QUT Law & Justice Journal Vol 3 No 1 2003

[Print-friendly version]

  Printer version of full article(pdf)

Competition in Information and Computer Technology Markets: Intellectual Property Licensing and s 51(3) of the Trade Practices Act 1974

Ian Eagles * & Louise Longdin **

I Introduction

Intellectual property licensing[1] presents competition lawyers and regulators with something of a puzzle. The practice is both pervasive and necessary and yet will often take forms which on their face appear to run foul of the sometimes broadly expressed prohibitions through which most jurisdictions enact their competition policy. While most of these prohibitions have long been tamed into rules of reason, they can, as currently in Australia still induce panic (a panic largely borne of confusion[2]) in right holders and licensees who confront them for the first time. Panic rises to even higher levels when intellectual property owners are threatened with the sudden removal of the statutory shield which up until now has protected most (but significantly not all) of their dealings with licensees from the attention of regulators and competitors. Feeling more than usually exposed by this proposed legislative disarmament are suppliers of computer software and other forms of information and communications technology, accustomed as they have been to the steady expansion of legal protection[3] (the lifting of parallel importation restrictions being one of the rare examples of retreat[4]) rather than the contraction they wrongly assume to be imminent unless fended off by intensive lobbying.[5] How justified are these fears? For the most part, not very. Only where the Trade Practices Act 1974 imposes per se liability do the lobbyists have a case and here the proposed changes improve the position of right holders rather than the contrary. Indeed if there is a criticism to be made of the suggestions of the Intellectual Property and Competition Law Review Committee's (the 'Ergas Committee') in its Review of Intellectual Property Legislation under the Competition Principles Agreement[6] and the government's response to them ('Government Response'[7]) is that the changes are prompted by long standing problems with the whole of Part IV itself (still under review at the time of writing).

II The Present State of the Intellectual Property Interface in Trade Practices Law

There are a number of exemptions to the raft of practices prohibited by Part IV of the TPA. Section 51(1) of the Act states in general terms that in deciding whether a person has contravened Part IV certain matters must be disregarded. These are 'anything specified in, and specifically authorised by an Act (not including an Act relating to patents, trademarks, designs or copyrights) or regulations made under such an Act.' Standing alone, this would mean that many dealings in intellectual property would be left unprotected against competition scrutiny. However, what the legislature took with its s 51(1) hand, it largely returns with its s 51(3) hand. The latter provision proceeds, albeit ambiguously, to provide a wide (if not unlimited) protection for intellectual property owners and their licensees. Section 51(3) protects the imposing or giving effect to conditions contained in licences and assignments to the extent that those conditions relate to the subject matter of the intellectual property statute in question. Where trademarks are concerned, the exemption is expressly limited to conditions that relate to the kinds, qualities and standards of goods bearing the mark.[8]

Section 51(3) is not a blanket exemption for all licences or assignments of intellectual property rights. It does not cover practices which infringe s 46 (misuse of market power); s 46A (misuse of market power in a Trans-Tasman market) or s 48 (resale price maintenance). The exemption extends only to conditions contained in or effected by licences or assignments which might otherwise be construed as one of the following anti-competitive practices:

• Collusive conduct under s 45. Contracts, arrangements or understandings in relation to goods or services[9] are prohibited where they contain an exclusionary provision or have the purpose and effect of substantially lessening competition. This includes price related contracts, arrangements or understandings under s 45A. Contracts, arrangements or understandings are deemed to have the purpose or effect of substantially lessening competition if a provision has the purpose or effect of fixing, controlling or maintaining prices, for, or providing discounts, allowances, rebates and credits in relation to, goods or services.

• Exclusive dealing under s 47. Supply of goods or services by a corporation to a third party is prohibited on terms constraining dealings with competitors of the corporation. (It appears that it is unnecessary for a specific competitor to be designated or identified in a licence or assignment for liability to be made out.[10])

• An acquisition resulting in a substantial lessening of competition in a market under s 50.[11]

• An acquisition outside Australia of a controlling interest in a corporation resulting in a substantial lessening of competition in a market under s 50A.

Putting aside for the moment the question whether intellectual property rights require this special treatment, the present s 51(3) is flawed even on its own terms. First the use of the nebulous 'relates to' (surely the vaguest and most open ended verb in the drafting tool box) to link right and exemption positively invites parties to expand the scope of the protection thus offered by sprinkling the same ambiguous terminology indiscriminately throughout their licences and assignments. Second, if there are sound policy reasons for leaving particular categories of rights and particular types of breach outside the protective umbrella of s 51(3) these have never been adequately articulated, much less defended.

