In a recent antitrust preemption appeal, the Second Circuit reaffirmed the fundamental and well-settled point that Section 1 liability requires an agreement. Specifically, in Connecticut Fine Wine & Spirits, LLC. v. Seagull, the Second Circuit reviewed a facial challenge brought by a liquor retailer against Connecticut, arguing that the state’s regulatory framework governing the distribution and sale of alcoholic beverages was preempted by Section 1 of the Sherman Act because the state’s regulations, in the plaintiff’s view, eliminated incentives for wholesalers to compete on price, thereby causing higher prices for consumers. Three parts of the regulatory framework were challenged:
- Post and hold provisions, which require wholesalers to “post” their unilaterally determined prices for each product being sold during the following month. Other wholesalers then have four days to match (but not beat) a lower posted price. The wholesalers are then required to “hold” that price for the following month;
- Minimum retail price provisions, which require retailers to sell at the wholesaler’s posted price, plus a markup; and
- Non-discrimination provisions, which prohibit wholesalers from giving volume or other pricing discounts.
For reasons that will be discussed further below, consistent with a number of prior court decisions within the Second Circuit and elsewhere, the Second Circuit found that none of the statutory provisions were preempted by the Sherman Act, allowing Connecticut’s regulatory regime to remain.
However, despite clear guidance from the Supreme Court, over the last three decades a small number of officials, academics, and a minority of circuit courts of appeals have pushed to reinterpret the “agreement” requirement in the context of antitrust preemption cases in a way that would signify a radical expansion in the reach of Section 1. In fact, Merrick Garland, current Attorney General of the United States, previously argued that, in the context of preemption challenges, the meaning of “agreement” is so “in doubt” that a violation of the antitrust laws can be found where a firm merely engages in entirely unilateral conduct that is compelled by state law. Academics Phillip Areeda and Herbert Hovenkamp have similarly urged that the preemption analysis should be “reconcile[d]” with the state action immunity test, such that liability may be found even where “there [are] no agreements at all.” These positions are not merely an academic matter—two courts of appeals have erroneously adopted the Areeda-Hovenkamp analysis.
Specifically, in the 1980s, in Miller v. Hedlund, the Ninth Circuit considered whether Oregon’s post and hold and price non-discrimination provisions for alcohol violated federal antitrust law. In 2001, in TFWS, Inc. v. Schaefer, the Fourth Circuit likewise considered an antitrust challenge to Maryland’s post and hold and price non-discrimination provisions. And in 2008, in Costco Wholesale Corp. v. Maleng, the Ninth Circuit decided a challenge to Washington’s liquor laws, again including comparable post and hold, price non-discrimination, and minimum retail price provisions. In all three cases, the central aspect of the states’ liquor pricing regulatory regimes—the post and hold requirement—was found to violate the Sherman Act, contrary to the Second Circuit’s decision in Connecticut Fine Wine & Spirits, LLC. This, of course, begs the question of what drove these two courts of appeals to reach their conclusions. The answer, we believe, is that the Fourth and Ninth Circuits have improperly conflated two separate doctrines—state action immunity and preemption—in a way that has permitted them to find Section 1 violations despite a lack of any evidence of agreement between competitors.
This article will explain why, in our view, efforts in these Circuits to expand the scope of the Sherman Act in preemption cases—by disregarding its plain text and clear Supreme Court guidance on the subject—are flawed; and why the Second Circuit’s analysis is most consistent with the Supreme Court’s state action immunity and preemption frameworks. Failing to enforce the agreement requirement, we think, risks a radical and legally unfounded expansion of potential antitrust liability.
