Double Aspect

In Loper Bright, the Supreme Court of the United States overturned the long-standing judicial review doctrine of Chevron deference. The decision has been met with all manner of criticism and celebration. There is no doubt that Loper Bright sends an important and valuable signal: it is the judicial role to decide questions of law, wherever they arise. But how this review is conducted, and what it portends for the American administrative state, is not so amenable to clear analysis. In this sense, I think there is a difference between what Loper Bright signals, and what it mandates as doctrine. Even so, there are some surprising Canadian connections that we should do well to heed on both the signal and doctrinal levels.

What Did It Decide?

Chevron’s initial two-step structure was simple enough. First, using the ordinary tools of statutory interpretation, courts would determine 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, and the court must give effect to the unambiguously expressed intent of Congress. If the court determines that Congress has not directly addressed the precise question at issue, the court will defer to the agency’s interpretation as long as that interpretation is based on a permissible construction of the statute.

In reality, Chevron’s demise was not a surprise. The Court had not deferred to a regulation under Chevron since 2016. Its domain, through the introduction of erratic distinctions, had been considerably narrowed. But the analytical path to ice Chevron was not exactly clear. Chevron, by now, had the vintage of stare decisis, and those committed to a strong theory of precedent might see it as a stable, entrenched rule that should be clarified and tailored, rather than totally jettisoned.

But for the majority in Loper Bright, Chief Justice John Roberts saw Chevron as inconsistent with fundamental constitutional and statutory law. Roberts CJ begins from the proposition that it is the role of the courts to say what the law is. Importantly, this commitment is reflected in the code that governs the American administrative state—the Administrative Procedure Act [APA]. APA,s.706 says: “To the extent necessary to decision and when presented, the reviewing court shall decide all relevant questions of law, interpret constitutional and statutory provisions, and determine the meaning or applicability of the terms of an agency action.” From this statutory prescription, Roberts CJ saw Chevron’s command of deference as inconsistent with this basic recognition of constitutional reality.

For Roberts CJ, the motivating justifications for Chevron deference were weak Roberts CJ notes that Chevron’s broad presumption was rooted, in part, in a recognition that “agencies have subject matter expertise regarding the statutes they administer…” (24). This broad assumption of expertise—with which Canadians should be familiar—is a weak reed on which to rest a broad presumption of deference. As Roberts CJ notes, not every question before an administrative agency will squarely engage its expertise; but “Chevron’s broad rule of deference… demands that courts presume just the opposite” (24). The concern relates to potential false positives: deference granted where there is no actual demonstration or existence of policy expertise.

The more likely candidate to ground Chevron was its reliance on a presumption of statutory interpretation. Chevron assumed that any “ambiguity” in a statute delegating power to an agency carried with it a congressional presumption that this ambiguity should be resolved by the agency. On this account, ambiguities are spaces for policy-making trusted to the agency. In a remarkable passage, Roberts CJ puts to rest this presumption: “[p]resumptions have their place in statutory interpretation, but only to the extent that they approximate reality. Chevron’s presumption does not, because ‘[a]n ambiguity is simply not a delegation of law-interpreting power. Chevron confuses the two’” [citations omitted].

From this, it may appear that any form of “deference” to administrative interpretations of law has been excised from the law of judicial review. However, Roberts CJ draws out two instances in which courts may, at the very least, grant non-binding weight to administrative interpretations of law. First, Roberts CJ gives so-called Skidmore deference a starring role in the new judicial review framework. Skidmore deference provides that an agency decision, informed by its expertise, may be “entitled to respect.” Roberts CJ endorses this non-binding form of judicial deference—triggered by judges where the situation is relevant—as consonant with the command of de novo review enshrined in the APA.

Secondly, and perhaps most importantly, Roberts CJ is alive to a key problem in all delegating statutes: the tendency for legislatures to delegate power in broad, qualitative terms. In a tantalizing passage, Roberts CJ opines that “[i]n a case involving an agency, of course, the statute’s meaning may well be that the agency is authorized to exercise a degree of discretion.” For example, sometimes—often—statutes delegated power to an agency to “regulate subject to the limits imposed by a term or phrase…such as ‘appropriate’ or ‘reasonable’.” In such cases, the APA’s command must be fulfilled differently.  Here, the court must independently fix the boundaries of the delegation, while ensuring that the decision-maker engages in reasoned decision-making within those boundaries—this is the function of so-called “hard look review” under the APA.

