My co-blogger Andy Grewal expresses skepticism that the Obama administration can cure the notice-and-comment problem in United States v. Texas by initiating notice-and-comment proceedings now and then promulgating its Deferred Action for Parents of Americans policy as a new final rule. In aprior post, I said it seems “pretty clear” that such an approach would cure the procedural deficiency with the DAPA guidance (assuming, arguendo, that the Administrative Procedure Act actually requires notice and comment in the DAPA case). Andy has persuaded me to modify that claim: it seems to me that the procedural question would go away once the Department of Homeland Security completed the 30-day notice-and-comment process, responded to comments, and then promulgated DAPA as a final rule (even while maintaining its position that the policy doesn’t fall within the scope of the notice-and-comment requirement in the first place). Other reasonable and knowledgeable observers disagree. So I’ll use this opportunity to explain why, doctrinally and normatively, I think mine is the better view.
For starters, it helps to distinguish among three different scenarios:
— Scenario 1: An agency issues rule R1 without notice and comment at time t. At time t + 1, the agency seeks to enforce the rule against a party. The party challenges the agency’s action on notice-and-comment grounds. The agency then completes notice-and-comment proceedings and, at time t + 2, promulgates rule R2. R2 is essentially identical to R1.
— Scenario 2: The agency issues rule R1 without notice and comment at time t. At time t + 1, the agency invites postpromulgation comments regarding the rule. The agency never promulgates a second rule. At time t + 2, the agency seeks to enforce the rule against a party. The party challenges the agency’s action on notice-and-comment grounds.
— Scenario 3: The agency issues rule R1 without notice and comment at time t. The agency then completes notice-and-comment proceedings and, at time t + 1, promulgates rule R2. R2 is essentially identical to R1. At time t + 2, the agency seeks to enforce the rule against a party. The party challenges the agency’s action on notice-and-comment grounds.
I agree that the notice-and-comment issue remains live in Scenario 1. At the time that the agency sought to enforce rule R1 against the party, rule R1 was invalid. I also agree that the notice-and-comment issue remains live in Scenario 2. Section 553(c) says that agencies should promulgate rules “[a]fter consideration of the relevant matter presented” in notice-and-comment proceedings. Promulgating a rule before the notice-and-comment proceeding does not satisfy the statute. [Let’s bracket the question of whether the agency would have a viable harmless-error argument in either Scenario 1 or Scenario 2.]
Scenario 3 is different. The agency satisfied the notice-and-comment requirement of section 553 before promulgating R2. The only possible basis for saying that rule R2 is procedurally invalid is that R2 is essentially identical to R1 and that R1 is procedurally invalid. But that’s no basis at all. The APA says that a reviewing court shall set aside a rule promulgated “without observance of procedure required by law.” The APA does not say that a reviewing court shall set aside a rule that is essentially identical to an earlier rule promulgated without observance of procedure required by law. And as the Supreme Court reiterated last year in Perez v. Mortgage Bankers Association, courts “lack authority” to impose procedural burdens on agencies “[b]eyond the APA’s minimum requirements.”
Any other approach to Scenario 3 would raise a raft of obvious problems. How close in content would R1 and R2 have to be in order for R1’s procedural invalidity to taint R2? And how close in time? If an agency issues R1 without notice and comment in 1947, then initiates notice-and-comment proceedings, and then promulgates R2 in 2016, should a court set aside R2 on the ground that it’s essentially identical to R1? Or does the procedural invalidity of R1 just mean that the agency can’t issue an essentially identical R2 for a limited time? And if so, how long? Until a new administration comes in?
Andy cites the D.C. Circuit’s decision in New Jersey Department of EnvironmentalProtection v. Environmental Protection Agency, 626 F.2d 1038 (D.C. Cir. 1980). That case seems to me like either a Scenario 1 case or a Scenario 2 case. In March 1978, EPA promulgated a “final rule” without notice and comment in which it required states to formulate implementation plans for national ambient air quality standards by January 1, 1979. EPA also opened up a postpromulgation comment period for 60 days. New Jersey challenged EPA’s action on notice-and-comment grounds. Then on January 26, 1979, EPA published a one-page statement in the Federal Register in which it “reaffirm[ed]” its March 1978 action. The D.C. Circuit said that the “post hoc” comment period and the January 1979 “reaffirmation” did not cure the notice-and-comment problem.