The Frameworks for Analyzing State Action Immunity and Preemption Issues
Over four decades ago, a unanimous Supreme Court first established the modern test for state action immunity in California Retail Liquor Dealers Ass’n v. Midcal Aluminum, Inc. There, the Midcal Court addressed a California regulatory regime which, among other things, permitted wine wholesalers to post a resale price schedule for their brands, and prohibited retailers from undercutting those prices. At the time the Court evaluated that regime, the substantive antitrust law was vastly different than today including, importantly, that resale price maintenance was still considered per se illegal. Relying on the since-overruled Dr. Miles decision, the Court stated conclusively that the “the Sherman Act’s ban on resale price maintenance” “plainly” applied to California’s regulatory regime. Notably, the Court did not conduct any analysis regarding whether the wholesalers actually entered into any agreements with the retailers to vertically fix prices, and there is no evidence that was the case. Instead, consistent with the state of antitrust law at the time condemning resale price maintenance, the Court simply assumed that the statute was illegal. The Court analyzed solely “whether the State’s involvement in the price-setting program is sufficient to establish antitrust immunity” under the state action doctrine. In doing so, it articulated the modern state action immunity test, sometimes referred to as the Midcal test: a statute may be saved where (i) the challenged restraint is “clearly articulated and affirmatively expressed as state policy;” and (ii) the policy is “ ‘actively supervised’ by the State itself.” The Court concluded that this standard was not met, and enjoined enforcement of the California law.
The Supreme Court addressed an antitrust challenge to another California liquor regulatory regime just two years after Midcal in Rice v. Norman Williams Co. But in Rice, the Court focused on the question of preemption rather than state-action immunity, applying a very different analytical framework. Specifically, the Court considered whether a so-called “designation statute”—prohibiting importers from accepting liquor unless the importers were designated by a liquor brand owner—was preempted by federal antitrust law. The Court explained:
a state statute, when considered in the abstract, may be condemned under the antitrust laws only if it mandates or authorizes conduct that necessarily constitutes a violation of the antitrust laws in all cases, or if it places irresistible pressure on a private party to violate the antitrust laws in order to comply with the statute.
The Court further instructed that:
condemnation will follow . . . when the conduct contemplated by the statute is in all cases a per se violation. If the activity addressed by the statute does not fall into that category, and therefore must be analyzed under the rule of reason, the statute cannot be condemned in the abstract.
Applying that framework, the Court held that because the designation statute imposed a non-price vertical restraint, it should be judged by the rule of reason under GTE Sylvania. Because the firms’ conduct under the California designation statute would not be a per se violation of the Sherman Act in all cases, the Court held that the law did not irreconcilably conflict with the Sherman Act and was not preempted. Notably, having found the law not preempted, the Court declined to consider state action immunity. The Court, therefore, drew a clear distinction between the state action immunity doctrine on the one hand and preemption on the other.
The Supreme Court’s next decision in this area was just a few years later in Fisher v. City of Berkeley, where the Court again applied the preemption framework to analyze a municipal rent control ordinance. As in Rice, the Court explicitly distinguished between state action immunity and preemption—it noted that its analysis was “limited to the antitrust preemption question”; that “consideration of state action is not necessary unless an actual conflict with the antitrust laws is established”; and that “[l]egislation that would otherwise be pre-empted under Rice may nonetheless survive if it is found to be state action immune from antitrust scrutiny under Parker v. Brown.”
Crucially, for preemption analyses under the Sherman Act, the Fisher Court made clear that there can be no preemption absent evidence of an agreement:
The distinction between unilateral and concerted action is critical here. Adhering to the language of § 1, this Court has always limited the reach of that provision to unreasonable restraints of trade effected by a contract, combination, or conspiracy between separate entities. We have therefore deemed it of considerable importance that independent activity by a single entity be distinguished from a concerted effort by more than one entity to fix prices or otherwise restrain trade. Even where a single firm’s restraints directly affect prices and have the same economic effect as concerted action might have, there can be no liability under § 1 in the absence of agreement.
The Court noted, “A restraint imposed unilaterally by government does not become concerted-action within the meaning of the statute simply because it has a coercive effect upon parties who must obey the law.”
The Court in Fisher distinguished unilateral conduct, which cannot form the basis for antitrust liability, with so-called “hybrid” conduct, which can. The Court explained that “hybrid” conduct is implicated where a regulatory regime “merely enforce[s] private marketing decisions” such that “private actors are . . . granted ‘a degree of private regulatory power.’” One example of hybrid conduct, the Court indicated, would be a “state- or municipality-administered price stabilization scheme [that] is really a private price-fixing conspiracy, concealed under a ‘gauzy cloak of state involvement.’” The Court wrote, by contrast, that the landlord’s activity in Fisher was not hybrid conduct, even though the landlords were required to set the same maximum rent, because “the rent ceilings they mandate have been unilaterally imposed on the landlords by the city.” Simply put, the landlords reached no agreement to fix rents.
