PERFORMANCE 'EVALUATION OF ELECTRICAL UTILITIES

Abstract

Performance evaluation becomes easy when only the desirable inputs and desirable outputs are considered. Moreover these inputs and outputs must be controllable. However in the real time scenario this is not the case. In addition to these desirable and controllable variables there are some undesirable variables as well as uncontrollable variables. There will be a lot of variation in the result when these undesirable variables and uncontrollable variables are taken into account. Some organizations simply neglect these undesirable variables to avoid complexity in the performance evaluation. But the results for those evaluations may not be accurate and it may lead them in a wrong way and as a result their performance will be affected. Also if the uncontrollable variables are neglected, the results will deviate much leading towards wrong decision. Here, we investigate the relative operating efficiencies for a set of electric distribution as well as generating companies in India considering both the undesirable as well as uncontrollable variables. In this study a set of 41 discoms operating throughout the country is taken and their relative performance was evaluated based on the models developed with BCC, CCR and ADDITIVE which incorporates both undesirable and uncontrollable variables. Similarly performance evaluation was done for the gencoms. But, as for as gencoms are considered, since there are no such undesirable and uncontrollable variables, analysis becomes very easy and it is evaluated by a simple method. For this analysis data were collected from the gencoms and discoms official websites. All the three model results were compared and a basic rule that a DMU is said to be efficient if and only if it is both CCR and BCC efficient. If this condition fails then that DMU was declared to be inefficient. This study not only contributes for the performance measurement methodology, but also has policy implications for the Indian power sector. Given difficulties in both model construction and data availability, very few published papers simultaneously consider the above two problems.