Framing the Analysis


Framing the Analysis


Finding the appropriate legal mechanism through which to mediate the interaction of intellectual property rights and competition values is a problem that at various times, and with varying degrees of urgency, has confronted policy makers, courts and regulators in more than one OECD economy. When the holders of these rights take active steps to exploit them by licensing or assigning them to others, the conceptual uncertainty surrounding the construction of the interface between the two areas of law is, to some degree, masked by the existence in most jurisdictions of legislative exemptions, administrative guidelines and safe harbours that grant such dealings varying degrees of special treatment.1 Where, however, the right owner simply sits tight and either actively or constructively2 refuses to license the right to actual or perceived competitors (or more remotely those who deal or wish to deal with such competitors3), the potential conflict between competition policy and intellectual property law has not been (and in some jurisdictions cannot lawfully be) disguised or softened in this way. The tug of war between right holders, competitors and consumers as to how putatively anti-competitive refusals to license might be approached is thus both more visible and more urgently in need of a set of guiding principles. In an ideal world such guiding principles would minimise forensic uncertainty and maximise economically defensible outcomes in individual cases. They would also be indifferent as to whether market power was sustained or exercised through tangible or intangible property, or indeed any other form of business advantage (proprietary or otherwise). To what extent these objectives are simultaneously attainable, and which should give way if they are not, are issues with which many competition regimes have had to grapple and with which, by and large, they are still grappling. Refusals to license intellectual property are thus not only important in themselves, they also place under the spotlight issues whose resolution is important to the internal coherence of competition law as a whole.


Anyone writing on this subject has to come to terms with the inescapable reality that in a world in which goods and services are freely traded across borders, real markets seldom correspond with the boundaries of any single legal system, and that any proffered solution to the refusal to license problem needs to be as workable and effective across jurisdictions (statutory language permitting) as within them. This immediately presents the writers with a problem of selection and organisation. Which jurisdictions should be surveyed, and how to make sense of the survey once conducted?4 Individual jurisdictions may throw up isolated examples of refusals to license which, however significant in their time and place, provide scant material from which to extrapolate a trend or pin down an analytical approach.5 Given that selection is both inescapable and inevitably arbitrary, we have chosen to make a virtue of necessity by concentrating on five jurisdictions, each possessing a sufficient body of case law or detailed enough legislative provisions to make trans-jurisdictional comparisons meaningful. The jurisdictions chosen are Australia, Canada, New Zealand, the European Union and the United States. We have, however, assumed that the last two competition regimes are better known outside their home jurisdiction than the first three, and have adjusted the level of detail accordingly. It is most emphatically not our intention in these pages to venture upon a reprise of the many excellent books and articles on Article 102 TFEU and section 2 of the Sherman Act.

If our trans-jurisdictional analysis is to work, however, it is important that our analytic framework should precede rather than follow detailed expositions of the law in the jurisdictions selected. With that in mind, we have organised our material by commencing in chapters one to four with the identification of the economic issues surrounding refusals to license intellectual property, and examining various policy choices when it is sought to convert these into legal rules. We then proceed, in chapters five through to eight, to survey the treatment of the problem in particular jurisdictions. In chapter nine, by way of conclusion, we offer some solutions and reject others.


One of the unfortunate features of the refusals to license debate (especially when conducted across jurisdictions) is that the language used by courts and commentators can sometimes muddy the waters in unintended but nevertheless unhelpful ways. Most commonly this occurs when a particular term used by a writer or decision maker to signify concept A can be (and sometimes is) understood by other participants in the debate to stand for the sometimes antithetical concept B. Most of the time, in most competition contexts, this is distracting rather than misleading. (It also tends to afflict lawyers more than economists, although the latter are not entirely immune; witness the loaded way in which ‘consumer welfare’ is used by some economists.) Where slippery words and phrases distort the analysis we address the issue in context and in detail. This does not happen very often. Of less immediate concern, but even so potentially disconcerting, are individual authorial idiosyncrasies. To minimise the effect of these, we think it useful simply to declare—Humpty Dumpty6 fashion—what we think we mean by certain key words and phrases used throughout the book.

