Refusals to License in the United States
United States anti-monopolisation law has few unifying features. Looked at in the round, jurisprudence under section 2 of the Sherman Act1 (employing few words and language old-fashioned enough to enable the courts to paint with a broad interpretive brush) presents as a series of discrete, highly-contextualised sub-rules, each built around a particular type of anti-competitive conduct: tying, predatory pricing and raising rivals’ costs, to list some examples. Unilateral refusals to deal, on this view, are part of the same phenomenon. The reasons for this contextual fragmentation are too complex to trace in detail here. Neither are we concerned with the wider consequences for United States antitrust law.2 What can be said, however, is that the separate evolution of these sub-rules over time, and the relative infrequency of their consideration (and reconsideration) by the Supreme Court, ensure that they embody very different attitudes to the efficacy of regulatory intervention (sometimes even falling on opposite sides of the rule of reason/per se divide at particular stages of their separate histories). This in turn has made it very difficult for the law to adopt or maintain an approach of regulatory neutrality between different manifestations of market power or different modes of exploiting it. In this fragmented environment the only unifying factors tend to be the rhetorical negatives outlined in chapter two, in which judicial aperçus have to do the conceptual heavy lifting for which they are particularly unsuited.3 Refusals to license intellectual property raise all of these issues in a very acute way.
The history of the intersection of antitrust and intellectual property law in the United States has been largely characterised by alternating cycles of over- and under-enforcement in which first antitrust and then intellectual property prevails.4 This recurring pattern, in which one regulatory regime moves into the front seat for a time, leaving the other relegated to the back, has prevented the two systems from interacting coherently or consistently. When cases have arisen, they have tended to involve patents rather than copyright or trade secrets. In periods of antitrust over-enforcement, this has had the effect of magnifying perceptions of the market power able to be wielded by patent owners. In times of under-enforcement, the emphasis has tended to be on the careful social bargain inherent in the patent system, under which ex ante disclosure and dissemination are supposedly traded off against relatively strong and unassailable protection. There has thus emerged a notion of a hierarchy of rights, but a reversible one in which the stronger or weaker intellectual property rights are balanced commen-surately by more or less forensic antitrust scrutiny.5
This jostling for supremacy between the intellectual property and antitrust laws was first evident during the early twentieth century, when patent holders, attempting to expand their traditional rights beyond the first sale of their products, ended by provoking a strong antitrust reaction, resulting in amongst other things judicial recognition of the doctrine of patent misuse.6 This period of ascendancy of antitrust law (or more properly antitrust concepts, since abuse of patent law existed in a formal sense outside the Sherman Act but drew heavily on jurisprudence under that statute) over patent rights came to an end in the 1950s, after patent rights were strengthened following the passage of the Patent Act 1952.7 Antitrust, however, regained the upper hand during the 1960s and 1970s, when the validity of many patents was successfully contested and patent owners were susceptible to a finding of antitrust liability if they attempted to enforce invalid patents or engaged in certain kinds of licensing practices.8 Antitrust lost ascendancy once again in the 1980s,9 when patents became increasingly easier to obtain and defend, and the Patent Misuse Reform Act 1988 narrowed the circumstances under which a defendant could claim patent misuse by way of defence.
Another factor at play during this era was the Reagan Administration’s express spurning of the regulatory authorities’ restrictive rules regarding intellectual property licensing practices. Fresh guidelines concerning such practices were issued in 1995 by the Department of Justice and the Federal Trade Commission during the Clinton Administration, and the late 1990s witnessed a flurry of regulator-initiated antitrust claims and cases, including United States v Microsoft Corporation.10 This resurgence of activity then dropped away dramatically with the advent of the George W Bush Administration in 2001,11 in which the Department of Justice issued a report on the issue of single-firm conduct under section 2.12 That report, issued not long before the incoming Obama Administration, was formally withdrawn soon afterwards,13 avowedly, in part, because of its ‘overly lenient approach to enforcement’.14
The current case law and regulatory practice on the appropriate interface between competition law and intellectual property in the United States, is best understood as a reaction against earlier unreflective assumptions by both courts and regulators equating legal with economic monopolies, and treating the former as an automatic source of market power to be approached with suspicion and sometimes outright hostility. While these views lasted (and the period of their ascendancy was lengthy15), a presumption of illegality hovered over all sorts of actions by intellectual property owners that would nowadays be seen as harmless or even pro-competitive. Per se liability was imposed on some licensing practices and the onus of proof reversed where others were present. The high water-mark of the presumed illegality approach was the formulation of a black list of putatively unlawful intellectual property licensing practices issued in 1971 by the Antitrust Division of the Department of Justice and its sister regulator, the Federal Trade Commission, a list referred to at the time as the ‘Nine No-Nos’.16
Fortunately for the intellectual coherence of United States antitrust law, this prescriptive approach was first eroded by the courts17 then abandoned by the regulators.18 New guidelines were issued proclaiming the principle of regulatory neutrality, under which intellectual property would be treated no differently from other forms of property for antitrust purposes. Unfortunately, however, the dominance of the former way of thinking persisted long enough, and generated sufficient controversy, to provoke its own over-reaction. Under-analysed presumptions of vice were replaced by equally under-analysed presumptions of virtue or claims for absolute immunity.19 Even where such extreme stances were not taken, neutrality proved in practice to be distinctly one-sided. There was an emphasis in both cases and commentary on the need to remove disabilities previously suffered by intellectual property owners. The possibility that there might be a downside in regulatory neutrality for these same owners was seldom adverted to, much less discussed. In this climate it was scarcely surprising that acts or omissions previously penalised might now be regarded as positively privileged. Logically this did not have to follow, as a few judges recognised. If there was an economic case for exempting or limiting the antitrust liability of intellectual property owners, it had to be made separately. It did not ineluctably flow from the goal of regulatory neutrality.20 Such responses became increasingly rare, however.
