Methodological Evaluation of Power-Distribution Equipment Systems in Kenya: A Difference-in-Differences Model for Risk Reduction (2000–2026)

"background": "Power-distribution infrastructure in many developing nations faces significant reliability and safety challenges. Ageing equipment and inadequate maintenance regimes contribute to elevated technical losses and public safety risks, necessitating robust methodological frameworks for evaluating intervention programmes. ", "purpose and objectives": "This study aimed to develop and apply a quasi-experimental econometric model to quantify the causal impact of a nationwide equipment modernisation programme on technical risk metrics within the Kenyan power grid. ", "methodology": "A difference-in-differences (DiD) model was employed, leveraging phased rollout data. The core specification was Y{it = \0 + \1 + \2 + \ (\) +, where Yit is the risk index for substation i in period t. Inference was based on cluster-robust standard errors at the regional level. ", "findings": "The modernisation programme caused a statistically significant reduction in the composite risk index. The average treatment effect on the treated (ATT) was -0. 18 units (95% CI: -0. 25 to -0. 11), equivalent to a 22% reduction relative to control groups. The parallel trends assumption was validated using pre-intervention data. ", "conclusion": "The methodological approach confirms the efficacy of systematic equipment upgrades in mitigating operational risks. The DiD model provides a rigorous framework for isolating programme effects from secular trends. ", "recommendations": "Utilities should adopt similar quasi-experimental evaluation methods for capital project appraisal. Policy should prioritise sustained investment in infrastructure renewal, informed by causal impact analysis rather than observational comparisons. ", "key words": "difference-in-differences, power distribution, infrastructure risk, causal inference, quasi-experimental design, technical losses", "contribution statement": "This paper provides a novel application of the DiD model for causal