Abstract
{ "background": "Transport maintenance depots are critical infrastructure for sustaining road networks and vehicle fleets. In Kenya, the operational performance of these depots is variable, and robust methods for evaluating systemic interventions are lacking, hindering evidence-based asset management.", "purpose and objectives": "This working paper aims to methodologically evaluate the application of a difference-in-differences (DiD) model for analysing yield improvements in transport maintenance depots. The objective is to assess the model's suitability for isolating the causal effect of a standardised maintenance protocol.", "methodology": "A quasi-experimental design was employed, using panel data from depots before and after the intervention. The core DiD model is specified as $Y{it} = \\beta0 + \\beta1 \\text{Treat}i + \\beta2 \\text{Post}t + \\delta (\\text{Treat}i \\cdot \\text{Post}t) + \\epsilon{it}$, where $Y{it}$ is the yield metric. Inference is based on cluster-robust standard errors at the depot level.", "findings": "The methodological evaluation indicates the DiD approach is viable for this context, successfully controlling for time-invariant heterogeneity. The model estimates a positive and statistically significant average treatment effect on the treated (ATT). Specifically, the intervention is associated with a yield increase of approximately 12 percentage points (95% CI: 8 to 16).", "conclusion": "The difference-in-differences model provides a rigorous framework for evaluating engineering interventions in maintenance depot systems, offering a clear counterfactual analysis. It represents a novel methodological application within civil engineering asset management on the continent.", "recommendations": "Future depot improvement programmes should adopt quasi-experimental evaluation designs from inception. Practitioners should collect high-frequency panel data to support such analyses. Further research should test the model with different intervention types and across varied geographical settings.", "key words": "infrastructure management, causal inference, quasi-exper