1.3.1 ‘Intellectual property’, ‘intellectual property right’ and ‘refusal to license’

The term ‘intellectual property right’ is used in what follows as shorthand for both proprietary and non-proprietary rules, except in those instances where the distinction leads to different legislative or judicial treatment in competition contexts. In such cases we use Professor Cornish’s term ‘allied rights’7 to refer to those non-proprietary obligations (or, as the Americans term them, liability rules) which sustain intellectual innovation and artistic creation while falling short of ‘true’ property rights (as usually conceived). We also accept that the term ‘refusal to license’ is not altogether apposite in cases when what is in issue may be nothing more than an advance indication that an unlicensed act will be treated as an infringement and responded to accordingly. Even mere silence in the face of a competitor’s demands we would, in some contexts, consider to be a refusal.

1.3.2 ‘Regulator’ and ‘regulation’

‘Regulator’ and ‘regulation’ are terms freighted with inevitable imprecision.8 Many (but not all) law and economics scholars reserve their use for those forms of public law intervention in the market that proceed by way of direct control over such things as prices and access fees. This form of regulation usually relates to a particular sector of the economy—most notably in what are generally referred to as public utilities.9 We have chosen to call this direct regulation. Absent this qualifier, ‘regulation’, for the purposes of this book, also embraces the ordinary processes of competition law enforcement. ‘Regulator’ then becomes a generic term for any enforcement agency or body, other than a court or tribunal, charged with bringing about competitive outcomes, or maintaining or fostering competitive conditions in a market. Our confessedly broad-brush usage necessarily ignores a great many fine institutional and structural distinctions, and does not seek to pinpoint the regulator’s exact place on a spectrum of judicial and administrative roles. ‘Regulation’ can also be used more widely, as well as more narrowly. Thus intellectual property may be viewed both historically and conceptually as a form of State intervention in the economy, sharing many of the social goals of competition law. While this particular aperçu is important in demystifying and hence de-privileging intellectual property (in competition terms), the usage is too idiosyncratic to be offered without qualification or explanation and so we do this where appropriate. (In this context there is of course no regulator. The situation is not too different from that obtaining when competition law is privately enforced.)

1.3.3 ‘Competition’, ‘antitrust’, ‘abuse of market power’ and ‘monopolisation’

‘Antitrust law’ and ‘competition law’ are used synonymously in this book, although strictly speaking the use of ‘antitrust’ should be confined to discussion of United States law. The same qualification applies equally to ‘monopolisation’ and ‘abuse [or misuse] of market power’.


One of the factors clouding the refusals to license debate is that it takes place at the intersection of two deeply flawed ideas. The first of these is the notion that unilateral10 refusals to act by holders of market power should be judged more leniently (if indeed they should be judged at all) by courts and competition authorities than overt acts by those same holders. Or, to put it another way, coercing a party with substantial market power into a relationship with rivals or potential rivals is inherently wrong and should be avoided wherever possible. The second equally unfounded notion is that dealings (or in this case non-dealings) in intellectual property are (or should be) to a greater or lesser degree privileged in competition cases compared with dealings or non-dealings in anything else.

Both of these statements (for they are no more than that) are normative. They seek to tell us about the proper shape of legal rules. Those rules are not plucked out of thin air, however. They rest, or at least purport to rest, on what economists have to say about the risks of over- (and less often under-) regulation, and the effect of both of these things on innovation and investment. The trouble is that conversations of economists on these subjects are also curiously normative, often disconnected from any empirically demonstrable reality. It is therefore not that surprising that some members of both professions are sometimes inclined to treat these propositions as free-standing statements of the obvious, quarantined off from the rest of competition law without any felt need for apology or justification.


