- The Legal Workshop - http://legalworkshop.org -

Is Pepsi Really a Substitute for Coke? Market Definition in Antitrust and IP

Posted By Mark A. Lemley On December 21, 2012 @ 1:01 am In Antitrust/Securities/Trade Regulation, Georgetown Law Journal, Intellectual Property | No Comments

Does Pepsi compete with Coke? It seems a straightforward question; perhaps you think it has a straightforward answer. Sure they compete, you might say; they are both colas, they are used for the same purpose, they are sold next to each other in the grocery store, and they cost about the same. In fact, however, the answer is far from clear. Whether you think Coke competes with Pepsi may depend on whether you drink one of them. If you do, ask yourself this question: how big a price increase would it take before you would switch from one to the other? If you are like a significant number of consumers, it would take a pretty dramatic change in price to get you to give up your preferred brand.

For antitrust, this insensitivity to price has a simple—and shocking—implication: Coke and Pepsi don’t compete in the same market. Antitrust defines markets by asking, in part, whether a small but significant nontransitory increase in price (or SSNIP) above marginal cost would cause customers to switch from one good to another. And by small, the regulators mean something like 5%. So if a bottle of Coke costs $1.60, unless a price increase of eight cents would send so many consumers running to buy Pepsi that Coke would lose money, the two don’t compete. And while supply substitution can also constrain market power, that won’t work here either; Pepsi-Cola can’t (legally) start making Coke rather than Pepsi just because the price went up. By our classic antitrust definition, then, Coke and Pepsi are not in the same market. And assuming Coke’s price change doesn’t send customers off to any other drink either, then Coke is in a market by itself, one in which it has market power.

Nor is this an isolated result. In market after market, intellectual property (IP)-based product differentiation and brand loyalty mean that small price changes have little or no effect on purchasing behavior. To antitrust law, all these products have market power in their own, individualized markets.

If we took market definition seriously, antitrust law would look very different as applied to IP. Monopolists are subject to a variety of constraints that ordinary competitors don’t face. To take just one example, exclusive dealing arrangements are subject to the rule of reason: they are unlawful if they foreclose a substantial share of the market. If any well-known brand has power in its own, individually defined market, exclusive deals or licenses signed by those brands would probably be illegal. That result would require a major change in business practice.
A variety of IP doctrines also require courts to define markets, though they often do so in a more ad hoc way than does antitrust. Here too, actual practice seems at odds with what antitrust’s approach to market definition requires. If generic drugs don’t compete with their branded counterparts, for example, then it makes little sense to award damages in patent cases on the assumption that sales of the infringing product cause the patentee to lose sales. In short, a world that recognized the market power apparently conferred by IP rights would impose many more restrictions on IP owners than our cases do today, as a matter both of antitrust doctrine and IP law. And, we would be a lot more worried about the anticompetitive effects of IP rights generally. The IP–antitrust relationship would look a lot more like a persistent conflict, and antitrust would limit IP rights more than it currently does.

We think that antitrust law needs to recognize that the point of IP rights generally is to create some form of differentiation, and therefore, some (usually constrained) power over price. Treating any product that does so as a monopoly can be counterproductive, both because it will put antitrust law into regular conflict with IP policy and because it may discourage interbrand competition more than it promotes intrabrand competition. At the same time, both IP and antitrust law are wrong to ignore the implications of product differentiation and the power it gives IP owners over customers who strongly prefer one brand or type of product. The law needs to acknowledge that power in setting IP policies, in awarding remedies, and in evaluating the competitive effects of IP licensing. And both IP and antitrust law need to focus less on artificial market boundaries and more on the actual effects of IP rules and IP owner conduct.

I. Market Definition In IP Doctrine

A wide variety of IP doctrines require courts to engage, explicitly or implicitly, in market definition, or at least something that closely resembles it. Yet despite the widespread reliance on market definition in IP, courts have developed no apparent methodology for defining markets. Indeed, courts frequently don’t even acknowledge that they are drawing difficult lines between markets.

