Introduction
What could be more flagrantly anticompetitive than bribing or threatening a business to deter it from becoming, or from dealing with, a competitor? Surely, one might think, any self-respecting antitrust system would come down very hard on a monopolist that tried anything of the kind.
But the practice is rife. Enforcers have sued digital platforms for offering an array of valuable benefits—from search preferencing to interoperability—to encourage their trading partners to steer clear of rivals and rivalry.
Agritech giants pay distributors to reject cheaper crop protection chemicals that would reduce farmers’ costs.
And pharmaceutical monopolists wield enormous rebates to induce customers to refuse cheaper generic alternatives that patients would value.
And when enforcers do challenge such practices—typically under the prohibition on “monopolization” in Section 2 of the Sherman Act
—courts often seem barely interested. For example, the D.C. Circuit has effectively shrugged at an allegation that a dominant social network leveraged valuable interoperability to deter other apps from developing competing functions,
while the Ninth Circuit showed little interest in allegations that a processor-chip monopolist used a patent license as a vehicle to surcharge customers’ dealings with rivals.
So what’s going on? What’s the point, one might ask, of having a monopolization statute if it doesn’t catch this kind of thing?
It turns out that antitrust does a staggeringly bad job at handling practices of this kind. These are all examples of conditional dealing: a monopolist treating other market participants more favorably when they refrain from (or limit) competition against the monopolist, or refrain from (or limit) dealing with its rivals. In a paradigmatic conditional dealing case, there is no actual agreement or commitment that the counterparty won’t compete or won’t deal with rivals. Instead, there is just an explicit or implied policy, unilaterally applied by the monopolist, that punishes competition and rewards “loyalty.” The inducement may be naked (e.g., cash payments or penalties) or it may involve differentiated terms of trade with the monopolist (e.g., granting or withholding access to a product or service, or offering better or worse prices, to encourage “loyalty”).
Conditional dealing falls into a troubling gap in antitrust doctrine. On the one hand, antitrust has fairly clear rules for agreements involving monopolists, including deals with rivals to avoid competition (“market allocation” agreements),
and deals requiring trading partners to cut off rivals (“exclusivity” agreements).
These rules provide for fairly close scrutiny. Agreements of the first kind are usually per se illegal;
agreements of the second kind are analyzed to determine whether rivals are being harmfully and unjustifiably foreclosed.
On the other hand, antitrust also has fairly clear rules for unilateral choices about pricing and supply. These rules, by contrast, are highly permissive, amounting to virtual immunity. Thus, above-cost pricing is usually per se legal, even if customers complain that a monopolist’s prices are too high or competitors complain they are too low.
And businesses can generally refuse to deal with their rivals at will.
In theory, it’s possible for a refusal to deal to violate the antitrust laws,
but the eye of that needle is so slender that no plaintiff has squeezed through it in decades.
Conditional dealing falls right between these two categories. It seems to present all the dangers of classically harmful agreements, regardless of whether an actual agreement exists, and regardless of how the threats and bribes are labeled, paid, or extracted. But it also seems to implicate all the liberty concerns that attend unilateral pricing and supply choices. After all, if there is no general antitrust duty to deal with the world, it seems to follow that a monopolist can choose to sell only to noncompetitors, or to sell to them on more favorable terms. That position even has some everyday intuitive appeal: Why should a business have to sell to its own rivals, or to businesses that choose to partner with its rivals?
Conditioning ruthlessly exposes a deep problem with the monopolization offense: its discomfort with practices that do not fall into its clean, shoebox-like categories (like exclusivity or tying) and that force courts to rely on monopolization’s elusive and uncertain first principles.
It’s all very well to say, as courts often do, that monopolization law asks whether conduct is “anticompetitive” or “predatory” rather than “competition on the merits,” but that kind of sloganeering is virtually no help in the real world.
So courts—and some commentators—tend to try to jam conditioning into an existing shoebox, often one that is subject to heavily pro-defendant rules. For example, when a coalition of states alleged that the Facebook personal social network dangled valuable interoperability to deter app developers from developing competing functions, the D.C. Circuit analyzed that practice as a simple refusal to deal: a practice that, as noted above, is virtually per se legal.
And when the FTC alleged that Qualcomm, a leading chip supplier, used patent licenses as a vehicle to tax customers’ dealings with its rivals, the Ninth Circuit analyzed the claim as a complaint primarily about excessive royalty rates, and automatic legality followed.
In still other cases, courts have been persuaded to apply an array of tests—“coercion,” below-cost pricing, the “predominance” of price, duration and terminability, and so on
—that have little or nothing to do with the underlying dangers, notwithstanding the rich economic literature protesting that these considerations are beside the point.
