Administrative law does not have to be hard. Follow the required procedures. Implement the statutes you have. Have a reasonable basis for your decisions. This is, really, it. Sure, there will always be questions about interpretive issues, or whether an agency has adequately addressed all comments, or the sufficiency of the agency’s evidence. And in some instances, these questions will be close calls. But, as a doctrinal matter, there really isn’t anything else. Know these three rules, law students, and you will do fine on your exam.
However, the waters seem to be getting increasingly muddied on the second command, “Implement the statutes you have,” which concerns the core questions of deference and authority. In particular, there is, on one side, an increasing skepticism of deference in general and especially on “major questions,” while on the other, we see a growing idea that existing statutes should provide the authority to take any action determined to be necessary unless such an interpretation is specifically foreclosed.
As always, when thinking through these issues, we should start with how the Supreme Court defined deference in Chevron:
First, always, is the question whether Congress has directly spoken to the precise question at issue. If the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress. If, however, the court determines Congress has not directly addressed the precise question at issue, the court does not simply impose its own construction on the statute, as would be necessary in the absence of an administrative interpretation. Rather, if the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency’s answer is based on a permissible construction of the statute. [internal citations omitted]Chevron, U.S.A. v. Natural Res. Def. Council, 467 U.S. 837, 842-43 (1984)
Although this precise language is typically quoted in formal documents, a “vulgar” Chevron doctrine is what is often said: if the statute can be considered ambiguous, the agency wins so long as their interpretation of the relevant text is at least reasonable. This framing tends to lead to a more radical view of the deference than Court’s original language in Chevron, which was heavy on Congressional intent. Instead, the focus is generally solely on the statutory text, and often a very narrow slice of that text. Agencies are then often viewed as making arguments to fit preexisting policy conclusions, in the same way that private litigants are assumed to make the arguments that best support their position, regardless of the objective strength of those arguments. Critics of deference argue that approaching the question in this way provides the executive branch too much power, as, so long as it has some half-way decent argument, it will prevail, and thus the central concept of deference ought to be abandoned. Others argue that the executive ought to have this authority, as the question, once there is litigation, is really whether an agency’s interpretation or that of a private litigant (or the court) should win, and the executive branch has the greater democratic legitimacy and accountability to Congress.
In reflection of this debate, the approach taken by most courts has generally focused on Chevron Step 1, a concession to the idea that Step 2 simply means the agency wins. Thus, when courts believe that an agency interpretation goes too far, they will often end up finding that the relevant language is unambiguous and prohibits the agency’s interpretation, even if doing so requires a significant degree of interpretive steps by the court, or, as seems to be increasing, simply not reference Chevron at all. And, recently, the major questions doctrine has arisen in response to criticisms that agencies are taking on too much authority based on more general statutes. Combined, the general criticism of deference and the “major questions doctrine” are potentially creating a world where courts may begin to approach agency interpretation from a position of doubt, particularly in important cases. We, therefore, find the law in a position where deference still exists as the basic approach, at least in theory, but is under increasing strain.
The debate, though, rests on the assumption that agencies approach statutory interpretation often from an ends-based type of reasoning. The hypothesis is that agency lawyers are told what the policy objective is and then they go about crafting the best legal argument that exists to support that objective, first in the rulemaking document and, then, if there is litigation, in briefing. Agencies are often viewed as simply acting like any other party and, if agencies act like any other party, why should they receive any sort of special treatment from a court? However, there is another way to think about how agencies should (and often do) approach statutory interpretation, which is one where agencies engaged in rulemaking implement the statutes they have, not based on any potential reading of the text, but on the agency’s best understanding of Congress’s intent behind the relevant provisions. Agencies are, after all, “creatures of statute,” and, thus, their authority is just what Congress gave them, no more and no less. Approaching statutory interpretation from the perspective of attempting to find the best reading accords with this role and, to some extent, reestablishes the view of agencies as at least semi-neutral parties, rather than advocates for one side or another. Viewed this way, statutory interpretation is only as “pro” or “anti” regulation as whatever the underlying statute is. This idea, which I’ll here call “fidelity” seeks to balance the contrasting views of deference by suggesting that agencies should seek to determine the best reading of the relevant statues by using all available resources, and, if they do conduct this analysis appropriately and base their action on the best available information, courts should defer to their judgment, regardless of the “size” of the issue. However, if agencies simply seek to make arguments that are merely convenient and supported by only overly clever or creative statutory interpretations, often to advance a particular policy goal, they should not receive any particular type of deference, even if those interpretations are not unambiguously foreclosed by a narrow view of the text.
