Diseases and pests pose risks to U.S. agriculture and forests, but regulations and quarantines to control these risks are costly. The U.S. Department of Agriculture issues rules to control these risks based on economic analyses that do not take adequate account of links between risks and policy outcomes. A benefit-cost analysis that fully incorporates both the risk of a disease outbreak and the effect of regulations and quarantines on such risk can yield quite different conclusions. We apply methods that combine probabilistic risk assessments with economic analysis. We show that if USDA had incorporated risk into its benefit-cost analysis of Karnal bunt, a disease affecting wheat, it would have reached different conclusions about the impact of its actions. We estimate that suboptimal regulatory decisions in the case of Karnal bunt cost between $350 million and $390 million per year. We recommend that USDA incorporate risk assessments into its economic analyses of proposed regulations.