Efficiency measurement and regulation in US telecommunications: A robustness analysis

The paper investigates the robustness of different efficiency measures that can support the implementation of diverse forms of incentive regulation in the context of US telecommunications. Comparisons, in terms of an output orientation, are considered for efficiency scores obtained from data envelopment analysis (DEA), distance function (with corrected ordinary least squares and a random effects model) and distance function embedded in a stochastic frontier framework (with time-invariant, time-varying efficiencies or with inefficiency effects). Similar to the previous empirical literature, one finds, in most cases, only a moderate consistency across the different approaches. In fact, the different spectrum of techniques imposed varied degrees of structure in the error term and indicated non-negligible discrepancies across the different measurement approaches in terms of the ranking structure, degree of persistence and best- and worst-practice patterns.