Taken together, Rice and Fisher established the rule that a state statute cannot be subject to a facial preemption challenge for conflicting with the antitrust laws unless (i) the conduct at issue is in all cases a per se violation of the federal antitrust laws; and (ii) the statute imposes a “hybrid” restraint, such that the private parties engage in concerted action. A law compelling firms to act unilaterally is not subject to preemption.
The Second Circuit Correctly Analyzed Preemption Under Rice and Fisher
In Connecticut Fine Wine & Spirits, LLC., the Second Circuit considered the argument that Connecticut’s post and hold provisions effectively brought about horizontal price fixing because, “[i]f a wholesaler were to drop its price on a particular product, its competitors would know immediately (from having seen the posted price), and would have four days to match the posted price.” The court began its analysis by rejecting plaintiffs’ position that no actual contract, combination or conspiracy need be shown for a state statute to be preempted by the Sherman Act. Notably, the court recognized, correctly, that Fisher stands for the principle that “[t]here need[s] to be concerted action,” calling that mandate a “significant” requirement. Although the court briefly noted that it “d[id] not take issue” with the district court’s holding that the post and hold law implicated hybrid conduct (we do), it ultimately said that it “doubt[ed] that such a law mandates or authorizes ‘concerted action’ among the wholesalers subject to it,” reasoning that “Connecticut’s prohibition on altering prices for a 30-day period is a purely negative restraint. It does not call for any private action, let alone concerted action.” The court explained: “Nothing about this [post and hold] arrangement requires, anticipates, or incents communication or collaboration among the competing wholesalers.” That focus on the actual conduct compelled by the statute is consistent with Rice and Fisher. Importantly, the court reiterated the importance of looking to the text of Section 1, rather than whether the statute simply has an anticompetitive effect: “‘§ 1 of the Sherman Act does not prohibit [all] unreasonable restraints of trade’ . . . . [T]hat conscious parallel conduct can create an equally uncompetitive market to parallel conduct achieved by agreement is of no moment. The gravamen of § 1 is an agreement among competitors.” Because the court determined that there was no evidence of any agreement between the firms, it held that the challenged liquor laws were not subject to preemption.
The Second Circuit’s decision in Connecticut Fine Wine & Spirits, LLC. is consistent with decades of jurisprudence construing post and hold provisions as unilateral restraints. In 1953, the Connecticut Supreme Court in Schwartz v. Kelly upheld a prior version of the Connecticut Liquor Control Act, reasoning that, “[i]n filing the schedule of minimum retail prices to be charged for their liquor, [wholesalers] . . . are merely complying with the law enacted by the General Assembly,” “not legislating” as “[t]here has been no delegation of legislative powers to the wholesalers.” Almost thirty years later, the District of Connecticut revisited the Connecticut post and hold provisions in two decisions, United States Brewers Ass’n, Inc. v. Healy and Serlin Wine & Sprits Merchants, Inc. v. Healy. In U.S. Brewers, the court upheld Connecticut’s post and hold provisions applicable to beer wholesalers on the basis that they required only “unilateral action” and they did “not exert irresistible economic pressure on the plaintiffs to violate the Sherman Act.” In Serlin, the court upheld the Connecticut Liquor Control Act against a preemption challenge, holding that “the most liberal application of the Sherman Act to ‘State related’ conduct itself does not apply to Connecticut’s pervasively state controlled liquor pricing law.” The court concluded that the Act is “facially valid” despite a limited role of private industry participants because “complete monopolization of the industry by the State” was not required.
The Second Circuit’s prior decision in Battipaglia v. New York State Liquor Authority similarly addressed whether New York’s post and hold regulatory regime “violates § 1 of the Sherman Act and [is] therefore invalid under the Supremacy Clause.” Judge Friendly, writing for the majority, quoted Rice for its instruction that, after a Court’s resolution of the preemption issue, it is not necessary “to consider whether the statute may be saved from invalidation under the doctrine of Parker v. Brown.” Finding that the New York post and hold regulation did not compel any agreement and instead “requires only that, having announced a price independently chosen by him, the wholesaler should stay with it for a month,” the Battipaglia Court held that the New York regulatory regime was not in all cases a per se violation of the antitrust laws, and therefore could not be preempted.