A.     Trademark

The most obvious context in which market definition plays a role in trademark law is in functionality. The Supreme Court appeared to downplay competitive need as determinative of functionality in TrafFix Devices, Inc. v. Marketing Displays, Inc,1 instead emphasizing the functionality doctrine’s role in channeling protection of useful product features to the patent system and therefore put much greater emphasis on expired utility patents as evidence of functionality. But TrafFix has not rendered competitive need irrelevant, and market definition therefore remains important.

First, some courts insisted that the availability of alternative designs remains relevant to the question of whether a feature is “essential to the use or purpose of the article,” because they maintain that functionality is fundamentally about competitive need.2 Competitive also remains an important part of functionality because the Supreme Court made clear that, in cases of aesthetic functionality, “[i]t is proper to inquire into a significant non-reputation-related disadvantage.”3 Since competitive disadvantage can only be determined in the context of markets, market definition remains critical in aesthetic functionality cases.

Sometimes courts explicitly talk about the market as part of the functionality analysis, though almost never by reference to any methodology. And sometimes the markets those courts define seem intuitively right to us. Our point is simply that, to the extent these decisions are based on conclusions about competitive need for a feature, courts must engage in market definition. The more broadly a court defines the relevant market, the less any particular design appears necessary; the more narrowly it describes the market, the more granting exclusive rights to particular design features seems to impose competitive harm. And courts appear to define markets in these cases entirely by their own intuition.

B.     Patent

Market definition plays a significant role in patent law in the context of damages. Patent law awards damages “adequate to compensate for the infringement.”4 The traditional measure of those damages is the patentee’s lost profits: the amount of money the patentee would have made selling the invention had the defendant not infringed. To determine that, we need to know which other competitors in the market might take the sale away from the patentee, what noninfringing alternatives might serve as substitutes for the patented good, how the presence of infringement affected the price the patentee could charge, and how an increase in price absent that infringement would have affected total demand. In some circumstances, we may also need to know how the patented product interacts with other, unpatented complements, and even how it interacts with unpatented non-complementary products sold by the patentee. The analysis of whether a patentee lost profits because of an infringement is accordingly quite sophisticated, and requires the estimation of supply and demand curves and cross-elasticities of demand. These are the classic hallmarks of antitrust market definition and power analysis.

An important category of patentees who are not entitled to lost profits are patentees who don’t sell their own products in the market for the patented goods. These patentees, like those who cannot prove lost profits, can fall back instead on a reasonable royalty for the licensing of their patent. Until recently, the reasonable-royalty analysis was comparatively less sophisticated than the lost-profits analysis. But developments in the last few years have required more careful attention to the question of what a willing buyer would have agreed to pay to license the patent from a willing licensor. Further, reasonable royalties, like lost profits, require consideration of available non-infringing alternatives, which means that courts must identify other technologies that are sufficiently different from the patent to avoid infringement but still close enough that they serve as functional substitutes for the patented product in the minds of actual consumers of that product.

C.     Copyright

1.   Distinguishing Ideas from Expression

It is black-letter copyright law that protection attaches only to expression and not to the ideas embodied therein. Courts must therefore distinguish ideas from expression both to determine what the plaintiff in a copyright case actually owns and to determine whether a particular defendant’s work infringes.

To take a modern example, when Mattel sued the makers of the Bratz line of dolls, claiming the dolls too closely resembled those depicted in preliminary sketches to which Mattel owned rights, the Ninth Circuit held that the district court erred because it failed to filter out the unprotected idea of “fashion dolls with a bratty look or attitude, or dolls sporting trendy clothing.”5 For the Ninth Circuit, without those elements filtered out, Mattel would gain monopoly control over a certain type of doll.