A 'Relates to' and the scope of grant problem

Most competition regimes around the world, when faced with the problem of how to treat intellectual property, take as their starting point an inquiry into the exact nature of the exclusive rights conferred on patent, copyright, trademark owners under national law: What is it that owners are allowed to do, prevent others from doing or licence those others to do? Actions within the four walls of the right are then given varying degrees of special treatment in the ensuing competition analysis. This approach, usually described as a 'scope of the grant' inquiry,[12] is intended to provide judges and regulators in jurisdictions which have no statutory exemptions for intellectual property with a rebuttable presumption of validity for actions by right holders which do not go beyond the bounds of whatever it is that the state has granted to them. Those who cannot bring themselves within these bounds have something to explain. For those who can, the onus remains on their opponent (be it competitor or regulator) to demonstrate anticompetitive intent or outcome.[13] In these jurisdictions it is the first part of a two part test. For jurisdictions like Australia which choose to regulate the competition law/intellectual property interface by statute, 'scope of grant' ceases to be purely evidentiary in its application and instead is used as a one step substantive test with which to divide applicants into saved saints who escape liability entirely and potential sinners to be subjected to the full rigours of an ordinary Part IV analysis. This is not the purpose for which the 'scope of grant' test was designed. Even if it were, it is not obvious from its terms that s 51(3) encapsulates the scope of grant test in any form that its European and American proponents would recognise. Nor has judicial exegesis illuminated what the drafters have left unclear. In Transfield Pty Ltd v Arlo International[14] two different interpretations of 'relates to' were provided by the High Court. On one view, that of Barwick CJ and Wilson J, the exception would extend to the terms of any licence or assignment as long as there is some nexus with the intellectual property right in question. Thus, irrespective of their impact on competition, exclusive licensing, price, quantity and territorial constraints would not infringe Part IV. On another view, that of Mason J, the scope of the restrictions right holders may properly impose on their licensees or assignees would not extend to gaining an advantage 'collateral' to those enjoyed directly under the grant of the right, a judicial nod in the direction of 'scope of grant.' [15]

B Unprotected rights and dealings

As it currently stands s 51(3) does not extend to rights whose only protection is at common law or equity. These include:

• Trade secrets and know how which lie outside a patent claim. These can only be protected by relying on contract or breach of confidence.
• The goodwill and reputation attached to names and identities which are not registered as trade marks. These are the province of passing off and Part V.

These omissions are deliberate, presumably on the grounds that non statutory rights are less certain in scope and application than 'true' intellectual property rights.[16] The absence of any reference to plant breeders' rights in s 51(3) on the other hand seems to be due to a failure to monitor and update within the bureaucracy.[17]

Another problem with s 51(3) in its present form is that it is inconsistent in its treatment of even the rights it does list. Assignments of patents, copyrights and registered designs are covered but the assignment of trade marks is left unprotected.

C Non exempt breaches

At present s 51(3) provides no protection for actual or putative transgression under s 46 (misuse of market power) or s 48 (resale price maintenance). This puts refusals to licence and attempts to control the price of licensed products (or spare parts for those products) at risk.

III The Changes Proposed by the Ergas Committee

The Ergas Committee is not the first consultative body to grapple with s 51(3). The National Competition Council had earlier recommended that it be retained but amended so as to remove horizontal dealings and price and quantity restrictions from its ambit.[18] This was not the road down which the Ergas Committee chose to go. Instead it took the view that:[19]

[Intellectual property statutes confer] upon the intellectual property right holder a series of exclusive privileges designed to promote innovation. Given that those rights are conferred by legislation, they should be able to be effectively exercised even when this involves (as it generally must) the exclusion of others. However, these rights should not be capable of being used to go beyond the market power those rights directly confer. That is the right holder should not be allowed to extend the statutory right into a wider right of exclusion with the effect of substantially lessening competition. [original emphasis supplied]

To this end the Committee suggested both widening and narrowing the protection currently provided by repealing the existing s 51(3) and replacing it with a provision which would:

• Extend the statutory shield to all the practices referred to in Part IV.

• Bring the more recently enacted forms of intellectual property within the exemption. (But not, one infers sub silentio,common law and equitable protections.)

• Subject all exempt dealings to a substantial lessening of competition override. This was to be an effects-based test not one of purpose. The Committee clearly intended to exclude all issues of anti-competitive intent from the intellectual property equation. Nor was any distinction to be made between per se rules and rule of reason. Both were to be equally entitled to the benefit of the statutory shield.

