Growing interest in the institutionalization of evaluation in the public administration raises the question as to which institutional arrangement offers optimal conditions for the utilization of evaluations. "Institutional arrangement" denotes the formal organization of processes and competencies, together with procedural rules, that are applicable independently of individual evaluation projects. It reflects the evaluation practice of an institution and defines the distance between evaluators and evaluees. This article outlines the results of a broad-based study of all 300 or so evaluations that the Swiss Federal Administration completed from 1999 to 2002. On this basis, it derives a theory of the influence of institutional factors on the utilization of evaluations. (Contains 3 figures and 3 tables.)