For our purposes, the important question is why the dolls were appropriately described at that level of specificity. Implicit in Judge Kozinski’s opinion is the notion that exclusive rights over “fashion dolls with a bratty look or attitude, or dolls sporting trendy clothing,” would have given Mattel power over a significant market segment—hence the concern about monopoly. But how does the Ninth Circuit know that exclusive rights to “fashion dolls with a bratty look or attitude” would impose significant competitive costs? The conclusion depends on an implicit judgment about what sorts of things compete in a relevant market. If non-bratty dolls were a good substitute for bratty dolls, there would be no reason to worry about giving Mattel protection over the latter, at least from a competitive standpoint. Put in copyright terms, we might define the idea at a higher level of abstraction, perhaps as “dolls.”

In fact, Mattel is fairly representative in this respect: the idea-expression dichotomy relies on an implicit market definition, and application of that rule tends to imply fairly narrow markets.

2.   Fair use and market harm

While the Supreme Court has insisted that consideration of the four statutory fair use factors6 is highly contextual and that no one factor is dispositive, empirical work suggests that the fourth factor—the effect on the market for the copyrighted work—is effectively determinative in most cases. Here, in contrast to the idea–expression dichotomy, courts tend to conceive of the market for a copyrighted work very broadly—so that not only do courts consider the extent of market harm caused by the particular actions of the alleged infringer, but also whether unrestricted and widespread conduct of the sort engaged in by the defendant would result in a substantially adverse impact on the potential market for the original. Moreover, according to the Supreme Court, the enquiry must take account not only of harm to the original but also of harm to the market for derivative works.

The inclusion of what are clearly separate derivative markets might seem to suggest that market definition doesn’t really matter in fair use analysis. In fact, however, how one defines the relevant derivative market in these cases has a significant effect on the outcome. If one believed, for example, that a particular rap song competed in the market with all other rap songs, and that each rap song was more or less interchangeable with the others, then a defendant’s use of a particular song might not seem to cause much market harm at all, except in the circularly-defined market for licensing. If the defendant were required to pay to use the plaintiff’s particular song, there’s no compelling reason to believe it would still choose that song (even assuming that the plaintiff would willingly license); plenty of other equally good alternatives are available, potentially at lower costs. If, however, the particular rap song is not readily substitutable, then, assuming all other factors are equal, uses by third parties are likely to cause greater harm, because uncompensated uses are more likely to take the place of uses for which the plaintiff would have been paid.

D.     The Problem with Implicit Market Definition

Lack of clarity about market definition in IP cases is to some extent a consequence of inconsistency in defining the terms of the debate. Courts and scholars alike sometimes talk about relevant markets in purely functional terms, so that markets are defined by the functional characteristics of the products at issue. Other times, however, markets are defined by consumer demand, which may or may not track functional considerations.

This framing choice, however, tends not to be random, but instead generally tracks normative views about the scope of the rights in question. When making the case for strong rights, advocates emphasize the need for incentives and the important role IP rights play in enabling creators to appropriate economic value. This is reflected in the damages phase of litigation, where the notion that there are plenty of available substitutes for the product covered by the applicable IP right—so that access to the particular feature is not particularly valuable—tends to disappear. In short, courts setting remedies define markets narrowly in order to emphasize the harm the plaintiff claims to have suffered.

Where the issue is concern about the scope of IP rights and their general effects on competitors or on the ability of others to speak, by contrast, parties and advocates of broad IP rights define markets broadly so as to downplay the consequences of IP rights. In copyright, for example, concerns about the costs of protection are brushed aside on the ground that copyright doesn’t create any market power since it allows others to create competing, functionally equivalent works. And in patent law, we have seen a consensus develop that patents do not confer market power—a view that depends on the belief that patented inventions compete with other products in more or less competitive markets.

Unlike IP law, antitrust law has a long-standing, well-developed methodology for defining markets. Accordingly, we turn in Part II to consider antitrust market definition and whether it offers a way to rationalize the conflicting understandings of market scope and market power in IP cases.

