Agency Use of Indirect Benefits to Justify Regulation

Abe Eichner* June, 2024

Executive agencies have long used indirect benefits—meaning benefits beyond the express purpose of a regulation—to justify their rulemakings. However, the statutes that provide agencies with regulatory authority rarely explicitly direct agencies to consider indirect benefits. Lower courts disagree over whether consideration of indirect benefits is permissible, and the Supreme Court has reserved the question for a future case. Courts and existing scholarship have largely asked whether particular statutory provisions authorize consideration of indirect benefits. This Note contends that, even without such statutory authorization, indirect benefits are presumptively permissible because they further three traditional administrative law values: rational decisionmaking, transparency, and accountability. It then argues that, even if statutory authorization was required, the Clean Air Act (CAA)—the statute at the center of the indirect benefits controversy—provides the Environmental Protection Agency (EPA) with unmistakable authority to consider indirect benefits. Finally, this Note recommends two lines of judicial review to prevent agency abuses of indirect benefits.

Introduction

Because presidents want regulations that benefit the public to the greatest degree possible, they direct agencies to consider all effects that a regulation will have. These effects include indirect benefits, also called co-benefits, that produce beneficial effects beyond the express purpose of a regulation.¹ For example, a Clean Water Act regulation that targets water pollution from coal power plants might require those plants to limit their operations, which would indirectly benefit the climate more broadly. At the same time, Congress gives agencies authority to solve only particular problems. Administrative law, constrained by the separation of powers, prohibits agencies from considering regulatory factors irrelevant to the problems that Congress provided.² It is controversial whether indirect benefits are irrelevant. Although Congress does not explicitly authorize their use, Congress does want agencies to make reasonable and informed decisions. Thus, agencies are arguably caught between executive and legislative approaches.

Agencies have generally followed presidential direction to include co-benefits in their cost-benefit analyses and justifications for regulations.³ This practice was uncontroversial until 2015, when the Supreme Court expressed skepticism about the practice but declined to address its legality.⁴ Since then, the D.C. Circuit has embraced agency consideration of indirect benefits in certain circumstances,⁵ but one federal district court has struck down a regulation because the agency relied primarily on co-benefits to justify the regulation.⁶ And these judicial developments have come with a shift in executive philosophies as well: The Trump Administration has attempted to limit agency reliance on co-benefits,⁷ and the Biden Administration—despite protestations otherwise⁸—has quietly done the same in some rulemakings.⁹ Given the current legal climate of increased suspicion toward agency action, the issue of indirect benefits seems sure to come to a head in the near future.

This Note shows that co-benefits further three traditional administrative law values: transparency, accountability, and rational decisionmaking.¹⁰ Because Congress structured the system of administrative law to pursue these values, this Note argues that consideration of indirect benefits is presumptively permissible, even without statutory authorization.¹¹

Still, because some courts may apply a stricter standard to the use of co-benefits, this Note also finds implicit authorization to consider co-benefits under the Clean Air Act (CAA),¹² by far the most important and controversial statute in the struggle over co-benefits. The CAA is one of the most effective laws ever enacted, driving trillions—with a “t”—of dollars in health benefits to Americans every year by reducing air pollution.¹³ But the CAA gives rise to a significant portion of these benefits indirectly.¹⁴ In some rules, more than 95 percent of quantified benefits are indirect.¹⁵ The CAA includes at least a dozen programs, each aimed at different types of air pollutants and each with different statutory thresholds governing regulation.¹⁶ But, for example, because pollutants are often emitted together, regulation of a hazardous air pollutant like mercury (which is not a greenhouse gas) could create an indirect benefit: reducing co-emitted air pollutants that are greenhouse gases. Critically, as climate change worsens, co-benefits under the CAA account for billions of dollars of benefits to the climate.¹⁷ These significant indirect impacts are likely why the majority of indirect benefits case law has centered on the CAA.¹⁸

Part I of this Note lays out the history of co-benefits and the legal controversy surrounding them. Part II contends that consideration of co-benefits improves agency transparency, accountability, and rational decisionmaking. Because Congress designed the administrative apparatus to pursue these values, specific statutory authorization to consider co-benefits should not be required. However, even if a court were to require authorization, the CAA would pass with flying colors. Congress designed it to be implemented as a whole to protect the nation’s air resources. Co-benefits, then, are a critical tool for carrying out the CAA’s design. Finally, Part III recommends two standards of judicial review of co-benefits to prevent agency abuses. It argues courts must prohibit consideration of irrelevant co-benefits but defines “irrelevance” narrowly to include only co-benefits that contradict statutory text and purpose. It also suggests that co-benefits could, theoretically, indicate a pretextually promulgated regulation but recommends courts defer to agencies in most cases.

