Die Dis cus si on Pape rs die nen einer mög lichst schnel len Ver brei tung von neue ren For schungs arbei ten des ZEW. Die Bei trä ge lie gen in allei ni ger Ver ant wor tung der Auto ren und stel len nicht not wen di ger wei se die Mei nung des ZEW dar.Dis cus si on Papers are inten ded to make results of ZEW research prompt ly avai la ble to other eco no mists in order to encou ra ge dis cus si on and sug gesti ons for revi si ons. The aut hors are sole ly respon si ble for the con tents which do not neces sa ri ly repre sent the opi ni on of the ZEW.Download this ZEW Discussion Paper from our ftp server:ftp://ftp.zew.de/pub/zew-docs/dp/dp07067.pdf
Non-Technical SummaryThis paper first provides a survey of methods for producing medium-term projections that are used by governmental bodies in the major industrial countries and by international institutions. Among the many techniques, the production function approaches which explicitly relate trend output to capital and labor input as well as to technology are by far most commonly used.In the second part of the study it is assessed whether projections from a widely employed version of the production function approach convey a reliable view about future economic developments. For this purpose an out-of-sample forecast exercise based on quarterly data from National Accounts is carried out and a framework that allows evaluation of forecast errors based on consistent test statistics is developed. Empirical implementation of the proposed test strategy is straightforward and conducted for three to five year cumulative projections of GDP growth for the G7 countries. In addition, available projections from national and international sources are included in the forecast evaluation.The evaluation of the forecast errors of the out-of-sample experiment for the observation period from 1985 to 2005 highlights the following: The production function approach with its assumption that output gaps close over medium-term forecast horizons yields unbiased projections of real GDP growth for most countries, but misses other important features of actual GDP developments. Root mean squared errors and mean absolute errors are large and there are long periods where the forecasts are not in line with actual GDP growth. However, projections from the production function approach are generally capable of beating naïve forecasts in terms of root mean squared error although the differences in accuracy are mostly not statistically significant. Nevertheless, this is a remarkable result considering the fairly long forecast horizons.The pseudo forecasts from the out-of-sample exercise serve as a sort of "status quo" or neutral benchmark which incorporates an assessment of the future economic outlook if factor contributions and total factor productivity follow past trends. The comparisons of these pseudo forecasts with projections from official authorities shows that the German government's and the IMF's future assessments of economic developments, in particular, tended to deviate systematically fro...