II. Market Definition and Product Differentiation

A.     Market Definition in Antitrust Law

Antitrust is about market relationships. It is designed to promote competition. Competition doesn’t occur in a vacuum; a company must compete with others in some market. As a result, the first step in virtually any antitrust case is the definition of the market in which the competitive harm is alleged.

Antitrust determines market power circumstantially by defining a relevant market and then computing the defendant’s share of this market. A relevant market is a collection of goods or services that can profitably be sold at a monopoly price. For that to be true, the goods in the collection must be effective substitutes for one another, and goods outside the group must not be effective substitutes for goods inside the group. A substitute is effective if it is desirable enough to customers for a good that a sufficient number of those customers will switch to the substitute if the good’s price rises somewhat above cost.

Antitrust market definition also considers supply substitution, which occurs when producers of a somewhat different product are able to shift their production into the price increaser’s market. Assume for a moment that cars and minivans could be manufactured using the same technology and distribution systems, and that a maker of cars can quickly and cheaply switch from the production of cars to the production of minivans. In this case, even if customers are unwilling to substitute away from high-priced minivans to cars, producers of cars will constrain minivan pricing by entering the minivan market, making more minivans and fewer cars.

Barriers to entry are a third consideration in defining an antitrust market. A barrier to entry is something that deters or delays the entry of new rivals even when the firm or firms in the market are already charging supracompetitive prices. If entry is so easy that new entrants can flood a market any time incumbents raise prices to supracompetitive levels, the products those incumbents sell cannot be effectively monopolized or cartelized, and it doesn’t make sense to talk about those products as a separate market. By contrast, if there is some structural barrier to entry, such as a high fixed cost of initial investment in a plant, a strong patent right, or a government licensing process that controls entry, supply substitution from new firms or existing firms making other products is less likely.

A relevant antitrust market is a grouping of sales for which both the cross-elasticity of demand with other products and the cross-elasticity of supply with other products are sufficiently low to warrant the conclusion that a significant price increase of that grouping to a supracompetitive level would be profitable and relatively durable. This typically means not merely that the increase will be immediately profitable, but that the monopoly price increase will be profitable for a significant length of time—say, two years.

Finally, the size of an antitrust market necessitates a policy judgment about how large a price increase above the competitive level should be tolerated. If we regard a price even one percent above the competitive level as intolerable, we would respond with a set of rules that defined markets very narrowly, because a one-percent price increase might not prompt much supply or demand substitutability. In a wide variety of product-differentiated markets, individual firms can profitably charge more than one percent above their marginal costs, and under such a rule we would conclude that each such firm is a “monopolist.” By contrast, if we regard prices fifty percent above cost as tolerable, the methodology produces a significantly broader market definition. Antitrust market definition generally includes in the same market firms that, while not perfect competitors, are able to hold one another’s prices to within five or ten percent above the competitive price. The federal merger guidelines refer to this test as a “small but significant nontransitory increase in price,” or SSNIP.7

B.     The Role of IP Rights in Defining Markets

It is an article of faith in antitrust law that particular companies do not define a relevant market because for most users different brands compete with one another. For example, antitrust might suppose customers are almost indifferent between Chiquita and Del Monte bananas, because most believe that there are no significant differences in quality, taste or other attributes. Antitrust is likely to define a market as “bananas,” not “Chiquita bananas.” But even when products clearly vary in quality—cars, say, or laptop computers—the basic assumption in antitrust law is that the products produced by individual companies compete in a larger market of (mostly) like goods. To an antitrust lawyer, brands aren’t markets.

Antitrust market definition hearkens back to the days of readily interchangeable commodities, like grain or plywood. Products and brands today, however, often are far more significantly differentiated. For while trademarks may telegraph information about product quality or seller reputation, they often do much more. Neuroscience research shows that brands convey emotional content as well as information about product characteristics, and indeed that people react to their favorite brands in ways that mirror their reaction to religious icons. When preferences created by that information or those attachments are substantial and rivals cannot readily attain the same status, then it is simply wrong to say that the brand does not constitute its own relevant market. People in the market for one kind of car may not be satisfied with another, and price variations reflect that fact.