To illustrate this point, I’ll discuss the two general types of statutory authority agencies rely on when issuing new regulatory requirements: (1) targeted statutes from Congress that direct an agency to conduct a particular rulemaking; and (2) broadly worded statutes that provide agencies with general powers. This is more of a continuum than a dichotomy, as broad grants of authority always come with some limits and statutory commands, while targeted statutes will generally include significant areas of ambiguity and gap-filling. But thinking about the two extremes helps clarify the two general approaches to review of agency interpretations that have arisen. Note also that many (and perhaps most) rulemakings simply modify existing regulations, which may have earlier been based on specific or general statutes; but those later rules generally do not touch on the validity of the underlying rule and will not be further discussed here.
First, many new regulatory requirements that an agency imposes are in response to clear statutory commands to conduct a rulemaking about a certain topic. These rules can have enormous economic and societal impacts and they can provide the relevant agency with some considerable discretion over crucial aspects of the rulemaking. And agencies can lose in litigation about these clearly-authorized rules and in ways that can set significant precedent, particularly regarding arbitrary and capricious review. But the bedrock questions that have animated the recent debate about “major questions” or delegation are not present; Congress has clearly and often very explicitly spoken on what the rule should be about, and the agency’s job is to follow that direction and issue a rule that meets the mandate and is based on the data. We can think of these rules as rules that would typically be analyzed under a “deference” framework, where the agency is generally engaged in relatively minor gap-filling and where the major policy decisions were clearly made by Congress in statute.
The second category has been the center of attention lately: new regulatory requirements that are based on broadly worded statutes that provide agencies with general powers. These types of rules are much rarer than the first group, and even when they occur, most of an agency’s new use of its general authorities is done on subjects that are very similar to prior uses. Some, though, are without question “big” rules, which have given rise to the growing body of caselaw on the “major questions” doctrine, the gist of which being that courts will come into an analysis of any “major question” doubting the agency’s position. At this stage, it is difficult to tell how far this doctrine will reach. It should be limited, as it would only apply to rules that have big effects based on laws that are broadly worded, which, even before the development of this doctrine, were rare rules. Its reach may also be limited because it is possible the doctrine will cause certain rules to never be issued in the first place. On the other hand, there is no clear answer about what is or is not a “major question.” The court’s formulation of the doctrine in West Virginia is somewhat helpful, but really only in retrospect, as it can only explain why the particular rule at issue qualifies (or, if the doctrine is not applied, does not). And this uncertainty is likely to be exacerbated in the near-term, as it is foreseeable that there may be some overuse of the doctrine as it develops in the lower courts, at least or until the Supreme Court steps in to ensure that it does not run wild.
These two types of regulatory actions present the two general judicial approaches to reviewing agency interpretations. On the one hand, we have the agency implementing a clear statutory requirement to conduct a rulemaking and, in those instances, we would expect courts to defer to the agency’s reasonable interpretation. And then, on the other, we have the “major questions” cases, where the court will review the agency’s legal argument from a position of doubt and skepticism, requiring that the agency prove its authority can support the rule at issue. Or to put it simply, we have cases where the court is likely to approach the authority questions from a posture of “deference” and those where it will approach those issues from a position of “doubt.”
The marginal cases are likely to be those rules that don’t fit neatly into one of these extremes, as agencies often conduct rulemakings that implement a provision that is specific in certain aspects but unclear in others. In those instances, agencies have both statutory guideposts but also significant discretion. (This, of course, is what proponents of the Clean Power Plan or the OSHA vaccine mandate argued supported the agency’s authority to issue rules in those instances, though, as noted above, we now know in retrospect that the Supreme Court considers those major questions.) We, therefore, have arrived at something of a dilemma, where the body administrative law seems poised between two contrary visions of how courts should view agency interpretations: whether a general posture of deference should continue or whether the doubt approach contained in the major questions doctrine (as well as the modern critics of Chevron) should inch further along, removing deference for more and more rules.
