Nigeria’s electricity supply has persistently lagged behind demand despite sector reform since the Electric Power Sector Reform (EPSR) Act of 2005. This study evaluates the operational efficiency of twenty-four grid-connected electricity generation companies (GenCos) over 2017–2024 using output-oriented Data Envelopment Analysis (DEA). The CCR model of Charnes, Cooper and Rhodes, which imposes constant returns to scale (CRS), and the BCC model of Banker, Charnes and Cooper, which allows variable returns to scale (VRS), are estimated with installed and available capacity as inputs and electricity generated as the output, using a pooled panel of 190 firm-year observations. Returns-to-scale regimes are classified through the CCR–BCC–Non-Increasing Returns to Scale (NIRS) decomposition. Mean CCR efficiency was 0.710, implying a potential output gain of 40.8 per cent through better use of existing assets. Mean scale efficiency of 0.985 and a CRS share of 64.7 per cent confirm that most GenCos already operate near their optimal production scale. A fractional logistic regression finds positive but statistically insignificant average marginal effects for private management (AME = 0.063) and hydroelectric fuel type (AME = 0.072); neither reaches conventional significance given the modest sample of 24 firms. Location and plant age show no material association with efficiency. Nigeria’s generation problem is primarily operational: improvements in plant availability, maintenance, and fuel procurement offer a more direct path to greater output than further capacity expansion.

Keywords: data envelopment analysis, electricity generation, operational efficiency, fractional logistic regression.