In dissent, Kagan J offered both principled and practical critiques of the majority’s reasoning. For her, Chevron was a stable background rule that generally approximated Congress’ expectations about the scope of its delegations:

That rule has formed the backdrop against which Congress, courts, and agencies—as well as regulated parties and the public—all have operated for decades. It has been applied in thousands of judicial decisions. It has become part of the warp and woof of modern government, supporting regulatory efforts of all kinds—to name a few, keeping air and water clean, food and drugs safe, and financial markets honest.

Kagan J was not convinced that the majority’s attempt to root the reviewing function in the APA—and bypass Chevron’s confusing “ambiguity” requirement—would lead to doctrinal stability. Leaning heavily on stare decisis, Kagan J sees the majority’s reliance on the APA as too weak to ground a new doctrine of judicial review. In heavy hitting words, she takes the majority to task:

Its justification comes down, in the end, to this: Courts must have more say over regula­tion—over the provision of health care, the protection of theenvironment, the safety of consumer products, the efficacyof transportation systems, and so on. A longstanding prec­edent at the crux of administrative governance thus falls victim to a bald assertion of judicial authority. The major­ity disdains restraint, and grasps for power.

Analysis

In my view, Loper Bright should be viewed as sending two messages. The first message is a signal about how judicial review courts should speak about the reviewing function. On this score, gone is the language of “ambiguity,” “deference,” and “policy-making space”; in is the new language of “fixing boundaries,” with a focused recognition on the role of courts in deciding questions of law. The second message sent by Loper Bright is the doctrinal message—how should courts, under this rubric, actually conduct review? Much remains to be settled on the mechanisms of the review.

When I say that Loper Bright sends a signal about how courts should conduct review, I do not mean to disparage the case. How courts speak of the reviewing function can be just as important as how it is actually conducted. In Canada, we are well aware of this problem—under the guise of “deference” and “reasonableness,” our Supreme Court has been notoriously inconsistent, engaging in disguised de novo review with a disturbing frequency. So, it matters that signals match doctrine, and understanding the signal is important.

The Loper Bright signal is clear: it is emphatically the judicial role to decide questions of law. Roberts CJ roots this conclusion in the US Constitution and Marbury v Madison, but the proximate justification for the conclusion is the APA, which appears to affix an important role for courts in deciding questions of law. To be sure, it is unclear whether the APA should be interpreted in this way, as Kagan J points out. But to the extent it should be, the signal sent by Loper Bright is crystal clear: courts cannot contract out of their role to police statutory boundaries through the language of deference, ambiguity, or expertise.

The justifying presumption for Chevron was always somewhat specious. This is because the trigger for Chevron—ambiguity—was always in the eye of the beholder. Judges more inclined to grant judicial deference to administrative interpretations of law may, in turn, be more inclined to find “ambiguity” in delegating statutes. Thus, it became possible for deference to be granted even when—with a bit more work—meaning could be extracted from the statute. The weakness of ambiguity as a trigger—a point about which I’ve written before—means that the deference doctrine was always a matter of judicial first impression.

It is tempting to take this signal, as Kagan J does, to say that modern government will be hobbled by Loper Bright, subject to the whims of an interventionist judiciary. But much like Kagan J’s similar comments in the non-delegation context, it is hard to imagine that modern administrative government will collapse with Loper Bright. For one, judicial review doctrine simply matters less than lawyers think. Judicial review is just the tip of the iceberg. Many (most) administrative decisions are never challenged. And even if they were, Canada now stands isolated among its Commonwealth and Anglo cousins in speaking of deference. The United Kingdom, Australia, and New Zealand, all conduct de novo review on questions of law. And in these countries, their administrative states nonetheless stand strong.

The doctrinal message sent by Loper Bright is a bit more nuanced, but the move from ambiguity to the scope of delegation is welcome.  As I’ve said, Loper Bright contemplates two ways in which the APA’s command for de novo review could be qualified. Skidmore deference could, in cases where there is a wide gap between the agency’s knowledge and a court’s, offer an “out” for a court to grant non-binding weight to administrative interpretations of law. This non-binding weight is consistent with the APA’s command. Nowhere does Loper Bright say that courts cannot benefit from agency expertise in particular cases—much like they may benefit from a party’s argument in a particular case. The important point is that the choice of whether to defer remains with the court—its hand is not forced by, say, a requirement of “ambiguity.” Whether and how this review will function remains to be seen.

