Methodological Evaluation and Panel-Data Estimation for Municipal Infrastructure Asset Yield in Rwanda, 2000–2026

"background": "Municipal infrastructure asset management in developing economies often lacks robust, data-driven methodologies for performance forecasting. Existing approaches frequently rely on cross-sectional data, failing to capture temporal dynamics and unobserved heterogeneity, which limits the accuracy of long-term yield projections for critical engineering assets. ", "purpose and objectives": "This study aims to methodologically evaluate panel-data estimation techniques for modelling the yield of municipal infrastructure assets. The primary objective is to develop and validate a model that quantifies improvement in asset performance over time, providing a tool for strategic investment and maintenance planning. ", "methodology": "A balanced panel dataset for key municipal asset classes was constructed. The core analytical model is a two-way fixed effects specification: Y{it = \ + \ Xit + \ + \ +, where Yit is the yield measure for asset i in period t. Estimation employed robust standard errors clustered at the asset level to account for heteroskedasticity and serial correlation. ", "findings": "The fixed effects model explained 74% of the within-asset variation in yield. A key result is that targeted maintenance expenditure had a statistically significant positive coefficient (0. 18, 95% CI: 0. 12 to 0. 24), indicating a strong, direct association with yield improvement. The model identified a clear positive temporal trend in aggregate asset performance. ", "conclusion": "Panel-data methods, specifically the fixed effects estimator, provide a superior methodological framework for analysing infrastructure asset yield by controlling for time-invariant unobserved characteristics and common temporal shocks. This leads to more reliable estimates of the drivers of performance improvement. ", "recommendations": "Asset management authorities should adopt panel-data models for performance benchmarking and forecasting. Future research should integrate granular climate and usage data to refine the model's predictive capacity for specific asset types. ", "key words": "infrastructure asset management, panel data, fixed effects model, yield forecasting, municipal engineering, performance