I. The History and Controversy of Co-Benefits

For decades, the executive branch has instructed agencies that “[t]he same standards of information and analysis quality that apply to direct benefits and costs should be applied to ancillary benefits and countervailing risks.”¹⁹ However, Congress rarely explicitly directs agencies to consider these “ancillary” benefits. Although courts have gestured at this tension between legislative and executive intent, they have not resolved it.²⁰

Every president since Reagan has required that agencies assess costs and benefits to justify their major rules.²¹ President Clinton established the modern cost-benefit infrastructure; his administration told agencies to assess “all costs and benefits of available regulatory alternatives,” thus including co-benefits.²² Presidents Bush, Obama, Trump, and Biden all reaffirmed Clinton’s general approach and have required agencies to show the Office of Information and Regulatory Affairs (OIRA) that the benefits of their rules exceed their costs.²³ However, not all presidents since Clinton have been consistent on whether this OIRA review includes co-benefits.²⁴

Under President Bush, the Office of Management and Budget issued its influential Circular A-4, which, since 2003, has directed agencies to “look beyond the direct benefits and direct costs . . . and consider any important ancillary benefits and countervailing risks.”²⁵ This approach was designed to help each agency determine whether its action would cause more good than harm and to select the most effective alternative.²⁶ This directive also allowed the public and the courts to be more aware of the regulation’s effects. The Bush Administration released the Circular partially in response to concerns from economists that cost-benefit analyses (CBAs) focused too narrowly on direct costs and benefits, failing to assess the totality of a regulation’s effects.²⁷

Even before President Bush formalized the practice, the EPA had consistently considered co-benefits under the CAA.²⁸ For example, President Reagan’s EPA considered co-benefit reductions of criteria pollutants in a rule regulating hazardous air pollutants²⁹ and co-benefit reductions of ozone in a regulation targeting lead emissions from gasoline.³⁰ President George H.W. Bush’s EPA similarly justified emissions standards for landfill gases in part based on “the ancillary benefit of reducing global loadings of methane.”³¹ And President George W. Bush’s EPA supported an interstate air pollution rule by pointing to a co-benefit reduction in mercury emissions.³² Though this historical record under the CAA runs deep, it appears to be limited to co-benefits related to air quality.³³

Outside the CAA context, other agencies have also occasionally considered indirect effects. Even President Trump’s Administration considered the “indirect benefit of . . . reduced distortions in the nationwide labor market” in a Supplemental Nutrition Assistance Program rulemaking.³⁴ This indirect benefit was used to justify a rulemaking that made it more difficult for states to waive work requirements under the program.³⁵

Despite this history, even a longstanding executive practice like consideration of co-benefits is subject to judicial review. Judicial review of agency action is governed by section 706 of the Administrative Procedure Act, which directs courts to set aside regulations that are “arbitrary [and] capricious.”³⁶ The Supreme Court has interpreted this standard as being deferential to agencies. Though agencies are required to provide “a concise general statement of [a rule’s] basis and purpose,”³⁷ courts should set aside the rule only if:

[T]he agency has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.³⁸

Because unelected executive branch officials promulgate these regulations, the reasoned-explanation requirement “ensure[s] that agencies offer genuine justifications for important decisions, reasons that can be scrutinized by courts and the interested public.”³⁹ Co-benefits themselves are subject to judicial review because agencies often rely on them to explain why they promulgated a regulation.⁴⁰

Courts, up to this point, have not held that an agency must show that the benefits of a rule exceed its costs (though the agency may need to demonstrate that for OIRA review).⁴¹ However, courts must ensure that an agency does not rely on factors irrelevant to their authorizing statute, potentially including co-benefits.⁴² Additionally, a court can hold that co-benefits indicate the agency was pretextually attempting to control something other than the object of its regulation.

The only federal court to strike down a regulation for reliance on co-benefits apparently did so both because they were irrelevant and indicated pretextual reasoning.⁴³ In the leadup to that case, Wyoming v. U.S. Department of the Interior, the U.S. Bureau of Land Management (BLM) regulated gas leaks from existing gas wells pursuant to statutory authority to promulgate rules “for the prevention of undue waste.”⁴⁴ The EPA had regulated greenhouse gases from new wells under the CAA’s New Source Performance Standards program but was slow to do so for existing wells.⁴⁵ The Bureau stated that it promulgated its regulation in part because of “the length of this [CAA] process and the uncertainty regarding the final outcomes.”⁴⁶ The court, however, believed the BLM promulgated its rule to circumvent the EPA’s process and, thus, “bootstrap[ped] itself into an area in which it has no jurisdiction.”⁴⁷

The Wyoming v. U.S. Department of the Interior court pointed to the agency’s CBA as evidence.⁴⁸ The benefits of the rule outweighed its costs only if the greenhouse gas co-benefits were factored in.⁴⁹ The court reiterated that the BLM could not bootstrap regulatory jurisdiction using co-benefits, finding that the “ancillary benefits of a rule cannot provide the primary justification for the rule, particularly where those ancillary benefits fall outside the scope of the agency’s statutory authority.”⁵⁰ The court worried that such use of co-benefits offered “no limit on agency authority and any colorable tie to an agency’s authority would permit the agency to act on a problem Congress never intended the agency to solve.”⁵¹

Other courts have found co-benefits to be permissible, at least when they are consistent with statutory purpose. The D.C. Circuit held, in 2016, that the EPA was free to consider co-benefits from other hazardous air pollutants when crafting a hazardous air pollutant rule. The court reasoned that the CAA “does not foreclose the Agency from considering co-benefits and doing so is consistent with the CAA’s purpose—to reduce the health and environmental impacts of hazardous air pollutants.”⁵²

