | |||||||
IM | |||||||
GDP |
17** |
18,5** | |||||
PC | |||||||
Real effective exchange |
GC IEQ IS |
79,5* |
26* |
16** | |||
EX |
80,5* | ||||||
IM | |||||||
GDP |
3*** |
23,5* | |||||
PC | |||||||
Share price index |
GC IEQ IS |
18** |
25* | ||||
EX |
21** | ||||||
IM |
17** | ||||||
GDP |
3*** |
19** | |||||
PC | |||||||
Share price index3 |
GC IEQ IS |
18** | |||||
EX |
21** | ||||||
IM |
17** | ||||||
Author’s computations. - |
1See eq. |
(10). For abbreviations see text and Annex. |
- Level |
of signifi- | |||
cance: *** 1%; ** 5%; * 10%. |
5. Conclusions
When evaluating their prediction errors, forecasters are in a dilemma to some
extent. If their projections are unbiased and efficient, they have done a good
job. But there is no lesson how to improve future forecasts, with no respect to
the past accuracy. If errors appear to be systematic, this signals that not all in-
formation has been taken into account in an appropriate way. But this offers
the chance to improve the quality in the future. This paper scrutinizes the in-
formation efficiency of German short term forecasts in order to evaluate,
whether past forecast errors provide lessons for the future.