• Mandate a regulatory safety net in the form of guidelines to be issued by the Australian Competition and Consumer Council ('ACCC') as to the manner in which it intends to enforce the new provisions.

What is envisaged here is a test which, although two step like its American and European progenitors, is substantive rather than evidentiary and would apply across the whole of Part IV. The second step in that test, the substantial lessening of competition override, is both workable (once supplemented by guidelines) and defensible to all sides of the debate currently raging among law and economics scholars as to how competition policy and intellectual property should interact in a modern economy.[20] It sensibly avoids the wilder shores of theory without unduly hobbling the regulator or subjecting owners, assignees and licensees to avoidable transaction costs. It is the first part of the two step test which causes difficulty on both conceptual and practical grounds. As regards the latter, the decision to retain the unsatisfactory 'relates' to set the parameters of exemption is simply to invite keen legal minds to endlessly parse the Transfield judgment in the interests of their clients until the High Court definitely opts for one interpretation over another. If it was intended to enact the scope of grant principle why not select language more suited to the task?[21]

But why did the Ergas Committee assume that the task has to be undertaken at all? Did they intend it to be a way of distinguishing between market power which is external to the right (and thus unprotected) and the power deriving directly from the right (which is assumed to be unassailable). If this was the Committee's intention then the two step test cannot work as they describe it because their substantial lessening of competition override makes no such distinction. Nor could it. Intellectual property rights may sustain market power derived from other sources without themselves creating that power, a point forcibly made by the seven strong Court of Appeals for the District of Columbia ('DC Circuit') in United States v Microsoft Corp.[22] In that case the defendant Microsoft had claimed that its intellectual property in its software gave it overarching power to use that property howsoever it wished and legally justified the entire raft of anticompetitive provisions contained in its licensing arrangements with original computer equipment makers. The DC Circuit brushed off this broad (and, in its view, frivolous) argument as 'no more correct than the proposition that use of one's personal property, such as a baseball bat, cannot give rise to tort liability'.[23]

If the boundaries of the right do not set the limits of liability why write them into the statute at all? They could of course be used as an evidentiary tool for resisting a finding of anti-competitive purpose but since the Committee rejects purpose tests this cannot be the explanation. The reason, one suspects, is to be found in their perceptive analysis of Part IV as a whole. While they concede that it was not their task to comment on, or seek amendment to, the wider provisions of the Act[24] they clearly felt uncomfortable both with the forms of business behaviour selected for per se proscription in Part IV[25] and the overall emphasis on enumerating practices even within those parts of the Act which proceed by way of rule of reason analysis.[26] The Committee was obviously sceptical about the degree to which any practice could be assumed to exploit consumers and lessen competition without further inquiry and thought intellectual practices particularly vulnerable to this way of thinking given its non rivalrous nature and the importance of licensing to its successful exploitation. The Committee's aim seems therefore to have been to remove intellectual property rights to a place of greater safety until such time as the rest of Part IV is overhauled, a job which they allow was outside their brief.[27] The scope of grant principle is there to provide the walls of this temporary sanctuary.

The difficulty with this approach is that it anticipates outcomes for the wider reform process which are far from assured. Are we really likely to see a Part IV entirely shorn of per se provisions (horizontal price fixing included) and purpose tests? If these things did not happen Australia would, if the Ergas proposals were enacted, be left with a competition regime which permanently treats intellectual property very differently from other rights, thus departing from the principle of neutrality encapsulated in the Microsoft judgment. As to the grounds on which such departures might be justified the Committee suggests three:

• The greater vulnerability of licensing practices to uncertainty.

• The transaction costs associated with seeking letters of comfort and resisting enforcement.

• A concern that the transition from exemption to scrutiny should not be too sudden or too brutal.

As to the first, absolute certainty means either absolute exemption or per se liability. As the American experience shows, the economic rationality which the rule of reason is meant to deliver carries its own uncertainty price tag as competing theories fight it out in the courts and the commentary. Relative certainty is obtainable only by guidelines. It cannot be legislated. Nor is it obvious that the absence of a special statutory regime for intellectual property in the United States or the European Union unduly inhibits investment in intellectual property licences in those jurisdictions.

As to the second, it is the proposed guidelines which will prevent mobs of apprehensive owners and licensees beating a path to the ACCC's door. Writing the scope of grant principle into the statute will do nothing to beat them back (especially while its operation continues to be made uncertain by the 'relates to' terminology).

