Economic Issues in Bell System Divestiture: A Bootstrap Application

Some of the issues that were debated at the time of the largest court‐ordered divestiture in the history of the Bell System in the United States were economies of scale, allocative efficiency of inputs, and overcapitalization. This paper applies statistical methods including Efron's (1979, 1982, 1985) bootstrapping to study these issues by utilizing data prior to divestiture. For this purpose we utilize least squares and ridge estimates of a translog production function to evaluate empirically the null hypotheses associated with these issues in terms of elasticity‐type functions. A special feature of this study is the utilization of the bootstrap resampling technique which relaxes the traditional normality assumption of regression errors in the construction of confidence intervals. Based on the bootstrap confidence intervals we conclude that economies of scale did exist for the Bell System and that the system did not resort to overcapitalization according to the Averch‐Johnson paradigm. We conjecture that the Bell System overspent on labour input in the form of human capital. These conclusions are consistent with some earlier studies, particularly those of Vinod (1972) and Sudit (1973).