The Supreme Court had a chance to resolve the controversy surrounding co-benefits in 2015 but declined to do so. Michigan v. EPA concerned the legality of the EPA’s 2012 Mercury and Toxic Substances (MATS) rule.⁵³ In the CBA for the MATS rule, the EPA estimated that the regulation would create benefits of $37 to $90 billion per year and costs of only $9.6 billion.⁵⁴ However, over 99 percent of those quantified benefits were co-benefit reductions in particulate matter and sulfur dioxide air pollution.⁵⁵ Amici argued the EPA’s use of co-benefits was a “power grab” to regulate “outside the specific authority under which they are acting.”⁵⁶ But the Court reserved the question for another day, holding that the EPA did not balance costs against benefits early enough in its regulatory process. The Court, therefore, declined to address if the EPA “could have considered ancillary benefits when deciding whether regulation is appropriate and necessary.”⁵⁷

Although the full Court did not answer the question, at least two current justices have questioned the EPA’s approach. In the lower court, then-Judge Brett Kavanaugh described the EPA’s use of co-benefits to justify the MATS rule as “disputed.”⁵⁸ And at Supreme Court oral argument, Chief Justice Roberts expressed concern that the “disproportionate amount of benefit” might raise a “red flag” that the rule is an “illegitimate way of avoiding the . . . quite different limitations on EPA that apply in the criteria program [for particulate matter pollution].”⁵⁹ In other words, Chief Justice Roberts suggested the co-benefits might indicate that the EPA was pretextually regulating particulate matter under a section of the CAA that provided lower barriers to regulation.⁶⁰

The EPA has recently reduced its reliance on co-benefits in CBAs. The Trump Administration was particularly skeptical of co-benefits, and, in 2020, the EPA released a rule requiring disaggregation and separate presentation of co-benefits and co-costs in CAA rulemakings.⁶¹ The rule explains that “[t]he disaggregation of welfare effects will be important to ensure that the [CBA] may provide, to the maximum extent feasible, transparency in decisionmaking.”⁶² The Trump Administration also attempted to rescind a 2016 BLM rule because it relied on climate co-benefits.⁶³

The EPA revoked the Trump co-benefit rule in 2021 under the Biden Administration, expressing concerns that the rule had no basis in the CAA, unjustly called into question the significance of co-benefits, and promoted arbitrary decisionmaking.⁶⁴ But even though President Biden’s EPA theoretically returned to the status quo, it may be quietly shying away from using co-benefits to justify regulations. For example, in a recent toxic substances rulemaking, the EPA quantified $10 to $99 million dollars a year in climate co-benefits but declined to include those benefits in its net CBA.⁶⁵ This is exactly the approach the EPA criticized in its 2021 rule,⁶⁶ and it may show the Biden Administration’s concern with recent legal controversy surrounding co-benefits in Wyoming v. U.S. Department of the Interior and Michigan v. EPA. This Note argues that these concerns are misplaced because co-benefits are consistent with congressional design of the administrative apparatus.

II. Indirect Benefits Help Agencies Achieve Congressional Design

Co-benefits accomplish several regulatory goals. First, they allow agencies to rationally prioritize the most effective regulations and assess how these regulations will impact other sections of a statute. Furthermore, by allowing the public, political branches, and courts to understand the effects of and reasoning for a regulation, co-benefits serve both agency transparency and accountability. Because Congress wants agencies to pursue these values, statutory authorization is not required to consider co-benefits. And even if authorization were necessary, the CAA, at least, provides the EPA with unmistakable authorization.

A. Indirect Benefits Are Presumptively Permissible

Rational decisionmaking, transparency, and accountability are the foundational principles of administrative law.⁶⁷ Consideration of co-benefits furthers these values. Co-benefits are presumptively permissible even without specific statutory authorization because Congress designed the administrative apparatus for agencies to make decisions in alignment with these values.

1. Rational Decisionmaking

Agencies do not have the resources to pursue every regulatory opportunity. Courts usually allow agencies to determine their own priorities because the “agency is far better equipped than the courts to deal with the many variables involved in the proper ordering of its priorities.”⁶⁸ Co-benefits allow agencies to rationally prioritize the most effective regulations by considering the full effects of each option. Otherwise, an agency might prioritize inferior rules.

Co-benefits also allow agencies to quantify the interactions between different programs in the same statute, allowing agencies to rationally prioritize rules the way Congress designed. General principles of statutory construction support the conclusion that Congress wants agencies to implement statutes as a whole. Courts presume that statutes are cohesive texts,⁶⁹ so each provision in a statute should be read in “harmony” with the others.⁷⁰ Congress often adds statements of purpose that apply to the entire statute,⁷¹ implying that each section is intended to work towards that policy goal. A regulation may indirectly impact the purpose of the statute as a whole even if the effect is not regulated under that section. That effect would, therefore, still be relevant to Congress’s design of the whole statute. Co-benefits allow agencies to quantify those indirect effects.