On the other hand, and more interestingly, Roberts CJ suggests that, sometimes, courts may only need to “fix the boundaries” of a delegation in order to give effect to the APA’s command. Broad delegating authority may mean that, in a particular case, there may not be much “law” to apply. Where, for example, a legislature delegates power to an agency to act in the “public interest,” Loper Bright strikes a good balance. It does not say that such delegations carry no legal content. Rather, what the “public interest” means in a a regulatory regime will depend on the text, context, and purpose of the delegating statute. At the same time, Roberts CJ is aware that in these cases, hard look review takes on greater importance to ensure that reasoned decision-making takes place within those boundaries. How this review will function also remains to be seen, but by shifting the focus to the scope of the delegation rather than ambiguity, the courts will at least be able to more clearly delineate cases in which the agency has wider scope to move from those where a term is narrower, with more legal content to constrain the decision-maker.

Loper Bright, in a manner of speaking, does not do away with a division of labour between courts and administrators, even putting aside Skidmore deference, which could rear its head more frequently. Rather, it puts the onus on the legislature. Where the legislature speaks in broad and qualitative terms, the court will apply the law to the extent it can extract legal content from the delegating language. Whether this turns into ambiguity redux is an open question–it depends on whether there really is a difference between Chevron’s ambiguity requirement and Loper Bright‘s recognition of broad delegations. But, at least at a conceptual level, there should be a difference. The term at issue in Chevron, “stationary source,” is much narrower, and more amenable to legal resolution, than a broad delegation to act “in the public interest.” Under Loper Bright, it is less likely that an agency would have flexibility to interpret “stationary source” than it would to interpret “public interest.” This does not rest on a contested presumption of statutory authority. Rather, it rests on a recognition of constitutionally-recognized legislative power to delegate to an agency the power to “fill up the details” of a statute. Whatever the policy problems with this approach, it is constitutionally unobjectionable. And because Chevron, as a rule, is no more, the court’s review function will be fixed by the variable ways in which legislatures express themselves in individual statutes. But this is a function of legislative choice, and the legislature undoubtedly carries the constitutional capacity to make these choices.

Connection to Canada

What does Loper Bright offer for Canada?

Let’s address, first, Loper Bright’s signal. In our law, it was once recognized that the law of judicial review is just a specialized branch of statutory interpretation. This memorable comment in Bibeault (Beetz J, citing de Smith) captured an elemental aspect of the law of judicial review: whether framed in terms of jurisdiction or statutory authority, the role of courts is to enforce the boundaries of the delegating statute, as best it can.  Through the pragmatic and functional era and the Dunsmuir era—and, continuing into Vavilov—we in Canada still speak of “deference” as a unified and distinct concept. Sometimes, we even say that our review is “deferential,” but also “robust.” It is hard to assign meaning to these adjectives. I have come to the view that speaking of deference in this way is liable to confuse. If—as I will point out below—the degree of “deference” is attributable to the scope of delegating power, then it may make sense to speak of statutory interpretation alone, rather than a special concept of deference. To the extent Loper Bright sends this signal, it is worth listening.

Incidentally, Roberts CJ’s opinion is rooted in the APA, which is a super-statute governing procedural and substantive incidents of judicial review. In Canada, there has been comparatively less systemic legislative attention given to the administrative state. Aside from British Columbia’s Administrative Tribunals Act, most statutes governing review and specifying procedures at the provincial and federal levels are quite weak. The Federal Courts Act, which does specify grounds of review was read by the Supreme Court in Khosa (wrongly, in my view) to essentially incorporate the common law standards of review. The Statutory Instruments Act, which purports to govern parliamentary review of delegated instruments, is notoriously weak. A Canadian APA could address all of these problems: it could more clearly specify, within constitutional limits,  the appropriate posture of review, ending the standard of review angst that has characterized our law; it could strengthen the anemic Statutory Instruments Act and deal with the problem of delegated legislation, which is the predominant mode of law-making in modern Canada. Standardizing review, at least at the federal level, could lead to a more predictable application of the standards of review.

