Description: Vocational education and training (VET) systems are expected to provide qualifications which are at least to some extent sustainable in the future. Therefore various kinds of approaches, methods, and instruments for anticipation the future development are utilised in VET policies. The paper is analysing effects of different national systems of anticipation of skills demand on indicators about the matching of supply and demand on the labour market. The research question is whether more formalised anticipation systems and/or the regular application of quantitative forecasting are related to differences in matching indicators. The purpose of the analysis is to shed some light on the question whether the quite cost intensive development of regular and formalised anticipation systems might improve the matching of supply and demand on the labour market. The conceptual framework is based on a approach about anticipation systems (Lassnigg/Markowitsch 2005) using the foresight paradigm (Keenan et al. 2003), a classification of anticipation approaches in EU member states (Feijen/Reubsaet 2002, 1996), and concepts of matching of supply and demand on the labour market (Lassnigg 2005). To the knowledge of the author this kind of comparative analysis has not been tried out so far. Therefore it is mainly an explorative exercise. Methodology: The research methodology is based on two cornerstones: - A typology of different anticipation systems is derived from a comparative analysis in EU 15 member states. The typology makes a distinction of a quantitative and a qualitative dimension (quantitative: regular use of quantitative forecasting - yes/no; qualitative: formal/informal/mixed methods of anticipation used for the development of VET programmes. - A set of matching indicators measuring different aspects of matching supply and demand in the labour market is used from European analyses (EC-Enterprise 2002). The types of anticipation systems are related to the matching indicators by descriptive statistics. More elaborate statistical methods (e.g., Kruskal-Wallis one-way analysis of variance, and/or extensions of the median test) will be employed to consolidate the descriptive picture. Conclusions: The question whether the development of VET programmes should be based on formal anticipation procedures, or rather on the market, has been frequently asked. So the idea of a more formal analysis came up in a regional project. The first contention has been that formalised systems and quantitative procedures would not show big effects over more informal or mixed systems. As expected, the descriptive results show small differences between the types. However, interestingly the small differences point rather clearly to the direction that formalised systems might have better matching indicators. Consequently, further analysis seems interesting. First, some other variables (e.g., growth, unemployment level) should be controlled for. Second, a statistical analysis focused on the second error might be interesting.