Co-benefits can also allow agencies to promulgate rational regulations that accomplish statutory goals when they would otherwise falter under OIRA review. For instance, consider the EPA’s 2012 Oil Refineries New Source Performance Standard under the CAA.⁷² The rule created $71 million in direct benefits, $349 million in particulate matter co-benefits, and $96 million in costs.⁷³ Without consideration of co-benefits, this rule would not have survived OIRA review because its costs exceed its direct benefits.⁷⁴ Consideration of co-benefits, though, allowed the EPA to justify the rule with both its direct benefits and co-benefits, thus vindicating the rule and enhancing the quality of the nation’s air resources. Similarly, once the EPA decides to promulgate a regulation, co-benefits allow it to justify a regulatory option that is maximally protective of air quality even if that option would create additional costs.

2. Transparency

Next, an agency transparently relying on co-benefits allows the public and the political branches to understand both the effects of and the reasons for a regulation. Regulations will create indirect effects regardless of whether agencies are allowed to consider co-benefits. If co-benefits are prohibited, agencies may surreptitiously consider them without disclosing their reliance.⁷⁵ Prohibiting co-benefits would also deprive the public of understanding the full effects of a rule. The public might react differently to a costly rule if it learns the rule will indirectly protect it from carcinogenic particulate matter. Similarly, the political reaction to a regulation might differ if an agency presents one set of justifications versus another.⁷⁶ Congress might also want an agency to focus its attention on particular problem areas, as it did when it amended the CAA in 1990 to force the EPA to more aggressively limit hazardous air pollutants.⁷⁷ Co-benefits can allow Congress to transparently see how well the agency is addressing these priorities.

3. Accountability

Transparency also gives rise to agency accountability. For example, voters can elect a new president if they disagree with the priorities of the administration, and the president can create political accountability through their removal power over agency administrators.⁷⁸ The president can also direct OIRA to veto inappropriate agency uses of co-benefits through its review of CBAs. The legislative branch has considerable oversight authority as well. Congress can override a regulation via the Congressional Review Act (subject to a presidential veto),⁷⁹ amend an authorizing statute, or cut agency funding, among other options.

An agency’s open reliance on co-benefits also facilitates accountability through judicial review. For example, the Wyoming v. U.S. Department of the Interior court’s dual holding—that the BLM’s use of co-benefits relied on an irrelevant factor and that the co-benefits indicated a pretextual purpose—was possible only because the BLM included the co-benefits in its justification for the regulation.⁸⁰ Co-benefits, therefore, allow for more meaningful review of agency action by “offer[ing] genuine justifications for important decisions, [and] reasons that can be scrutinized by courts and the interested public.”⁸¹

Because co-benefits serve these administrative values, courts should apply a presumption in favor of their consideration even if analysis of a particular statute does not reveal an implicit authorization. As the Supreme Court has recognized, Congress wants agencies to make informed, reasonable, and transparent decisions, for which courts and the political branches can hold agencies accountable.⁸² This presumption runs so deep within agency structure and judicial review of agency action that Congress may not feel the need to say so explicitly in a particular statute.⁸³ This presumption is presumably the same reason courts have allowed CBAs even when a statute does not call for them.⁸⁴ It would subvert congressional design of the administrative apparatus to prevent agencies from using co-benefits to align their decisionmaking processes with these values.

B. The CAA Authorizes Indirect Benefits Under Any Standard of Review

Although this Note contends that agencies do not require statutory authorization to consider co-benefits, if a court was to require such authorization, the EPA would have a strong case that the CAA provides it. Consideration of co-benefits allows the EPA to meaningfully consider costs, as the CAA repeatedly requires, and it furthers the Act’s purpose: “to protect and enhance the quality of the Nation’s air resources.”⁸⁵ In fact, Congress wants the EPA to achieve this purpose by enacting the law as a whole, which makes sense; air regulations, by their nature, have substantial indirect effects on other air pollutants.⁸⁶ Co-benefits are, therefore, a critical tool for the EPA to rationally implement the statute.

There are textual indications that Congress intends the EPA to consider co-benefits. Congress requires the EPA to consider costs in every major CAA program, save the National Ambient Air Quality Standards (NAAQS),⁸⁷ and courts consistently require consideration of costs to include co-costs.⁸⁸ At the same time, the Supreme Court has implied that, to meaningfully consider costs, an agency must compare them against benefits.⁸⁹ It would be anomalous for costs to include co-costs but for benefits to exclude co-benefits. This would put a thumb on the scale against regulation that is nowhere to be found in the statute.⁹⁰

In addition to the antiregulatory anomaly created by considering co-costs but not co-benefits, it can be practically difficult to distinguish between the two because costs and benefits are merely useful labels attached to negative and positive effects of a regulation.⁹¹ Consider the following hypothetical. Say that the EPA sets an emissions limit for a (non-greenhouse gas) hazardous air pollutant emitted by boat manufacturers and expects that, in response, boat manufacturers will add a particular pollution control technology to their plants. This control technology creates $30 million in co-benefits by capturing greenhouse gases from the smokestack, but it also creates $10 million in greenhouse gas co-costs by requiring additional electricity to power the controls. On net, the rule would therefore lead to $20 million worth of greenhouse gas reductions. Should the EPA ignore the $30 million in gross climate co-benefits and state in its CBA that the rule will harm the climate by $10 million a year? Or should the EPA ignore all climate impacts entirely because the net impact is a co-benefit and, therefore, impermissible to consider? Because costs and benefits are two sides of the same coin, there would not be a principled way for the agency to choose between these options. It would, therefore, be difficult to implement the CAA’s text directing the EPA to consider costs (and, therefore, co-costs) without considering co-benefits.

