<p>Swedish district courts have undergone a major mergers and acquisitions program between 2000 and 2010 to centralize activity in larger and fewer courts. The purpose of this contribution is to conduct an efficiency analysis of these courts to identify the eventual efficiency gains. Distinguishing mainly between technical and scale efficiency and determining the returns to scale of individual observations, we try to find the potential rationales behind this merger wave. We are to the best of our knowledge the first to combine traditional convex with nonconvex nonparametric frontier methods to calculate efficiency before and after the mergers. It turns out that the nonconvex methods provide a more cogent ex post explanation of this historical merger wave aimed at increasing the size of operations. A battery of recent test statistics rejects convexity in favour of nonconvexity.</p>