This economic literature leaves no doubt that conditioning can enable a monopolist to do just what the Sherman Act abhors: inflict welfare harms by excluding rivals in ways that contribute to monopoly power and are not justified by offsetting benefits.
But antitrust’s ability to respond to harmful conditioning is being hobbled by the structure of monopolization doctrine: specifically, its heavy reliance on analytical categories that were designed to respond to other, rather different, practices.
In other words, this is a problem that the law has created for itself. Economists seem perfectly clear-eyed about the effects and dangers of conditional dealing.
Monopolization’s failure to reckon with conditioning is holding back efforts to deal with some of the most pressing concerns on the antitrust agenda. This includes, for example, concerns about platform monopolists proffering a benefit (like interoperability or better search rankings) to induce their trading partners to disfavor rivals;
agricultural monopolists using conditional discounts and “exit penalties” to prevent rivals from getting a foothold;
and pharmaceutical monopolists using rebates to keep lower-cost competitors down or out in markets for life-saving treatment.
In these and other areas, monopoly conditioning may threaten worse harms than just higher prices.
* * *
This Article argues that we should meet conditional dealing on its own terms by recognizing a new category of monopolizing conduct. Anticompetitive conditioning or conditional dealing is the application by a monopolist of conditions that punish others for competing with it (horizontal conditioning) or for trading with its rivals (vertical conditioning).
This Article unfolds in three parts. Part I presents the problem: conditional dealing by monopolists. It surveys the uses and dangers of this practice in a selection of critical tech, agriculture, and healthcare markets, then synthesizes the rich body of economic scholarship on conditioning and its effects.
Part II sets out the Article’s primary contribution: a new analytical framework for courts and others analyzing conditioning under Section 2 of the Sherman Act. This includes a test for identifying conditional dealing (the “hold-constant” test) and a doctrinal framework for assessing its legality. This involves assessments of whether a condition has an exclusionary incidence on rivals (i.e., whether a horizontal condition significantly impairs the incentives of one or more rivals to meet demand, or whether a vertical condition significantly impairs the ability of one or more rivals to do so by substantially foreclosing their access to inputs, distribution, customers, or complements); whether the exclusion is reasonably capable of contributing significantly to monopoly power; and whether the practice is justified by offsetting welfare benefits.
This Part also explains why many factors often emphasized by courts and others—from price-cost measures and coercion tests to doctrines of de facto exclusivity—should have no place in this analysis. And, using conditional dealing as a vehicle to explore some broader questions of principle, this Part proposes some more general course corrections for antitrust: a modest regrounding of the concept of substantial foreclosure; a clarification that free riding in an antitrust case is, without more, a neutral fact, not a trump card for a defendant; and the long-overdue recognition that an unconditional refusal to deal is per se lawful, notwithstanding the agonizing refusal of courts to say this out loud. Part II also points out some landmarks in antitrust’s precedential canon that are best understood as conditioning cases.
Part III sketches two further ideas to reinforce monopolization’s frontier. The first idea is what might be called quick look monopolization. It draws on a doctrine developed under Section 1 of the Sherman Act that, in clear cases, a plaintiff may establish a prima facie case by reference to the basic nature and context of the agreement without having to piece together evidence of actual effects or impacts.
This approach has never been applied in monopolization law, but it should be, because its logic applies equally in that setting. It provides a principled way to sharpen monopolization doctrine in a small subset of clear cases, including conditioning cases lacking plausible justifications.
The second idea is a reinterpretation of the offense of attempted monopoly maintenance. Conventional accounts present the attempt offense as a sort of mini-monopolization: that is, a ban on conduct by a near-monopolist that has provably resulted in actual exclusion and actual contribution to power. But the offense also bears a second, neglected reading: a prohibition of conduct by an actual monopolist that is intended to cause welfare harms by suppressing rival ability and incentive to compete and which is dangerously likely to have led to that outcome—regardless of whether it really did have that effect. It provides a principled way to deter intentional misconduct, even in our most complex and dynamic markets.
Ultimately, this Article’s central claim is a simple and intuitive one. A monopolist’s use of an explicit or implicit condition to punish trading partners for competing, or for dealing with competitors, is a distinct form of antitrust wrongdoing, not an edge-case example of a more familiar practice like tying or predatory pricing. Just like other familiar kinds of violations, conditioning presents clear, well-understood risks of consumer harm, and it can be scrutinized by courts without unreasonable intrusion on the freedom of businesses to run their affairs. When a monopolist uses such a practice to exclude rivals and augment its monopoly, courts should demand evidence of justification—and should impose liability if it is not forthcoming. Such claims should not be shrugged off for failure to fit neatly into a handful of doctrinal boxes that were crafted with very different practices in mind.
It is time to close antitrust’s conditioning loophole.