"background": "Chronic power outages and unreliable electricity supply remain significant impediments to economic development in many African nations. In Tanzania, governance of the power-distribution infrastructure is a critical policy concern, yet empirical analyses linking equipment-level data to system-wide reliability are scarce. ", "purpose and objectives": "This policy analysis evaluates the effectiveness of current infrastructure governance by quantifying the determinants of power-distribution system reliability. The objective is to provide an evidence-based methodological framework for prioritising maintenance and investment. ", "methodology": "A multilevel regression analysis is employed, modelling failure rates across hierarchical data: transformers and circuit breakers (level 1) nested within regional grids (level 2). The core statistical model is \ () = \0j + \1X{1ij + eij, with \0j = \00 + \01Z1j + u0j. Robust standard errors are used for inference. ", "findings": "Equipment age and lack of scheduled maintenance were the strongest predictors of failure. A one-year increase in transformer age was associated with a 7. 3% increase in failure rate (95% CI: 5. 1% to 9. 5%). Regional disparities in technical capacity accounted for 31% of the variance in system reliability. ", "conclusion": "System reliability is predominantly driven by equipment-level factors exacerbated by uneven regional governance capacity. Current policy frameworks are insufficiently granular to address these multilevel challenges. ", "recommendations": "Policy must mandate data-driven, condition-based maintenance schedules. Infrastructure governance should be decentralised, with resources allocated based on predictive risk models and regional capacity-building programmes established. ", "key words": "infrastructure governance, power distribution, reliability engineering, multilevel modelling, predictive maintenance, Tanzania", "contribution statement": "This study provides a novel, hierarchical modelling framework for power-system reliability analysis in a Tanzanian context,
Kavishe et al. (Wed,) studied this question.