Who Made a Vague Law Vague?

On Friday, the Court unexpectedly ordered reargument in Johnson v. United States, which had been argued back in November. Whereas the case originally asked a question of statutory interpretation, the reargument will address whether the “residual clause” added to the Armed Career Criminal Act  (or “ACCA”) in 1986 is unconstitutionally vague. This re-argument order is quite extraordinary, particularly because no party had raised the vagueness issue before the Court (though it had been raised and adjudicated below). Last year, the Court seemed unusually skeptical of federal prosecutors in cases like Bond and Yates, and that trend seems to be carrying over into the new year.

The reargument order itself is reminiscent of Citizens United, which likewise involved new argument on a constitutional challenge that called for overturning precedent and striking down a federal statute. “Essentially,” Justice Stevens wrote in his Citizens United dissent, “five Justices were unhappy with the limited nature of the case before us, so they changed the case to give themselves an opportunity to change the law.” Whether for good or ill, that description seems to match what has happened in Johnson.

But if the residual clause is so vague as to be unconstitutional, why did it take the Supreme Court roughly 30 years to figure that out? It’s not as though the statute hadn’t come to the Court before—quite the contrary. The residual clause has been a regular source of business for the Justices, and even the specific issue of vagueness has previously been aired—and not just once before, but repeatedly. Yet only now does it appear that a majority of the Court is prepared to call vagueness. What gives?

A little bit of background: ACCA’s residual clause calls for a harsh 15-year mandatory minimum sentence for federal firearm offenders who have previously been convicted of three “violent felonies.” The statute defines “violent felony” in part as a crime punishable by more than a year that “is burglary, arson, or extortion, involves use of explosives, or otherwise involves conduct that presents a serious potential risk of physical injury to another.” But, what unites the listed offenses? And how are courts to discern “conduct that presents a serious potential risk of physical injury to another”? These are the tricky questions that the Court has struggled to answer. For example, the Court has held that vehicular flight from a law-enforcement officer is a violent felony, even though driving under the influence and failure to report to penal confinement aren’t. In Johnson, the original question was whether possession of short-barelled shotguns qualified as a "violent offense."

The question for reargument in Johnson is whether the residual clause is so vague that it wouldn’t “give a person of ordinarily intelligence fair notice” of what is subject to the enhanced mandatory minimum. The due-process concern here is most naturally understood as an objection to arbitrary punishment, and recent residual-clause jurisprudence has indeed become quite arbitrary. In a roughly four-year period starting in 2007, the Court heard a new residual-clause case about every year. And each of the Court’s four decisions applied a different analysis. In Sykes, the Court’s last residual-clause case, Justice Scalia's dissent drilled this point. The majority, Scalia explained, “adds to the ‘closest analog’ test (James), the ‘purposeful, violent, and aggressive’ test (Begay), and even the risky-as-the-least-risky test that I had proposed [in James] as the exclusive criterion, [but] has not made the statute’s application clear and predictable.” Note that Justice Scalia candidly acknowledged that he himself had previously proposed a test to avoid the vagueness problem—but, citing the lessons of experience, he has now given up the game.

When Scalia first started expressing vagueness concerns in James back in 2007, the Court responded by flatly rejecting that worry. Indeed, the Court then felt that the residual clause wasn’t any more vague than the many criminal laws that (like all uses of language) are fuzzy at the boundaries. Here’s the Court’s explanation:

The statutory requirement that an unenumerated crime “otherwise involv[e] conduct that presents a serious potential risk of physical injury to another" is not so indefinite as to prevent an ordinary person from understanding what conduct it prohibits. See Kolender v. Lawson, 461 U.S. 352, 357, 103 S.Ct. 1855, 75 L.Ed.2d 903 (1983). Similar formulations have been used in other federal and state criminal statutes. See, e.g., 18 U.S.C. § 2332b(a)(1)(B) (defining “terrorist act” as conduct that, among other things, “creates a substantial risk of serious bodily injury to any other person”); Ariz.Rev.Stat. Ann. § 13-2508(A)(2) (West 2001) (offense of resisting arrest requires preventing an officer from effectuating an arrest by “any ... means creating a substantial risk of causing physical injury to the peace officer or another”); Cal. Health & Safety Code Ann. § 42400.3(b) (West 2006) (criminalizing air pollution that “results in any unreasonable risk of great bodily injury to, or death of, any person”); N.Y. Penal Law Ann. § 490.47 (West Supp. 2007) (“[c]riminal use of a chemical weapon or biological weapon" requires "a grave risk of death or serious physical injury to another person not a participant in the crime”).

