Do mergers really increase output? Evidence from English hospitals

Abstract

The English hospital sector underwent a major restructuring program between 2000 and 2008 to centralize activity in fewer and larger hospitals. The aim of this paper is to evaluate the effects of such consolidations on hospital outputs. As mergers occurred in a staggered way, treatment could start and end at every time and treatment duration varied over the years. As every time is a mix of hospital pre-treatment, treatment and post-treatment phases, the canonical difference-in-differences assumption of homogeneous policy effects is not only meaningless but also misleading, raising doubts about the appropriateness of the methods previously used in this literature and consequently the accuracy of its results. We instead adopt a new matching and difference-in-differences approach, the flexible conditional difference-in-differences approach, developed by Dettmann et al. in 2020, more appropriate for causal analysis of treatments characterized by varying start dates and varying treatment duration. Our results suggest that mergers downsize hospital activities, especially the most expensive ones. If the goal of hospital mergers is to gain efficiency by centralization of activity, our findings suggest this restructuring programme is not the most successful policy to pursue. Mergers reduce the scope for competition between hospitals and do not create any incentive for poorly performing hospitals.