There is, though, a potential way out, which speaks less to how courts should review agency interpretations or even how agencies should defend rules in litigation, but more to how agencies interpret their authorities during the rulemaking process itself. As is often the case, it is worth going back to the source, which here is the Chevron decision itself. Before it was a doctrine, Chevron was just a case involving agency rulemaking, like many others. In Chevron, the Court set forth its now famous formulation as if it were simply explaining existing law, and, now, most of us stop there (unless you are interested in the Clean Air Act itself). But we should keep reading because how the Court looks at this question, “[b]ased on the examination of the legislation and its history,” Chevron, 467 U.S. at 845, can help point a way out of the dilemma administrative law has increasingly found itself in.
We have now come to expect that statutory construction will, more or less, not only start but also functionally end with the statutory text. At most, other statutory text may be discussed, but reference to extra-statutory materials is generally discouraged and viewed as an inherently weak argument. But the Chevron court was from an earlier time. As such, the Court examined closely the House and Senate reports, as well as a statement from Sen. Muskie, treating them all as serious sources that should be considered when determining how to interpret the statute. See id. at 851-53. In doing so, the Court was also generally guided by the overriding policy of Congress to balance economic growth with environmental improvement, which was made most clear in the Reports. See id. at 853. Of course, had the Court found that the statutory text foreclosed the “bubble concept,” none of this would have mattered, but the Court in Chevron clearly informed its view that Congress intended the agency to have some latitude in interpreting the core term (“source”) with a broad look at the formal legislative history and the more general historical context.
The decision in Chevron itself, which would become the standard for review of agency rulemakings, was not a “textualist” decision, at least not in the sense of how that term is generally used now. Instead, it took a full look at the context of statute before it, including the official House and Senate reports, and concluded that those purposes would be best served by allowing EPA significant discretion in filling the gap at issue. It is a somewhat ironic historic turn, then, that the Chevron doctrine would come to be applied during a strong “textualist” era throughout the Federal courts.
But is that era over? As Justice Kagan wrote in her dissent in West Virginia, “Some years ago, I remarked that ‘[w]e’re all textualists now.’ . . . It seems I was wrong.” Her criticism seems fair, even if one agrees with the majority’s decision, as that opinion is not based on the type of narrow, textualist approach to the specific text, but, rather, focused more on how this interpretation fits, generally, with the Court’s broader understanding of the Clean Air Act. It was not, “Can this language be read to support the government’s position,” but, “Did Congress really give the agency this type of authority?” But, if true, that leads to the core question here: so what? That is, the major questions doctrine, instead of simply being an attack on agency authority, as many of its critics contend, may be an acknowledgement that a pure “textualism” that focuses just on a portion of the statute without any broader considerations has limits, at least in the context of rulemaking. Perhaps, then, we are seeing a return to the more fulsome inquiry into Congressional intent that had been dominant in the past.
Such a shift may be a major change in how these questions have been addressed in litigation and in decisions, but we have never truly all been textualists, and certainly not in the narrow way. In particular, at the rulemaking level, interpretation of statutes is a far more dynamic activity than simply looking only at a small slice of a statute. Legislation is often an iterative process between Congress and the executive branch: there is technical assistance, official Administration proposals and positions; and discussions involving agency and Congressional staff. Throughout this process, a Congressional record is developed, as bills take shape and are debated, culminating in a statute, yes, but also, in most cases, a formal report from one or both chambers, as well as the relevant committee and subcommittees. And when implementing a statute, agencies will generally review and be aware of this extra-statutory information and history, as well as any ongoing discussions with Congressional staff. None of this is “binding” in the way that an enacted public law is binding or completely dispositive, and, to the extent there is any conflict between even an official report and a statute, the statute will prevail. Further, there may be times when there is good reason to disregard some or all of the legislative record, such as if there was a conflict between the House and Senate reports. Other times, agencies may find that a view expressed in the legislative record is inconsistent with broader statutory considerations, such as how an interpretation would fit within the agency’s broader authorities and programs. Further, not all legislative history should be treated equally, as the more formal statements should generally be viewed as having more persuasiveness than less formal history, such as member statements, records of debate, and failed amendments. But the legislative record does matter, as this history, as recognized by the Court in Chevron, helps determine how to interpret the statute consistent with “congressional intent,” which is to say, Congress’s purpose in enacting a particular statute.
