Decomposition of inefficiencies in a vertically integrated, hydro-dependent electricity sector: A two-stage DEA–Tobit framework

We examine sector-level efficiency patterns within Zambia's hydro-dependent, vertically integrated electricity sector using annual national-level data from 1990 to 2024. A two-stage Data Envelopment Analysis (DEA)–Tobit framework is employed to decompose overall technical, pure technical, and scale efficiency. The results indicate that scale inefficiency constitutes the dominant constraint, whereas pure technical inefficiency remains limited. This pattern is descriptively consistent with persistent structural capacity constraints relative to system demand. Second-stage Tobit estimates report conditional associations between efficiency scores and contextual variables. Several counterintuitive associations emerge, particularly regarding rural electrification, fossil-fuel use, and regulatory quality. These findings document persistent system-level efficiency patterns and their statistical correlates over the long horizon within a vertically integrated electricity sector. The framework provides a replicable benchmarking approach for comparable hydro-dependent, vertically integrated electricity sectors in developing-country settings. • Scale inefficiency dominates Zambia's hydro-dependent electricity sector. • Pure technical inefficiency remains limited over 1990–2024. • Rural electrification shows counterintuitive associations with efficiency scores. • Regulatory quality is negatively associated with CCR and BCC efficiency. • A two-stage DEA–Tobit framework supports sector-level benchmarking.