The third ground does have more force but the answer lies in education. The transition can either be effected or delayed until after the Dawson Committee reports and the legislative fate of its recommendations is clear. There is no obvious case for a temporary via media to lessen a pain which must eventually be suffered

These are minor cavils, however. Taken in the round, the Ergas Committee's proposals are a thoughtful and well reasoned response to the problems of integrating intellectual property and competition policy. Given that their hands were largely tied by having to work within the existing ss 45, 47 and 48, the Committee's recommendations would, if enacted, have been a vast improvement on the present position. The same, alas, cannot be said of the Government's sketchy and conceptually incoherent response.

IV The Government Response

While the Government Response endorses the need for intellectual property rights to be given distinctive treatment under Part IV there is little sign that this is based on any deep understanding of what led the Ergas Committee to adopt this stance. Worse, it introduces further layers of complexity to an already convoluted s 51(3) on the basis of policy distinctions which it fails to state, much less explain. As outlined to date, the government's intentions are to bring intellectual property licensing in general under Part IV but retain s 51(3) in some form to deal with the problems raised by the existing per se provisions in ss 45, 45A, 47 and 4D, that is those concerned with horizontal price fixing, third line forcing and exclusionary boycotts. This residual exemption would then be subject to a substantial lessening of competition test. Byinference this override would not need to apply to the 'purpose and effect' rules of reason in the rest of ss 45 and 47 because they already require the application of such a test. (A crucial departure from Ergas which may not have been fully thought through.) One of the perhaps unintended consequences would be that licensing practices intended to bring about an anti-competitive outcome could continue to be at risk irrespective of whatever outcome was achieved or even achievable. The residual s 51(3) would seemingly not apply to breaches of ss 46, 46A and, importantly for this discussion, the resale price maintenance provision in s 48. Whether the government would retain the shapeless 'relates to' is as yet unclear.

Where does this lead the scope of grant principle? It would have no relevance to the rule of reason provisions in ss 45 and 47. Licence terms attacked under these sections would neither be less vulnerable within grant or more vulnerable outside it. Scope of grant would be equally irrelevant to the operation of s 48. Once the terms of s 48 were met and the supply of goods or services withheld or threatened to be withheld liability would attach without any need for an inquiry into the boundaries of the right being licensed. Only in relation to third line forcing, horizontal price fixing and exclusionary boycotts would scope of grant continue to matter. In these cases, practices outside the grant would not have the benefit of the proposed substantial lessening of competition override.

What is puzzling about the choices the government has made here is that they do not seem to be based on an evaluation of the content and scope of existing per se rules (whether in general terms or as applied in intellectual property licences). Given that all of these breaches are now authorisable the continued relegation of resale price maintenance to the licensing sin bin needs explanation. (Horizontal price fixing would be a better candidate, if one were needed, there being few economists willing to find glimmers of pro-competitive gold among its universally conceded anti-competitive dross.)

Considerably less controversial is the government's decision to include plant breeders' rights within the residual exemption and to continue existing protections for licences and arrangements entered into before any amendments take effect. The role and mode of operation of the guidelines is also spelt out in more detail in the Government Response than in the Ergas Committee's Report. This detail, however, masks hidden complexities.

V Designing the Guidelines

In its response to the Ergas Committee's proposals the government indicated [28]that it wants three things from the guidelines which the ACCC is to issue.

• The guidelines are to indicate which conduct is likely to be within the new s 51(3) and which is likely to be outside its protection. (We can call these the 'scope of grant guidelines.')
• The guidelines must also give reasonably clear indications of the kind of market behaviour likely to fall foul of any substantial lessening of competition test override (the 'SLC guidelines').
• The ACCC must be able to tell market participants, if only in broad terms, the factors it will take into account in deciding whether a particular intellectual property dealing (or non dealing) should be authorised (the 'authorisation guidelines').

The ACCC does have overseas models to assist it in carrying out those three quite distinct tasks. Not all of these models point in the same direction. Nor do they use the same analytical techniques. All have their blind spots. Stripped to their essentials the choice of methodologies is as follows:

First, the guidelines could be industry specific. Much as practitioners and industry participants might welcome information and communications technology specific rules this seems to have been ruled out by the Ergas Committee who thought that Part IIIA and Part XIC were not well designed for handling intellectual property rights.

Secondly, the guidelines could be practice specific, the approach used in the European Union in relation to patents and know how.[29] This approach proceeds by listing restrictions and conditions in intellectual property licences and classifying them on a scale of yes,[30] no[31] and maybe. While useful in answering scope of grant questions, practice specific guidelines will always under or over regulate because they ignore market power.

