Estimation and Identification of Merger Effects: An Application to Hospital Mergers

Advances in structural demand estimation have substantially improved economists' ability to forecast the impact of mergers. However, these models rely on extensive assumptions about consumer choice and firm objectives, and ultimately observational methods are needed to test their validity. Observational studies, in turn, suffer from selection problems arising from the fact that merging entities differ from non-merging entities in unobserved ways. To obtain an accurate estimate of the effect of consummated mergers, I propose a combination of rival analysis and instrumental variables. By focusing on the effect of a merger on the behavior of rival firms, and instrumenting for these mergers, unbiased estimates of the effect of a merger on market outcomes can be obtained. Using this methodology, I evaluate the impact of independent hospital mergers between 1989 and 1996 on rivals' prices. I find sharp increases in rivals' prices following a merger, with the greatest effect on the closest rivals. The results for this industry are more consistent with predictions from structural models than with prior observational estimates.