A Quasi-Experimental Evaluation of Power-Distribution System Reliability in Rwanda, 2000–2024

"background": "Reliable power distribution is critical for economic development, yet many low-income nations face persistent challenges with system performance. There is a paucity of longitudinal, methodologically robust evaluations of distribution network reliability in such contexts, limiting evidence-based infrastructure investment. ", "purpose and objectives": "This short report aims to methodologically evaluate the reliability of power-distribution equipment systems using a quasi-experimental design. The primary objective is to quantify the causal impact of targeted equipment upgrades on system reliability metrics. ", "methodology": "A quasi-experimental difference-in-differences design was employed, comparing reliability indices in treatment and control regions before and after a large-scale intervention. The core statistical model is Y{it = \0 + \1 (Treati \ Postt) + \ + \ +, where Yit is the System Average Interruption Duration Index (SAIDI) for region i in period t. Robust standard errors were clustered at the regional level. ", "findings": "The intervention significantly reduced SAIDI. The coefficient \1 was estimated at -12. 7 hours per customer per year (95% CI: -18. 3, -7. 1), indicating a substantial improvement in reliability. This represents a 34% reduction in outage duration compared to control regions. ", "conclusion": "Targeted upgrades to distribution infrastructure can yield statistically significant and practically meaningful improvements in system reliability. The quasi-experimental approach provides a robust framework for evaluating engineering interventions in real-world settings. ", "recommendations": "Utilities should adopt similar quasi-experimental methodologies for programme evaluation. Investment planning should prioritise regions with baseline reliability indices above the national median, where marginal returns are likely highest. ", "key words": "power distribution, reliability engineering, quasi-experimental design, difference-in-differences, infrastructure evaluation", "contribution statement": "This paper provides a novel application of a quasi-exper