Justin (Gus) Hurwitz of Pennsylvania and Gregory M. Dickinson of Stanford have written The Consumer Protection Paradox: Learning, Feedback, and the Limits of Consumer Protection Law, forthcoming in the Emory Law Journal. Here’s the abstract:
Modern consumer-protection law is built on a simplifying picture: The consumer and her vulnerabilities are treated as static over time. That model fits consumer protection’s regulatory posture, which evaluates injuries at the moment of choice. But it fits markets poorly. Consumers differ in sophistication and-more importantly-they can become more sophisticated through experience. This Article develops the Consumer Protection Paradox, the concern that when law ignores heterogeneity and learning, even well-intentioned interventions can increase long-run vulnerability. By suppressing exposure even to manageable risk and complexity, regulation can suppress exploration, distort the feedback linking choice to consequence, and reduce the returns to attentiveness. It can also reshape market composition by inducing capable consumers to disengage, weakening the demand-side discipline that constrains opportunism. The Article’s central contribution is to elevate feedback quality as a missing dimension of consumer-protection analysis and to show how that variable changes both diagnosis and remedy in modern settings, including retail finance design, privacy consent pop-ups, and price obfuscation. Drawing on cognitive learning models, it explains why some harms can double as consumer “tuition”-bounded costs that teach-and why others cannot, because they are structured to defeat learning through delay, opacity, or strategic noise. The Article translates this insight into a harm feedback matrix meant to guide regulatory instrument choice-when law should preserve learning, when it should restore feedback, when it should contain tail risk without ritualized choice, and when strong paternalism or prohibition is warranted because learning is implausible. The aim is protection that reduces short-run harm without quietly replacing capability with dependency.