IP rights contribute significantly to that product differentiation, and hence make it more likely that two products that we think of in the same broad class are not in fact effective substitutes for each other from an antitrust perspective. By definition, IP rights attach to things that are differentiated from their putative competitors. Further, the existence of IP rights means that the things that differentiate my product from yours cannot be copied. Put another way, an IP right is a barrier to entry that prevents the sort of quick and easy supply substitution that might undermine market power.

This IP-induced product differentiation can sometimes have the effect of excluding other goods in the same functional category from an antitrust-defined market, despite our intuition that the goods must compete. Tylenol and Advil, for instance, are brand names for acetaminophen and ibuprofen, respectively. But neither is the subject of a current patent. So anyone is free to copy the chemical formula of those drugs exactly and to produce an acetaminophen or ibuprofen product that competes with Tylenol or Advil. And they do. Curiously, however, those generic alternatives sell at a significant discount compared to the price of brand-name Tylenol or Advil, often as much as 50% less. Even though generic companies can and do advertise that they have the same ingredients as Tylenol or Advil, and even though they are often placed next to each other on pharmacy shelves, the brand name is sufficient to cause many consumers to pay a higher price for the same basic good.

These brand-driven price differences have an interesting antitrust consequence: just as Advil and aspirin might not compete with each other in a relevant antitrust market, Advil and generic ibuprofen might not compete with each other on traditional antitrust analysis. Indeed, a number of courts have held that brand name drugs and their generic equivalents are in separate markets for antitrust purposes.

The implications aren’t limited to pharmaceuticals. IP rights can create their own antitrust market in any product in which consumer loyalty to the brand is strong enough, or in which IP rights make the products different enough. We see brand-based price differentiation among goods of undifferentiated quality in a wide variety of consumer markets. Some brands of gasoline, such as Shell, systematically charge more than their competitors, even when those competitors are across the street. Others, like Arco, charge less. Consumers willingly pay substantially more for Clorox-brand bleach than for generic bleach, even though the essence of bleach is a simple chemical formula. And sports fans won’t be willing to replace a hat with their team’s logo with a hat from a competing sports franchise.

Brands clearly create significant differentiation in the minds of consumers, whether or not that belief is based on functional differences in the goods being sold. Survey evidence suggests that across all products, a significant number of consumers would be willing to pay a substantial premium for their brand of choice. And the self-confessed goal of marketing is precisely to build such insensitivity to price. More famous brands inspire still more loyalty. Cola drinkers are notoriously loyal to their brands, notwithstanding the fact that they seem unable to distinguish among different colas in blind taste tests. The evidence doesn’t support the conclusion that all brands confer market power. But it seems quite likely that for many of the most famous brands, and for a variety of goods in which price is a mixed or even affirmatively positive signal of value, trademark owners can easily raise the price of their goods 5–10% over the cost of an alternative product without their customers deserting them for that product. For those brands, other products of similar type are not in the same market as antitrust defines that term.

A wide variety of copyrighted works also command significant willingness to pay. If you want to read the final book in the Harry Potter series, you are unlikely to be satisfied by the latest Stephen King novel. And indeed people line up in advance to get access to the newest Harry Potter novel, iPhone, or movie, while ignoring a variety of other creative products of the same type available with no wait and frequently at the same or a similar price. It may be hard to quantify the power over price in a case like that, but it’s surely there.

Many product patents do little more than create relatively minor enhancements in a product that make it distinctive to one group of customers, and competitors in that product are likely to have their own offsetting patented enhancements. As a result, the markets for automobiles, vacuum cleaners, cleansers, and pharmaceuticals are characterized by numerous patents, most of which suffice to make their products somewhat distinctive in a product differentiated market. For antitrust law, the question is how distinctive those products are. If a small price increase won’t dislodge purchasers who want a hybrid gasoline–electric engine in their car, hybrid cars aren’t even in the same market as non-hybrid cars with the same brand name.

