Time series clustered benchmarking in data envelopment analysis: Attainable efficiency enhancement approach for an entity

Efficiency evaluation and making valiant efforts for improvement are a mandate to be successful, for every entity across various industries and Data Envelopment Analysis (DEA) is a comprehensively and extensively accepted technique for efficiency evaluation and benchmarking of Decision-Making Units (DMUs). Although DEA has been applied in almost all domains, for efficiency evaluation in a single time period as well as in time series analysis, using along with other techniques, but the DEA in its conventional form has certain drawbacks like consideration of a single time period only and assignment of different benchmarks for every time period, which dilutes the purpose of analysis. To consider these problems, this paper suggests a Time Series Clustered Benchmarking (TSCB) DEA, which incorporates conventional DEA models and cluster analysis, to analyse multi-period time series data and thus assignment of benchmarks to DMUs with a similar structural genealogy, to result in achievable and feasible goals, for efficiency enhancement. For functional procedure of the proposed approach, the data related to twenty-four Indian public sector banks (PSBs) has been used, taken as decision making units (DMUs). It has been found that the proposed TSCB DEA approach categorizes the DMUs among nine clusters, identifies benchmark for each cluster separately, thus for each inefficient bank there is a benchmark found, in more effective way.