The discussion that follows is based on a very different set of normative principles about what an ideal competition regime should look like.11 Underpinning these principles is the concept of regulatory neutrality, under which competition enforcement operates even-handedly across all sectors of the economy and all forms of business activity. If an economic justification needs to be found for regulatory neutrality, it is best sought in a particular application of investment displacement theory, under which the shielding of intellectual property dealings (or non-dealings) from competition scrutiny and the favouring of inaction over action has the potential arbitrarily to suck investment out of one form of economic activity and into another for reasons of legal form rather than efficiency. We, however, would prefer to rest the case for regulatory neutrality (Rawlsian fashion) on the need to preserve the integrity of the legal system by ensuring that regulatory intervention commands the widest possible degree of support amongst market players before it is applied to them, and without the need for any ability on their part to predict their own future behaviour. Neutrality is not, we would stress, a cure-all for all competition ills. One could imagine a regime in which courts and regulators came to share the scepticism of some commentators as to the efficacy of restrictions on the unilateral exercise of market power in general. There would then be neutrality in the sense that refusals to license would become part of a wider problem (or evidence of a non-problem), but that is not the book we are writing. Our aim is not to lay down a blueprint for the harmonisation of substantive rules across jurisdictions (an unattainable objective given the political difficulties attending amendments of core legislative provisions in the United States and Europe, and one that the World Trade Organisation (WTO) has now formally abandoned12). Still less do we suggest that the application of similar principles (economics-based or otherwise) will inevitably lead to similar regulatory outcomes on the same or similar facts.13 Our objective is the much more modest one of distilling from judges’ and regulators’ decisions in selected jurisdictions a list of desiderata on which most commentators could agree even if individual competition regimes fall short in detail—for the purposes of discussion. Such a list would look something like this:

a) Competition law necessarily elevates substance over form. Dominant market players should not be able to game the regulatory system by manipulating bright-line rules. Indeed, bright-line rules are to be preferred to rules of reason only where the transaction costs avoided by adopting them are both large and easily demonstrable. This remains true whether or not the bright-line rule in question imposes or deflects liability. Shifting the onus of proof by adopting presumptions of vice or virtue for particular transactions merely disguises the choice between substance and form. It does not eliminate it.

b) To be effective, competition law needs to be able to cut across private law rights and obligations, whether those rights and obligations be based in contract or property. There can be no private law shields against regulatory action.14 Opinions may differ as to whether there should be a self-denying presumption by regulators as to their ability to get things right or the need to recognise that temporarily anti-competitive situations may self-correct over time. These are matters on which competition regimes (and economists) may rationally disagree. What all agree on, however, is that there will be occasions on which private ordering has to give way to administrative action. To hold otherwise would be to render competition law entirely nugatory; indeed it would question the need for its very existence.15 Deference to property rights becomes especially problematical when the boundaries of the property right are themselves unclear,16 or if their content varies across jurisdictions between which goods and services are traded freely.

c) All economic activity should equally be grist to the regulator’s mill. It is not the role of competition law to play favourites and prefer one business strategy or structure over another.17 From this it follows that neither property in general nor particular forms of property should confer any immunity or advantage on owners when used as instruments of market power. Conversely, market advantages which cannot be propertised (or which have yet to be propertised) should not be treated less gently by competition regulators than those instruments of market power to which the law grants property status.

d) High-technology markets and industries based on creative content do not need to be subjected to more (or less) onerous competition rules than dealings or refusals to deal in fence posts18 or access to bridges,19 ports20 or print distribution networks.21 It has never been the function of competition regulation to direct investment to any particular form of economic activity, and it is a measure of its failure if it does.

e) Competition regimes may draw the line between rules of reason and strict liability in different places and by different means but, once drawn, that line should not need to be moved again in order to accommodate particular types of market or particular categories of property. Similarly, while there may be room for doubt as to the efficacy of probing anti-competitive intent in competition cases,22 and a consequential preference for tests of liability based on proof of anti-competitive outcomes, such doubts should not resonate more or less loudly just because the alleged source of market power, or the mechanism for its misuse, is an intellectual property right.23

f) Regulatory outcomes should not depend upon the form of the conduct complained of. This means that when assessing putative abuses of market power, action on the part of the supposed perpetrator is not presumptively more virtuous or more vicious than inaction.24

g) Economic opinion may be divided on predicting particular market outcomes from particular market behaviours, or the effect of regulatory intervention in particular cases. Such divisions only rarely and accidentally coincide with the jurisprudential categorisations of the instruments through which market power is sought to be exercised.

h) Competition law exists to protect the competitive process—not individuals, nor the collective interests of consumers or producers.25 Any of these persons may be the indirect beneficiary of regulatory activity, but that is not the aim of the exercise. Neither is the substantive objective of protecting competition in any way diminished by the evidentiary truth that in particular cases outcomes which are not passed on to consumers may be neither substantial nor sustainable.