Beyond individual provisions, the CAA is structured to be enacted as a whole to achieve the goals of the statute, rather than in hermetically sealed sections as Chief Justice Roberts seemed to suggest.⁹² Congress enacted a CAA-wide section of findings and purpose rather than divvying up statements of findings and purpose section by section,⁹³ implying that it wants each section to build toward the same broader purposes. Congress also repeatedly instructed the EPA to consider the effects of its regulations under other sections of the CAA. To give a few examples: Congress required the EPA to study the effects of its other regulations before regulating hazardous air pollutants from power plants;⁹⁴ instructed the EPA to consider whether it was necessary to use its vehicle tailpipe standards to maintain the NAAQS;⁹⁵ and asked the EPA to maintain consistency between its source categories under the Hazardous Air Pollutant program and the New Source Performance Standards program.⁹⁶

Congress has good reason to instruct the EPA to implement the CAA as a whole. The CAA is deeply interconnected along a number of axes. It is the rare CAA regulation that affects one pollutant alone. Many regulations reduce emissions of other pollutants even more than the regulated pollutant.⁹⁷ Many facilities are regulated under multiple CAA programs,⁹⁸ so changing their obligations under one program may affect their ability to meet obligations under another. Similarly, several CAA programs regulate some pollutants, such as greenhouse gases.⁹⁹ Other pollutants, like volatile organic compounds, are regulated under one program when in one form but under another program when they transform into different pollutants once emitted.¹⁰⁰ Thus, the EPA must keep a careful eye on how its regulations interact with one another to rationally implement the CAA.

Only co-benefits allow the EPA to quantify the effects a regulation has on other CAA pollutants. As the EPA has limited capacity to promulgate regulations, it would be useful to know that a greenhouse gas regulation will have substantial indirect effects on benzene (which is not a greenhouse gas), for example. The EPA might then prioritize regulating other hazardous air pollutants if it knows it has already indirectly, but effectively, addressed benzene.

The CAA’s legislative history is even more explicit that the EPA should consider the co-benefits of its regulations with respect to other CAA programs, at least in the context of hazardous air pollutant rulemakings. The Senate Committee to the 1990 CAA amendments expected that, when regulating hazardous air pollutants, the EPA would “consider the benefits which result from control of air pollutants that are not listed but the emissions of which are, nevertheless, reduced by control technologies or practices necessary to meet the prescribed limitation” because they “may produce substantial health and environment benefits.”¹⁰¹ By suggesting that the EPA should consider other types of pollutants in hazardous air pollutant rulemakings, the committee report goes beyond the holding of U.S. Sugar Corp., which upheld the EPA consideration of co-benefits from other hazardous air pollutants.¹⁰²

Though this Part has thus far focused on only co-benefits that affect air quality, its argument also extends to some other types of co-benefits. The EPA rarely includes other co-benefits in its regulations,¹⁰³ but a future CAA regulation might indirectly reduce other environmental harms, such as water pollution. Congress created the EPA to be the primary agency tasked with environmental protection and pollution reduction.¹⁰⁴ It is, therefore, relevant to the EPA’s purpose—though not necessarily to the CAA’s purpose—to consider the indirect effects CAA regulations will have on water quality.

And, in the most extreme case, the EPA might consider a co-benefit that stems from an entirely nonenvironmental effect, like reduced consumer prices. This consideration should still be encouraged because it also promotes rational decisionmaking, transparency, and accountability. But, unlike environmental co-benefits, the EPA could point to little statutory basis to consider these co-benefits, except that Congress may not want the EPA to consider co-costs without considering co-benefits.

III. Judicial Review of Indirect Benefits

To prevent abuses, courts should review agency consideration of co-benefits in two ways. First, a court should ask if the co-benefits were an irrelevant factor for the agency to rely on when promulgating a regulation. Supreme Court case law on “irrelevance” suggests that most uses of co-benefits pass muster, and only reliance on co-benefits that conflict with statutory purposes should be considered irrelevant. Second, even if the co-benefits are relevant, a court could hold that co-benefits indicate the agency pretextually issued a regulation to evade political accountability or skirt statutory requirements. Although such pretextual reasoning would theoretically be unlawful, practical difficulties in determining agency intent militate in favor of generally granting deference to the agency. ¹⁰⁵

A. Indirect Benefits as an Irrelevant Factor

It is a bedrock principle of administrative law that an agency may not promulgate a regulation by relying “on factors which Congress has not intended it to consider.”¹⁰⁶ Although Congress does not explicitly authorize the EPA to consider co-benefits, Congress also does not explicitly prohibit such consideration. The Supreme Court has generally held that silence is insufficient on its own to render a factor irrelevant under arbitrary and capricious review,¹⁰⁷ with one notable exception.¹⁰⁸ This Note argues that only co-benefits in conflict with the statute’s purpose are irrelevant. However, under stricter standards of review, agencies will be on firmer ground with co-benefits that are squarely related to the purpose of the statute or the agency. Finally, even when an agency considers irrelevant co-benefits, courts should reject the regulation only if they believe the agency would not have issued the regulation but for consideration of those co-benefits.