The way in which this lopsided notion of neutrality gradually took hold is perhaps most clearly seen in the different approaches adopted over time by judges and regulators to the boundaries of intellectual property rights. In the era of the Nine No-Nos, any step by intellectual property owners beyond what the intellectual property statute allowed was apt to be seen as liability creating in itself.21 This was of course an economic nonsense, and was ultimately rejected as such. It was only a short step, however, from treating what happened outside the right as presumptively anti-competitive (or worse deeming it to be such) to regarding what happened within the parameters of a particular intellectual property right as actually or presumptively immune from antitrust attack (the scope of grant principle discussed above at 1.6.2). As we shall see, that step was increasingly taken.
The essential facilities doctrine had its origins in the early years of the twentieth century, and thus long pre-dated the explicit reception of economic theory in United States antitrust law. Its beginnings (if not its name) are usually traced back to United States v Terminal Railroad Association,22 a classic bottleneck dispute in which it was claimed that the Terminal Railroad Association (TRA), a group of railroad owners, was in breach of section 2 of the Sherman Act by denying non-cartel competing railroad services access to and from the city of St Louis via a terminal and rail bridges built over the Mississippi River, facilities controlled by the cartel. Because of the peculiar topography of the region and the narrowness of the valley through which all rail lines had to pass, the Supreme Court found the TRA’s refusal oppressive, and held that the TRA was required to act impartially in allowing access to competitors on reasonable terms or break up the association.23 In coming to its decision, the Court introduced a concept plucked somewhat uncritically from the common law doctrine of prime necessities,24 a doctrine which even at its most fully developed did not require the sharing of resources which were merely useful.
If matters had been left where they lay in Terminal Railroad, the impact on United States antitrust law would have been minimal. Terminal Railroad was, after all, no more than a collective boycott, one not much different from others of its kind. The Supreme Court in Associated Press v United States25 went on to apply similar logic to a news cartel that was, in the words of the court,26
a vast, intricately reticulated organisation, the largest of its kind, gathering news from all over the world, the chief single source of news for the American press [and] universally agreed to be of great consequence.