Thirdly, the guidelines can be outcome directed, that is they look at gains and losses to both the competitive process and efficiency. This approach is adopted in the United States[32] and Canada.[33] While admirably suited to dealing with both the application of the substantial lessening of competition test and the authorisation question they have the drawback that, being drafted as aspirational objectives, they provide little detailed guidance to anyone faced with a client who demands to know right now without ifs or buts: 'Does this clause breach the Act?'

Bearing these models in mind how should the ACCC design its intellectual property guidelines? Each phase of the task will need to be approached differently.

(a) Scope of grant guidelines The object here will be to, first, allocate a particular restriction to the appropriate 'saved' or 'dammed' pen and then to indicate the likely fate of the latter when subjected to the full rigour of the per se prohibitions and purpose based liability imposed by the rest of Part IV and how it might be redrafted to avoid this result. The skills required for this purpose will be those of the intellectual property lawyer who has a thorough understanding of what each type of intellectual property right authorises owners to do or prevent.[34] They will face two problems:

• The language of intellectual property does not distinguish between horizontal and vertical arrangements. As long as an arrangement is vertical in form it will pass the 'scope of grant' test even though it may be horizontal in substance.[35]
• Many licences will not distinguish between exempt and non exempt rights. Patent licences will refer to know how. Franchise and distribution agreements will often be framed in terms of what is being sold or licensed and draw no distinction between trademarks, passing off and Part V protections for this purpose.

(b) Substantial lessening of competition guidelines These will need to indicate the factors used to:

• Define the relevant market remembering that markets are delineated by substitutability not the boundaries of the right and that future innovation may both constrain market power and be constrained by it.[36] Again the market in which the substantial lessening of competition occurs may be upstream or downstream from that market (it being generally accepted that power in one market may be leveraged into other markets). Care will also need to be taken to distinguish between markets for rights and markets for goods or services channelled through those rights.
• Identify the conditions which might make it harder or easier for substitutable products not protected by the right in question to enter the market. (It may be hard to avoid a certain circularity here. The right itself may sometimes be construed as a barrier to innovation.[37])
• Determine the market share and concentration thresholds which are to act as safe harbour surrogates for market power. This measure has to be crude if it is to work. Finely grained economic analysis has to await litigation.[38] In the United States the Department of Justice and the Federal Trade Commission will not challenge terms in an intellectual property licence where the collective share of licensor and licensee of the market or markets 'significantly affected' by the restraint does not exceed twenty per cent.[39] In Canada the figure chosen is thirty five per cent.[40]As regards concentration, under the United States guidelines regulators will look benignly on a licensing practice where there are four other firms unconnected with the licensor or licensee who have the capacity and incentive to engage in the kind of research and development which could lead to substitutable innovations. The Canadians would exclude players who cannot switch the necessary production respond within a year or who need to invest in large amounts of new research and development from the supply side of the market definition equation. Capacity tests of this kind downplay the ability of widely defined intellectual property rights to pre-empt competitive or derivative innovation in some fields for what might be considerable periods of time. They should be used with caution especially in the absence of hard evidence as to the pace[41] and direction of technological change in a given industry.

(c) The authorisation analysis Here the core issue is likely to be the generation of countervailing efficiencies. In the intellectual property environment one would expect there to be a greater emphasis on innovative efficiency than its allocative or productive brethren. It would be sensible for the guidelines to require applicants to quantify efficiency gains[42] although this may be difficult with new and unproven technology. Factors which the ACCC may take into account include:

• The extent to which the licence increases inter brand competition even if it decreases or eliminates competition between licensees.
• Whether the claimed efficiencies can be achieved in other competitively neutral (or even pro-competitive) ways taking into account cost and commercial viability.[43]
• Any saving in transaction costs as with collective administration of software licences for example.[44]
• The effect on other intellectual property rights held by the parties to the arrangement.

