Estimating Time-Varying Technical Inefficiency for a Panel of Telecommunications Operators: A Distance Function Approach

Estimators of relative efficiency commonly employed for regulatory purposes treat efficiency as fixed over time or constrain its time pattern to a given function. This paper aims to explore empirically the robustness of these assumptions. With this objective, the study compares results from standard panel estimators with estimators that allow for efficiency to vary freely over time. Evidence from a panel of 27 U.S. telecommunications operators over the period 1990 - 2003 shows significant variation of efficiency over time, suggesting that an estimator which constrained such variation could potentially produce misleading results. Results also confirm that standard panel estimators tend to overestimate inefficiency.