All that makes sense. Even if the January 1979 action amounted to a separate R2, it came after EPA sought to apply R1 to New Jersey. And if we consider the January 1979 action to be a reaffirmation of R1 rather than a separate R2, then the procedural invalidity of R1 is even starker: the only opportunity for comment came after R1 was issued.
Contrast the New Jersey case with Natural Resources Defense Council, Inc. v. Nuclear RegulatoryCommission, 680 F.2d 810 (D.C. Cir. 1982), in which the Nuclear Regulatory Commission sought to amend the license for a naval nuclear fuel facility in Tennessee. At the time, commission regulations provided for a full adjudicatory hearing on license amendments. Rather than holding such a hearing, the commission promulgated a rule (I’ll call it R1) without notice and comment that authorized alternative hearing procedures for matters involving “military and foreign affairs functions.” The NRDC challenged R1 on the ground that the commission had failed to go through notice and comment. While NRDC’s challenge was pending, the commission initiated notice-and-comment proceedings and then promulgated a final rule (R2). The D.C. Circuit found that R2 was “essentially the same” as R1.
Yet despite the fact that R2 was essentially the same as R1, the D.C. Circuit saw no procedural problem with R2. A unanimous panel held that the notice-and-comment issue was now moot. As Judge Harry Edwards wrote:
NRDC has maintained that the promulgation by NRC of the “military functions” rule was unlawful because it deprived NRDC and other members of the public of the right to participate in the rulemaking process by not providing notice and an opportunity to comment on the rule. The Commission undeniably provided that opportunity, however, when it repromulgated the rule in accordance with the requirements of section 553 of the APA. Petitioner does not contend that there was any infirmity in the Commission’s repromulgation of the rule; rather, it continues to focus its attack on the initial adoption of the rule without notice and comment. Even if this attack was originally well-founded, we can hardly order the NRC at this point to do something that it has already done.
So what if the Department of Homeland Security publishes a proposed rule in the Federal Register that’s essentially identical to the previously announced DAPA policy, allows 30 days for comment, and then promulgates a final rule (R2)? I think we’d be in the territory of Scenario 3. Texas still might argue that the original policy (R1) was procedurally invalid, but that argument wouldn’t get Texas very far. The district court in this case stayed the policy even before it took effect, so Texas has suffered no harm as a result of R1 until now. The harms of which Texas complains are all future harms. So if the Department of Homeland Security promulgates DAPA as R2 before the injunction against R1 is lifted, all harms suffered by Texas will be harms suffered as a result of R2—and R2 is procedurally unproblematic.
Importantly, I am not arguing that agency action always should be set aside in Scenario 1 or Scenario 2 if a court determines that R1 is a legislative rule requiring notice and comment. As my co-blogger Nicholas Bagley notes, section 706 of the APA instructs that courts reviewing agency action should take “due account” of “the rule of prejudicial error.” And as the Supreme Court held in Shinseki v. Sanders, that language places the burden on the party challenging agency action to show that the agency’s error affected the party’s “substantial rights.” It may well be that some of the cases in Scenario 1 and Scenario 2 are cases in which the prejudicial error standard applies. Moreover, section 553(b) carves out an exception to the notice-and-comment requirement “when the agency for good cause finds . . . that notice and public procedure thereon are impracticable, unnecessary, or contrary to the public interest.” Some of the cases in Scenario 1 and Scenario 2 may fall within the “good cause” carevout as well.