In short, the evidence does not support the conclusion that IP rights automatically confer market power. But it does suggest that a surprisingly large number of IP rights do in fact give their owners power over price. Under traditional antitrust principles, those products are not in the same market as the goods that would seem to be their closest competitors.

C.     Implications of Market Definition in IP Cases

It is difficult to overstate just how shocking this conclusion ought to be to antitrust lawyers. Consider the implications of concluding that Coke (or Harry Potter books, or Shell gasoline, or Clorox bleach, or brand-name pharmaceuticals) really is in an antitrust market by itself. Antitrust law draws sharp distinctions between unilateral conduct, agreements with competitors, and agreements with noncompetitors. Agreements between horizontal competitors to create cartels that restrict price or competition get the harshest condemnation; they are illegal per se. Agreements between parties in a buyer–seller relationship (vertical agreements) are treated much more leniently. And agreements between companies that are not in a competitive relationship at all generally get no antitrust scrutiny at all. Unilateral conduct is not condemned at all unless the actor is a monopolist. If it is, a number of otherwise legal acts become antitrust problems if they have the purpose or effect of acquiring or maintaining that market power.

A world in which individual books and individual brands define their own markets and hence confer market power turns that world upside down. To oversimplify, if IP rights define their own markets, everyone is monopolizing and no one is cartelizing. On the one hand, a number of accepted business practices become illegal if the firm that engages in them has market power. For instance, exclusive dealing arrangements, in which a party agrees to sell to or buy from only one party, are subject to antitrust’s rule of reason. They are illegal if entered into by firms with market power and if they foreclose a “substantial share” of the market. But if the market is defined not as “novels” or even “fantasy novels” but “Harry Potter books,” J.K. Rowling entered into an illegal exclusive dealing arrangement when she allowed only Scholastic to produce and sell her books. Grants of exclusive franchise territories by McDonald’s look like a market allocation scheme if we don’t think Burger King competes in the same market. And so on. Antitrust spent the last thirty-five years undoing its rules restricting vertical restraints, in large part on the theory that it was worth sacrificing “intrabrand competition” among dealers or franchisees in the service of promoting more “interbrand competition.” But if interbrand competition doesn’t work—if the brands aren’t really competing in the same market at all—the last forty years of antitrust history have been a colossal mistake.

Concluding that many IP rights define markets turns a wide array of companies into monopolists, and so in many ways it would dramatically expand the role of antitrust law. But in other respects it would shrink it. Suppose Pepsi and Coke were to enter into a horizontal market division agreement, in which Pepsi agreed to sell only north of the Mason–Dixon Line, and Coke agreed to sell only south of that line. If Pepsi and Coke compete in the same product market, that agreement would be illegal per se, just as it would be if they both agreed on the price they would charge. But if they aren’t competitors, traditional antitrust analysis doesn’t have much to say about agreements they enter into, any more than it would object to Coke’s agreeing with the two of us on the price it might charge. Similarly, if (as some courts have concluded) brand name drugs are not in the same market as their generic equivalents because the branded drugs command different prices than the generics, an agreement to fix the prices of the branded and generic drugs, or to pay the generic drug company to leave the market, would not seem to be an antitrust problem. And companies that make one of only two drugs for a particular disease could acquire the other, avoiding competition concerns by arguing that the drugs are differentiated and are prescribed by different doctors in different circumstances.

In short, if we live in a world in which IP rights regularly confer market power, antitrust law has been doing pretty much everything wrong—finding illegal practices we shouldn’t care about, and ignoring things that should be quite troubling.

