Twenty-five years ago, the Administrative Conference of the United States (ACUS) brought the technique of “direct final rulemaking” to the attention of the administrative law community. Since that time, agencies have used the technique thousands of times to adopt noncontroversial regulations on an expedited basis. But its legality depends on a creative reading of the Administrative Procedure Act (APA). A recent D.C. Circuit case, applying the APA in a manner that overlooked the distinctive features of this device, has exposed this vulnerability and may well have seriously undermined the viability of the practice.
The EPA invented direct final rulemaking in the early 1980s. In 1995, ACUS followed up with Recommendation 95-4, which explained how it could be used by other agencies. To use this procedure, an agency publishes a rule in the Federal Register with a statement that the rule will become effective unless someone files an objection—in ACUS’s version, a “significant adverse comment”—within a specified period. If no one opposes the rule, it can go into effect immediately, without further deliberations on the agency’s part. If anyone does file an adverse comment, the agency will withdraw the rule and, if it chooses, propose it again using the regular notice-and-comment process. The purpose of the direct final rulemaking procedure is to streamline the rulemaking process for rules that are expected to be entirely noncontroversial. As an industry lawyer said at the time, this procedure “is in everyone’s interest. It just saves time.”
As outlined by ACUS, the legal justification for direct final rulemaking rests on § 553(b)(B) of the APA, which exempts rules from notice-and-comment obligations when the agency finds for good cause that those procedures would be “unnecessary.” As the committee reports on the APA explained, this means “unnecessary as far as the public is concerned, as would be the case if a minor or merely technical amendment in which the public is not particularly interested were involved.” In direct final rulemaking, the agency makes a provisional judgment that this criterion is satisfied, but it then uses the objection period as a reality check. If the agency doesn’t receive a significant adverse comment within the allowed period, it can infer that its prediction was correct. Conversely, if someone does object, the objection demonstrates that the premise of the good cause finding was mistaken.
That reasoning was rather elegant. (Of course I would say that, because I proposed the theory myself when I was the ACUS consultant for Recommendation 95-4.) Still, the Achilles heel in the legal case for direct final rulemaking has always been the fact that the APA does not explicitly authorize the procedure. For twenty-five years, that vulnerability didn’t seem to matter. Agencies made use of the procedure, on the order of 100-300 times per year in recent years, and there was no significant controversy about it. Bur there was always a potential for breakdown if participants in the process didn’t understand its premises. And that time bomb has now exploded, as the D.C. Circuit’s July 2 decision in Milice v. CPSC demonstrates.
The Agency’s Missteps
The problems began at the agency level. The underlying dispute grew out of regulations that the Consumer Product Safety Commission (CPSC) has prescribed to ensure the safety of bath seats for babies. Those regulations, like thousands of other federal regulations, incorporate by reference many provisions in a voluntary code written by an industry association, in this case the American Society for Testing and Materials (ASTM). The regulation does not itself contain the operative language of the code. Moreover, the code is copyrighted, so anyone who wants to obtain a copy of it would have to purchase it from ASTM.
In 2019, when ASTM made minor revisions to its code, the CPSC sought to update its regulation with a direct final rule, but the New Civil Liberties Alliance (NCLA) filed an opposing comment. NCLA’s letter argued that the anticipated CPSC rule was bad policy and would deprive members of the public of due process rights, because the provisions incorporated into it by reference were not freely available to the public. The Commission recognized that similar contentions have been advanced before. (Indeed, eminent scholars of administrative law have urged essentially the same position.) However, the Office of Federal Register, which oversees incorporation by reference in federal regulations, has steadfastly disagreed, insisting that the industry groups’ proprietary interests in their copyrighted codes must be respected, and the CPSC adhered to that position in its response to NCLA.
For the sake of discussion, we might assume that the CPSC would have prevailed if it had adhered to its views at the culmination of a regular rulemaking proceeding. In this context, however, the Commission asked itself the wrong question. Its reasoning fundamentally lost sight of the purpose of the objection period in a direct final rulemaking proceeding. The purpose is to ascertain whether the proposed rule is so uncontroversial that notice and comment proceedings would be “unnecessary.” NCLA’s comment showed conclusively that it wasn’t.
