Reducing overhead is a target for governments across Europe that are looking for ways to economize. This study contributes to our understanding of overhead levels in different types of agencies. Regression and fuzzy set Qualitative Comparative Analysis (fsQCA) are applied to examine the relationship between an agency's overhead level and its formal autonomy, result control, size, and task. The results support the claim that greater formal autonomy leads to a higher overhead, especially when combined with performance-based target setting or when frequent monitoring is absent. The relationship, however, is nonlinear. Distinct effects for the different subdimensions of result control are found. Agency size and task mediate the relationship between formal autonomy, result control and overhead, but only in specific combinations.