This study examines the determinants of regulatory efficiency in the U.S. Environmental Protection Agency's Toxic Substances Control Act (TSCA) Section 5 New Chemicals Review process, focusing on why review durations vary substantially across cases that reach a “not likely to present an unreasonable risk” determination.
The analysis combines case-level chemical risk attributes, program-level institutional indicators (staffing, budget, workload, leadership) and inferred latent organizational factors for new chemical cases reviewed between 2016 and 2024. Regulatory efficiency is operationalized as review duration and analyzed using a Random Forest classifier with SHAP-based explainability to capture nonlinear interactions and heterogeneous effects across review-time categories.
Short-duration reviews are primarily driven by low chemical hazard, persistence and exposure profiles, indicating risk-proportional decision-making. Longer and mid-range reviews are less strongly explained by chemical risk and are increasingly associated with workload pressures, timing and internal coordination. Resource increases alone do not consistently shorten review times, suggesting institutional and organizational frictions play a significant role.
The analysis is limited to publicly disclosed §5(a)(3)(C) cases, which may differ from confidential determinations. Findings are descriptive and highlight areas for future causal and mixed-methods research using restricted-access data.
Improving regulatory efficiency requires management-focused reforms such as enhanced triage, workflow transparency and internal coordination alongside scientific rigor and adequate resourcing.
The study provides one of the first quantitative, explainable machine-learning analyses of TSCA §5 review timelines, empirically operationalizing regulatory “black box” dynamics through a co-production framework that integrates scientific, institutional and organizational dimensions.
