"background": "Industrial machinery fleets in developing economies are critical for productivity, yet systematic, data-driven evaluations of their operational yield are scarce. Existing assessments often lack rigorous econometric frameworks, leading to imprecise estimates of performance improvements. ", "purpose and objectives": "This study aims to develop and apply a panel-data econometric methodology to quantify yield improvement within Senegal's industrial machinery sector. The primary objective is to isolate the effect of systematic maintenance and upgrade programmes from other operational factors. ", "methodology": "A novel unbalanced panel dataset was constructed from maintenance logs, output records, and operator reports for a fleet of heavy machinery. Yield was modelled as a function of machine age, capital investment, and maintenance intensity using a fixed-effects estimator: Y{it = \ + \1 Ait + \2 Iit + \3 Mit +, where \ denotes machine-specific unobserved heterogeneity. Inference is based on cluster-robust standard errors. ", "findings": "The analysis reveals a statistically significant positive relationship between targeted capital investment and yield. A 10% increase in dedicated upgrade expenditure is associated with a 3. 2% yield improvement (95% CI: 1. 8% to 4. 6%), controlling for machine age and standard maintenance. ", "conclusion": "The methodological framework provides a robust tool for evaluating machinery fleet performance. The results confirm that structured capital upgrades, beyond routine maintenance, are a key driver of yield gains in this context. ", "recommendations": "Fleet managers should adopt panel-data tracking systems to inform investment decisions. Policy should incentivise data collection and analysis to optimise national industrial asset management strategies. ", "key words": "panel data, fixed effects, industrial machinery, yield, maintenance, econometrics, asset management", "contribution statement": "This paper introduces a novel panel-data model tailored to the constraints of industrial data in developing economies, providing the
Diallo et al. (Sun,) studied this question.