"background": "Chronic power outages and unreliable electricity supply present a significant constraint to economic development in many African nations. The governance and technical performance of power-distribution infrastructure are critical, yet analyses often lack robust, quantitative frameworks linking system characteristics to reliability outcomes. ", "purpose and objectives": "This policy analysis aims to evaluate the reliability of Kenya's power-distribution system using a multilevel modelling approach. It seeks to quantify the influence of equipment age, maintenance regimes, and governance structures on failure rates, thereby informing evidence-based infrastructure policy. ", "methodology": "A multilevel regression analysis was conducted on a longitudinal dataset of distribution system performance. The core statistical model is specified as \ () = \0 + \1X{ij + uj +, where is the failure rate for transformer i in region j, Xij are transformer-level covariates, and u₉ are region-level random effects. Inference is based on robust standard errors. ", "findings": "Transformer age exhibited a non-linear relationship with failure probability, with units over 15 years old showing a 34% higher likelihood of failure (95% CI: 22% to 48%) compared to newer assets. Regional governance quality, measured by audit frequency and budget autonomy, accounted for significant variance in reliability outcomes at the second model level. ", "conclusion": "System reliability is determined by a confluence of technical asset condition and higher-level governance factors. Policy focusing solely on equipment replacement without addressing institutional governance gaps is likely to yield suboptimal improvements in service continuity. ", "recommendations": "Policy should mandate integrated asset-management frameworks that couple targeted infrastructure renewal with capacity-building for regional electricity boards. Regulators should adopt performance-based models linking funding to reliability metrics disaggregated by asset age and region. ", "key words": "infrastructure reliability, multilevel modelling, power distribution, asset management, regulatory policy",
Mwangi et al. (Mon,) studied this question.