Much evidence suggests that IRBs make decisions badly. For example, when multiple IRBs review the same research protocols, they treat them inconsistently. The IRB system makes error inevitable. First, bureaucratized event-licensing overtaxes IRBs with too much work to handle and too much paperwork to manage. Second, IRBs necessarily lack the expertise to understand the hyper-specialized studies they license. Third, IRB incentives distort decisions. Unlike the FDA, which weighs a drug’s benefits against its costs, IRBs just protect research subjects. Also, IRBs that say no to research are safe; if they say yes, they risk trouble like institutional disgrace, lawsuits, and federal sanctions that have included (briefly but unforgettably) closing down research at distinguished universities.