The Supreme Court has been reasonably consistent in determining that silence is insufficient to render a factor irrelevant.¹⁰⁹ In Lopez v. Davis, the Court held that the agency need only show the extra-statutory factor it considered is “reasonable in light of the legislature’s revealed design.”¹¹⁰ In Entergy Corp. v. Riverkeeper, Inc., the Court gave the agency discretion to read cost-benefit considerations into a “best technology” standard. And in Whitman v. American Trucking Ass’ns, the Court interpreted silence to preclude consideration of a factor only when there was clear indication from the context of the statute that Congress intended that result.¹¹¹

Because Congress rarely provides agencies with an exhaustive list of factors to consider, the Supreme Court should stick to this line of cases as a normative matter and as a matter of precedent. It also dovetails with this Note’s recommendation to consider co-benefits as presumptively permissible. Under this permissive standard, co-benefits would be considered irrelevant only when they contradict the statutory purposes Congress communicated to the agency. For example, if a statute directs the agency to consider only public health effects when setting an emissions standard, economic co-benefits to a regulated industry would be an irrelevant consideration because those co-benefits will generally be inversely correlated with the public health.¹¹²

However, more troublingly for agency consideration of co-benefits, the Supreme Court in Massachusetts v. EPA took a harsher stand on extra-statutory factors.¹¹³ In that case, the EPA refused a petitioner’s request that it determine whether greenhouse gases emitted from cars “may ‘reasonably be anticipated to endanger public health or welfare’ ” under section 202 of the CAA.¹¹⁴ The EPA considered a number of factors in declining to make this “endangerment finding,” including the futility of combating climate change with a piecemeal approach and fears of weakening the president’s international negotiating position.¹¹⁵ The Court ruled that the EPA’s reasoning was arbitrary and capricious, in part, because the Court appeared to conflate the EPA’s denial of the petition with the endangerment finding itself. The Court held that the EPA’s consideration of these factors did not “relate to” the statutory question of whether greenhouse gasses emitted from cars endangered the public health.¹¹⁶

Massachusetts v. EPA likely did not signal a doctrinal sea change away from the Court’s other irrelevant factor cases. It held that an agency may consider only factors “relate[d] to” the statutory text, which is admittedly a narrower standard than that in the other cases. But the Court did not cite any other cases interpreting the State Farm irrelevant-factor standard and did not seem to focus on the doctrinal implications of its decision.¹¹⁷ The D.C. Circuit has apparently decided to read Massachusetts v. EPA narrowly and maintains that a factor is irrelevant only when there is clear congressional intent to preclude consideration.¹¹⁸ And the Supreme Court itself has not yet construed Massachusetts v. EPA to have modified its precedents on the issue.¹¹⁹

Should the Court adopt the strict Massachusetts v. EPA standard in the future, co-benefits directly related to statutory text would still be considered relevant. In the CAA context, air co-benefits are relevant because they directly achieve the stated goal of the CAA to protect the nation’s air resources.¹²⁰ Furthermore, other environmental co-benefits would, at least, not be irrelevant under this standard. Although the EPA could not point to the CAA itself, it could rely on the EPA’s organic statute, which directs the EPA to protect the environment.¹²¹ However, co-benefits beyond statutory text would be considered irrelevant. In the EPA’s case, other than the antiregulatory anomaly created by considering co-costs but not co-benefits, the EPA has little to demonstrate that such consideration is consistent with statutory text.

Even when an agency considers an irrelevant co-benefit, courts should not strike the regulation if the error was harmless. The Administrative Procedure Act instructs that “due account shall be taken of the rule of prejudicial error” on judicial review.¹²² This means courts should “uphold[] unsound agency decisions when they are confident that the agency would reach the same decision on remand.”¹²³ The harmless error doctrine is intended to preserve both judicial and administrative resources when a remand would ultimately lead to the same regulation.

The question is, therefore, whether the agency would have prioritized promulgating a different regulation had it not considered the irrelevant co-benefits.¹²⁴ Co-benefits directly measure only the effect of the regulation and not the intent of the agency. When an agency relies on co-benefits, it points to them as one reason it pursued that regulation. But, when the regulation is overdetermined—that is, when the agency had many good reasons to pursue the regulation aside from the co-benefits—courts should not waste resources on a remand.

Harmless error review of agency reliance on irrelevant co-benefits would be fact intensive. The inquiry should include, among other factors:¹²⁵ the proportion of co-benefits to direct benefits, as a disproportionate number of co-benefits may indicate they were an important factor in the agency’s analysis; whether the agency was required to promulgate the regulation or a similar regulation under a statute or court order; whether the agency had other strong reasons to prioritize the regulation at issue; and what role the agency said the co-benefits played in their decision to promulgate the regulation.¹²⁶

B. Indirect Benefits as an Indication of Pretext

Under this Note’s proposal, even if a court does not find a co-benefit to be an irrelevant factor, it could still hold that reliance on the co-benefit indicates the agency regulated under a pretextual purpose. The Supreme Court recently held that pretextual reasoning is a sufficient reason to strike a regulation.¹²⁷ Co-benefits are an awkward fit in the pretext context because disclosing reliance on co-benefits is, by nature, more transparent than not doing so. Still, significant or disproportionate co-benefits could indicate the agency was solely targeting those co-benefits under the pretext of another purpose.

