Context: Vocational education and training (VET) is expected to be designed for creating learning outcomes which meet the needs for skills and competences in the labour market. Hence, identifying current and upcoming skill requirements and ensuring that these requirements are incorporated into education has long been the subject of academic and policy discussion. Governance processes keeping VET systems up-to-date have been more recently addressed as `feedback mechanisms'. The term broadly summarizes the interplay of institutions, actors and processes which allows the continuous renewal of VET provision (i.e. by creating new qualifications or updating curricula). The aim of the paper is to enhance the understanding of cross-national variations in formally institutionalised `feedback mechanisms' between VET and the labour market. Method: The research builds on a comparative analysis of case studies in 15 European countries. The paper presents examples for four different `formal feedback mechanisms' in Germany, France, England, and Austria. Results: Four main types of formal mechanism have been identified: 1) The liberal model explained by VET in England and Higher VET in Austria; 2) The statist model explained by school-based VET in Austria; 3) the participatory model explained by VET in France and 4) the coordinated model explained by apprenticeship training in Germany and Austria. Conclusions: Existing approaches in the economic sociology of labour markets, the varieties of capitalism approach as well as comparative research on welfare states are useful in predicting whether particular VET systems are likely to be predominant. However, they do not provide an alternative in describing differences in VET systems which the concept of formal feedback mechanism does. Moreover, by analysing formal feedback mechanisms, it is possible to demarcate where a VET sub-system ends and another VET sub-systems begins. In this sense research presented here also asks for new standards for comparative VET research as it suggests that entities to be compared are not countries' overall VET systems, but their potential sub-systems.