Some agencies, such as the Environmental Protection Agency (EPA), allow a single objector to derail a direct final rule by merely stating an intent to submit adverse comments. Under that practice, NCLA’s comment obviously would have halted the CPSC’s baby bath seat proceeding. The ACUS model, however, provides that a would-be objector must at least submit a “significant adverse comment.” ACUS included this requirement because it feared that members of the public might sometimes submit bad-faith objections in order to block the proceeding, rather than because they wanted to place any issues before the agency. (Experience has borne out that apprehension, especially during the Trump administration. A recent Brookings Institution study found that about half of the objections to EPA direct final rules in 2017 were “spam” comments, apparently intended merely to annoy the government.) Nevertheless, ACUS intended the definition of “significant adverse comment” to be quite lenient. It said that a comment that “explains why the rule would be inappropriate” would suffice. It also suggested that, “[i]n determining whether a significant adverse comment is sufficient to terminate a direct final rulemaking, agencies should consider whether the comment raises an issue serious enough to warrant a substantive response in a notice-and-comment process.” Obviously, if the CPSC had conducted such a process in this instance, it would have been expected to address NCLA’s position, not completely ignore it.
For these reasons, NCLA would have had a strong case on judicial review in Milice that the CPSC had misapplied the good cause exemption, especially in light of the courts’ usual practice of construing rulemaking exemptions narrowly. Unfortunately, the D.C. Circuit never reached that issue, because it too failed to take account of the realities of direct final rulemaking.
The Court’s Missteps
Here I need to explain some key dates in the litigation. The CPSC issued its direct final rule on September 20, 2019, inviting significant adverse comments within thirty days and stating that the rule would take effect on December 22 if no such comments were received. NCLA submitted what it thought was an adverse comment on October 21. However, it heard nothing until the CPSC notified it on February 6, 2020, that it did not consider the comment significantly adverse. Milice (a consumer of products for babies, who was represented by the same lawyers but probably had a stronger basis for standing to sue) promptly appealed, but the court held that the appeal was untimely, because the September 20 announcement had been a final agency action, and Milice had not petitioned for review within 60 days of that date.
The linchpin of the court’s conclusion that Milice’s suit was time-barred was that the CPSC rule had been “promulgated” on September 20, so that the sixty-day statute of limitations clock began running as of that date. In reaching this conclusion, the court said that the language of the preamble to the rule certainly didn’t sound tentative. The court also cited circuit precedent for the proposition that publication of a rule is final agency action “notwithstanding the possibility that the Commission might reconsider and change its standard in the future.” Under most circumstances, the court’s argument would have been persuasive, because an agency might always develop second thoughts after it issues a rule, and the courts cannot let that bare possibility prevent the rule from becoming appealable (and thus from letting the limitations period begin running as of that time).
But direct final rulemaking is quite different. The whole point of this technique is that the agency uses the opportunity for post hoc objection as a reality check to confirm its initial impression that the rule would be uncontroversial, rendering public comment “unnecessary” under § 553(b)(B). In this sense, an agency’s publication of a direct final rule is inherently provisional and should not be considered “definitive” for finality purposes. In this instance, although the CPSC did self-confidently say that it was using direct final rulemaking because it did not “expect” adverse comments, this was qualitatively different from declaring that the exemption applies, regardless of what anyone may say to the contrary. The court’s reasoning would have been more apt if the agency had simply invoked the good cause exemption outright when it published the rule. But, by treating this case as though such a categorical invocation had occurred, the court effectively treated the direct final rulemaking procedure as meaningless.
So, if the court erred by considering September 20 the rule’s promulgation date, what date should the court have identified instead? The Milice opinion relied on its prior decision in Horsehead Resource Development Co. v. EPA. There the court had said that the “default rule” is that promulgation means Federal Register publication—but it went on to explain why that date made sense in context. In Horsehead, the plaintiff filed its petition before the rule was published, so it needed to find a reason why the petition should not have been dismissed as premature. To that end, it argued that the date on which the agency had distributed signed copies of the rule to Horsehead and other interested persons should be considered the promulgation date, even though the rule didn’t appear in the Federal Register until ten days later. The court rejected this argument, observing that “an agency must give some notice of ‘the substance’ of its final action before that action can be deemed ripe for judicial review.” It quoted the Supreme Court’s opinion in Skelly Oil Co. v. Phillips Petroleum Co.: “[S]urely [an order] cannot be said to have been issued for purposes of defining rights and the seeking of reconsideration by an aggrieved person if its substance is merely in the bosom of the commission. Knowledge of the substance must to some extent be made manifest.”
