ein zeitreihenmodell fuer die oesterreichische wirtschaft

abstract: the investigation of time series models for economic data represents an interesting task for the econometric model builder. this paper concentrates on the creation of such a model as well as on a discussion of the fundamental problems involved. answers are searched for technical as well as for theoretical questions. following an introductory overview, chapter 2 starts with the preparation of causality tests. notes about the selection of the economic variables involved and about the transformations of those variables used for the remainder of the paper pave the ground for a thourough discussion of the wiener-granger causality concepts. along these lines, the test procedure is put into practice and the results are displayed in a graphical diagram. chapter 3 is based on the foregoing causality tests and aims for an operational time series model for forecasting. again the background is depicted before the multivariate autoregressive model is represented in all its details. the model is compared with univariate models, it produces forecasts until l990, and it is finally revised using a robust filtering procedure due to martin. the forecasts based on this revised model complete the paper.;