Agencies must better quantify the uncertain economic effects of proposed regulations.
Since the Reagan Administration, agencies have used benefit-cost analysis to determine whether a proposed rule will yield net benefits for the public. This not only helps to determine whether to regulate, but also helps courts determine whether an agency considered all relevant factors or made errors in judgment. An important assumption of this practice is that costs and benefits, collectively referred to as “welfare effects,” can be monetized so that conceptually disparate impacts are commensurable. When effects are uncertain or are mediated by factors outside of an agency’s control, a neat-and-tidy benefit-cost estimate is less straightforward.
Agencies have often responded to this methodological challenge by reporting certain effects qualitatively. In a 2025 report to the U.S. Congress, the White House Office of Management and Budget (OMB) noted that of the 30 major non-transfer rules—rules that impact which entities receive federal funds and therefore reallocate welfare, necessarily producing offsetting gains and losses—reviewed by OMB in fiscal year 2023, 36.7 percent did not quantify either costs or benefits.
University of Chicago professors Jonathan S. Masur and Eric A. Posner argue that this lack of quantification “may be leading to both under- and over-regulation,” and in some cases, “agencies may be justifying regulations that do not pass a cost-benefit test on the basis of unquantified benefits, without any way of knowing whether those unquantified benefits exceed the known costs.” Without a commensurable rubric to measure costs and benefits, benefit-cost analysis becomes less reliable as a tool for quantitative evaluation.
This is not to say that all welfare effects are easily monetized. Circular A-4, the guidance document for federal agency regulatory impact analysis, has long acknowledged that some benefits and costs are difficult to quantify and that qualitative descriptions may be appropriate at times. The guidance, however, emphasizes that agencies should monetize effects whenever reasonably possible. The problem is not the existence of qualitative impacts in these analyses, but a failure to take advantage of available empirical evidence when it exists, particularly in cases where agencies explicitly identify economic effects but do not quantify them.
A growing body of empirical literature suggests that nonquantification disproportionately occurs for benefits. For example, a study published in the American Journal of Evaluation studied RIAs for education regulations and found that although the majority of RIAs reported costs (93%) and benefits (82%), only 71% provided dollar values of costs and 4% provided dollar values of benefits. An earlier study published by the American Enterprise Institute–Brookings Joint Center for Regulatory Studies found that agencies only quantified net benefits for 29% of rules.
This lack of information leaves open the possibility for disparate treatment of one group of effects compared to another. Regulators can treat these qualitative welfare effects as either insignificant or decisive This treatment matters because regulatory impact analyses function as a regulatory oversight mechanism and as inputs into agency decision-making. Without a method for quantifying welfare effects in the benefit-cost analysis equation, how estimates are used becomes a black box, with no clear guidelines to understand when or how a particular effect is considered.
We can look to the past to see what this may lead to. In U.S. Chamber of Commerce v. Securities and Exchange Commission, the U.S. Court of Appeals for the D.C. Circuit reviewed a U.S. Securities and Exchange Commission (SEC) rule imposing new mutual fund governance requirements. Although the SEC acknowledged the costs associated with compliance, it argued that it lacked a “reliable basis for determining how funds would choose to satisfy” the rule’s requirements and therefore found it “difficult to determine the costs.”
The court rejected that argument, noting that while uncertainty might be a limit on the precision of the estimates, it “does not excuse the Commission from its statutory obligation to determine as best it can the economic implications of the rule it has proposed.” The court explained that an agency must “exercise its expertise to make tough choices about which of the competing estimates is most plausible, and to hazard a guess as to which is correct.” In this way, reliance on qualitative welfare effects may end up interpreted as a weakening of the benefit-cost analysis framework required for reasoned decision-making.
This is not to say that qualitative judgements have no place in the development of a regulation or even the decision to finalize one. Indeed, it is fair to say that benefit-cost analysis was never meant to be the sole basis for regulating, but rather one of a suite of tools to inform agency decision-making. The point is that regulatory impact analyses need not provide precise point estimates to be useful. Indeed, incorporating an element of uncertainty, while not necessarily convenient from an interpretive perspective, is more methodologically sound. When overall point estimates are infeasible, the default judgment should not be to provide point estimates for one set of effects and qualitative descriptions for another within the same benefit-cost analysis framework.
Instead, agencies should opt for alternative approaches, such as breakeven analysis, which estimates the minimum level of benefits or the maximum amount of costs at which a regulation would yield zero net benefits. The agency would then identify the requisite conditions in which the regulation would be able to yield that minimum.
The breakeven approach has been used with success in the past. A 2015 rule from the U.S. Food and Drug Administration that required food facilities to implement risk-based preventive controls and maintain records documenting the movement of food through the supply chain relied on a breakeven analysis to evaluate whether the rule’s compliance costs would be justified by reductions in foodborne illness and contamination risks that were difficult to quantify ex ante.
The use of breakeven analysis in the regulatory impact analysis was effective in this case because it matched the nature of the problem. Whether the rule would in fact achieve this level of illness reduction was inherently uncertain, as foodborne outbreaks and intentional contamination events are sporadic and influenced by complex supply chains. Under these conditions, breakeven analysis allowed the agency to evaluate the rule without assuming precise probabilities for rare but high-impact events, instead making explicit the level of effectiveness the rule would need to attain to justify its costs.
Uncertainty should not excuse agencies from engaging in some form of quantification. When some welfare effects are quantified, and others are not, regulatory impact analyses end up skewing towards monetized effects, underscoring the weakness of using point estimates. As the example from the FDA illustrates, agencies can do better at monetizing impacts and putting important effects on more equal footing. Courts have made it clear that uncertainty does not excuse agencies from estimating economic effects to the best of their ability, suggesting that reasoned decision-making requires additional effort towards quantification.
