<![CDATA[This study aims to evaluate the relative operational efficiency of Bank Jakarta branches and identify the operational factors influencing branch-level efficiency using a two-stage DEA–Tobit approach. By utilizing exclusive internal branch-level data and a pooled frontier framework, the study provides empirical evidence on operational efficiency at the micro-operational level of a Regional Development Bank during the digital transformation era. This study utilizes a two-stage analytical approach. The first stage employs Data Envelopment Analysis (DEA) with the BCC model (Variable Return to Scale) and an output orientation. To establish a consistent efficiency benchmark, a pooled frontier approach is applied to a total of 224 observations, which are derived from 112 branch offices observed over two annual periods (2024 and 2025). The second stage employs the Tobit Regression model to identify the determinants of efficiency, including the number of employees, branch class, service type, geographic location, branch age, and a year dummy. Bank Jakarta branches currently operate at approximately two-thirds of their potential output, with inefficiency primarily driven by structural scale sub-optimality rather than purely managerial failures. Findings underscore that human capital remains a vital complementary driver of productivity alongside digital services, while branches located in the Jabodetabek area leverage a dominant structural advantage through their proximity to the Jakarta government ecosystem. Conversely, the negative impact of branch age suggests that senior units struggle with organizational inertia and aging infrastructure. These results necessitate a shift from uniform top-down targeting to data-driven benchmarking using DEA-based projections to establish realistic, unit-specific performance targets and address the network's significant performance gaps.]]>
