I’ve recently been blogging about my new article, The Inherent-Powers Corollary: Judicial Non-Delegation and Federal Common Law, which I’ve posted to SSRN.
Here’s a link to my third post, where I discussed my basic thesis — that the non-delegation has an “Inherent-Powers Corollary”, under which delegations are judged more charitably if they are into an area where the delegate already has inherent power. (The Inherent-Powers Corollary also has an “Interlinking Principle”, under which all that’s necessary is that the delegated power be interlinked with (not necessarily actually within) the delegate’s inherent powers.)
Here’s a link to my fourth post, where I started applying this principle to actual delegations; my first example was delegation to the courts to make procedural rules. Here’s a link to my fifth post, where I used the Inherent-Powers Corollary to discuss the constitutional foundations of the Erie doctrine. And here’s a link to my sixth post, where I discuss various other inherent judicial powers, including the power to make common law in subject-matter “enclaves” like admiralty or foreign affairs, craft defenses, and design remedies.
Today, I’ll talk about the type of delegation to courts that consists of Congress passing an ambiguous statute — an implicit delegation to courts of interpretive power.
First, I discuss the threshold question of whether ambiguity really is a delegation at all: is it really right to characterize the interpretive power that courts use when faced with ambiguous statutes as an exercise of delegated power?
Next, I discuss the problem of delegated lawmaking: if a statute invites the development of common law, or incorporates a common-law test — or any test whose meaning is known because it’s been fleshed out over a series of past cases — does that mean the courts have to freeze a preexisting test in place, or can they continue to develop the caselaw?
Finally, I apply the Inherent-Powers Corollary to four illustrative statutes: the residual clause of the Armed Career Criminal Act, the Religious Freedom Restoration Act, the Alien Tort Statute, and section 1 of the Sherman Act.
* * *
Now, I’ll examine the implicit “delegation” to courts that occurs every time Congress passes a statute with an ambiguous term that courts must interpret. Cases of common-law-style and policy-laden judicial development of vague statutory standards are of course pervasive—I’ll limit myself here to a handful of cases.
Is Ambiguity a Delegation?
First, is it really a delegation for Congress to pass an ambiguous statute? As I’ve noted above, it’s perfectly conventional to refer to certain open-ended statutes—chiefly the Sherman Act, but also other statutes like the Alien Tort Statute and securities laws—as delegations to the judiciary. Moreover, when Congress passes an agency-administered statute with an ambiguous term, the Chevron rule (when it applies) interprets the ambiguity as an implicit delegation of interpretive power to agencies. If that’s the case when an agency is doing the interpreting, it seems that the same ambiguity in a judicially administered statute is an equivalent delegation of interpretive power to courts.
The “ambiguity as delegation” perspective shows up in some judicial opinions and is quite common in the scholarly literature. Dan Kahan, for instance, discusses at length how, even though—in theory—there are no federal common-law crimes, in fact Congress has delegated the creation of substantial criminal law to the courts by means of open-ended statutes like RICO or the mail fraud statute. Working against the courts’ exercise of their delegated criminal lawmaking authority is the rule of lenity, which—again in theory—by choosing the more lenient of competing interpretations of statutory language, is delegation-blocking. The rule of lenity sits uneasily with delegated lawmaking, but this tension isn’t widely appreciated by those who don’t recognize that ambiguity is a delegation.
Justice Scalia (who is, incidentally, one of the foremost proponents of the rule of lenity) disagrees with this perspective. He writes that, unlike agencies, which can choose any reasonable interpretation of ambiguous text:
[c]ourts . . . must give the statute its single, most plausible, reading. To describe this as an exercise of “delegated lawmaking authority” seems to me peculiar—unless one believes in lawmakers who have no discretion. Courts must apply judgment, to be sure. But judgment is not discretion.
One can respond to Justice Scalia in at least three ways.
First, why must a delegate have discretion—why is it unthinkable to imagine that “lawmakers who have no discretion” might be exercising delegated power? Delegation doesn’t definitionally require that the delegate have discretion; even if an act is perfectly ministerial, Congress can delegate the power and duty to perform it to the President or anyone else. Congress sometimes directs an agency to do a very specific task—for instance, write regulations whose content is already largely specified in the statute. The non-delegation doctrine has no trouble handling discretionless delegations. They’re clearly valid because the presence of an intelligible principle is obvious. But they’re still delegations.