And, to be clear, I also am not arguing that United States v. Texas fits within Scenario 3. I still fight the premise that DAPA is a legislative rule requiring notice and comment. My point is simply that even ifDAPA is a legislative rule requiring notice and comment, the Department of Homeland Security can initiate notice-and-comment proceedings tomorrow and then promulgate a new final rule a month later that embodies DAPA and cures any notice-and-comment problem. (Maybe a bit more than a month—the Department of Homeland Security needs time to read the comments and respond. But if a presidential administration can get the First New Deal through in 100 days, it should be able to complete a top-priority rulemaking in far less time.)
While I’m at it, I would remiss if I failed to mention an excellent new article in the Cornell Law Review out this month that is very much relevant to the topic of this post: “Open Minds and HarmlessError: Judicial Review of Postpromulgation Notice and Comment,” by Kristin Hickman and Mark Thomson. Hickman and Thomson look at the various ways courts have handled questions involving postpromulgation notice-and-comment, and they end by proposing “a set of factors for courts to employ in evaluating postpromulgation notice and comment case by case.” I agree with much of what Hickman and Thomson say. In my view, though, the Scenario 1/Scenario 2/Scenario 3 distinction adds to the postpromulgation analysis.
Hickman and Thomson begin with three case studies of postpromulgation notice and comment. The first case study involves the EPA regulations addressed in D.C. Circuit’s New Jersey decision—a case that, as mentioned above, I think falls into either Scenario 1 or Scenario 2. The second case study focuses on the Attorney General’s application of the Sex Offender Registration Notification Act of 2006 (SORNA). SORNA requires convicted sex offenders to register with state governments and imposes criminal penalties on offenders who fail to do so. One provision, codified at 42U.S.C. § 16913(d), authorizes the AG “to specify the applicability of the [registration] requirements” to sex offenders convicted before July 2006 and “to prescribe rules for the registration of any such offenders.” In February 2007, the AG promulgated an interim rule (R1) specifying that offenders convicted before July 2006 must register. The AG promulgated the interim rule without notice and comment but invited public comments afterwards. The AG then issued a final rule (R2) in July 2008 after notice and comment. R2 used “almost the exact same language” as R1.
As Hickman and Thomson chronicle, cases arose in several circuits involving individuals convicted of sex offenses before 2006 who faced penalties under SORNA for failing to register between the time of R1 and the time of R2. These cases fall squarely within Scenario 1. A circuit split developed, with the Fourth, Fifth, and Eleventh Circuits holding that the lack of prepromulgation notice and comment for R1 was either harmless error or excused for good cause, and the Third, Sixth, Eighth, and Ninth Circuits holding that R2 was procedurally invalid. Note, though, that the Sixth, Eighth, and Ninth Circuits all now accept that R2 is itself procedurally unproblematic. That is, the Justice Department can enforce R2 against individuals convicted of sex offenses before July 2006 who failed to register after the promulgation of R2 in July 2008. See United States v. Stevenson, 676 F.3d 557, 565-66 (6th Cir. 2012); United States v. Manning, 786 F.3d 684, 687 (8th Cir. 2015); United States v. Mattix, 694 F.3d 1082, 1084-85 (9th Cir. 2012); accord United States v. Whitlow, 714 F.3d 41, 47-48 (1st Cir. 2013);United States v. Lott, 750 F.3d 214, 219-20 (2d Cir. 2014). (As far as I know, the Third Circuit—which also ruled that R1 was procedurally invalid—has yet to address the procedural validity of R2.) Put differently, whereas the circuits were split as to the procedurally validity of R1 in Scenario 1, they are unanimous as to the procedural validity of R2 in Scenario 3.
