<![CDATA[Purpose - This study evaluates cost efficiency in PLN electricity distribution by distinguishing managerial inefficiency from structural constraints across 22 distribution units during 2014–2024. Methods - The study applies stochastic frontier analysis with a four-component error structure (SFA-FCE) using panel data. Inefficiency is decomposed into persistent and transient components through a restricted translog cost function estimated in Stata 17 with a customized sfbook module. Findings - The average cost efficiency is 58.74%. Transient efficiency is high at 92.88%, while persistent efficiency is much lower at 63.19%. This indicates that inefficiency is driven mainly by long-term structural constraints rather than short-term managerial performance. The LR test shows that transient inefficiency is statistically insignificant, whereas persistent inefficiency is significant. Regional characteristics, electrification gaps, and customer density significantly influence persistent inefficiency, while service reliability and network losses are not statistically significant. Java distribution units show higher inefficiency despite higher customer density, suggesting diseconomies of density due to network complexity. Research implications - The findings imply that uniform benchmarking across PLN distribution units may be misleading because each unit operates under different structural conditions. Regulators and policymakers should design differentiated policies that account for regional characteristics, electrification gaps, customer density, and network complexity. Originality - This study offers rare empirical evidence on cost efficiency at the PLN distribution-unit level using internal data. It applies the SFA-FCE framework to separate persistent structural inefficiency from transient managerial inefficiency, providing a more reliable basis for electricity distribution benchmarking.]]>