III. Lessons for IP

In our view, the difficulty courts have experienced with market definition in IP cases stems at least in part from their (perhaps intuitive) resistance to relying on the antitrust understanding of markets. The antitrust approach derives from a classical conception of competition involving undifferentiated goods where producers compete solely on price and quality. But markets for goods encumbered by IP rights are different: these goods are, by definition, differentiated, either in terms of some product feature or because of the product’s branding.

This is not to say, however, that all IP rights create market power that antitrust should care about, or even that all IP rights have the same effect on substitution. Indeed, we suspect that the effects of an IP right are likely to be quite different in a variety of respects. The general point is that the competitive costs of IP rights vary with the extent to which they enable the owners of those rights to exclude close substitutes, and courts should therefore conceive of competition in the IP space in terms of the quality of available substitutes. Rather than considering whether a particular product is or is not in the relevant market, courts must at a minimum think of competition as existing along a spectrum. The further away one product is from another on that spectrum (that is, the more imperfect it is as a substitute), the less we can see the product as effectively constraining the price of the first product or offering a realistic alternative to many consumers. Indeed, there may be multiple dimensions to the differentiation, including price, quality, and (for IP goods) brand association.

Of course, even if courts take the approach we suggest and focus on the relative closeness of available substitutes, determining when the substitutes are sufficiently close that we can count them as competing goods entails a judgment about how much competition we want in this space. At the level of specific cases, we think that IP law can learn something from a more flexible antitrust market definition, despite the many flaws in antitrust’s approach. Hence, while IP courts should think more coherently about the nature of competition between IP-protected products, they might well make different policy judgments in different circumstances about when products are sufficiently close that they raise competitive concerns.

At the more general level of the appropriate scope of IP rights, the question is when an IP right creates or enables so much differentiation that the costs of IP protection are too high to justify their benefits. Resolving that question requires policymakers to ask how much market power we want an IP owner to have, and that determination is bound up with the policy justifications of IP rights in the first place. We don’t purport to answer the question of “how much market power is sufficient” in this Article. Our point is that IP rights are not free. Though IP and antitrust doctrine and scholarship have minimized the presence of a conflict between the two in recent decades, IP rights do in fact confer significant power over price in a surprising array of cases.

If we wish to restrict the extent to which IP rights confer market power, and therefore minimize the range of conflict between the doctrines, we should focus on IP doctrines through which differentiation could be reduced.

A.     Copyright

In the copyright area, courts could require much greater similarity between a plaintiff’s and defendant’s works in order for the defendant’s work to be considered “substantially similar” and therefore infringing. Likewise, we could decrease copyright owners’ power by significantly curtailing or eliminating the derivative work right, thereby allowing better substitutes in the market. To put it more concretely, if consumers do not regard Stephen King novels as good substitutes for a new Harry Potter book, then we would better reduce the market power of Harry Potter books by weakening the derivative work right and allowing new Harry Potter-related works in at least some contexts. To the extent derivative works are complements rather than substitutes, we might want to focus more on limiting the scope of the copying right than on restricting the derivative works right. Among derivative works, we might care more about controlling power over that subset of works likely to serve as a substitute for the existing work, rather than derivatives that expand into a new market or act as complements to the original work.

B.     Trademark

If our goal is to reduce the differentiation afforded by a trademark and make other products better substitutes, then we should focus in the trademark area on other parties’ ability to get closer to the claimed trademark. One means of accomplishing this in the trade dress context is a robust functionality doctrine, which would prevent parties from claiming certain features as trademarks at all. But in the broad range of cases, the best way to accomplish this would be for courts to put more significant weight on the similarity of the marks factor of the likelihood of confusion test.

In terms of the scope of rights, we should pay attention to circumstances in which a mark doesn’t merely reflect differences in the products but instead is the difference. If the goal then is to decrease differentiation, doctrine must decrease mark owners’ ability to control that brand meaning. Under current law, mark owners protect this brand meaning primarily through broad sponsorship or affiliation claims in contexts far beyond those in which consumers could reasonably believe the mark owner was responsible for the quality of the defendants’ goods or services. But they also protect this meaning more directly through dilution claims, which are intended to protect the singularity of a mark’s meaning. To reduce brand owners’ ability to differentiate along this brand meaning dimension, both of these types of claims could be eliminated or at least substantially limited.