Put another way, if Congress simply legislated the residual clause—penalizing felonious “conduct that presents a serious potential risk of physical injury to another”—it’s hard to say that the result would be unconstitutional. And, indeed, legislatures have enacted many laws just like that, without causing any jurisprudential meltdowns. So why has the residual clause fared so much worse?

Part of the answer is that the residual clause is linked to a maddeningly esoteric list of exemplary offenses: “burglary, arson, or extortion" and crimes "involv[ing] use of explosives.” Law clerks have been driven to distraction staring at that catalogue, looking for hidden patterns and clues. For a while, the Court, too, tried to extract lessons from this cipher (even though it's set off from the residual clause by the word "otherwise"). And that effort seemed especially necessary in order to honor the term being defined—“violent felony”—as well as to mitigate ACCA’s harshness. For many jurists, it was particularly hard to believe that a severe law expressly concerned with violent gun crime should come into play because the defendant drove under the influence—even though driving under the influence is quite obviously “conduct that presents a serious potential risk of physical injury to another” (and so fits the actual text of the residual clause). Alas, the Court failed to translate that intuition into a workable test and then stick to it.

To a great extent, however, the residual clause’s downfall stems from the way that the Court supervises the circuit courts. In adopting the “categorical” and then “modified categorical” approach to figuring out what qualifies as a violent felony, the Court effectively transformed the residual clause into a machine for making circuit splits. Under these approaches, whether an offense qualifies as a predicate offense depends on its legal elements, as assessed in the abstract. Thus, the categorization of each and every predicate offense is a pure question of law—and there are gobs of candidate offenses, each with tricky regional variations. Meanwhile, prosecutors keep pushing the envelope, to the point that crimes of absence (like failure to report) or possession (as in Johnson itself) have triggered mandatory minimums. ACCA’s circuit-split engine is always running.

By contrast, the Court often resolves the case before it with a squishy test and then kicks the all-important job of “factbound” application back down to juries, trial courts, and courts of appeals. While divergent outcomes may result, there are usually enough factual nuances to preclude clean circuit splits. Sure, the Ninth Circuit may trend one way while the Eleventh trends another, but they can usually do so without adopting contrary legal rules. In this way, the Court frequently avoids having to revisit knotty legal issues for years, even decades. The Court denied itself that luxury in residual-clause cases. Instead, it created a situation where it had to examine and reexamine a high-stakes, complex legal issue that lent itself to fact-dependent sympathies. Even without changes in the Court’s composition (which, of course, have also happened), this is a formula for mushy jurisprudence.

And so here we are. The smart money is that the Court will take Justice Scalia’s suggestion and simply strike the residual clause, thereby leaving only the listed offenses. This would have the salutary effect of putting the proverbial ball back in Congress's court, allowing for a prospective legislative "fix." Still, the Court could conclude that the residual clause is vague as currently understood and then try to save it. In fact, something similar has happened before. The last time that Justice Scalia called for voiding a federal statute for vagueness, he had honest services fraud in his sights. But instead of downing his target, he only winged it. Afraid of obliterating a measure that had served to convict countless malefactors, the Court in Skilling narrowed the honest-services law down to a constitutionally permissible form, thereby allowing the government to preserve many of its past convictions. A similar law-and-order dynamic might arise in connection with the residual clause. Indeed, when it comes to ACCA we should by now expect the unexpected.

When the Court next hears argument in Johnson, there will be a lot of talk about artless congressional drafting. But if the residual clause is unconstitutionally vague, it was the Court—not Congress—that made it so.

Cross-posted from Re's Judicata.