Abstract
{ "background": "Municipal infrastructure asset systems in many developing nations face chronic inefficiencies, yet robust, data-driven diagnostic tools for their evaluation are scarce. Existing methods often fail to account for the hierarchical structure of asset data and inherent uncertainties in performance measurement.", "purpose and objectives": "This study aimed to develop and validate a novel Bayesian hierarchical model to diagnose the technical efficiency of municipal infrastructure systems, providing a probabilistic framework for identifying performance drivers and potential gains.", "methodology": "We formulated a Bayesian stochastic frontier model, $\\ln(y{ij}) = \\beta0 + \\beta X{ij} + v{ij} - u{ij}$, where $u{ij} \\sim \\text{Half-Normal}^{+}(0, \\sigmau^2)$ represents inefficiency for asset $i$ in municipality $j$, and $v{ij}$ is stochastic noise. The model was applied to a unique panel dataset of water supply, road, and drainage assets from multiple municipalities.", "findings": "The model identified significant latent inefficiencies, with a posterior probability of 0.92 that the median municipality operates below 65% of its potential technical capacity. Municipal governance quality was the strongest predictor of efficiency, with its effect size having a 95% credible interval of [0.15, 0.31].", "conclusion": "The Bayesian hierarchical model provides a robust diagnostic tool, quantifying efficiency with associated uncertainty. It reveals that systemic underperformance is prevalent and strongly linked to institutional factors rather than purely technical or financial constraints.", "recommendations": "Municipal authorities should adopt probabilistic performance diagnostics to prioritise interventions. National policy should focus on enhancing institutional governance capabilities, as this yields the greatest marginal efficiency gain across multiple asset types.", "key words": "Infrastructure asset management, efficiency analysis, stochastic frontier analysis, Bayesian inference, urban services, developing countries", "contribution statement": "This paper presents the first application of a Bayesian hierarchical stochastic frontier model to multi-asset municipal