Modeling a nonradial measure in nonconvex global technology for banking sector: Evidence from commercial banks in China

This study proposes a nonconvex global slack-based directional distance function that incorporates a nondiscretionary input to improve the accuracy of bank efficiency measurement. First, the model treats fixed asset depreciation as a nondiscretionary input, thereby capturing its influence on the production possibility set. Second, the nonconvex global technology assumption is adopted to avoid infeasible input–output combinations across periods and to better reflect technological heterogeneity. Third, the global Malmquist index and its decomposition components are employed to assess efficiency changes over the sample period and identify their key drivers. The results show that the overall efficiency of the 16 sampled banks declined in the latter part of the study period, with widening disparities indicating an increasingly polarized banking sector. During the observation period, pure efficiency change exhibited a fluctuating downward trend, while scale efficiency change showed moderate improvement. These findings suggest that improving productivity in China's commercial banks requires more than internal technical adjustments. Policymakers should focus on encouraging leading banks to engage in continuous fintech and management innovation, while promoting the sector-wide adoption of proven strategies.