In the result, the Court determined that by-laws allowing the cartel to prevent its members from selling its news reports to non-members were in breach of the Sherman Act because, while those news reports might be attributable to the association members’ own ‘enterprise and sagacity’, they were clothed ‘in the robes of indispensability’. Since the press association controlled 96 per cent of news circulation in the United States, it was, in the view of Frankfurter J, akin to a public utility. The finding, at first instance, went unchallenged that it was ‘a business infused with the public interest, required to serve all’, and that a need existed for the maximum flow of information and opinion.27
It was not until Otter Tail Power Co v United States28 that the essential facilities doctrine can be said to have assumed its modern form. In that case, the Supreme Court was for the first time faced with applying the doctrine in the context of a purely unilateral refusal. On the facts, the Otter Tail company not only enjoyed a monopoly over the transmission of electricity in its local area but the company was also vertically integrated. Upstream it generated the power that it supplied to its own grid, while downstream it distributed electricity at retail rates to customers in the local area by way of selling off monopoly franchises. Otter Tail refused to supply the power that it generated itself at cheaper wholesale rates to several municipalities who sought to set up their own retail distribution systems. It also denied transmission of electricity over its grid to municipalities who could have acquired power that had been independently generated. Since the Court considered the refusal went further than merely imposing a handicap on potential new competitors in the downstream market for retail electricity and precluded all possibility of competition in that market, it ordered Otter Tail both to sell its own electricity to municipalities at wholesale rates and to transmit other independently-generated wholesale electricity to those municipalities who wanted it for their own retail distribution systems.29
Otter Tail paved the way for a later generation of American judges to take hold of the essential facilities doctrine and extend it to refusals to deal by a single dominant firm, often comparable to a public utility of some kind but not necessarily also vertically integrated. The first case in which the concept of ‘essential facility’ was judicially aired as a term of art was Hecht v Pro-Football Inc.30 The claimed essential facility to which access had been denied was a sizeable stadium used for professional football matches. In finding the facility essential and the denial anticompetitive, the Court expounded on the circumstances under which the doctrine could successfully be invoked31:
[If] facilities cannot practically be duplicated by would-be competitors, those in possession of them must allow them to be used on fair terms. It is [anticompetitive] to foreclose the scarce facility … To be ‘essential’, a facility need not be indispensable; it is sufficient if duplication of the facility would be economically infeasible and if denial of its use inflicts a severe handicap on potential market entrants.
A later, much-cited decision in which the parameters of the essential facilities doctrine were more clearly articulated was MCI Communications Corporation v AT&T.32 MCI had challenged AT&T’s refusal to interconnect its telephone lines with AT&T’s then nationwide telephone network on the grounds that interconnection was essential if MCI were to compete in the long-distance telecommunications market. It was held that it had been ‘technically and economically feasible for AT&T to have provided the requested interconnection, and that AT&T’s attempt to thwart MCI constituted an act of monopolization’. In finding AT&T’s refusal governed by the doctrine, the Court stated in general terms33:
[A] monopolist’s control of an essential facility (sometimes called a ‘bottleneck’) can extend monopoly power from one stage of production to another, and from one market into another. Thus the antitrust laws have imposed on firms controlling an essential facility the obligation to make the facility available on non-discriminatory terms.
There is no hint here of any doubt as to the existence, rationality or durability of cross-market leveraging, something that previously could only be teased out of the facts of cases like Otter Tail.34 Indeed the two-market factor was clearly of pivotal importance in MCI itself, in which the Seventh Circuit found that AT&T had complete control over the local distribution network to which MCI required interconnection in order to offer long-distance services. What was not clear, alas, for precedential purposes (because on the facts it did not have to be clarified) was whether cross-market leveraging was the only situation in which the essential facilities doctrine could operate; or to put the question another way, whether the need to identify two markets was a necessary or merely a sufficient test of liability.35
The Seventh Circuit did, however, proceed to lay down a four-point, conjunctive checklist to be satisfied before the essential facilities doctrine could apply. There had to be established36:
a) control of the essential facility by a monopolist;
b) a competitor’s inability practically or reasonably to duplicate the essential facility;
c) the denial of the use of the (essential) facility to a competitor; and
d) the feasibility of providing (access to) the (essential) facility (to a competitor).
That the first two factors are closely interrelated (ie that the second factor has a great deal to do with determining what amounts to an essential facility in the first place) was a point made by the Ninth Circuit in City of Anaheim v Southern California Edison Co.37 It sensibly solved the chicken-and-egg conundrum by addressing the first two factors concurrently,38 here building on what it had said in its own earlier decision in Ferguson v Greater Pocatello Chamber of Commerce Inc 39:
[The essential facilities doctrine imposes] on the owner of a facility that cannot reasonably be duplicated and that is essential to competition in a given market a duty to make that facility available to its competitors on a non discriminatory basis.
This succinct restatement can perhaps lay claim to being the broadest summation of the essential facilities doctrine at its zenith. Two aspects of that summation are worth emphasising:
a) the reference to non-discriminatory access necessarily carries with it the seeds out of which a notion of constructive refusal could grow; and
b) the fact that the Ferguson court’s paraphrasing of the doctrine contained no allusion to the need for the existence of two markets.
Later decisions progressively parsed and qualified the criteria in the MCI checklist. Thus, the Ninth Circuit in Metronet Services Corporation v US West Communications40 explicitly recognised that a constructive refusal could be inferred from situations in which access is not refused outright but where the imposition of unreasonable terms and conditions results in practical denial of access.41 An important gloss was also put on the fourth criterion relating to the feasibility of granting access to an alleged essential facility when the Ninth Circuit upheld on appeal a District Court’s finding in Paladin Associates Inc v Montana Power Co42 that no legitimate business justification could be found on the facts for refusing access, a question that the District Court had expressly factored into its analysis, thereby recognising, if only in a backhanded way, what was to become an exculpatory exit from the application of the doctrine.
