An Evaluation of Revisions and Quality Aspects of Austrian Quarterly GDP Publications
This study presents results of a revision analysis of the Austrian Quarterly National Accounts using data between 1999 and 2008. In order to analyse different aspects of the revision pattern, a broad set of indicators is calculated and evaluated on the basis of a real time data set. Considering QNA data, adjusted and non-adjusted for working day and seasonal effects, in nominal and real terms, we find that mean revisions of GDP growth rates are unbiased over all observed revision windows. Revisions to aggregate GDP series are generally found to be smaller than those of its components. Some GDP aggregates (especially foreign trade and investment components) show relatively large and biased revisions. We also observe that changes in the release calendar and the introduction of flash estimates did not affect the reliability of first GDP estimates. In contrast, revisions were observed to decrease continuously over the sample period.