Notice & Comment

Cost-Benefit Analysis: Just One of the Tools in the Toolbox, by Amy Sinden

*This post is part of a symposium on Modernizing Regulatory Review. For other posts in the series, click here.

I was pleased to see the new proposed Circular A-4 acknowledge right up front that the “[r]egulatory analysis [it] describe[s] does not supplant any analytic requirements . . . set out in the statutes that authorize or require agency action.” (Draft A-4, at 3)  This is perhaps stating the obvious—it’s a matter of high school civics that an executive order can’t trump a duly enacted statute.  But it’s a point that often gets lost in discussions of regulatory review.   

Relatedly, it was good to see the new A-4’s acknowledgment that formal cost-benefit analysis (CBA) doesn’t always work—sometimes agencies are unable “to quantify and express all of the important effects of a regulation in monetary units,” thus precluding any meaningful calculation of net benefits.  (Draft A-4, at 3)  Moreover, the new A-4 rightly recognizes that formal CBA, even if it does work, focuses only on one relevant value (social welfare maximization) among the array of values potentially relevant to agency decision making—things like equity, fairness, rights, and human dignity. (Draft A-4 at 19-20, 62, 64) (Livermore & Revesz, 2008, at 15)  Indeed, in one of its more surprising but welcome moves, the new A-4 even goes so far as to acknowledge that sometimes one or more of those values should trump welfare maximization, warranting an agency decision to “select a regulatory alternative with lower monetized net benefits over another.” (Draft A-4 at 64)

Finally, I was pleased to see the new A-4, in recognition of these various shortcomings, describe and elaborate in some detail additional decision tools—including cost-effectiveness analysis, break-even analysis, and stand-alone distributional analysis—and direct agencies to use these tools instead of or alongside formal CBA in instances where the information environment or the relevance of competing values make these tools appropriate. (Draft A-4 at 5-7, 46-47, 61-64)

The necessary implication of all of this is that formal CBA is but one tool in the toolbox of decision frameworks that agencies have at their disposal and, in the words of Justice Scalia, it is “up to the agency to decide how to account for costs [and benefits]” by choosing among the wide array of tools available. (Sinden, 2022)  While this point is implicit in the proposed A-4, OIRA should make it explicit and give it far more emphasis throughout the document.  Instead, the proposed A-4 continues to insist that formal CBA is “the primary analytical tool used for regulatory analysis.” (Draft A-4 at 2)  While the term “cost-benefit analysis” can mean many different things and, as elaborated below, there are many ways that an agency can assess regulatory costs and benefits in order to avoid issuing a rule that imposes disproportionate costs for little benefit, the term is clearly used in the new A-4 to describe a very specific type of CBA that has come to dominate OIRA’s regulatory review process.  That is, a highly formal and technical method grounded in welfare economics and aimed at achieving economic efficiency that attempts to fully quantify and monetize all of the social costs and benefits of a whole range of regulatory alternatives in order to select the alternative that maximizes monetized net benefits. The new A-4’s continued insistence that this kind of formal CBA should be “the primary analytical tool used for regulatory analysis” is misguided for three reasons.

First, as the proposed A-4 itself recognizes, the economic efficiency at which formal CBA is aimed is only one of a series of goals and values Congress has tasked various agencies with promoting.  Some statutes aim at goals, like “protecting civil rights and civil liberties, . . . safeguarding democratic institutions,” (Draft A-4 at 20) as well as promoting equity and distributional fairness, that formal CBA, with its focus on aggregate social welfare, is simply not aimed at evaluating. Additionally, some aim to promote values, like dignity, individual rights, the preservation of indigenous cultures, or the safety of young children, that cannot be reduced to monetary terms. (Draft A-4 at 44)  Where such goals and values are relevant, agencies must use other decision tools to evaluate them.  Indeed, in his Day 1 memo on Modernizing Regulatory Review, President Biden notably did not make social welfare maximization “primary,” but, instead, very pointedly put it on a par with a list of other values that formal CBA is simply not designed to evaluate—specifically, “racial justice, environmental stewardship, human dignity, equity, and the interests of future generations.”

Second, it’s just odd to insist that formal CBA is the “primary tool” for regulatory analysis in light of the refresher course in basic civics we began with.  In the realm of environment and public health, at least, Congress has only rarely chosen formal CBA as the basis for regulatory decision making. (Sinden, 2015, at 129-34)  The statutes largely direct agencies to use other decision tools instead.  So making formal CBA the “primary tool” for regulatory analysis just puts the agencies in the awkward position of being pulled in two directions at once.