Unfortunately perhaps for our purposes, it is only first instance courts that have had to address directly the issue of whether intellectual property may be thought of as an ‘essential facility’. For example, in Bellsouth Advertising v Donnelley,43 the defendant, who had copied the organisation and headings of the plaintiff’s yellow pages directory and was sued for copyright infringement, raised an antitrust objection to denial of access by way of counterclaim. While the District Court found that the defendant had infringed Bellsouth’s copyright, it proceeded to hold that there was a genuine issue for trial as to whether Bellsouth’s copyrighted yellow pages directory was an essential facility to which it should grant access. The Court was unfazed by the fact that the doctrine of essential facilities had been applied only to tangibles in the past, and saw no reason why it could not also apply to information wrongfully withheld. In the Court’s view, the result in both situations is the same: a person is denied something essential to compete.44 The intellectual property issue was later side-stepped by the Federal Circuit in Intergraph Corporation v Intel,45 when it was faced with a claim by Intergraph seeking not a licence under Intel’s patents and copyrights, but a reinstatement of its preferred position as the recipient of:
(i) Products that embodied that intellectual property before these products were made commercially available to the public; and
(ii) Access to trade secrets … [ie] technical information that is not generally known, samples of new products before they are available to the public, and individualized technical assistance.
In its review of the essential facilities doctrine (invoked by Intergraph), the Court commented on the absence of competition between the two parties and lack of any flow-on consequences for a downstream market. It noted that although the protected products and trade secrets were considered essential by Intergraph to maintain its ‘strategic customer’ position, it was not the role of courts to ‘transform cases involving business behavior that is improper for various reasons … into treble-damages antitrust cases’. It was, it went on, to say ‘inappropriate to place the judicial thumb on the scale of business disputes in order to rebalance the risk from that assumed by the parties’. Simply because Intergraph had previously been supplied with the withheld products and information could not in itself be determinative of a valid antitrust claim, however relevant it might be in a claim for breach of contract, it said.
One possible reason why the essential facilities doctrine evolved in the way in which it did was that it was seen as a way for courts in the United States to extract themselves from the corner into which they had previously painted themselves on refusals to deal. Instead of focusing on the remedial problems associated with pricing and monitoring access in technologically complex situations,46 they preferred to express themselves in terms of what sounded like an absolute right not to deal even for dominant players. This approach was dubbed the ‘Colgate principle’ after the decision in United States v Colgate & Co,47 in which the Supreme Court delivered itself of the view that:
In the absence of any purpose to create or maintain a monopoly, the [Sherman Act] does not restrict the long recognized right of a trader or manufacturer engaged in an entirely private business freely to exercise his own independent discretion as to parties with whom he will deal …
Instead of questioning this over-extended and seemingly unqualified statement of principle (after all, a refusal to deal is just one of many ways of exerting market power where it exists), a later generation of American judges simply tacked the essential facilities doctrine on to Colgate in order to soften any anti-competitive outcomes that might flow from it in ‘hard’ cases. The ad hoc rolling-back of the Colgate principle in an attempt to achieve an intuitively fair result was the sort of thing courts tended to do before they became economically literate.
Another qualification of the Colgate principle is to be found in the decision of the Supreme Court in Aspen Skiing Co v Aspen Highlands.48 There the issue was how to respond to the situation where the monopolist/gatekeeper, by refusing, had ceased participation in a previously cooperative venture. In the result, the Court found that a unilateral termination of a voluntary (and presumably profitable) course of dealing was dispositive, suggesting a willingness in the monopolist to forgo short-term profits to achieve an anti-competitive end.