To get back to the beginning of this essay: “Implement the statutes you have.” We have seen the development of two very different views of what this means. On the one hand, the agency should win unless it’s interpretation of the statutory text is essentially just bad. On the other, the agency should generally lose unless it can clearly show its view is correct. Neither view, though, seems correct. Agencies are close to Congress and have greater democratic accountability, through the President, than either the courts or the private litigants challenging a rule. Agencies are also home to sophisticated subject matter experts and are generally required to engage in public comment before issuing a rule. With these considerations, some amount of deference ought to be given to the agency. However, that level of deference should not be limitless, as all statutory grants of authority are themselves inherently limited and, thus, significant expansions of existing authority ought to require some type of convincing showing. There is also a compromise position (which is probably most consistent with the caselaw as it exists at this very moment) that deference will be given for “small” decisions, while the major questions doctrine, and thus doubt, will apply to “major” decisions. But this compromise ends up being nothing more than a constant struggle to define a decision as small or large, without any clear indication of what that means, and does not seem like it can survive this tension long, particularly in light broader criticisms of any type of deference.
What I suggest here is another way to resolve the need to defer to agencies in some manner, while asking the “really?” question for others, which is what I call “administrative fidelity” (only because I could not think of a good “d” word). Agencies are provided significant powers by Congress, and the purpose of all statutory interpretation in a rulemaking is not to come up with a plausible reading that supports or opposes a policy decision, but to determine and implement Congressional intent. When the statute is completely clear, then that’s the end of the issue. However, in the many instances where there is at least some ambiguity, the agency should seek to use all tools of statutory interpretation, starting with, of course, the text, but also considering the key legislative and Congressional history, to, as best as it is able, determine the best interpretation of the statute. Not simply a reasonable one, but the one that fits all relevant evidence the best. This would not be some sort of “unbounded” purposivism, as it would take seriously the usual “step 1” question to consider interpretations that are simply and clearly inconsistent with the statute, and it would further acknowledge that the “step 2” question is an actual inquiry and not just an automatic win for the agency, which is, in many ways, a return to the original analysis in Chevron. There will be cases where, after conducting that sort of analysis, the agency will be able to say, as the Court in Chevron found for EPA, that its reading is allowed and consistent with the statute’s overriding purpose, even if another interpretation would also be allowed. But, in order to reach that conclusion, the agency will need to “show its work.” The Administrative Procedure Act, after all, is primarily a procedural statute, and thinking about statutory interpretation as a process the agency should engage in to receive deference from courts fits within that broader framework. Deference, then, would be granted as a matter of the agency going through the often difficult task of doing its best to find actual intent to the extent possible, and having to defend that argument in both the proposed and final rules. The statutory text would still control but the agency would need to engage seriously with the legislative record, even if it ends up arguing it is not persuasive. Rulemaking is different from litigation, as it is the process of the agency determining the scope of its authority, as compared to challenging or defending a decision already made, and agencies should exercise that authority carefully in all circumstances, regardless of the size of the rulemaking and the likelihood or existence of a judicial challenge.
“Fidelity” would also work to constrain the major questions doctrine from growing too far and swallowing the entire concept of deference, as the question would not turn on whether a rule is large or how specific the statutory text is, but rather whether the totality of the evidence supports the agency’s interpretation. For claims of very broad authority, the agency may be pressed to come up with an argument that its claim is actually right, and the results would be the same as under the current major questions doctrine. However, there would be cases where the agency could show that, despite vague text, Congress’s clear intent was to provide the agency with the ability to issue “major” rules to address new circumstances. Though I suppose some would at this point argue that this would be where the “non-delegation” concept would enter into the analysis, that is an argument for another day.
Finally, I want to stress that the concept of “fidelity” is not intended in any way to take sides on the substantive policy questions of any particular rule or even the more general question of “should there be more or less regulation?” Chevron itself involved a deregulatory rule, and administrations of both parties certainly made deference arguments when they were available or likely to be persuasive. Further, while the “fidelity” idea would in some instances make it harder for agencies to claim a new authority, it would also limit in other circumstances the ability to avoid or postpone obligations or deregulate based on certain types of legal theories. Instead, the point is to stress that the role of agencies is to implement the will of Congress to the best of their ability and, to the extent they do that in a rule, they should receive deference. In contrast, interpretations that seek to interpret text that is not consistent with a complete examination of the relevant context should be viewed as more similar to the arguments of any other party, which a court can choose to agree with or not.
“Implement the statutes you have.” What else, really, is there?
Tim Mullins is an attorney with the U.S. Department of Transportation. The views expressed in this publication are the author’s own and do not necessarily represent the views of the U.S. Department of Transportation or the United States.