This month, two judges made a strong bid — perhaps the strongest of the century — to limit when agencies may make law through adjudication rather than rulemaking. In Brown-Forman Corp. v. NLRB, the Sixth Circuit declined to enforce a bargaining order that relied on Cemex Construction Materials Pacific, LLC, which announced a new standard for remedial bargaining orders. 

The Sixth Circuit didn’t evaluate Cemex’s substance. Instead, it held that the National Labor Relations Board should have adopted the new remedial bargaining standard through notice-and-comment rulemaking. 

According to Brown-Forman, an agency cannot announce a new rule through adjudication unless it is “motivated” by “resolving the parties’ dispute.” Because the Board’s new remedial-bargaining-order standard did not change the outcome in Cemex — the Board there had already found that a bargaining order was warranted under its old standard — Cemex arguably failed that test. And because the Board’s order in Brown-Forman relied on Cemex, it, too, was invalid.

Some scholars have urged courts to rein in agency lawmaking through adjudication. Rather than strengthening their case, Brown-Forman exposes pitfalls in that project. Its suggestion that agencies may only announce standards through adjudication when doing so changes the outcome would forbid ordinary dicta and entrench reasoning the agency no longer finds persuasive. Meanwhile, its ultimate holding — that because the Cemex rule was invalidly adopted, the bargaining order in Brown-Forman cannot be enforced — appears plainly inconsistent with Supreme Court precedent.

In 2022, a majority of employees at a Brown-Forman bourbon facility signed union cards with the Teamsters. Shortly before the election, Brown-Forman gave everyone an unprecedented raise to deter unionization. The Board found that Brown-Forman committed an unfair labor practice and, applying Cemex, ordered the employer to bargain with the Teamsters. But the Sixth Circuit refused enforcement, deeming Cemex a rule of general applicability that should have gone through notice and comment.

To see why, we have to return to Cemex (which I’ve written about in more detail here and here). In that case, Cemex had committed more than twenty unfair labor practices, including threatening to terminate workers who organized. The Board ordered Cemex to bargain with the Teamsters, but under two alternative rationales. Applying its existing framework, the Board concluded that “the possibility of erasing the effects of the . . . misconduct and ensuring a fair rerun election by the use of the Board’s traditional remedies is slight.” The Board then also adopted a new rule: it will now issue a bargaining order whenever an employer commits more than de minimis misconduct in the lead-up to an election.

According to the Sixth Circuit, the new rule was an improper exercise of adjudicative authority because “[th]e Cemex Board created the standard for means unrelated to resolving the parties’ dispute.” 

What does that mean? The opinion offers only hints. One clue is its charge that the Cemex standard “was based not on the case-specific facts, but on general observations from ‘[d]ecades of experience . . . in other adjudications.” But the Sixth Circuit cannot mean that agencies may consider nothing outside the factual record — which would include the law.

A more plausible reading is that, under Brown-Forman, an agency cannot announce a new standard when doing so would not change the outcome. On this view, the Board may consider materials outside the record, but only insofar as they help resolve the dispute before it.

But this approach is hard to defend. First, it would effectively forbid dicta. For example, an agency could never identify limits to its holding — the familiar “although today’s rule covers X, it doesn’t cover Y” would become impermissible. The suggestion that all dicta is incompatible with adjudication flies in the face of the Brown-Forman opinion itself, which posits, without a hint of irony, “that even where the Board properly exercises its adjudicatory authority, reviewing courts [ought to] make a point to re-establish the contours of that policymaking power.” 

Second, a must-change-the-outcome requirement would handcuff adjudicators to reasoning they no longer find persuasive. Brown-Forman implicitly acknowledges this problem in its discussion of another Board case involving changed standards, Alstate Maintenance: “In Alstate, the modification to the standard did not change the ultimate outcome, but the modification was still prompted by the nuances of the specific facts, designed to better adjudicate the parties’ particular dispute because the previous standard was insufficient.” Here, “insufficient” cannot mean insufficient to reach the holding: the same outcome was available under the old standard. (Alstate overruled a case holding that “an employee who protests publicly in a group meeting is engaged in initiating group action,” even though Alstate itself did not involve a group meeting.) The only remaining explanation is that the prior standard was “insufficient” in that it rested on reasoning the Board no longer found persuasive. But if the Alstate Board was entitled to reconsider reasoning it no longer found persuasive — even though doing so did not change the outcome — why could the Board in Cemex not do the same? Brown-Forman offers no answer.

Instead of the Sixth Circuit’s requirement that adjudicative rules be outcome determinative, here’s a better, more modest principle: a rule can be announced in adjudication only if it is applied in the case. A plurality of the Supreme Court suggested as much in NLRB v. Wyman-Gordon Co. It invalidated the Board’s prior adoption, in Excelsior Underwear, Inc., of a rule requiring employers to furnish employee lists before elections, on the basis that the Excelsior Board “did not even apply the rule it made to the parties in the adjudicatory proceeding.” 

The logic is that formal adjudication and rulemaking rely on particular guardrails to promote reasoned decision-making and protect against interest-group capture. In adjudication, the chief safeguard is the adversarial process. That security weakens when the agency decides an issue that operates only prospectively, giving the parties no concrete stake in it. When the new standard is applied in the case, even as an alternative holding, that concern largely fades.

Wyman-Gordon underscores the point. After finding that Excelsior had invalidly created a prospective-only rule, the plurality nonetheless enforced the Board’s order applying Excelsior to Wyman-Gordon because it was “an order in the present case that the respondent was required to obey.” Although the Wyman-Gordan plurality did not explain its logic, we can fill it in: unlike in Excelsior, the parties in Wyman-Gordon had skin in the game. The same logic should have applied in Brown-Forman.

To be sure, even when a rule is applied in a case, the adversarial safeguard collapses when the agency decides issues the parties did not litigate. Courts seeking to discipline adjudicatory rulemaking might therefore insist on stricter adherence to the party-presentation principle. In Cemex, the parties litigated the framework for remedial bargaining orders after the General Counsel asked the Board to reconsider it. Alstate, by contrast, adopted a new standard without either party’s request. If either required notice and comment, it was Alstate, not Cemex.

Ultimately, if courts are to police the boundary between adjudication and rulemaking, they will need a more coherent framework than the one Brown-Forman offers. Far from justifying limits on adjudicatory policymaking, the Sixth Circuit’s difficulties make a better argument for judicial restraint.