In Aspen Skiing, the defendant owned three of the four mountains in the Aspen, Colorado ski area, while the plaintiff owned the fourth mountain. For many years they had jointly offered a ‘joint ticket’ in the form of a multiple-day, multiple-area pass, providing skiers with admission to all four mountains. Purchasing this ticket tended to be a cheaper option for skiers than purchasing multiple single-day tickets. Under the joint ticket system, profits were split between the two parties commensurate with the percentage of use of each of the four mountains made by buyers of the joint ticket. The defendant effectively discontinued the joint ticket arrangement by making the plaintiff an offer to continue it only on the basis that the plaintiff accepted a fixed percentage of joint ticket revenues considerably lower than the historical average based on actual usage of the plaintiff’s mountain. When the plaintiff rejected the offer, the defendant proceeded to sell a joint ticket admitting skiers to its three mountains only. The plaintiff then attempted to market its own multiple-day, multiple-area package, by offering ski passes to the fourth mountain along with vouchers, each equal to the retail price of a single-day ticket to one of the defendant’s mountains. The defendant, however, refused to accept these vouchers and to sell any lift tickets to the plaintiff at retail price. In upholding a jury verdict in favour of the plaintiff on its claim that the defendant had monopolised the market for downhill skiing services,49 the Court, while avoiding the language of ‘essential facility’, imposed what could be (and subsequently was) read as an obligation upon monopolists to continue dealing with those with whom they had dealt in the past50:
[T]he monopolist did not merely reject a novel offer to participate in a cooperative venture that had been proposed by a competitor. Rather, the monopolist elected to make an important change in a pattern of distribution that had originated in a competitive market and had persisted for several years.
Although subsequently interpreted as an essential facilities case by lower courts, Aspen Skiing always sat rather uneasily within that paradigm. Nevertheless, evidence of anti-competitive animus of the kind punished in Aspen Skiing persuaded some courts in refusal to deal cases subsequently to find antitrust liability. Thus, at first instance in Intergraph Corporation v Intel Corporation,51 the District Court (before being overruled on appeal) concluded that Intel, the owner of patent rights and other intellectual property rights in the microprocessors and sophisticated technical information with which it had ceased supplying Intergraph, had
no legitimate basis [to] refuse to supply … since Intel has been doing so for the last four years on a mutually beneficial basis.
It was no justification for severance of the long-standing licensing arrangement in the view of the Court that Intel had immediately ceased supply after learning that Intergraph was suing it for infringement of patents owned by Intergraph.52 This fits in with the earlier finding of the Supreme Court in Eastman Kodak v Image Technical Services53 that the right to refuse ‘exists only where there are legitimate competitive reasons for the refusal’.
The expansion of the doctrine provoked Professors Areeda, Hovenkamp and other eminent commentators54 into arguing strongly for its abolition, and it has now been heavily downplayed (if not quite extinguished) by the Supreme Court in Verizon Communications Inc v Law Offices of Curtis V Trinko LLP),55 in which the Court somewhat coyly neither confirmed nor denied whether the doctrine was still part of United States law, pointing out only that they themselves had never expressly endorsed it.56 The Court reaffirmed that there was no general obligation on single firms (as opposed to a consortium) possessed of market power to deal with others unless they had previously dealt or co-operated with them. Neither would the monopolist be under an obligation to offer any explanation for its refusal where there had been no prior dealing between the parties. By so holding, the Court made plain that its earlier decision in Aspen Skiing57 now stood simply for the principle that terminating an existing and voluntary business relationship without a legitimate efficiency reason might (not, be it noted, must) be unlawful in some unstated circumstances.
The Trinko Court also emphasised that there was no room for whatever was left of the essential facilities doctrine to apply where a parallel regulatory structure had been set up by statute to mandate and control terms and conditions of access.58 It recoiled from any notion that antitrust courts could be involved in enforcing sharing, a task that the Court said would require them ‘to act as central planners, identifying the proper price and quantity, and other terms of dealing, a role for which they are ill suited’.59 (As far as the Court was concerned, the Telecommunications Act 1996 had already provided the machinery for access to local telephone services60 and, furthermore, a regulatory provision compelling nearly global interconnection as well as the very connection to independent local telephone service providers that was at issue in the case.) The Trinko Court also poured cold water on earlier Supreme Court authority61 suggesting that leveraging could be anti-competitive if done by a monopolist. There had to be something over and above the leveraging, it said, and that something could not be a refusal to deal.62
Whatever its views on essential facilities, the Trinko Court was at pains to affirm Colgate, finding an added reason for doing so in the possibility that forced negotiation might ‘facilitate the supreme evil of antitrust collusion’.63 This, if taken literally, would greatly inhibit the ability of first instance courts to nudge parties towards agreements on the price of access to technology protected by intellectual property rights.64
Just where Trinko leaves the supposed ‘obligation’ to continue dealing remains unclear.65 Some commentators have read the case as continuing Aspen Skiing for this limited purpose.66 Others refuse to concede it even this narrow role, pointing to the existence of cases such as Intel67 where discontinued relationships triggered no liability.68 A small minority continue to offer a spirited defence of the doctrine’s ongoing relevance and usefulness.69 More pragmatically, and as demonstrated in Bell Atlantic Corp v Twomby,70 some antitrust lawyers sought to sidestep the problems posed by Trinko