"background": "The reliability of manufacturing plant systems is critical for industrial productivity and economic development. In Kenya, a lack of robust, quantitative methodologies for evaluating system performance and the impact of interventions has hindered evidence-based maintenance and investment decisions. ", "purpose and objectives": "This study aims to develop and apply a rigorous econometric framework to evaluate the reliability of manufacturing systems. The primary objective is to quantify the causal effect of a systematic maintenance programme on plant operational uptime. ", "methodology": "A quasi-experimental difference-in-differences (DiD) model is employed, using panel data from multiple plants. The core specification is Y{it = \ + \ (Treati \ Postt) + \ + \ +, where Y₈ₓ is the monthly uptime percentage for plant i in period t. Robust standard errors are clustered at the plant level to ensure valid inference. ", "findings": "The intervention significantly increased system uptime by an average of 7. 3 percentage points (95% CI: 5. 1 to 9. 5). This effect was persistent and showed no evidence of diminishing over the observation period. The parallel trends assumption, tested via event-study analysis, was satisfied prior to the intervention. ", "conclusion": "The DiD approach provides a statistically robust method for assessing engineering system reliability, isolating the causal impact of specific interventions from confounding temporal trends. The results demonstrate the substantial operational benefits of structured maintenance programmes. ", "recommendations": "Manufacturing plant managers should adopt similar quasi-experimental evaluation frameworks for capital project appraisals. Policymakers are encouraged to support the collection of standardised, high-frequency operational data to facilitate such analyses industry-wide. ", "key words": "system reliability, difference-in-differences, causal inference, maintenance engineering, manufacturing, quasi-experimental design", "contribution statement": "This paper provides the first application of a difference-in-differences model to isolate
Ochieng et al. (Mon,) studied this question.