Forecasting Cost Efficiency of Banks in India with Integrated DDF-Based Network DEA and SVR Approach

Cost efficiency is a significant measure to assess an organization’s ability to utilize the resources to achieve the desired output at the lowest possible cost. The estimation and prediction of the cost efficiency are vital for decision-makers to strategically allocate resources, optimize operations, and forecast future financial health. This enables the identification of inefficiencies, the facilitation of strategic planning, and the enhancement of overall organizational performance. Addressing this necessity, the present study introduces a methodology to estimate the cost efficiency of decision-making units (DMUs) by employing the directional distance function (DDF)-based network data envelopment analysis (DEA) approach. To enhance the predictive capabilities, the methodology integrates a support vector machine for regression (SVR), which significantly reduces the computational resources required for re-executing the network DEA model when new units are added, especially in large and expanding datasets. The study applied this methodology to the Indian banking sector to measure the cost efficiency across two distinct divisions, ‘productivity and profitability’ within the network architecture. The cost efficiency scores were estimated for the banks operational in the financial period 2012–22 and predicted for the latest year 2021–22. The proposed hybrid approach demonstrates exceptional precision and accuracy in predicting cost efficiency scores when compared to actual efficiency scores, thus confirming the robustness and reliability of the integrated proposed methodology.