Hickman and Thomson’s final case study involves a temporary Treasury Department regulation (R1) promulgated without notice and comment in September 2009 stating that the statute of limitations on deficiency assessments extended six years rather than three years under certain circumstances. Treasury then promulgated a final rule (R2) in December 2010 adopting the same position after notice and comment. Taxpayers in several circuits who were assessed before December 2010 challenged R1 on notice-and-comment grounds. The Tenth and Federal Circuits held that the promulgation of R2 mooted the notice-and-comment question with respect to R1. See Salman Ranch, Ltd. v. Commissioner, 647 F.3d 929, 940 (10th Cir. 2011); Grapevine Imports, Ltd. v. United States, 636 F.3d 1368, 1380 (Fed. Cir. 2011). The D.C. Circuit likewise rejected a notice-and-comment challenge from a taxpayer who was assessed before R2 but whose appeal was adjudicated afterwards. See Intermountain Insurance Service of Vail v. Commissioner, 650 F.3d 691, 709-10 (D.C. Cir. 2011). The Supreme Court later held in United States v. Home Concrete & Supply, LLC that Treasury’s interpretation of the statute of limitations was invalid—but on entirely different grounds.
The question in the tax statute-of-limitations cases was the Scenario 1 question: whether subsequent promulgation of an essentially identical and procedurally valid rule R2 can cure a procedural problem with respect to R1. That’s an interesting and difficult question. Yet it strikes me as a separate question from the Scenario 3 question: whether a rule R2 promulgated according to APA notice-and-comment requirements and subsequently applied to a party is valid notwithstanding the fact that it is essentially identical to a rule R1 previously promulgated without notice and comment. And recall that in the SORNA context, circuits addressing this latter question have answered it unanimously in the affirmative.
Regardless of whether my characterization of the Scenario 3 rule is right as a doctrinal matter, is it desirable as a normative matter? One concern might be that agencies will use postpromulgation notice and comment to circumvent section 553’s requirements. But that concern seems to me misplaced: after all, the agency in Scenario 3 is not circumventing any of section 553’s requirements with respect to R2. And the agency still has an incentive to go through notice and comment before promulgating R1. If it doesn’t, then it may lose the ability to enforce R1 between time t and time t + 1.
Andy also raises the concern that postpromulgation notice and comment may be a charade: “Anyone who practices before an agency knows that comments are most likely to be taken seriously before the agency has made up its mind, not after it has already publicly bound itself to a position.” But in the normal course, an agency stakes out its tentative position in a notice of proposed rulemaking at the outset of the notice-and-comment process; there too, the agency reveals what it’s thinking before any comments come in. To be sure, the agency is free to change its position between the proposed rule and the final rule (at least within the limits of the “logical outgrowth” doctrine). But by the same token, the agency is free to change its position between time t and time t + 1 in Scenario 3 on the basis of the comments it receives in the run-up to R2.
Some courts have suggested that an agency must keep an “open mind” throughout the notice-and-comment process or else the agency’s action will be set aside. See Advocates for Highway & Auto Safety v. Federal Highway Administration, 28 F.3d 1288, 1292 (D.C. Cir. 1994). Yet this open-mindedness requirement is a made-up one too. As the Federal Circuit has observed, “[t]he APA does not expressly require agencies to keep an ‘open’ mind, whatever such a subjective term might mean.”Mortgage Investors Corp. v. Gober, 220 F.3d 1375, 1379 (Fed. Cir. 2000). The Supreme Court has never adopted the open-mindedness requirement—and rightly so. While we should be concerned about closed-mindedness on the part of agencies, we should be even more concerned about courts trying to read the minds of agencies (assuming that agencies even have “minds” apart from the many individual minds of their officers and employees).
In sum, a rule R2 promulgated after notice and comment is (and should be) procedurally valid regardless of whether the rule is essentially identical to a previous rule R1 promulgated without notice and comment. If the agency applies R2 after it is promulgated, an individual who challenges the application of R2 on notice-and-comment grounds has no claim at all. This seems from a cursory examination to be the approach that circuit courts are taking, even while they disagree as to the effect of postpromulgation notice and comment under other scenarios. And given the Supreme Court’s admonition that lower courts should stop conjuring up new procedural requirements found nowhere in the APA, the approach seems like a wise one—at least insofar as lower court judges are averse to reversal. As for whether the Homeland Security and Justice Departments thought this all through in the same way when plotting the administration’s strategy in United States v. Texas, though: I’m no better equipped to read the minds of agencies than courts are.