C.     The Scope of Patent Rights

The scope of patent rights is a function of the language of the patent claims. Accordingly, any effort to modulate the market power conferred by a patent needs to start with the doctrines of claim construction and infringement. One reason patents often now are able to cover broad swaths of conceptual space is that patent lawyers frequently claim their inventions in functional terms. Rather than limiting the patent to a particular device, patentees claim any device that performs a particular function. This will often, though not always, coincide with market power, since control over a particular function may make the development of an effective substitute impossible.

One way to limit the reach of patent claims, especially in software, is to take seriously the old idea that functional claiming is not permitted, and to limit the patentee to what she actually invented or equivalents thereof. Indeed, some (including one of us) have suggested going even further: returning to the days of central claiming.

The control that patentees have over markets is also a function of a power that patents confer that copyrights and trade secrets don’t: the power to prevent independent development. Indeed, patents today are overwhelmingly asserted not against accused copyists but against independent developers. Some have suggested that patent law, like copyright, should limit its reach to those who obtain the idea from the patentee. That may or may not be a good idea. But if adopted, it would substantially reduce the number of cases in which patents conferred market power by allowing an important source of substitutes.


Courts in a variety of IP contexts engage more or less explicitly in market definition as a predicate to some doctrinal determination. They do this, however, without any apparent methodology, and that has the pernicious effect of making market definition simply a tool for reaching desired results. Antitrust law has a well-developed methodology for defining markets, but as we argued, that methodology does not work well in the context of IP goods. IP goods differ precisely because they are differentiated; indeed that is largely the point of IP rights—to enable owners to prevent others from copying too closely so that the IP owners can reap a profit sufficient to provide incentives to invent or distribute (or, in the case of trademarks, to enable consumers to differentiate between goods of different source).

Traditional market definition is too rigid to accommodate this reality, because it asks whether particular goods are in or out of the market rather than focusing, as we would, on the closeness of available substitutes, taking into account the inevitability of some differentiation. In other words, in a market full of imperfect substitutes, it makes more sense to ask just how imperfect the substitutes are. To the extent we want to reduce IP owners’ market power, there are doctrinal levers within IP that can be used to enable more and better substitutes.dingbat


Copyright ©2012 Mark A. Lemley & Mark P. McKenna.

Mark A. Lemley is the William H. Neukom Professor, Stanford Law School and partner, Durie Tangri LLP. Mark P. McKenna is Professor of Law and Notre Dame Presidential Fellow, Notre Dame Law School.

This Legal Workshop Editorial is based on: Mark A. Lemley & Mark P. McKenna, Is Pepsi Really A Substitute For Coke? Market Definition in Antitrust and IP, 100 GEO L.J. 2055 (2012).

  1. 532 U.S. 23 (2001).
  2. See Valu Engineering, Inc. v. Rexnord Corp., 278 F.3d 1268, 1276 (Fed. Cir. 2002).
  3. TrafFix, 532 U.S. at 33.
  4. 35 U.S.C. § 284 (2006).
  5. Mattel, Inc. v. MGA Entm’t, Inc., 616 F.3d 904, 916 (9th Cir. 2010).
  6. 17 U.S.C. § 107 (2006).
  7. U.S. Dep’t of Justice & Fed. Trade Comm’n Horizontal Merger Guidelines, 57 Fed. Reg. 41, 552 (Sept. 10, 1992).

Article printed from The Legal Workshop: http://legalworkshop.org

URL to article: http://legalworkshop.org/2012/12/21/is-pepsi-really-a-substitute-for-coke-market-definition-in-antitrust-and-ip

Copyright © 2009 The Legal Workshop. All rights reserved.