The Second Circuit’s decision was also consistent with authority from some other courts that have reviewed post and hold statutes and correctly denied preemption claims as the provisions did not compel any agreement. For example, the Supreme Court of Minnesota reviewed Minnesota’s post and hold statute, which required liquor wholesalers to file prices on a monthly basis and allowed a five day period to match a lower price. The court found that the statute did not “mandate conduct violative of the Sherman Act” because the statute “does not, on its face, compel any sort of agreement among wholesalers” “[n]or is Minnesota’s price posting system the equivalent of an agreement.” The Supreme Court of Missouri, too, has upheld a similar post and hold regime, finding that “[t]he price posting laws neither authorize nor sanction individual wholesalers to contract, combine or conspire to fix the prices that they or any other wholesaler or retailer charge. Rather, the price posting laws mandate unilateral action on the part of every wholesaler.”
Nor is the Second Circuit alone among the circuit courts in recognizing the fundamental requirement of an agreement in preemption cases reviewing state liquor laws. In Massachusetts Food Ass’n v. Massachusetts Alcoholic Beverages Control Comm’n, the First Circuit reviewed a Massachusetts statute limiting the number of retail liquor stores that could be owned by a single entity, which supermarket chains had argued is preempted by the Sherman Act. The First Circuit concluded that the statute was not preempted. The court reasoned that “there is no private agreement or arrangement between retailers as to the number of retail outlets and therefore no violation to be shielded.” The court also rejected the argument that the “statute is preempted because it produces an effect that could not be produced by agreement of private parties without violating the antitrust laws,” concluding that is true of “much direct government regulation prohibiting one form of economic activity or requiring another.”
The Second Circuit’s decision in Connecticut Fine Wine & Spirits, LLC. is consistent with this authority, and the Supreme Court’s decisions in Rice and Fisher.
Efforts to Conflate Preemption and State Action Immunity to Eliminate the Agreement Requirement
As set out above, the Fisher court was explicit in differentiating between the state action and preemption, and required evidence of an agreement for preemption to be successful. After that decision came down, certain individuals, however, sought to limit Fisher’s holding. For example, in 1987, Merrick Garland, who was then a law firm partner, published an article concerning the state action immunity doctrine and preemption challenges. While recognizing “Fisher’s requirement of an ‘agreement’” for preemption, he criticized that requirement as “impos[ing] yet another barrier to an antirust plaintiff’s success.” In his article, he sought to do away with the agreement requirement by offering an implausibly strained interpretation of Fisher. He wrote that “what saved rent control in Fisher was not the absence of abstract agreement, but rather the fact that the ordinance ‘place[d] complete control over maximum rent levels exclusively in the hands of the [city’s] Rent Stabilization Board.’” That, of course, ignored the Court’s actual holding in Fisher—set out two paragraphs later—that what saved the rent controls was that “Berkeley’s Ordinance lack[s] the element of concerted action needed before they can be characterized as a per se violation.” It also ignored the Supreme Court’s clear holding in Fisher that “there can be no liability under § 1 in the absence of agreement.” Garland proposed that courts essentially ignore the agreement requirement by collapsing the preemption analysis called for by Fisher and Rice into the older “the two-pronged Midcal test” for assessing state action immunity.
Garland’s proposal to collapse the state action immunity and preemption analysis and dispense with the agreement requirement—with which we respectfully disagree—gained traction among a small number of influential scholars. Areeda and Hovenkamp, for example, admit that, while Rice and Fisher “might be read to distinguish Parker [state action immunity] from preemption analysis,” they would nonetheless “attempt to reconcile” the cases. To cover the explicit distinction Rice draws between preemption and state action, Areeda and Hovenkamp seek to narrowly cabin its holding, writing that, although Rice “might suggest that ‘preemption’ and ‘state action’ analysis are totally different, all it need mean is that the Court was not yet obliged, in this ‘facial’ attack on the state statute, to apply the [Midcal test].” Similarly, when considering Fisher’s agreement requirement, Areeda and Hovenkamp ignore Fisher’s plain text, writing: “Although the Court thus seemed to hold that an illegal agreement was a prerequisite to preemption, it did not do so in fact, for it readily affirmed the preemptions found in Schwegmann and Midcal.” In other words, like Garland, Areeda and Hovenkamp acknowledge and then choose to disregard Fisher’s clear instruction that “there can be no liability under § 1 in the absence of agreement.”
