This paper presents an empirical test of the anticompetitive effects of joint ownership by examining the operation of three nuclear plants in Sweden. Since maintenance is the main conduit explaining variation in output, I formulate a model
of optimal maintenance allocation given three behavioral assumptions: i) maximal
collusion, where all owners’ profits on both nuclear and other output are jointly maximized; ii) Cournot competition, where the majority owners’ profits on both nuclear and other output are maximized; and iii) a divested solution, where all owners’ profits on nuclear output are maximized, but no weight is given to non-nuclear output. The behavior that fits the data best is a “hybrid” model where maximal collusion is only achieved during periods of the year when regulatory oversight is less strict. During the remainder of the year, data is instead most consistent with the divested solution.