The refusals to license debate is all too frequently clouded by resort to rhetorical exaggeration and distortion, usually in the form of punchy sound bites that glide easily over the complexity of the issues involved. Sometimes these incisive one-liners will be unproven in the sense that they rest on empirically untested assumptions. As such they are no more than attempts to hijack the refusal to license debate before it has even begun. (Indeed, the claims made will often be so broad as to be beyond easy empirical verification by anyone.) On other occasions the rhetorical posturing may be true as far as it goes but is of no assistance in resolving the policy choices a judge or regulator is being asked to make. Set out below are the most commonly asserted of these fallacies and red herrings.

1.6.1 Ownership carries with it the right to exclude others from the thing owned

As a statement of one of the key indicia of property this is unexceptional enough, although not all might agree that it is property’s defining attribute and opinions may differ as to whether it is a precondition or a consequence of property status. When invoked in a competition context, however, it becomes a claim for an absolute right to use one’s property entirely free of regulatory constraints. Such, for example, was the basis for Microsoft’s claim in United States v Microsoft Corporation26 that it was free to impose whatever restrictions it liked on computer hardware manufacturers because the computers operated Microsoft’s software in which Microsoft had copyright and other forms of intellectual property right.27 Once intellectual property rights had been acquired, so the argument went, their subsequent exercise by Microsoft could not give rise to antitrust liability. So large a claim invited rejection on the obvious ground that, once allowed, it could not be confined to any particular type of property. Rejected it duly was by the seven-strong District of Columbia Court of Appeals, who rather mordaciously pointed out that the next step for litigants would be to argue ‘that use of one’s personal property such as a baseball bat cannot give rise to tort liability’.28 The Court might well have added that acceptance of such arguments would introduce an economically irrational distinction into competition law under which those liability rules allowed to shelter under the intellectual ‘property’ umbrella are subjected to the full rigours of competition scrutiny, while anti-competitive actions involving the exercise or non-exercise of ‘true’ proprietary rights would be to a greater or lesser degree protected against that same scrutiny. The distinction would also be unworkable when the property status of a right is challenged, as it is in many jurisdictions in relation to trade secrets and other valuable forms of information.29 Again, when the property status of a right varies across jurisdictions (as it often does30) then competition outcomes would also vary even when the matters complained of spanned jurisdictional boundaries and concerned the actions or non-actions of a single corporate.

1.6.2 What the State has expressly granted it shall not take back by stealth

This statement is a more refined version of the ‘property is privileged’ argument. It is based on the assumption that because intellectual property takes the form of an express legislative grant of a bundle of rights exclusive to the right holder, only the legislature can remove what it has explicitly bestowed, a function which it is not for the courts or regulators to usurp. Sometimes referred to as the ‘scope of grant’ principle, this view of things allows the boundaries of the right to determine the jurisdiction of both court and regulator. What the right owner or licensee does or refuses to do within the four square walls of the right cannot therefore be challenged on competition grounds.

‘Scope of grant’ theories rest on two logical fallacies. The first takes the unexceptionable (but in the past sometimes forgotten) truth that the legal monopoly represented by an intellectual property right does not always represent market power, and inverts that truth so that it becomes the empirically untested proposition that intellectual property can never do so. Economic and legal monopolies need not coincide to reinforce each other, either within or across markets. The second fallacy is a legal one. By inflating the principle of generalis non specialibus non derogant into a rule of invariable and rigid application, its proponents argue that regulatory intervention is excluded as a matter of law. An express legislative grant, they would say, requires an equally express revocation. Absent such revocation, the competition regulator must observe the statutory ‘keep out’ sign. The obvious difficulty with this line of reasoning is that it characterises intellectual property statutes as ‘general’ and competition jurisdiction as ‘special’. This is to prejudge the position. Competition statutes are not noticeably more or less ‘general’ in their terms than intellectual property statutes.

As well as being conceptually flawed, the scope of grant approach also has enforcement downsides. The principle is the logic on which the European Union Technology Transfer Block Exemption31 rests, and it also underlies statutory exemptions for proactive licensing arrangements in some countries such as Canada and New Zealand. If applied to refusals to license it would effectively grant refusers immunity from competition scrutiny, provided they were acting within the confines of an identifiable statutory right. If correct it would mean that much European jurisprudence on the subject32 could be explained (as indeed some commentators have argued33