Behavioral economics and policy evaluation

Behavioral economics posits a number of cognitive biases and limitations, which raises questions as to whether revealed willingness to pay equals true willingness to pay. If so, benefit-cost analysis, with a number of methodological advantages, would need to be replaced. Prior analyses of the issue by Sunstein, Sugden, and Bernheim and Rangel fail to offer guidance that would avoid substituting centralized judgments for decentralized information on benefits and costs. Alternatives including using post-implementation valuations, libertarian paternalism, and direct democracy on policy issues also have conceptual or practical limitations. A tentative suggestion is democratic delegation, somewhat appealing because it is already applied to cope with bounded rationality and non-efficiency values. Viewing benefit-cost analysis as a market analogue, and restricting the domain of behavioral economics to uninformed consumers, may be useful guides. The most important guidance may be to require very strong evidence of substantial choice failure before abandoning benefit-cost analysis.