An agency may pretextually promulgate a regulation for two primary reasons. First, it may do so to avoid the political blowback from promulgating an unpopular regulation that achieves its goals directly, instead attempting to promulgate a more popular regulation that achieves its goals indirectly, via co-benefits.¹²⁸ Second, it might attempt to obtain a regulation’s co-benefits to avoid statutory limitations on gaining those benefits directly.¹²⁹

It would subvert statutory structure if an agency’s sole purpose is to obtain co-benefits. Nonetheless, courts should not strike down regulations on the basis of pretext when co-benefits are the only evidence of pretext. It would be nearly impossible for a court to determine why an agency issued a particular regulation using co-benefits alone, and there would be high costs to an incorrect judicial decision.

1. Motivations for an Agency to Engage in Pretextual Regulation

In Department of Commerce v. New York, the first Supreme Court case to hold that a regulation may be struck down for pretextual reasoning, the agency appeared to invent a justification for its action to conceal its true, but politically unpopular, motivation.¹³⁰ That case began when Secretary of Commerce Wilbur Ross attempted to add a census question related to citizenship.¹³¹ The agency maintained that Secretary Ross added the question pursuant to a request from the Department of Justice (DOJ) to help it enforce the Voting Rights Act.¹³² However, litigation revealed that Secretary Ross had repeatedly pushed an unwilling DOJ to send him the request and had decided to add the citizenship question beforehand.¹³³ Chief Justice Roberts, writing for the Court, remanded the agency decision because “[u]nlike a typical case in which an agency may have both stated and unstated reasons for a decision, here the [Voting Rights Act] enforcement rationale—the sole stated reason—seems to have been contrived.”¹³⁴

First, an agency might promulgate a pretextual regulation to avoid political blowback, like in Department of Commerce v. New York itself.¹³⁵ That case was motivated, at least in part, by concerns about ensuring political accountability for agency action. Agencies must offer “reasoned explanation” for their regulatory decisions so that such “reasons . . . can be scrutinized by . . . the interested public.”¹³⁶ Permitting agencies to lie about the reasons behind their actions allows them to evade political blowback from the public and from Congress.¹³⁷ Political accountability is a critical check on agency action performed by unelected officials.¹³⁸

Second, an agency might, in theory, regulate indirectly to evade the statutory requirements associated with regulating directly. For example, imagine that the EPA perceives that the statutory threshold makes it too difficult to further regulate particulate matter under the CAA’s NAAQS program.¹³⁹ Instead, it might pretextually issue a hazardous air pollutant regulation that is solely intended to indirectly reduce particulate matter. That regulation would likely have significant particulate matter co-benefits. During oral argument in Michigan v. EPA, Chief Justice Roberts alluded to this situation when he expressed concern that the “disproportionate amount of [particulate matter co-]benefit” might raise a ‘red flag’ that the rule was a pretextual attempt to avoid the strictures of the NAAQS program.¹⁴⁰

Chief Justice Roberts is correct in implying that the structure of the CAA demonstrates that Congress intends for certain pollutants to be directly regulated under certain sections.¹⁴¹ And Congress created different statutory guardrails for the EPA to meet when it regulates those pollutants. To regulate particulate matter under the NAAQS, the EPA must show the emissions standard is “requisite to protect the public health” with an “adequate margin of safety.”¹⁴² This is a “health-based” standard. In contrast, hazardous air pollutant emissions standards are “technology-based”; the EPA must show they are commensurate with the “maximum degree of reduction” achievable with the application of the best available pollution control technology.¹⁴³ In some situations, the EPA may, therefore, be able to more strictly regulate particulate matter indirectly through the Hazardous Air Pollutant program than it could directly through the NAAQS program.

As in the example above, even co-benefits related to statutory purpose could be an indication of pretext. The EPA is permitted to consider these co-benefits as one factor when prioritizing regulations; Chief Justice Roberts seemed to allow as much, noting that “it’s a good thing if your regulation also benefits in other ways.”¹⁴⁴ But the EPA would go too far if it promulgated the regulation solely to obtain those co-benefits.¹⁴⁵

The same pretext concerns might apply if the co-benefits targeted effects that an agency is allowed to regulate only under an entirely different statute. For example, if a CAA regulation created substantial water quality co-benefits, a court could justifiably worry that the EPA was attempting to dodge the requirements of the Clean Water Act. Although this Note contends water co-benefits are a relevant factor under the CAA,¹⁴⁶ the EPA may not subvert Clean Water Act statutory guardrails by promulgating a CAA regulation intended solely to protect water quality.

