Rejoinder—Natural Monopoly and the Bell System: Response to Charnes, Cooper and Sueyoshi

Our study of the Bell System cost function shows that it is possible to reject the hypothesis that AT&T was a natural monopoly. Our study is based on a regression analysis of a translog cost function estimated on 1947–1977 data. Charnes, Cooper, and Sueyoshi (Charnes, A., W. W. Cooper, T. Sueyoshi. 1988. A goal programming/constrained regression review of the bell system breakup. Management Sci.34 1–26.) claim that they reverse our conclusion when they use goal programming estimators of a translog cost function estimated on exactly the same data that we use. This claim is false. There is no basis for comparing our estimates with their estimates because they, in fact, use different data than we use and estimate a different cost function than we estimate. Moreover, when goal programming estimates and regression estimates based on the same data and similar cost function specifications are compared, they yield similar estimates and produce the same inference about natural monopoly.