A similar argument was advanced, unsuccessfully, in Judge Winter’s dissent in Battipaglia. In that dissent, Judge Winter called for application of the Midcal state action test to strike down the New York regulatory regime for reasons paralleling Garland, Areeda, and Hovenkamp:
The arrangement in question . . . would be per se illegal if accomplished solely through private agreement, and our decision, therefore, is governed by Midcal . . . . The relevant issue under the supremacy clause has to do with the degree to which a state may bring about the very anti-competitive arrangements the Sherman Act was designed to avoid . . . . [T]he fact that the state compels a private cartel offers no reason to exempt the legislation from scrutiny under the supremacy clause.
Thus, Judge Winter, like Garland, Areeda and Hovenkamp, looked for a way to find a Sherman Act violation based on supposedly anticompetitive outcomes, regardless of whether the firms’ conduct actually violated Section 1.
The Fourth and Ninth Circuits Have Erroneously Dispensed with the Agreement Requirement
Garland, Areeda and Hovenkamp, and Judge Winter’s Battipaglia dissent have apparently gained purchase, we think erroneously, with the Fourth and Ninth Circuits. In Costco, the Ninth Circuit explicitly cited Garland, and Areeda and Hovenkamp, for its determination that “there is such substantial overlap between the active supervision and hybrid inquiries that they effectively merge.” The “active supervision” requirement, of course, comes from the Midcal state action test, while the “hybrid” inquiry comes from the preemption framework set out in Rice and Fisher. The Fourth Circuit, too, has adopted Areeda and Hovenkamp’s position in the TFWS case, favorably citing Areeda and Hovenkamp’s approval of Judge Winter’s dissent in Battipaglia, discussed above. These courts’ adoption of the outcome-driven effort to combine the state action and preemption frameworks, we believe, is central to understanding the circuit split between the Second, and Fourth and Ninth, circuits.
The Fourth Circuit, in TFWS, Inc., analyzed a similar facial challenge to a post and hold law, but did so instead under the Garland, and Areeda and Hovenkamp framework. There, as in Connecticut Fine Wine & Spirits, LLC., the court considered whether “the State’s [post and hold] pricing scheme restrains competition by allowing wholesalers to do two things: (1) match each other’s prices at artificially high levels and (2) maintain those high prices.” Unlike Connecticut Fine Wine & Spirits, LLC., however, the Fourth Circuit began its analysis by asserting that “[t]he framework for considering whether a state’s liquor pricing regulations can be successfully challenged under the Sherman Act was established by the Supreme Court in [Midcal].” The court then proceeded to merge the state action and preemption frameworks, finding that “the post-and-hold system is a classic hybrid restraint” because “the State requires wholesalers to set prices and stick to them, but it does not review those privately set prices for reasonableness.” Recall, however, that Rice and Fisher dictate that the “hybrid” preemption analysis is a determination of whether firms engaged in concerted action, while Midcal’s state action inquiry instructs that a statute is subject to state action immunity if the state “actively supervises” its regulatory regime—the TFWS court combined the two tests. After finding that the Maryland post and hold requirement was hybrid, given the lack of state “active supervision,” the Fourth Circuit next concluded that the Maryland post and hold regulation was in fact a per se violation of Section 1 by looking at the anticompetitive effect of the statute, rather than the actual conduct the firms engaged in, writing that “[i]f liquor wholesalers entered into private agreements to accomplish what is required (and allowed) under the Maryland scheme, a per se Sherman Act violation would result.” The court wrote that the statutory scheme was, therefore, “essentially a form of horizontal price fixing.”
The Ninth Circuit adopted a similarly flawed approach. Miller v. Hedlund addressed a facial challenge to Oregon’s liquor laws, which the plaintiffs argued had “the effect of stabilizing and maintaining the prices of beer and wine in Oregon in violation of the Sherman Act,” including similar post and hold, and price non-discrimination provisions. Like the Fourth Circuit, the Ninth Circuit opted to extend its analysis beyond the determination of whether the conduct compelled by Oregon’s law implicated an agreement among firms, simultaneously admitting that “it is true that there is no agreement or concerted activity among the wholesalers,” but concluding that “[s] imply ending the analysis because of the lack of concerted activity among the wholesalers fails to take into account the presence and effect of the state’s involvement in the matter.” In fact, citing Midcal, the court wrote that “a showing of concerted activity among the Oregon wholesalers is not necessary to establish an antitrust violation.” There is no way to square that statement with Fisher’s instruction, just one year before Miller, that “[e]ven where a single firm’s restraints directly affect prices and have the same economic effect as concerted action might have, there can be no liability under § 1 in the absence of agreement.”