Second, the exercise of judgment in finding the single best meaning does in fact often involve significant discretion. Justice Scalia himself wrote, in Mistretta, that:
a certain degree of discretion, and thus of lawmaking, inheres in most . . . judicial action, and it is up to Congress, by the relative specificity or generality of its statutory commands, to determine—up to a point—how small or how large that degree shall be. Thus, the courts could be given the power to say precisely what constitutes a “restraint of trade” . . . because that “lawmaking” was ancillary to their exercise of judicial powers.”
So Justice Scalia has recognized that at least interpreting “restraint of trade” from the Sherman Act involves “a certain degree of discretion,” and therefore some lawmaking. It would be strange otherwise, given the extent of judicial exegesis in antitrust. Likewise, the proliferation of substantive canons of interpretation has been an exercise of judicial discretion. One might say the same of Chevron, which is likewise a rule of statutory interpretation that judges adopted in their discretion, in part for policy reasons.
Third—and relatedly—there are many interpretive methods out there. Justice Scalia is a textualist who justifies his textualism in part by the need to control judicial discretion. But others—including purposivists, believers in dynamic statutory interpretation, and even Justice Scalia sometimes—allow for consideration of policy consequences and to that extent introduce discretion into statutory interpretation. Moreover, textualism itself isn’t fully determinate. But even if one imagined a hypothetical judge who was always applied his textualism in a way that lacked discretion—call him Scalia-Prime—that would be Scalia-Prime’s choice and isn’t inherent in the exercise of statutory interpretation.
So interpretation of vague terms is a sort of judicial discretion that, while perhaps often less amenable to overt policy-driven choices, isn’t different in kind from the delegations with discretion that we see for executive agencies.
It’s clear, then, that I’m not wild about Justice Scalia’s theory of why statutory interpretation isn’t an exercise of delegated power. But he may be partly right after all. Just because judges are using discretion in interpreting statutes doesn’t mean that they’re necessarily using a delegated discretion. The status of statutory interpretation is unclear and somewhat metaphysical. Congress sometimes gives specific interpretive instructions, but often doesn’t. When explicit guidance is lacking, do we as judges choose an interpretive method because Congress has ordered us to do so, or because we believe it’s true? The former situation is a delegation; the latter is just an exercise of our own inherent power to figure out what words mean.
Both ideas—congressional intent as to judicial interpretation, and judicial interpretation not tied to congressional intent—are common in federal law.
The fiction that Congress intended to implicitly delegate interpretive power to agencies is essentially the theory of Chevron and Skidmore deference. Our current free-wheeling regime of antitrust policymaking is also often justified as stemming from congressional intent to give this power to courts.
On the other hand, the choice of interpretive method isn’t always justified by reference to actual congressional intent: it’s often justified on the grounds that it’s the only permissible method, that it provides a useful baseline that Congress could take into account (if it chooses) when it legislates, that textualism (say) is the inherently correct way to determine what words mean, and so on. For many substantive canons, the presumed congressional intent is fictitious—and some canons can’t possibly reflect congressional intent, for instance a canon of constitutional avoidance applied to avoid a constitutional problem that wasn’t recognized by courts until after the statute was passed. And different judges can take different attitudes toward statutory interpretation: though some interpretive methods (for instance, Chevron) have been mandated by Congress or the Supreme Court, much of the time the choice of interpretive method isn’t really considered “law” (though of course the bottom-line interpretation of the statute is law).
As a result, though there’s plenty of discretion in statutory interpretation, it’s often unclear whether this discretion is really delegated—and different observers will see it differently. Thus, it’s unclear whether the non-delegation doctrine really applies. Some might speak of vague statutes (to the extent they’re interpreted as mandating excessively indeterminate interpretive methods) as being unconstitutional delegations to the judiciary; others might speak instead of excessively indeterminate interpretive methods being beyond judges’ power to adopt.
These are two different but more or less equivalent ways of looking at a statutory interpretation problem. One of them doesn’t implicate the Inherent-Powers Corollary, and the other one does:
- The statute doesn’t say, or even imply, how it should be interpreted. The judge uses his own judgment as to which method to use. Here, there’s no delegation. Of all the interpretive methods out there, some might be within the judge’s inherent power, and some might not be. This inquiry may involve the illegitimacy of excessive judicial policymaking, but that will be because of Article III concerns, not Article I concerns. If the judge uses a permissible method, so much the better; if not, we can say that he’s acting ultra vires.
- The statute doesn’t explicitly say how it should be interpreted, but one can infer it using interpretive techniques. Now that’s a delegation. To the extent Congress is implicitly mandating a method that’s already within the judge’s inherent power, the Inherent-Powers Corollary says all is well, even if that method is quite indeterminate. Even if the method is slightly beyond the judge’s inherent power, but “interlinked” with it, the Interlinking Curlicue says all is still well.
