« Must the Law Be as All-or-Nothing as Leo Katz Claims? | Main | A Digital Fix to Alleviate Some Angsting Stress? »

Friday, April 20, 2018

Gundy and Non-Delegation: Which of Several Non-Delegation Doctrines Should Apply to SORNA?

Marty Lederman recently asked me in a comment to an earlier post on Gorsuch’s concurrence in Dimaya how I thought Gundy v. United States should and would be decided. For those who are not hopeless Non-Delegation Doctrine dorks, SCOTUS recently granted cert in Gundy to hear argument on the question of whether SORNA (the federal Sex Offender Registry Law) violates the so-called “non-delegation doctrine” (NDD) by delegating to the Attorney General the decision about whether SORNA should apply retroactively to persons convicted of covered sex offenses prior to SORNA’s effective date.

In other words, Gundy is one of those exciting cases in which SCOTUS might actually revive a moribund constitutional doctrine that has been flitting ghost-like through the US reports since 1935, making a noise but not a difference in the outcomes of cases. Although the prospect of resurrecting the dead naturally draws a crowd, I think that revival of the NDD is unlikely. Despite Justice Gorsuch’s enthusiasm for it, the NDD writ large, as a general legislative duty to cabin executive power with an “intelligible principle,” is a non-starter. The problem is that the NDD contains no intelligible principle for determining when a statutory principle is intelligible. SCOTUS would have to be immune to irony to confer on itself unbridled judicial discretion to decide whether an agency has unbridled executive discretion. In its lack of intelligibility, one might say that the NDD is a doctrine that violates itself.

So revival of the NDD writ large seems unlikely to me. But what about the NDD writ small? After the jump, I will offer a few thoughts about whether SCOTUS might revive a mini-NDD based on the various non-delegation canons described by as Cass Sunstein ‘way back when. These canons include the ideas that (1) agencies do not get deference for their interpreting statutes to be retroactive and (2) prosecutors do not get any deference for their interpretations of criminal law (as Scalia asserted but did not really explain in his Crandon concurrence). SORNA delegates to the Attorney General the power to impose an arguably retroactive effect with an arguably criminal statute. Perhaps the SCOTUS can and should create a mini-NDD to deal with this specific sort of delegation.

Gundy had been convicted in Maryland in October 2005 of a state sex offense. This conviction violated the conditions of his federal supervised release, so, after serving time in Maryland prison, he was shipped off to Pennsylvania to serve his federal sentence in a federal facility. From Pennsylvania, Gundy was shipped off to a community release facility in the Bronx to serve out the last part of his sentence, presumably because the feds judged that he was not a big risk to the community. While still officially in federal custody (although he was unescorted on a bus on furlough), Gundy crossed state lines to travel from Pennsylvania to New York. He did not register as a sex offender under Maryland law, NY law, or SORNA after arriving in New York. The feds argue that he had a duty to register after arrival, because he had crossed state lines. In the lower courts, Gundy argued that his interstate travel did not qualify under SORNA because he was in custody when he crossed state lines and the “travel” required by 18 U.S.C. Section 2250(a)(2)(B) must be voluntary travel. The Second Circuit declined to reach that statutory question, because there was sufficient evidence to indicate that Gundy’s travel to the Bronx was “voluntary.”

Gundy’s underlying Maryland sex conviction occurred in October of 2005, a few months before SORMA became effective. Gundy is covered by SORNA, therefore, only if the statutory delegation to the AG to decide the retroactive scope of the statute is consistent with the NDD. That delegation, codified at 18 U.S.C. Section 16913(d), provides that the “Attorney General shall have the authority to specify the applicability of [SORNA] to sex offenders convicted before [SORNA’s enactment].” The AG carried out this delegation in 2008, after providing an opportunity for notice and comment, by publishing “National Guidelines” applying SORNA to “sex offenders who remain in the system as prisoners, supervisees, or registrants, or who reenter the system through a subsequent criminal conviction” even if their underlying sex offense convictions predated SORNA’s effective date. (An earlier 2007 interim rule had run into some trouble because of non-compliance with the APA’s rule-making procedures). In responding to critics who argued that applying SORNA so broadly constituted an unconstitutional Ex Post Facto law, the AG argued that the Ex Post Facto clause did not apply to the duty to register, because, under Smith v. Doe, that duty was not “penal” but merely “civil” in character, designed to prevent future harm rather than punish past crimes. (See pages 38035-36 of the “National Guidelines”). The AG also noted that, as a matter of policy, “fairness does not require that an offender, at the time he acknowledges his commission of the crime and pleads guilty, be able to anticipate all future regulatory measures that may be adopted in relation to persons like him for public safety purposes.”