On the other hand, the doctrinal mechanics of Loper Bright review, compared to Vavilov, are not as jarring as one might think. True, Vavilov tells us to start with the reasons and again it wraps the review in the language of deference. But as mentioned above, Loper Bright does not do away with agency flexibility. Given that the APA governs the review, and that delegating statutory language is the locus of that review, Vavilov‘s laser focus on institutional design choices–like rights of appeal–is apposite. Indeed, as Stratas JA says in Alexion, legislative language is the “biggest constraint of all” on administrative action (para 27). This is borne out in the post-Vavilov cases. So, though our review is characterized by words like deference, the best instances of Vavilov’s application centre on paras 108-110, and 117-118 of that decision, which outline the following propositions:

1. The “governing statutory scheme is likely to be most salient aspect of the legal context relevant to a particular decision.”
2. The language of the delegating language conditions the review, and administrators–like courts–are bound by the ordinary rules of interpretation: “If a legislature wishes to precisely circumscribe an administrative decision maker’s power in some respect, it can do so by using precise and narrow language…Conversely, where the legislature chooses to use broad, open-ended or highly qualitative language…it clearly contemplates that the decision maker is to have greater flexibility in interpreting the meaning of such language. Justification, as I note below, is an ethic of the Vavilov framework, but it is subordinate to the statutory language: “It will, of course, be impossible for an administrative decision maker to justify a decision that strays beyond the limits set by the statutory language it is interpreting” [110].

Taken together, and as applied, these propositions are also at the heart of Loper Bright. In each case, courts are instructed to fix the boundaries of the delegated language, paying attention to the breadth of that language, while conducting reasonableness review within the boundaries of the language. As in Loper Bright, Vavilov rejects expertise as a binding reason for deference—again, because of the potential for “false positives” (see Vavilov, at para 30). The only justification for reasonableness review is the very fact of delegating power itself. And note that the trigger of ambiguity is absent from Vavilov, as it is now from the American review framework. Instead, the posture a court takes on review should be conditioned by the scope of the language itself, rather than functional or practical reasons for deference, which may influence but cannot bind.

Perhaps the best example of this sort of review in Canada took place in the Emergencies Act case. In that case, the question was whether the Governor in Council’s invocation of the Emergencies Act met the structured and narrow provisions of that statute. Here, rather than a broad delegation of authority to the Governor in Council to declare emergencies, the statute was rather more prescriptive. Finding that the GIC’s implied interpretation of that language did not meet its specific constraints, the Court did not grant deference to the GIC in the abstract. Rather, the “amount” of “deference” was fixed by the statutory language, which erected narrow space for the GIC to operate. One can also point to the Federal Court’s recent plastics case for an example of this sort of review.

To be sure, and because Vavilov’s signal is not as strong as Loper Bright’s, this view of Vavilovian reasonableness is contentious. Others could easily glom on to other aspects of the decision that say something else. For example, it is true that justification remains an important aspect of the Vavilov framework. The conceptual relationship between Vavilov‘s focus on institutional design choices and justification is fraught: for example, can we attribute a legislative presumption that when a decision raises particular stakes, legislatures expects courts to conduct stricter review? What of the focus on reasons? For now, it is enough to say as in Loper Bright, that justification and responsiveness occur within the boundaries of the statutory language. For example, determining whether or not a particular exercise of administrative authority is reasonable will depend on the degree to which the decision-maker was responsive to particular submissions that might condition the exercise of its legal power, as well as whether the interpretation itself falls within the enacting language. The focus on reasons is still appropriate here: it was the agency that was delegated the power to render a decision at first instance, and so to understand what that decision was, a focus on the reasons is appropriate. As Vavilov says, no justification can be offered that goes beyond that language.

Moreover, it is true that Vavilov sometimes speaks out of both sides of its mouth. Vavilov was clearly a compromise decision. But to the extent we can work out a rational mode of review from that compromise, it appears clear that the Vavilov majority wanted to tighten up reasonableness review on questions of law. If that is so, then it should not be a surprise to see Loper Bright review and Vavilov review sometimes coalescing in justification and function.

Of course, and again, how we speak of the law of judicial review matters. Still, we speak of presumptions of deference in Canada; we speak of reasonably robust reasonableness review that is nonetheless deferential; and we say that administrative justice does not need to look like judicial justice. These notions are part of our law, unlikely to go anywhere. But they should be put in their place: these turns of phrase risk distracting from the core task on judicial review, reflected in Vavilov‘s actual doctrinal mechanics, which leave room–in appropriate cases–for flexibility in the interpretation of an enabling statute. Loper Bright, in so many words, does the same thing. In this way, perhaps Canadians and Americans could learn from each other in calibrating our respective review frameworks.