These two rhetorics of statutory interpretation are more or less equivalent, but not completely: the main difference between them is the applicability of the Interlinking Curlicue. Under the delegation view, there could be some methods that judges can’t use on their own but that are interlinked enough with judicial power that Congress could require the judges to use them.
Nick Rosenkranz, for instance, discusses “starting-point rule[s]” or “default rule[s],” which only Congress can change; perhaps, for instance, judges are constitutionally required to apply the rule of lenity or the canon against altering the state-federal balance or horizontal statutory stare decisis, but Congress is constitutionally permitted to override those methods; and maybe those overrides are interlinked with judges’ core powers. If the method is interlinked, Congress is subject to the relaxed version of the non-delegation doctrine, under the Inherent-Powers Corollary. Congress might also be able to mandate other methods that aren’t interlinked with judges’ core powers; this is fine, but then Congress is subject to the ordinary non-delegation doctrine and can’t mandate a method that’s too indeterminate.
If you think that any interpretive method is within judges’ power—or at least interlinked with it—then by the same token, Congress could implicitly or explicitly mandate any method without violating the non-delegation doctrine. But if you think that some methods are beyond judges’ inherent power—and not even interlinked with that power—then, to the extent they’re too indeterminate, Congress might violate the non-delegation doctrine when it implicitly or explicitly mandates those methods in a statute. Perhaps one example might be if the Sherman Act candidly told judges to develop antitrust law “in the light of reason and experience.”
Because, as I noted earlier, it’s not clear what exactly “interlinking” involves, it’s not clear how much leeway the interlinking possibility gives Congress—and thus how much leeway this gives judges in reading implicit interpretive directions into statutes.
* * *
The Problem of Delegated Lawmaking
Let’s step back for a moment and observe a potential difficulty in applying the Inherent-Powers Corollary to statutory interpretation: the Inherent-Powers Corollary requires us to judge congressional delegations by reference to the powers that the courts would have had in the absence of the delegation. But the power to interpret statutes is, by definition, contingent on a statute’s having already been passed. (This problem has the same structure as the problems of defenses and remedies, which I discussed earlier.)
We can’t rely on a necessity argument (“Courts must interpret federal statutes once they’ve been passed, therefore they may interpret federal statutes”). Clearly courts have some ability to determine the meaning of words in English—this much must be part of the meaning of the phrase “Judicial Power.” But that doesn’t tell us whether Congress has to express itself with the specificity required of the classic non-delegation doctrine, or whether it can be unspecific in areas of preexisting judicial power, as the Inherent-Powers Corollary would allow.
Fortunately, this question needn’t trouble us too much: the judicial power to interpret is prior to congressional statutes. Courts can and must interpret the Constitution (this is applying “the usual principles . . . to an unusual text”), state statutes (often not considering state courts’ interpretive methods binding on federal courts), and treaties (the applicability of the non-delegation doctrine to treaties is an open question). So interpretation is an inherent—i.e., statute-independent—power that can support an exercise of the Inherent-Powers Corollary.
The harder it is to determine semantic meaning or congressional intent, the more statutory interpretation shades into lawmaking. At one extreme are statutes where an elaborate common-law-style edifice has been constructed on the barest of text. I’ll discuss four such statutes: the residual clause of the Armed Career Criminal Act, the Religious Freedom Restoration Act, the Alien Tort Statute, and § 1 of the Sherman Act.
The last three of these statutes have been conceptualized as invitations to the judiciary to develop a federal common law. The recurring question is how the courts should treat such an invitation. If a statute that merely states a heretofore undeveloped standard and tells courts to develop it “in the light of reason and experience,” it would seem hard to justify it by reference to an inherent power to interpret words; such a statute would be automatically constitutional under the Inherent-Powers Corollary (plus the Interlinking Curlicue) only if it were interlinked with a power courts already had in that area—for instance, as noted above, the delegation to courts of the power to develop evidentiary privileges or statutory defenses. Without the possibility of relying on inherent powers, we’d need an intelligible principle, and the phrase “reason and experience,” as a naked appeal to the judgment of the courts rather than any judgments made in the first instance by Congress, can’t be good enough.
Now let’s consider delegations that are less stark: statutes that state an existing legal test. RFRA states the pre-Smith test for religious exemptions; the Alien Tort Statute refers to the preexisting concept of torts that violate the law of nations; the Sherman Act refers to the pre-1890 notion of unreasonable restraints of trade; Federal Rule of Evidence 501 (despite its “reason and experience” language) was enacted against a background of prior caselaw on evidentiary privileges. In the area of tort law, the Federal Tort Claims Act incorporates the tort law of all 50 states, and—more dramatically—the Federal Employers’ Liability Act has been taken to direct courts to develop a federal tort law.
