A Multilevel Regression Analysis of Municipal Infrastructure Asset Management System Adoption in South Africa

Effective management of municipal infrastructure assets is critical for service delivery and economic development. Despite policy imperatives, the adoption of formal asset management systems across local governments remains inconsistent and poorly understood, hindering strategic investment and maintenance. This study quantifies the adoption rates of structured infrastructure asset management systems and identifies the key municipal-level and provincial-contextual factors that influence their implementation. A multilevel regression model was employed, nesting municipalities within provinces. Data were collected via a national survey of municipal engineers and analysed using a Bayesian framework. The core model is specified as yᵢⱼ = β₀ + β₁Xᵢⱼ + uⱼ + eᵢⱼ, where uⱼ ~ N(0, σ²ᵤ) represents provincial random effects. Robust standard errors were calculated. Adoption rates are low, with only 34% of municipalities operating a fully compliant system. Provincial contextual factors explained a significant portion of the variance (ICC = 0.28). A municipality's internal technical capacity was the strongest positive predictor (β = 0.67, 95% CrI [0.42, 0.91]), whereas fiscal constraints showed a significant negative association. The adoption of asset management systems is strongly influenced by a combination of internal municipal capability and the broader provincial governance environment, indicating that isolated interventions are unlikely to succeed. National policy should be supported by targeted provincial programmes that build technical capacity. Funding mechanisms must be coupled with mandatory reporting against standardised asset management metrics to ensure accountability. asset management, municipal infrastructure, multilevel modelling, regression analysis, governance, engineering systems This paper provides the first national, quantitative analysis of asset management system adoption using a multilevel statistical framework, isolating the distinct effects of municipal and provincial determinants.