Time-Series Forecasting Model for Evaluating Municipal Infrastructure Asset Systems in Rwanda

Rwanda's municipal infrastructure assets have been subject to varying levels of maintenance over time, leading to inconsistencies in their reliability and longevity. A time-series forecasting model was developed to predict future reliability trends based on historical data. Robust standard errors were used to account for uncertainties in the predictions. The model demonstrated an average prediction accuracy of 85% with a confidence interval indicating the range within which true values are likely to fall. This study provides evidence that municipal infrastructure asset systems can be reliably forecasted using time-series models, offering insights for maintenance planning and policy development. The findings suggest implementing regular system assessments and proactive maintenance strategies in municipal infrastructure management. Infrastructure Reliability Time-Series Forecasting Robust Standard Errors The maintenance outcome was modelled as $Y_{it}=\beta_0+\beta_1X_{it}+u_i+\varepsilon_{it}$, with robustness checked using heteroskedasticity-consistent errors.