Interpreting the statutes as freezing the earlier caselaw in place might seem as though it doesn’t pose non-delegation problems, because then Congress isn’t inviting courts to make common law. Daniel Farber and Brett McDonnell suggest, for instance, that a textualist approach to antitrust should favor a static incorporation of the common law of restraints of trade; Kahan notes that the federal courts could have “tied the meaning of the mail fraud and conspiracy statutes to some finite schedule of deceptive practices that existed at the time these statutes were enacted, thereby forcing Congress to enact additional statutes to deal with any new forms of dishonesty or deception”; Kontorovich argues for static incorporation of Founding-area law-of-nations torts.
But it might not be possible to avoid making common law—even in criminal law, despite the supposed prohibition on common-law crimes and despite the limiting effect of the rule of lenity on statutory interpretation. Even before Smith, people might have disagreed on the direct (vs. indirect) implications of existing religious exemptions caselaw. It’s actually difficult to distinguish between participating in the ongoing development of caselaw and merely limiting oneself to finding what was there the whole time. (Indeed, this problem regularly causes trouble in AEDPA or § 1983 qualified immunity cases where the question is whether a right is “clearly established.”)
Rather than trying to grasp that elusive distinction, we should instead focus on whether the prior caselaw was well-enough developed at the time of enactment. Perhaps, if no prior caselaw existed, a command to develop a common law on the subject would lack an intelligible principle. But given a determinate enough baseline of prior caselaw, developing that caselaw in common-law style either satisfies the intelligible principle—or satisfies the Inherent-Powers Corollary because it’s an exercise of the federal common-lawmaking power to interpret terms.
For instance, we might take the FTCA’s incorporation of the tort law of all 50 states as the paradigm of well-enough developed caselaw: some states have been developing their common law for centuries, and even Hawaii has received “[t]he common law of England, as ascertained by English and American decisions,” by statute. (The FTCA can also fit into the enclave of torts involving the federal government.)
If a case comes up that’s somewhat similar to prior cases but somewhat different, the common-law method consists of analogizing and distinguishing the prior cases to come up with a decision that “fits” the overall structure of the caselaw. This is similar to the common-law decisionmaking that agencies use when they operate by adjudication instead of by rulemaking—for instance, when the NLRB develops a nationwide statutory common law of employees and independent contractors rather than incorporating state-by-state definitions. If such agency-based common-law decisionmaking satisfies the non-delegation doctrine, it’s hard to say that it should violate the doctrine when courts do.
Admittedly, I’ve argued that the non-delegation doctrine should be applied more strictly when courts rather than agencies are delegates, so the mere fact that agencies can do it doesn’t necessarily mean that courts can do it. But another factor tending to validate this sort of common-law decisionmaking is that it’s very similar to traditional tools of statutory interpretation, like the canons ejusdem generis and noscitur a sociis. Determining whether one case is close enough to previous cases is similar to determining whether an item asserted to be part of a non-exhaustive list is similar enough to the items already listed.
Does this approach prevent Congress from (explicitly or implicitly) mandating any methods of statutory interpretation? Some forms of “dynamic statutory interpretation,” which involve overt judicial policymaking and reliance on subsequent legislative history, or even just aggressively applied (static) purposivism, might be illegitimate. So there are at least some outer bounds to Congress’s ability to give policymaking authority to judges in the guise of statutory interpretation.
As I’ve already suggested, for these sorts of statutory interpretation cases, it’s hard to distinguish between the following two formulations: (1) under the ordinary non-delegation doctrine, we have an intelligible principle (because the term was well-enough defined given existing caselaw at the time of enactment); and, (2) under the Inherent-Powers Corollary, we don’t need an intelligible principle because of the federal common-lawmaking power to interpret terms.
The first formulation sounds intuitively better than the second: if we didn’t need an intelligible principle because we could rely on the judicial power to interpret terms, and that judicial power existed in this case because the terms could be determined by ordinary statutory interpretation, then wouldn’t that be an intelligible principle?
But both formulations have something to add. In the first place, it’s still an open question how strictly the intelligible principle test should apply—recall my suggestion that delegations to the courts should be policed more actively (but how much more?) than delegations to agencies. So the power to interpret terms functions as a backstop to prevent the intelligible principle test from becoming so strict as to rule out clearly permissible interpretation. But in the second place, it’s also an open question what interpretive methods are permissible. So, in case one has an extremely restrictive view of the common-lawmaking power to interpret terms (such that the Inherent-Powers Corollary will rarely apply), one can always fall back on the presence of an intelligible principle in the statute and apply the ordinary non-delegation doctrine.