On the facts of Milice, the reasoning of Horsehead and Skelly points in the opposite direction. As I just explained, in direct final rulemaking the initial finding that public comment is “unnecessary” is, by definition, always provisional. It doesn’t become definitive until the agency has waited to see whether it has received a “significant adverse comment.” At the same time, it would be insupportable to say that the limitations statute begins running before the agency makes its determination known. The CPSC had said that, in the event of a negative determination, it “would withdraw the 2019 Rule before its effective date and publish notice in the Federal Register.” Therefore, the earliest date on which the public could have known that the rule had been “promulgated” was on the effective date (December 22), when an interested person could observe that the Commission had failed to withdraw the rule. That date should be considered the promulgation date. Milice filed her petition for review 59 days later, so her petition was timely.
Some might insist, however, that only the date of a rule’s publication can be considered its promulgation date. Even so, the court could have found that Milice’s petition was timely for a different reason. The letter that NCLA submitted on October 21 could be regarded as tantamount to a petition for reconsideration. A reconsideration petition, of course, suspends the running of the statute of limitations. As I have written elsewhere, an adverse comment in a direct final rulemaking proceeding “resembles other reconsideration requests in that it brings new information to the agency’s attention and demonstrates that the agency’s earlier decision (i.e., that public proceedings were ‘unnecessary’) was based on a misapprehension.”
In Milice, however, the court declined to reach the reconsideration theory, because it said that reconsideration is a “party-based” objection. That is, even if NCLA arguably sought reconsideration, Milice certainly did not. Yet this is another doctrine that works in ordinary cases but not in the context of direct final rulemaking. The rationale of the general rule seems to be that if an order applies to A and B, and A seeks reconsideration, that move tolls the clock for A, but B should be free to go to court immediately and not have to wait around for the agency to dispose of A’s petition. In the direct final rulemaking context, however, the significance of NCLA’s (constructive) reconsideration petition wasn’t that it gave the agency an occasion to respond to concerns voiced by NCLA (but not Milice). Rather, the very existence of NCLA’s adverse comment showed that notice and comment wasn’t “unnecessary” after all. If the CPSC had recognized the comment as “significantly adverse,” it presumably would have conducted a regular rulemaking proceeding, at which it could have heard from NCLA, Milice, and other interested persons (including the nine well-known administrative law scholars who submitted an amicus brief on the incorporation by reference issue in the court of appeals). But the agency failed to recognize it as such. This was a procedural error that Milice could raise on judicial review, as litigants typically can do when an issue was raised in a rulemaking proceeding by someone other than themselves.
Direct final rulemaking may well continue to operate normally in the kind of situations for which it was designed. Either the rule will survive due to the absence of a bona fide objection, or the filing of such an objection leads the agency to withdraw the rule for further consideration. Sometimes, however, agencies seek to exploit the direct final rulemaking process in situations that it was not designed to handle. The system needs a stable legal framework with which to respond to those situations. After all, this is one of the core purposes of administrative law.
The expected constraints on direct final rulemaking break down when an agency adopts a rule without notice and comment by claiming that such procedure is “unnecessary,” where the agency has received a serious, good-faith objection to the anticipated rule but happens to disagree with the objection. Similarly, the expected constraints break down when a court reads its judicial review statute to mean that an interested person’s ability to file suit against a direct final rule in court may expire before the person has any way to know that the agency has rejected its objection to the rule.
Furthermore, the Milice decision threatens to subvert the process of direct final rulemaking itself. Under the D.C. Circuit’s reasoning, a direct final rule is considered final and ready for judicial review from the day when it is first announced. Yet, if the judicial review record is closed as of that date, the agency has much less incentive to pay attention to any post-publication comments it may receive. Indeed, the court’s finality reasoning may imply that the agency cannot heed such comments except by commencing an entirely new rulemaking proceeding.
In sum, the Milice decision has cast a pall over the entire practice of direct final rulemaking, which agencies have used thousands of times without significant controversy ever since ACUS recommended it for general use in 1995. Regrettably, the court’s opinion does not even discuss these ramifications. Perhaps the court saw only bathwater in the case and didn’t realize it was throwing out a baby. If the Milice precedent stands, however, Congress may need to step in. Indeed, a 2017 Senate bill did propose to codify direct final rulemaking. At the time, such legislation didn’t seem necessary, but that assessment may now have to be reevaluated.
Ronald M. Levin is the William R. Orthwein Distinguished Professor of Law at Washington University in St. Louis.