Each policy process is in need of knowledge. Policy research is supposed to provide it, but it does so only in part. One of the reasons is that a policy researcher tends to produce new knowledge, although the need is more general. Knowledge management is considered to be a necessary prerequisite to nourished policymaking. Apart from the production of knowledge it includes activities like translation, structuring, interpretation, and so on, of both existing and newly produced knowledge elements. The primate of the two-cultures approach is rejected. Knowledge types are articulated, rules for handling knowledge are explored. A framework is developed for harmonizing the requirements of both pol icymaking and research.