* * *
Some Ambiguous Statutes
Having gotten these preliminaries out of the way, we’re ready to look at particular statutes. Below, I discuss four statutes. I argue that the residual clause of the Armed Career Criminal Act, which was struck down under the Due Process Clause, could just as well have been struck down under the non-delegation doctrine. RFRA’s standard is adequately specified in the statute; and recall that it could be justified under a more general theory if one takes a broad view of courts’ power to develop substantive canons and clear statement rules that function like defenses. The Alien Tort Statute doesn’t incorporate a comparably clear caselaw, so the statutory interpretation power wouldn’t be able to support much more than the core law-of-nations torts that existed in 1789. But as noted above, a broader ATS passes non-delegation scrutiny based on the federal courts’ common-lawmaking power in international affairs (though there might be objections based on other constitutional provisions). The Sherman Act is the reverse of the ATS: antitrust isn’t an area involving “uniquely federal interests” sufficient to suppose an independent common-lawmaking power; but in light of the pre-1890 caselaw, the standard is adequately specified.
1. The Armed Career Criminal Act
In Johnson v. United States, the Supreme Court considered the Armed Career Criminal Act, which increases criminal penalties for a “violent felony.” “Violent felony” is defined to include certain listed crimes and crimes with certain elements, but the definition ends with a so-called “residual clause” that includes crimes that “involv[e] conduct that presents a serious potential risk of physical injury to another.”
The Court held that the residual clause is unconstitutionally vague and thus violated due process. In past cases, the Court had held that whether a crime is a violent felony under the Act doesn’t depend on whether there was violence in the specific case. Rather, one has to apply the “categorical approach” first recognized in Taylor v. United States—that is, “picture the kind of conduct that the crime involves in ‘the ordinary case,’ and . . . judge whether that abstraction presents a serious potential risk of physical injury.” But imagining “the ordinary case” is speculative and subjective. What materials do we use: “A statistical analysis of the state reporter? A survey? Expert evidence? Google? Gut instinct?” Evaluating how “serious” is the potential risk posed in this hypothetical ordinary case introduces still more subjective elements. As Justice Scalia wrote:
Each of the uncertainties in the residual clause may be tolerable in isolation, but “their sum makes a task for us which at best could be only guesswork.” Invoking so shapeless a provision to condemn someone to prison for 15 years to life does not comport with the Constitution’s guarantee of due process.
The Court was right to decide this case under due process: the void-for-vagueness doctrine should apply against both state and federal governments. But the language quoted above could just as well establish that the statute—interpreted in light of the Taylor “categorical approach”—violates the non-delegation doctrine. Inviting judges to imagine the ordinary case lacks an intelligible principle; nor does the Inherent-Powers Corollary help here, because the notion that federal judges have an inherent common-law power to define crimes was rejected in the early 19th century.
To the extent the categorical approach was invented by courts without regard to Congress’s intent, it was just ultra vires. But in Taylor, Justice Blackmun argued that, though the matter was ambiguous, on balance Congress intended the categorical approach. If we can extend that congressional intent to the specific application of the categorical approach involved in Johnson—the “imagine the ordinary case” rule for applying the residual clause—then we can conclude that Congress was responsible for that rule. And that delegation is so indeterminate, and so distant from courts’ inherent powers, that it violates the non-delegation doctrine.
Is this analysis worthwhile, since we already have a holding on the issue rooted in due process? Non-delegation and due process are related, but obviously not identical. A disadvantage is that the non-delegation theory applies only to statutes passed by Congress, not to state statutes.
But non-delegation does have some coverage that the due process void-for-vagueness theory lacks. Due process applies to all state action, but its void-for-vagueness aspect applies only when criminal statutes or fundamental rights are at issue: Justice Scalia’s Johnson opinion stressed the injustice of “condemn[ing] someone to prison for 15 years to life” on the basis of such vagueness. By contrast, a doctrine rooted in the non-delegation doctrine would apply whenever there was a vague term, whether the statute was civil or criminal, and even if it didn’t implicate a “life, liberty, or property” interest within the meaning of the Due Process Clause. The required degree of specificity would vary in a more general way depending whether the delegation had a significant effect, but it wouldn’t be triggered by the same circumstances as the due process analysis.