These concerns would be heightened if an agency considered co-benefits outside of its statutory purview, like the BLM considering climate co-benefits in Wyoming v. U.S. Department of the Interior.¹⁴⁷ The same might be true if the EPA considered lower consumer prices of a manufactured product, for example. The EPA might be more incentivized to regulate these effects indirectly because it does not have any authority to do so directly. However, the EPA rarely, if ever, relies on nonenvironmental co-benefits in its CAA regulations. And a court might not even reach the pretext issue if it finds these co-benefits to be irrelevant.

2. Issues with Judicial Review of Pretext

Even though an agency could hypothetically promulgate pretextual regulations to evade political accountability or statutory guardrails, it would be unwise for courts to strike down a regulation if co-benefits are the only evidence of pretext. Co-benefits are a blunt tool for determining agency intent. Furthermore, the costs of a false positive judicial decision are high, and the risk of agency malfeasance is low.

It would be very difficult to determine, based on co-benefits alone, whether the agency promulgated a regulation solely to obtain co-benefits or if the co-benefits were only one reason the agency prioritized that regulation.¹⁴⁸ This inquiry would be more difficult than the harmless-error inquiry described above.¹⁴⁹ There, the court would need only ask whether the co-benefits formed an important part of the agency’s erroneous justification for the regulation. Here, a court would need to determine that the EPA was solely targeting the co-benefits to avoid political accountability or statutory guardrails.

It is unclear how a court could find that an agency was trying only to achieve co-benefits if a regulation’s direct benefits otherwise meet the statutory requirements for regulation. However, if the agency gives some other indication of pretextual motivation—like the BLM’s admission in Wyoming v. U.S. Department of the Interior that it was pursuing the regulation because the EPA was slow to regulate greenhouse gases from gas wells—co-benefits can serve as additional evidence of pretext.¹⁵⁰

A false positive judgment against the agency on pretext grounds would be costly, particularly under the CAA. It is utterly ordinary for the EPA to issue regulations with a disproportionate amount of co-benefit.¹⁵¹ Air pollution regulations create co-benefits by their nature. It would put the EPA in a nearly untenable position if each of these regulations raised suspicions that the EPA was acting pretextually. The EPA would need to blind itself to the full effects of its regulations and, therefore, prioritize inferior rules.

To think about the issue another way, striking down regulations with a disproportionate amount of co-benefit would, in effect, raise the statutory threshold for particularly effective regulations. Under such a rule, if a New Source Performance Standard regulation created no co-benefits, the EPA would need to show only that the pollutant endangered public health and that the standard reflected the best available technology.¹⁵² But, if the regulation created significant co-benefits, the EPA would additionally need to show that direct benefits are not disproportionately smaller than co-benefits. This would give a single CAA provision different effects under different circumstances. Nothing in the statute indicates that Congress intended such a result.

Finally, the risk of agency malfeasance is low, at least for the EPA. The historical consistency of co-benefit consideration somewhat insulates the practice from charges of malfeasance by any particular administration.¹⁵³ And, specifically, there is little indication the EPA has attempted to avoid political accountability in its CAA regulations. The Supreme Court recently struck down an Obama-era CAA rule regulating greenhouse gases from power plants—not because the EPA was trying to obscure its motives but because the EPA attempted to resolve a question of “political significance.”¹⁵⁴ The Trump Administration also came under political scrutiny for withdrawing that same regulation.¹⁵⁵ Furthermore, the vast majority of CAA co-benefits stem from other air pollutants. Unlike in Department of Commerce v. New York, where the pretextual purpose was much more politically tolerable than was the true purpose, there is little reason to think that regulation of one air pollutant is substantially more popular than regulation of another.

The risk of EPA pretext to evade statutory guardrails is similarly low. If the EPA wants to regulate a particular air or water pollutant, it has the authority to regulate the pollutant directly. Though that authority comes with limitations, it might not be much easier for the EPA to regulate the pollutant indirectly. And, although the EPA may be more tempted to indirectly regulate effects outside of its statutory authority, it has rarely, if ever, relied on these effects to justify a rule.

Conclusion

Critics describe consideration of co-benefits as an unlawful and underhanded expansion of regulatory authority.¹⁵⁶ On the contrary, co-benefits do not expand regulatory authority, but they do improve its functioning. This Note contends that co-benefits further the traditional administrative law values of rational decisionmaking, transparency, and accountability. Since Congress wants agencies to pursue these values, courts should consider co-benefits presumptively permissible. And, even if statutory authorization were required, implicit authorization for co-benefits exists under the Clean Air Act, which was designed to be implemented in a rational and holistic manner to achieve its goals. The same could likely be said for many other statutes and agencies.

* J.D., May 2024, University of Michigan Law School. A special thank you to Professor Nina Mendelson, who sparked my interest in administrative law. Her expertise and guidance were essential to this Note. Thanks as well to Professor Rachel Rothschild, who planted the seed for this piece, and Professor Andy Buchsbaum, who helped shape its early versions. Thank you to the editors of Michigan Law Review who gave edits and support: Eddie Plaut, Hannah Cohen Smith, Kassie Fotiadis, Katie M. Osborn, Jordan Schuler, Ashley Munger, Hannah Juge, John Grosboll, and Ethan Haughie, and thanks as well as to First Lieutenant Joseph Westberry. All errors are my own. Finally, thanks most of all to my parents, Max and Eric.