Third, the proposed A-4’s insistence on the primacy of formal CBA continues the conceptual mistake that the old A-4 made—a mistake which (unfortunately, in my view) infects much of the current debate on CBA.  That is to say, it begins from the standpoint of theoretical welfare economics—with its ideologically freighted and myopic focus on economic efficiency—and then views the real-world information gaps that put constraints on agencies’ ability to carry out that theory in practice as mere after-thoughts—as exceptions to the rule, issues that might require minor tweaks to the theory at most.  If what we care about are real decisions in the real world, this approach is exactly backwards.  A more effective and meaningful approach to regulatory review would instead begin with a sober assessment of the information environments the agencies find themselves in and build decision tools specifically adapted to those environments. This would involves, among other things, letting go of the perhaps intellectually beguiling but impractical idea that there is a one-size-fits-all tool that can be applied to all regulatory contexts—recognizing instead that there is “not a single hammer for all purposes.” (Gigerenzer, 2001, at 7) 

The new A-4 takes a few baby steps toward recognizing the backwardness of its approach by acknowledging that benefits are sometimes unquantifiable and that alternative decision tools may sometimes be appropriate. But OIRA would do well to recognize just how frequently this may turn out to be the case.  My empirical study of EPA’s CBAs of major rules during much of the Bush II and Obama administrations revealed that the agency was unable to monetize significant categories of regulatory benefit 83 percent of the time, making any calculation of net benefits impossible.  Other empirical studies make similar findings.  This suggests that the relevant decision environment and the prevalent information gaps agencies face need to be the starting point in the design of decision tools, not an afterthought.

Fortunately, Congress has often taken a more pragmatic approach to the design of decision tools grounded in an appreciation of these severe epistemic challenges. (Sinden, 2022, at 931-32) Accordingly, many of the agencies’ organic statutes direct them to use decision tools for the evaluation of regulatory costs and benefits that take into account and work around information gaps and constraints in the decision environment in order to deliver more useful and meaningful results.  

A number of major environmental statutes, for example, pair a health-based trigger for deciding whether to regulate with a feasibility analysis for deciding where to set the standard. This pairing can be thought of as a “sequential cost-benefit analysis.”  This decision tool performs a fairly granular and often quantified analysis of the relevant pollution control technologies, gauging their costs by comparing them to the overall financial capacities of the industry. Meanwhile, in order to work around information gaps, it takes a more cursory look at regulatory benefits, without the detailed level of explication or quantification a formal CBA would require, but looking carefully enough to ensure that there is some harm worth regulating, i.e., that regulation will produce significant benefits.  

In this way, a sequential cost-benefit analysis (employing feasibility analysis in conjunction with a health-based trigger) is designed to work within epistemic constraints and avoid the most information-intensive and controversial aspects of formal CBA—its requirements that regulatory benefits be precisely measured, quantified, and translated to monetary terms in order to allow a direct comparison to costs. (Sinden, 2022, at 925, 931-32)  Indeed, it was congressional experience in the 1970s and 1980s with health-based standards and the regulatory paralysis that often resulted from their comparatively voracious information demands that led lawmakers to employ these forms of sequential cost-benefit analysis instead. (Houck, 1991)  And much of the dramatic success in cleaning up air and water pollution in the 1970s, 80s and 90s has been attributed to the standards set under these decision tools. (Wagner, 2000, at 94-97; Houck, 1991, at 10,531-39) 

In sum, rather than continuing to insist that CBA is “the primary analytical tool used for regulatory analysis,” OIRA should make explicit and further emphasize what is implicit in the current draft—that formal CBA is only one of an array of decision tools in the agencies’ toolbox; and that this toolbox also includes cost-effectiveness analysis, breakeven analysis, and distributional analysis, along with various additional tools for considering costs and benefits specified in the agencies’ organic statutes.  Accordingly, rather than continuing the pretense that CBA can or should operate as a one-size-fits-all tool across all contexts, the new A-4 should direct agencies to use the context-specific methods specified in their authorizing statutes for considering costs and benefits, informed and elaborated by guidance provided in Circular A-4 on a rich variety of decision tools.

Amy Sinden is a Professor of Law at Temple University’s James E. Beasely School of Law.  She is also a Member Scholar at the Center for Progressive Reform.


Gerd Gigerenzer & Reinhard Selten, Rethinking Rationality, in Bounded Rationality: The Adaptive Toolbox 1, 7 (2001)

Robert W. Hahn and Patrick M. Dudley, How Well Does the U.S. Government Do Benefit-Cost Analysis?, 1 Rev. Envtl. Econ Pol. 192, 200 (2007)

Oliver A. Houck, The Regulation of Toxic Pollutants Under the Clean Water Act, 21 Envtl. L. Rep. 10,528, 10,529 (1991)

Michael A. Livermore & Richard L. Revesz, Retaking Rationality: How Cost-Benefit Analysis Can Better Protect the Environment and Our Health (2008)

Jonathan Masur & Eric Posner, Unquantified Benefits and the Problem of Regulation under Uncertainty, 102 Cornell L. Rev. 87 (2016)

Amy Sinden, All the Tools in the Toolbox:  A Plea for Flexibility and Open Minds in Assessing the Costs and Benefits of Climate Rules, 39 Yale J. on Reg 907 (2022)

Amy Sinden, The Problem of Unquantified Benefits, 49 Envtl. L. 73 (2019)

Amy Sinden, A ‘Cost-Benefit State’? Reports of Its Birth have Been Greatly Exaggerated, 46 Envtl. L. Rep. 10,933, 10,943-47 (2016)

Amy Sinden, Formality and Informality in Cost-Benefit Analysis, 2015 Utah L. Rev. 93, 172–73 (2015)

Wendy Wagner, The Triumph of Technology-Based Standards, 200 U. Ill. L. Rev. 83, 95-97 (2000)

Print Friendly, PDF & Email