Replication and Methodological Evaluation of a Difference-in-Differences Model for Industrial Machinery Fleet Cost-Effectiveness in Uganda

{ "background": "Difference-in-differences (DiD) models are increasingly applied in engineering economics to evaluate the impact of interventions, such as fleet management system upgrades. A seminal study applied a DiD framework to assess the cost-effectiveness of industrial machinery fleets in Uganda, but its methodological robustness and replicability in a low-resource setting remain unverified.", "purpose and objectives": "This study aims to replicate the original DiD analysis and conduct a rigorous methodological evaluation of its design, identifying potential biases and assessing the reliability of its cost-effectiveness conclusions for engineering asset management.", "methodology": "We executed a full computational replication using the original specification: Yit = \α + \β (Treati \× Postt) + \ + \ + \εit, where Yᵢₜ is the maintenance cost per operating hour. We then performed a suite of robustness checks, including parallel trends assessment, placebo tests, and estimation with cluster-robust standard errors at the fleet level.", "findings": "The replication successfully reproduced the original point estimate, indicating a 12.7% reduction in average hourly maintenance costs. However, the methodological evaluation revealed that the parallel trends assumption, critical for DiD validity, was not supported by pre-intervention data (p-value < 0.05). Subsequent placebo tests on pre-period data generated statistically significant false effects.", "conclusion": "While the replicated numerical result matches the original study, the methodological evaluation undermines the causal interpretation of the reported cost-effectiveness. The findings are likely confounded by pre-existing differential trends between treatment and control groups.", "recommendations": "Future applications of DiD in engineering economics must rigorously test and report on the parallel trends assumption. We recommend the use of event-study plots and sensitivity analyses as a minimum standard for similar fleet management studies.", "key words": "replication study, difference-in-differences, engineering asset management, cost-effectiveness, methodological evaluation, industrial machinery", "cont