Lower courts have generally upheld SORNA against these sorts of NDD challenges by citing the general idea that Congress can delegate broad powers to executive officials under Schecter Poultry. But questions of statutory retroactivity might be different from the run-of-the-mill delegation because retroactive application of statutes qualifies as a “major question” of policy. One might reasonably expect Congress to decide this sort of question for itself or, at the very least, specify some criteria for the AG to apply when deciding the question in a rule-making process. SCOTUS is fond of declaring without much explanation that executive officials should not get Chevron deference on Benzene-MCI-Brown & Williamson canon of statutory construction “major questions.” The same reasons that preclude judicial deference to agencies on “major questions” also suggest constitutional limits on delegations to decide them when the magnitude of the question’s importance increases. One might argue, therefore, that, when a question becomes really, really “major,” then, just for those “major, major questions,” the non-delegation canon ought to be elevated to a full-fledged constitutional doctrine rather than a mere interpretative canon, a sort of “mini-NDD” requiring extra-specific criteria in the statute itself rather than a judicial opinion controlling the agency.

I am not sure what constitutes a “major, major question,” but I am pretty sure that questions of statutory retroactivity ought to qualify if anything does. Quite apart from any constitutional problems suggested by the Ex Post Facto clause, the retroactive application of legal duties strikes at the heart of Lon Fuller-style morality of law. One might, of course, argue that Herman Gundy was not really subject to any retroactive penalty, because he traveled to New York after the date of the AG’s Guideline. It seems frivolous to me, however, to argue that a penalty ceases to be retroactive just because it is imposed on the basis of conduct that both pre- and post-dates the statute imposing the burden. Admittedly, some “continuing offenses” can safely “straddle” the enactment date of a statute. If I begin a conspiracy to do X before it is outlawed and continue that very same conspiracy after it is outlawed, it is hard to say that punishing me for the entire conspiracy inflicts some sort of unfair surprise on me. Gundy, however, may have pled guilty to a crime in Maryland without any knowledge that thereafter his plea would later be used as a basis for burdening his travel between states. The AG brushed aside this worry with the glib statement that, at the time of conviction, one need not “be able to anticipate all future regulatory measures” that could be later imposed against you. Maybe not — but the question of whether it is fair play to pile extra “public safety measures” on to a past conviction seems to me to be the sort of “major” issue that deserves a bit more explanation than such a one-line brush-off.

Of course, one might simply cite Smith v Doe for the proposition that merely protective, non-penal, “civil” burdens are not limited by the same constitutional constraints barring Ex Post Facto laws. My response to this sort of argument is colored by my belief that Smith was wrongly decided. The notion that SORNA imposed its extraordinarily punitive public shaming on “sex offenders” for purely protective rather than vindictive purposes seems to me not just untenable but risible. Moreover, the category of “sex offender” is ridiculously broad, sweeping up in its moral panic all sorts of people who pose no special public safety threat or risk of recidivism.

But Smith is the law, so let’s put aside the Ex Post Facto clause. Quite apart from EPF, these sorts of retroactive burdens give rise to weighty Due Process and non-constitutional policy implications that deserve serious — meaning congressional — consideration. If Congress does not want to do the heavy lifting, then it could delegate the decision to the AG by requiring some sort of findings before SORNA is extended to cover anyone convicted of a sex offense prior to SORNA’s enactment. But simply handing the question off to the AG without any statutory criteria whatsoever — not even a duty to make findings about “fairness” or “reasonability” — seems a little frivolous.

In sum, if the Court wanted to resurrect a mini-NDD to prohibit Congress from making completely standardless delegations of retroactivity issues to executive officials, then Gundy might be a good vehicle with which to do so. I am not betting that they will, but the likelihood of the Court’s imposing such a mini-NDD on Congress seems to me much higher than the odds of any wholesale resurrection of the general non-delegation doctrine.

Posted by Rick Hills on April 20, 2018 at 01:29 AM | Permalink


Does it really matter? Fairness and intelligibility didn't stop us from locking up the Japanese during WWII. Given the hatred towards sex offenders in the current climate, a climate that #metoo has only made worse, does anyone actually think that adding a few pleasing words to the law would have made any difference in the eventual outcome? I don't. Congress tipped its hand on what type of result it wanted when it delegated that authority to the Attorney General, not exactly a neutral party when it comes to matters of criminal justice. So a "mini-NDD" doesn't actually get one anywhere other than putting some constitutional theory lipstick on the public panicked pig.

Posted by: James | Apr 20, 2018 4:55:13 PM

"But simply handing the question off to the AG without any statutory criteria whatsoever — not even a duty to make findings about 'fairness' or 'reasonability' — seems a little frivolous."

Isn't this a good way of saying that SORNA lacks an intelligible principle, and indeed an apt demonstration of the intelligibility of the intelligible-principle rule? What's so hard about deciding if Congress has handed off a question to an agency without any statutory criteria whatsoever? I guess it's been said that the intelligible principle is contained in SORNA's purpose section, but does that limit or really go to the AG's discretion at all? If the AG opted for minimal retroactive coverage in spite of the purpose section and without reference to it, no one would say that that violates the statute. It's always seemed to me that the statute permits absolutely any choice -- which I suppose might be said of statutes that authorize regulation in the public interest -- but that further, it contains no principle that constrains the reasons the AG has to give in support of whatever decision it makes. Whereas with a public-interest delegation, perhaps any degree of regulation can be justified in public-interest terms, but they still must be justified in public-interest terms.

Posted by: Asher Steinberg | Apr 20, 2018 10:16:10 AM

The comments to this entry are closed.