2. The Religious Freedom Restoration Act
Recall the text of RFRA: “[The federal g]overnment may substantially burden a person’s exercise of religion only if it demonstrates that application of the burden to the person (1) is in furtherance of a compelling governmental interest; and (2) is the least restrictive means of furthering that compelling governmental interest.” This seems awfully vague.
In Employment Division v. Smith, Justice Scalia wrote that:
[T]o say that a nondiscriminatory religious-practice exemption is permitted, or even that it is desirable, is not to say that it is constitutionally required, and that the appropriate occasions for its creation can be discerned by the courts.
In other words, courts didn’t know how to reliably apply the then-prevailing substantial burden/compelling government interest test.
Scalia did state that legislative exemptions for religious practice could be permissible. And indeed, some legislative exemptions might turn out to be relatively easy for judges to implement: Scalia himself listed statutes from Arizona, Colorado, and New Mexico making “an exception to their drug laws for sacramental peyote use.”
But Congress not only took Scalia’s hint but went so far as to adopt a blanket exemption regime covering all past and future statutes—a regime that substantially reenacted the pre-Smith religious exemption framework, the very test that Scalia had argued was too hard for courts to reliably apply. Congress was quite clear about its intentions: one of the expressed “Purposes” of the statute was “restor[ing] the compelling interest test as set forth in Sherbert v. Verner and Wisconsin v. Yoder and . . . guarantee[ing] its application in all cases where free exercise of religion is substantially burdened.”
One way of reading Scalia’s note in Smith—that judges can’t discern the appropriate occasions for the creation of a non-discriminatory religious practice exemption (even using the then-current test)—is that the principle behind it is insufficiently intelligible. If so, reenacting that principle in a statute does nothing to make it more intelligible. When judges were making the doctrine up (so the argument might go), they were engaged in illegitimate judicial lawmaking; the passage of RFRA removed that source of illegitimacy but created a new one: Congress gave away some of its legislative power by specifying an excessively vague test.
That’s one way of reading Smith, but not the only way. Perhaps the pre-Smith test was difficult, but not so difficult as to be unconstitutional. Had it really been constitutionally required, the Court would have gritted its teeth and tried to make some sense of it; the real reason to abandon the pre-Smith test was that it simply wasn’t constitutionally required. If that’s so, Congress could still reenact the test as a statutory matter.
And indeed, RFRA’s “substantial burden”/“compelling governmental interest” test is actually pretty intelligible. Whether a means is least restrictive doesn’t seem all that problematic, and “substantial burden” has been pretty well hashed out in the 25 years of pre-Smith caselaw. “Compelling governmental interest” is a lot more in the eye of the beholder, but we also have long experience with that concept, both from the pre-Smith standard and from elsewhere in constitutional law, like Due Process, Equal Protection, and Free Speech Clause jurisprudence. (Alternatively, as discussed above, we could say not that the statute has an intelligible principle, but that it doesn’t need an intelligible principle because of the federal common law power to use legitimate methods of statutory interpretation. Recall, also, the discussion above, giving an independent possible justification of RFRA in terms of courts’ power to craft defenses, clear-statement rules, and substantive canons.)
3. The Alien Tort Statute
The ATS gives district courts “original jurisdiction of any civil action by an alien for a tort only, committed in violation of the law of nations or a treaty of the United States.” The question is what conduct violates the “law of nations.”
I’ve already argued above that international relations is an area involving “uniquely federal interests” where federal courts have a common-lawmaking power. And international-law torts are either within that area or interlinked with it. So even a broad reading of the ATS survive non-delegation scrutiny under the Inherent-Powers Corollary (though there might still be arguments against the broad ATS based on other provisions, like the Offenses Clause). In this section, though, the question is whether—even if you disagree that the ATS falls within the foreign-affairs enclave—the statute is valid purely because of the judicial statutory-interpretation power.
Kontorovich writes that “the breadth of the delegation in the ATS is troubling. The statute leaves it to the courts, without any statutory guidance, to identify and adopt causes of action for torts in ‘violation of the law of nations.”
In the ATS, the subject matter of the delegation is as broad as the Offenses Clause itself. . . . The only limits (aliens, torts) are jurisdictional, not substantive; there is no policy determination at all in the statute. Congress has not specified any particular offenses or even kinds of offenses, let alone their elements, that can serve as a basis for liability under the statute. Rather, Congress has left all of the defining to the judiciary.
In the early-19th-century case United States v. Smith, the Supreme Court had said that a statute providing the death penalty for the crime of “piracy, as defined by the law of nations” was an appropriate exercise of Congress’s Offenses Clause power because piracy had a precise definition in international law. By comparison, the Alien Tort Statute doesn’t refer to a comparably precise concept. Thus, Kontorovich writes, Justice Souter’s limiting of the statute, in Sosa v. Alvarez-Machain, to cover only the historical offenses of “violation of safe conducts, infringement of the rights of ambassadors, and piracy” was not just a good prudential decision but also a form of constitutional delegation avoidance.
