Published November 12, 2025 | Version v1

A Difference-in-Differences Framework for Evaluating Manufacturing Systems Risk Reduction in South Africa (2000–2026)

  • 1. Department of Mechanical Engineering, University of the Witwatersrand
  • 2. Department of Sustainable Systems, University of the Free State
  • 3. Department of Mechanical Engineering, University of the Free State
  • 4. Department of Civil Engineering, University of Pretoria

Description

{ "background": "The South African manufacturing sector faces persistent systemic risks, including infrastructure instability and supply chain disruptions. Existing engineering risk assessment methods often lack robust counterfactual analysis, making causal attribution of intervention efficacy difficult.", "purpose and objectives": "This article presents a methodological framework for the causal evaluation of engineering interventions aimed at reducing systemic risk in manufacturing plants. The objective is to provide a rigorous, quasi-experimental design suitable for longitudinal plant-level data.", "methodology": "We detail a difference-in-differences (DiD) model for panel data. The core specification is $Y{it} = \\alpha + \\beta (Treati \\times Postt) + \\gammai + \\deltat + \\epsilon{it}$, where $Y_{it}$ is a composite risk index. The parallel trends assumption is tested using event-study plots, and inference relies on cluster-robust standard errors at the plant level.", "findings": "The framework, applied to a simulated dataset, demonstrates that a hypothetical intervention targeting energy resilience is associated with a 15% reduction in the composite risk index. The key methodological finding is the critical importance of pre-trend validation; violations of the parallel trends assumption can lead to a 95% confidence interval for the treatment effect that erroneously includes zero.", "conclusion": "The DiD framework provides a statistically rigorous methodology for isolating the causal effect of engineering interventions on manufacturing systems risk, moving beyond descriptive correlation.", "recommendations": "Practitioners should adopt this DiD model for retrospective programme evaluation. Future work should integrate high-frequency sensor data into the risk index and explore synthetic control methods for single-unit case studies.", "key words": "difference-in-differences, causal inference, risk assessment, manufacturing systems, quasi-experimental design, industrial engineering", "contribution statement": "This paper provides the first formalised difference-in-differences econometric framework tailored for evaluating engineering risk reduction programmes in an industrial context, explicitly linking intervention timelines to plant-level panel data."

Files

zenodo.18972332.pdf

Files (239.4 kB)

Name Size
md5:dab4910ecedb2d599017115150b6bbbc
28.4 kB Download
md5:bfdaaed0b3d0e95b7d6cf887c41ed879
211.0 kB Preview Download

Additional details