Abstract:
This paper investigates the existence of any systematic relationship between preliminary estimates and subsequent revisions of GDP growth rates in Macedonia. Accordingly, we use various statistical tools for testing the "news" and "noise" hypotheses and empirically assess if GDP data revisions are unbiased and efficient and vice versa. The results based on different empirical approaches in general provide mixed results about the predictability of GDP revisions, although the empirical evidence tends to incline that the long-term and final GDP revisions are efficient and contain new information that is in line with the "news" hypothesis. We conclude that each subsequent revision is unpredictable.