Based on empirical findings on the effects of cluster policies in Germany, this paper scrutinises the available knowledge on cluster policies impact. There is a growing body of insights on direct effects of policy measures on cluster actors, cluster organisations and innovation networks of the promoted clusters. For some industries such as biotechnology, there are indications that cluster policies had a substantial influence on the formation of new firms and emerging sectoral structures. While the available information seems to support the hypothesis that cluster policies can provide positive impulses for the development of clusters, the actual knowledge on far-reaching impacts of cluster policies on economic structures and processes is still rather limited. The paper asks for the reasons of this knowledge gap between expectations placed in cluster policies and the available evidence on their impact. We identify five reasons: (i) problems in addressing the systemic nature of cluster policy interventions and their effects, (ii) deficiencies regarding the methodologies used, (iii) a lacking informational basis, (iv) practical contexts (e.g., a lack of interest of policy makers) leading to deficiencies in incentive mechanisms and (v) the limited transferability of evaluation results to other cluster policy contexts. For future evaluations, we propose among others the use of system-related approaches to impact analyses based on mixed-method designs as well as comparative case studies based on new methods like process tracing. In order to improve the incentives for evaluators, an increasing awareness of policy makers about the relevance of evaluation studies would be important.