Certainly, a hyper-minimalist view—that the ATS freezes law-of-nations torts in place where they were when the ATS was passed in 1789 (i.e, offenses against ambassadors, violation of safe conducts, and piracy)—satisfies the non-delegation doctrine. But the ATS doesn’t need to be constitutionally limited to those three torts: incorporating a living common-law tradition, given sufficiently determinate caselaw, can also satisfy the non-delegation doctrine (as discussed above, either because there is an intelligible principle, or because, thanks to the Inherent-Powers Corollary, there doesn’t need to be an intelligible principle).
Apparently the law wasn’t well-enough developed in 1789 to support torts other than the classic three—so the inherent interpretive power can’t support a broad ATS, if we’re limited to static interpretive methods. But a lot depends on one’s view of federal courts’ inherent statutory interpretation powers. If these include the power to use dynamic statutory interpretation, then courts can rely on developments since 1789 that have made the doctrine more developed. This analysis would be consistent with Justice Souter’s opinion, which allows courts to recognize further law-of-nations torts. “[T]he door is still ajar subject to vigilant doorkeeping, and thus open to a narrow class of international norms today,” Justice Souter wrote; he thus looked at present-day international-law norms, but determined that the specific conduct complained of didn’t violate them.
And, of course, even if dynamic statutory interpretation is impermissible, a broad reading of the ATS may well be ultra vires, but provided that the use of this method stems from a judge’s independent preference for the method (rather than an interpretation of the statute as requiring the method), it doesn’t violate the non-delegation doctrine.
4. The Sherman Act
Section 1 of the Sherman Act declares “[e]very contract, combination . . . or conspiracy, in restraint of trade or commerce” to be illegal. Since the late 1970s, the Supreme Court has been designing antitrust doctrine to comport with its views of economic efficiency, and this sort of freewheeling policymaking has been frequently attacked by opponents of unmoored judicial discretion, even those who agree with the result on policy grounds.
Lemos, for instance, writes that “the Sherman Act stands on shaky constitutional ground. It delegates virtually boundless discretion to the federal courts to craft substantive antitrust rules, and that broad delegation is difficult to defend in practical terms.” “[I]f the nondelegation doctrine were applied to courts,” she writes, “the Sherman Act would be a likely candidate for constitutional invalidation.”
Three aspects of antitrust law are important to the non-delegation inquiry. The first is that the statute itself bans all contracts in restraint of trade, which on its face is reasonably clear and doesn’t raise non-delegation concerns. Unfortunately, this hyper-literal reading would also ban a huge range of contracts. The Supreme Court wrote in 1911 that the banned classes of acts were “broad enough to embrace every conceivable contract or combination which could be made concerning trade or commerce or the subjects of such commerce.” It wrote similarly in 1918: “Every agreement concerning trade, every regulation of trade, restrains. To bind, to restrain, is of their very essence.” And Ronald Coase’s revolutionary insight in 1937 was that the suppression of competition in favor of hierarchy was the very nature of the firm. Nobody thinks the Sherman Act generally bans firms and contracts.
The second is that the Supreme Court has long taken the view that its common-law-style policy development was part of the intent of the enacting Congress. In 1978, near the beginning of the modern era of economics-inspired judicial antitrust policymaking, the Court wrote:
Congress . . . did not intend the text of the Sherman Act to delineate the full meaning of the statute or its application in concrete situations. The legislative history makes it perfectly clear that it expected the courts to give shape to the statute’s broad mandate by drawing on common-law tradition.
Recall that judicial elaboration of a vague statute using an impermissible method can’t be considered a violation of the non-delegation doctrine unless Congress has mandated that method; otherwise, it’s just judges acting beyond their rightful powers. Right or wrong, (the Supreme Court’s view of) the legislative history brings us into non-delegation land (at least, by the Supreme Court’s own standards).
The third aspect is that there was already, at the time of the enactment of the Sherman Act, an existing body of caselaw elaborating what was an unreasonable “restraint of trade.” This is part of why no one interprets the Sherman Act literally (that, and the absurdity of the position). We should interpret the statute in light of that existing law. This caselaw is well-enough developed that the Sherman Act has an intelligible principle (or, under the Inherent-Powers Corollary, doesn’t require an intelligible principle because of the courts’ power to interpret terms).
