Simulating Hospital Merger Simulations

David J. Balan, Federal Trade Commission
Keith J. Brand, Federal Trade Commission

Abstract

In recent years, researchers have developed a number of new methods for predicting the price effects of hospital mergers. Though there are several variants, the basic steps are the same. First estimate a discrete choice model of hospital choices; then use these estimates to generate a hospital-level measure of market power (a large part of the innovation was in creating new market power measures that have certain attractive properties); and then use the market power measure as an independent variable in a hospital price regression. Finally, use the estimated relationship between the market power measure and price to simulate the effects of mergers. In this paper, we seek to test the accuracy of these simulation methods. To do this, we set up a simple model of hospital competition which can, for any given values of the parameters of the model, generate the “true” effects of a merger between any two hospitals. These “true” effects are then compared to the effects predicted by the simulation methods described above. We repeat this exercise 32,400 times and, using each of several market power measures, derive results regarding the conditions under which the simulation method does or does not generate predicted effects that are close to the “truth.” Our preliminary results suggest that the simulation methods slightly under-predict merger effects on average, and that this under-prediction becomes more pronounced as the diversion between the merging hospitals increases.