It is not surprising, then, that when the Ninth Circuit revisited post and hold regulations in 2008, in the context of a facial challenge to Washington’s liquor laws in Costco Wholesale Corp. v. Maleng, the court again employed the same outcome-based analysis. The court based its reasoning on a purported “uncertain relationship between the ‘active supervision’ inquiry under Midcal . . . and the ‘hybrid/unilateral’ inquiry under Fisher.” But that distinction between state action immunity and preemption, in fact, is quite clear, for the reasons discussed above. In any event, the court went on to cite Garland, Areeda and Hovenkamp to argue that “there is such substantial overlap between the active supervision and hybrid inquiries that they effectively merge.” Adopting this view of the case law allowed the court to again import the Midcal active supervision analysis into the determination of whether a firm acts unilaterally: “The rule to be taken from these cases is that state statutes or local ordinances creating unsupervised private power in derogation of competition are subject to preemption.” Perversely, then, the Ninth Circuit’s determination of what constituted unilateral or hybrid conduct did not consider the firms’ actual conduct; instead it looked at the state’s role in policing its regulations. The Ninth Circuit expressly adopted Judge Winter’s dissent in Battipaglia, and Areeda and Hovenkamp’s focus on outcomes: “A leading treatise on antitrust law suggests that ‘[g]iven the great danger that agreements to post and adhere will facilitate horizontal collusion, the [Battipaglia] dissent’s position is more consistent with Midcal.’” The court ultimately concluded that the Washington post and hold statute was subject to preemption, because “[a]lthough each wholesaler is only required to adhere to its own posted price and is not compelled to follow others’ pricing decisions, the logical result of the restraints is a less uncertain market, a market more conducive to collusive and stabilized pricing, and hence a less competitive market.” That focus on outcomes, we believe, wrongly ignores Rice’s instruction that a “state statute is not preempted by the federal antitrust laws simply because the state scheme may have an anticompetitive effect.”
In our view, the Second Circuit analysis in Connecticut Fine Wine & Spirits, LLC. faithfully adheres to the Supreme Court’s precedent and the text of the Sherman Act. The preemption analysis must necessarily begin with the text of Section 1, which prohibits agreements in restraint of trade. If concerted “hybrid” action can be shown, further consideration may be given to whether the conduct at issue is a per se violation. Whether or not the state actively supervises its regulations is simply inapposite to the preemption question. Of course, statutes that are prone to preemption may be saved by the state action doctrine, but that is a distinct analysis. Yet the Fourth and Ninth Circuits have erroneously avoided the Sherman Act’s limited prohibition of only agreements in restraint of trade—and consequently, have in effect expanded their ability to strike down state laws specifically designed to limit competition—by collapsing the state action immunity and preemption frameworks into a single lens. Those courts have erroneously adopted the point of view of a small number of scholars that the determination of what constitutes unilateral or hybrid conduct should look not to whether or not firms have engaged in concerted action, but instead, to whether or not the state actively supervises that conduct.
Jettisoning the agreement requirement from Section 1, as the Fourth and Ninth Circuit have done, risks significantly expanding the scope of antitrust liability. Although Garland derided “Fisher’s requirement of an ‘agreement’” as “impos[ing] yet another barrier to an antitrust plaintiff’s success,” that is precisely the crucial function the agreement requirement serves. Indeed, without Section 1’s contract, combination, or conspiracy language, Section 1 would be enormously expanded as follows: “Every restraint of trade or commerce among the several States, or with foreign nations, is declared to be illegal.” Such a low bar for finding antitrust liability would, as a consequence, risk penalizing a whole range of private conduct heretofore beyond the reach of Section 1. Rent control statutes—which were saved by the agreement requirement in Fisher—would be struck down in cities across the country; unilateral corporate decisions to reduce output or raise prices above the competitive level might become per se violations; and single-firm monopolization conduct that would formerly pass Section 2 muster may now be subject to additional review under Section 1. Failing to adhere to the agreement requirement in Section 1 cases has consequences that extend well beyond the preemption context.