The pre-1890 common law is well reviewed in the Areeda-Hovenkamp treatise, so I won’t rehearse it here. And the early Sherman Act cases discussed this caselaw—for instance, then-circuit judge Taft’s opinion in United States v. Addyston Pipe & Steel Co., and the Supreme Court’s opinions in Dr. Miles Medical Co. v. John D. Park & Sons Co. and Standard Oil Co. of New Jersey v. United States.
In 1988, Justice Scalia discussed the common law again in Business Electronics Corp. v. Sharp Electronics Corp. His discussion is worth quoting at length:
In resting our decision upon the foregoing economic analysis, we do not ignore common-law precedent concerning what constituted “restraint of trade” at the time the Sherman Act was adopted. But neither do we give that pre-1890 precedent the dispositive effect some would. The term “restraint of trade” in the statute, like the term at common law, refers not to a particular list of agreements, but to a particular economic consequence, which may be produced by quite different sorts of agreements in varying times and circumstances. The changing content of the term “restraint of trade” was well recognized at the time the Sherman Act was enacted.The Sherman Act adopted the term “restraint of trade” along with its dynamic potential. It invokes the common law itself, and not merely the static content that the common law had assigned to the term in 1890. . . .Of course the common law, both in general and as embodied in the Sherman Act, does not lightly assume that the economic realities underlying earlier decisions have changed, or that earlier judicial perceptions of those realities were in error. It is relevant, therefore, whether the common law of restraint of trade ever prohibited as illegal per se an agreement of the sort made here, and whether our decisions under § 1 of the Sherman Act have ever expressed or necessarily implied such a prohibition.
And he proceeded to discuss the pre-1890 common law as reflected in Dr. Miles. It’s clear, then, that the Supreme Court has long understood itself, and continues to understand itself, as participating in the ongoing development of a preexisting common-law action on Congress’s command.
Oldham attacks the statute in at least two ways. On the one hand, he uses the literal language of the statute as a non-delegation argument: Perhaps banning all contracts and firms has an intelligible principle, but that’s not what Congress intended. And explicitly banning all contracts and firms while implicitly wanting judges to exempt whatever conduct they like lacks self-evidently grants too much power to courts and lacks an intelligible principle. But this concern of Oldham’s rests entirely on taking seriously the statute’s general ban on contracts and firms. And this literal reading is clearly incorrect, since textual interpretation properly reads terms of art (like “restraint of trade”) in their legal context.
On the other hand, Oldham writes, the statute would violate the non-delegation doctrine “even if the Sherman Act means what modern interpreters assert.” He grants that a standard like promoting consumer welfare would qualify as an intelligible principle under the delegation standards that apply to the executive branch, but argues (as I do) that judicial delegations should be policed more tightly. Perhaps that’s true, if Congress had truly written a statute providing that “courts shall ban any economic transactions or business practices that do not promote consumer welfare.” And perhaps some judges really do think about antitrust that way.
But the blockquote above from Justice Scalia shows that there’s another possibility—one that the Supreme Court itself has repeatedly endorsed: Congress meant to refer to the prior restraint-of-trade caselaw; that caselaw is determinate enough to guide us; and we continue to develop that caselaw. But because the statute bans a result, not particular practices, antitrust rules can legitimately change with new understanding.
This is crucial for determining what’s violated if judges have it wrong. Suppose the Court is wrong that the Sherman Act is about economic efficiency and/or consumer welfare; perhaps we should give more credence to other principles present in both the legislative history and the common law, like “fairness and economic independence,” protecting competitors (even less efficient ones) from being forced from the business, and preventing undue concentrations of political influence. Perhaps modern-day doctrine can’t really be justified as a series of steps since 1890 developing the original common-law standard.
But if so, that’s the courts’ fault, not Congress’s. It’s a problem of judges acting ultra vires, not of Congress delegating away its legislative power. Congress gave the courts a coherent common-law standard; at that point, the non-delegation doctrine was satisfied. Any later misapplication of that standard by others implicates different norms.
I argued above that antitrust isn’t one of the areas involving “uniquely federal interests” like admiralty or international relations, where courts have a federal common-lawmaking power independent of statute, so the Inherent-Powers Corollary doesn’t automatically validate the statute on that ground. But fortunately for the Sherman Act, it has an intelligible principle (or, equivalently, doesn’t need an intelligible principle because its meaning is adequately determined by conventional methods of statutory interpretation). In this sense, the Sherman Act is the mirror image of the Alien Tort Statute, which lacks an ascertainable meaning beyond covering a few core historical torts, but falls within the federal courts’ international relations common-lawmaking power.