UC Berkeley

Cost-benefit analysis is today a cornerstone of American administrative law. Congress has sometimes asked it of administrative agencies, the Army Corps of Engineers being an early and notable example. But it took a series of orders, issued by President Reagan and revised and reaffirmed by Presidents Clinton, Bush, and Obama, to firmly establish cost-benefit analysis as a staple of regulatory practice. Most recently, some legal scholars have suggested that a rule is presumptively “arbitrary and capricious,” and hence liable to be set aside under the Administrative Procedure Act, if it cannot survive a cost-benefit test. The chapters in this dissertation examine cost-benefit analysis in its institutional context by drawing on perspectives from law, philosophy, political science, and economics.

The Supreme Court’s decision in Michigan v. EPA (2015) contrasted formal (or quantitative) approaches to informal (or qualitative) ones. The distinction is, however, distracting, and even misleading, because monetization and aggregation, not quantification, is at the heart of cost-benefit analysis. The first chapter articulates three interpretations of monetization and aggregation, and hence, cost-benefit analysis. In its welfarist mode, cost-benefit analysis serves as an indicator of a rule’s impact on overall well-being. In its replicative mode, cost-benefit analysis tries to arrive at the same destination as another mechanism, a frictionless market being the most salient example. Finally, in its rationalizing mode, cost-benefit analysis seeks to demonstrate that there is a set of numbers, satisfying certain structural and substantive conditions, that makes the regulation a rational one. These interpretations of cost-benefit analysis have implications for how monetization and aggregation are carried out. The regulatory state should therefore strive for reflective equilibrium between how cost-benefit analysis is defended and how it is practiced under a system of executive review. While the academic debate over the details of regulatory reform remains unsettled, lower court judges applying the Administrative Procedure Act’s standard of arbitrary and capricious review can, under existing doctrine, at most demand of agencies cost-benefit analysis that is rationalizing.

Although cost-benefit analysis can be a mechanism for compelling agencies to reveal the information that they have to their principals, they are, for theoretical reasons and as an empirical matter, susceptible of obfuscation. The second chapter thus studies the informative-ness of cost-benefit analysis under a system of regulatory review by the Office of Information and Regulatory Affairs (OIRA) by introducing a review technology into Crawford and Sobel (1982)’s cheap talk model. I find that depending on the cost of OIRA review, three types of communicative outcomes – no review, universal review, and selective review – may be obtained in equilibria. Moreover, communications become less informative, ex ante, as the ideological distance between the agency and the White House increases and, assuming a regime of selective review, as OIRA review becomes more efficacious at recovering the information that the agency has. The reason for the latter is that since the agency and the White House desire regulation that is based on actual social, economic, and political conditions, the agency has incentives to invite scrutiny from OIRA. This reduces the informative-ness of representations that the agency can credibly make. In addition, communications between the White House and the agency are more precise if they are observed by a relatively moderate body that can influence the stringency of the final rule. Conversely, the agency’s communications are less precise if they are observed by a relatively extreme intervener. An agency that is perfectly aligned to the White House fudges under such circumstances because of the divergence between itself and the other member of the audience. The stylized model that is described here furnishes a theoretical basis for studying the informative-ness of cost-benefit analysis under a system of regulatory review. It hints at positive explanations of the quality of regulatory impact analyses performed by agencies and also adumbrates some of the normative considerations that should inform calls for more rigorous cost-benefit analysis in the administrative state.

Finally, the public’s reaction to cost-benefit arguments is relevant to issues of institutional design. It is generally thought that policies involving “taboo” trade-offs, such as those that balance risks to life against financial costs, may be rejected by the public if their utilitarian logic becomes too apparent. The third chapter asks whether these attitudes are moderated by the type of body that is engaged in cost-benefit reasoning. This survey experiment employs a two-factor design that randomized the source of a policy and the presence of a quantified cost-benefit argument to explore this question. Analyzing responses from 1004 Amazon Mechanical Turk workers, I find no evidence that the Federal Motor Carrier Safety Administration or the fictitious National Federation of Trucking Companies have an advantage or disadvantage, vis-à-vis Congress, at persuading through cost-benefit analysis. Surprisingly, however, the presence of a cost-benefit argument did not diminish, and may even have strengthened, support for the policy. An exploratory search for heterogeneous treatment effects suggests that this contrarian result is unlikely to be due to differences in observed demographic variables between the sample and the population. As some of the treatment conditions used by prior research in the literature did not quantify the considerations on both sides, attention to the influence that a quantitative, rather than qualitative, description of costs and benefits has on opinion may improve our understanding of the how cost-benefit reasoning is received in public life.