Is Garmon on its last legs? First came the Supreme Court’s 2023 decision in Glacier Northwest, Inc. v. Teamsters, where five Justices called the labor law preemption doctrine “unusual” while two others invited the Court to reconsider the “strange[] . . . Garmon regime.” Then came Loper Bright, ending Chevron deference. Add in attacks on the NLRB’s constitutionality and a quorum-less Board, and many of us concluded Garmon’s days were numbered — and that we should look exclusively to the states to build the new labor-law order.

Maybe it’s time for Garmon to go. But whatever one thinks of Garmon as a matter of policy, the Sixth Circuit’s recent decision in Rieth-Riley Construction Co. v. Operating Engineers Local 324 offers an occasion to revisit the doctrine’s legal foundations. The upshot: those foundations hold.

Rieth-Riley clarifies Garmon by isolating its primary jurisdiction logic.

After Glacier was decided, I argued that Garmon is best understood not as a novel or anomalous doctrine, but as the combined application of two ordinary ones: primary jurisdiction and conflict preemption. Under that account, Garmon’s sequencing rule (yield to the Board first) reflects the Board’s primary jurisdiction to decide labor law disputes; the resulting displacement of state law (once the Board has spoken) reflects ordinary Supremacy Clause preemption. The two components typically travel together because the usual Garmon case involves a state-law claim.

But in Rieth-Riley, there was no preemption issue — there was no state law issue at all. Rieth-Riley and several employees sued fringe-benefit funds and their trustees under ERISA (a federal law) after the funds stopped accepting the employer’s contributions. 

So why did the Sixth Circuit apply Garmon? Because of primary jurisdiction, standing alone: as pleaded, the ERISA claims required a court to decide a disputed, unresolved NLRA question — whether the funds had a statutory duty under the NLRA to keep accepting contributions while bargaining continued — before the Board had spoken on it. The question in Rieth-Riley was whether the NLRB needed an opportunity to resolve that antecedent labor law question in the first instance before a court could adjudicate the ERISA claims. 

The answer, under Garmon, was yes. When conduct is “arguably subject” to Section 7 or 8, “the States, as well as the federal courts, must defer to the exclusive competence of the National Labor Relations Board.” Garmon also explains why: Congress assigned “primary interpretation and application” of those rules to a specialized tribunal, with Board adjudication first and judicial review later. 

This has nothing to do with preemption; it has more in common with exhaustion or abstention doctrines, which are about defining the order of review rather than limiting a court’s ultimate jurisdiction. In this way, Rieth-Riley clarifies Garmon by isolating the primary jurisdiction part of the doctrine.

Loper Bright does not destabilize Garmon.

In her concurrence, Judge Hermandorfer suggests that Loper Bright may cast doubt on using Garmon to “shunt statutory interpretation questions from courts to the Board.” I imagine this isn’t the last time this argument will be made, but it doesn’t work. Loper Bright is about deference; Garmon is about sequencing. If anything, the two doctrines reinforce each other.

First, primary jurisdiction is itself the product of courts reading the NLRA’s text de novo. The statute assigns initial resolution of unfair labor practice disputes to the Board, not to any court. Garmon and its predecessors independently derived the primary jurisdiction rule by construing the NLRA, as Loper Bright demands.

Second, the Board’s “primary” jurisdiction does not deny courts their ultimate jurisdiction. As Garmon recognized, the Board initially decides labor law questions, “subject to appropriate federal judicial review.” This sequences the procedures for adjudicating a claim, requiring labor law disputes to move through the Board first, akin to administrative exhaustion requirements found in many statutes. Courts still have the final word.

Third, Garmon prescribes no rule of deference. Loper Bright requires courts to exercise independent judgment, rather than deferring to agency interpretations of law, once a disputed question properly reaches the court. But even a constitutionalized Loper Bright would cast no doubt on Congress’s ability to require Board review of a question as a precondition for judicial review. Otherwise, virtually all administrative exhaustion procedures would be impugned.

Garmon is a flexible and practical doctrine.

Judge Hermandorfer eventually acknowledges that Garmon could be read “as a subset of primary jurisdiction doctrine.” But this leads her to a litany of questions about Garmon’s application: What if the Board lacks a quorum? What if the Board cannot decide the underlying labor law issue because of some other practical barrier? What if a party asserts that the Board is unconstitutionally constituted and thus lacks power to adjudicate the issue? What if a party unreasonably delays Board proceedings? 

On these questions, I agree with Judge Hermandorfer: courts should take a functional approach. They should “resist applying ‘the Garmon guidelines in a literal, mechanical fashion,’” as the Supreme Court wrote in Sears, Roebuck & Co. v. Carpenters. There, the court observed:

The primary-jurisdiction rationale justifies pre-emption only in situations in which an aggrieved party has a reasonable opportunity either to invoke the Board’s jurisdiction himself or else to induce his adversary to do so. In this case, Sears could not directly obtain a Board ruling on the question whether the Union’s trespass [while picketing] was federally protected.

Sears further noted that the Garmon question would turn out differently if the Board “provided an effective mechanism whereby an employer could obtain a determination . . . as to whether picketing is protected or unprotected.” It just so happened that “an employer faced with ‘arguably protected’ picketing is given by the present federal law no adequate means of obtaining an evaluation of the picketing by the NLRB,” and that practical hurdle was relevant for deciding that Garmon did not apply.

So, yes, not every labor law question embedded in a federal case should trigger Garmon. The labor law issue must be unresolved. Resolving it must be necessary to deciding the federal claim. And the Board must have a real opportunity to resolve the issue. 

All of this flows from the NLRA’s text and, insofar as Congress left some procedural minutiae unresolved, from “a practical understanding of legislative intent.” The underlying task is figuring out what procedural arrangement Congress created for resolving the questions at issue. That requires courts to harmonize the institutional roles assigned by overlapping statutes and to ask what each scheme contemplates. That is ordinary, even prototypical, judicial work. 

Whether Garmon remains good policy is a separate debate. But inconvenient as it may be for the future of labor law experimentation, the doctrine continues to rest on a sound premise. Reports of its death may be premature.