Toxic Rules: Chemical Regulation, International Trade, and the Epistemic Consequences of Standardized Practices

Abstract

This dissertation is a study of regulatory epistemology: the practices of managing knowledge, expertise, and evidence in regulatory decision-making. Using chemical regulation as a case study, it examines science at the intersection of commodity trade, regulation, and safety. I demonstrate how the desiderata of bureaucratic efficiency and free-flowing trade have shaped how the U.S. and the Organisation for Economic Co-operation and Development (OECD) developed norms and practices to know and regulate risk.

Since 1900, over 10 million new chemicals have been synthesized; 150,000 have been used in plastics, drugs, foods, and pesticides. Many are known to be harmful. Yet regulatory action has generally been weak and slow, despite robust evidence demonstrating the environmental and health consequences of chemicals in everyday products. This dissertation investigates why many governments have failed to regulate chemicals widely considered toxic: because regulatory agencies including the U.S. Food and Drug Administration (FDA) and Environmental Protection Agency (EPA) have relied on partial knowledge. They have done so to make chemicals governable and to facilitate international trade, reflecting the fact that the epistemic demands of governance and market capitalism are different from those of public health and environmental protection.

In the 1970s, fraudulent studies submitted to regulatory agencies cast doubt on the integrity of chemical safety data—and hence the safety of approved drugs, pesticides, and other chemical products. U.S. regulators formulated quality assurance and reporting standards called Good Laboratory Practices (GLPs). The rules were ostensibly adopted to assure the integrity of toxicology data. This dissertation shows that, at least in the U.S., government officials knew they would never achieve this goal. The aim was not to make laboratory practices and the resultant knowledge good, but rather to make them governable.

As part of the goal of governance, a question that was previously understood as the domain of science—“how do we determine if this chemical is safe?”—became reframed as a matter of economic policy and international governance. In 1981, the twenty-four industrialized nations of the OECD adopted GLPs and Test Guideline protocols as part of their Mutual Acceptance of Data framework. The OECD hoped to protect public safety by ensuring data quality, but their primary aim was to facilitate international trade by creating uniform standards. Namely, the OECD hoped to avoid trade barriers due to conflicting national rules, which would otherwise require companies to test substances multiple times, impeding economic growth. The standards achieved those narrow aims. But over time, reliance on them has restricted the evidence for regulatory action to a subset of studies: standardized studies, generally performed and funded by industry. Most academic studies have been excluded. In many cases, this has meant ignoring evidence from precisely those studies that suggest a need for stronger regulation.

This dissertation highlights a problem with science in regulation writ large: a weakness in how the state knows. Regulatory bodies have invoked procedural fairness and rational management as defenses against charges of public interference in private affairs. Those rationales, in turn, informed how regulators know. The way that they do so—based on standard protocols and credentialed experts—benefits business and government. But does it make scientific sense, and does it protect people or the environment? I argue that the answers are “no.” These frameworks constrain the evidence and expertise used to govern by impeding the uptake of knowledge about and urgently needed action on toxics.