VI Potentially Assailable Practices

How would licensing practices fare under the as yet virtual guidelines described above? As many commentators have observed, the process of digitisation associated with the convergence of content and communication technologies in the information and communications industries has brought about new configurations of power and behaviour.[45] The following discussion focuses on constraints and conditions in software licences and assignments that may be caught under s 51(3) if and when it is amended. The analysis necessarily takes into account two important and closely related features of information and communications technology. The first is the fact that collaboration among players in the industry is far from uncommon. (Indeed 'teaming up' and the formation of alliances to combine technological inputs by multiple right holders are strategies often highly conducive both to innovation itself and returns obtained on that innovation.) The second is that intellectual property rights (particularly copyright and patents) do not always neatly translate into software and other associated goods or services.[46] Software (especially that depending on the interoperability of operating system platforms, application systems, application programming interfaces and internet and digital communications technology) will often incorporate works or processes protected under a whole raft of copyrights and/or patents (frequently with multiple owners). The ensuing cross licensing thicket may be the only effective means of ensuring that pooled expertise and the intellectual property rights that go with it actually end up in the hands of those who can best exploit the software.[47] That said, application of the substantial lessening of competition test to any agreement or arrangement in the computer or any other industry mandates a search for anti-competitive effects. In that search particular attention will have to paid to arrangements which although vertical or diagonal in form are horizontal in outcome. (This is not to say that horizontal arrangements should always be unlawful, merely that they have a greater capacity to foreclose markets, particularly narrowly defined innovation markets.) They are particularly likely to be found in cases where but for the software licensing agreement or arrangement in issue the parties would have been competitors. Where such an agreement or arrangement facilitates a firm's ability to exercise market power on its own account or in concert with another party and has (or is likely to have) an effect on the price, quantity or quality of the software or raises a rival's costs of producing a substitutable or interoperable product or process it will be at risk under the new test.

1 Abuse of standard setting process by competitors

Virtually all firms involved with the information and communications technology infrastructure will have occasion to co-operate or collaborate in standard setting or comply with standards set by others, standards that fundamentally alter the nature of competition. Where a standard becomes widely adopted the market shrinks to encompass only those products that adhere to that standard. Where there is no existing standard, competition 'for the standard' will be intense especially in those markets characterised by network effects[48] precisely because they are likely to tip to a single format or technology.[49]After the battle is won the standard setter will hope to enjoy less competition from competing technologies.[50] Suppose then, that parties agree among themselves to make only products or provide services that comply or are interoperable with a certain standard. When is such an agreement likely to pass the substantial lessening of competition test? For a particular standard setting exercise among competitors or potential competitors to remain under both competition and efficiency thresholds, its outcome would have to be considered to have strong pro-competitive potential and/or confer a public benefit by ensuring products are compatible with each other. The agreement would, however, be likely to be found anti-competitive if diverts and/or sinks all or most of a rival's research and development resources. A case in point here is Dell Computer Corporation.[51] There Dell agreed with its competitors to work towards a standard for the Video Electronics Standards Association. All participating firms assured each other (on more than one occasion) that they held no intellectual property rights which would block others from developing towards the standard. Only when the standard was successful did Dell claim it infringed on its patent and seek royalties from its competitors. As Australian law now stands the agreement would be anti-competitive under s 45A if revealed as being largely a device for price control or fixing or demonstrating an anti-competitive purpose under s 45. Post Ergas, the standard would have to be shown to be both pervasive and irreversible.

2 Pooling Software Patents

Software patent pooling arrangements between competitors may be struck down if they foreclose future innovation without conferring pro-competitive benefits. These may arise where blocking patents are cleared and expensive infringement litigation is avoided or when the agreement facilitates the integration of complementary technologies and the parties are able to thereby reduce their research and development costs.

One useful pointer as to how far horizontal patent pooling may lawfully go is contained in the Canadian Competition Bureau Intellectual Property Enforcement Guidelines (2000)[52] which identify as vulnerable[53] an arrangement whereby two firms agree to pool patents (the same logic also applies to software copyright pooling) but only one firm is likely to infringe the other's intellectual property right when it puts its product on the market. In the Bureau's view such a patent pooling practice would amount to an agreement to prevent price competition between the parties. Their guidelines suggest that two firms in a patent pooling arrangement will only escape having their practice declared anti-competitive where they both are blocked from using their own technology because it infringes on the other's patent and neither is able to 'invent around' the other's blocking patent. Should the end result be that if the patent pooling arrangements in question are not necessary for new technology to enter the market they will, on this way of looking at things, have no pro-competitive saving graces. Rather worrying, however, is the suggestion implicit in the Canadian guidelines that the competition authority might in some circumstances be prepared to jump in itself and assess the strength of the patent infringement claim before considering whether the potentially infringing party can invent around the patent. How well is a competition authority placed to carry out this kind of complex analysis? Courts and patent offices find it difficult enough to test the merits of patents and copyrights claimed over software, computer implemented business methods, computer screen displays and interfaces.[54] That a regulator would be equipped to do better seems unlikely.

3 Third line forcing and bundling

Third line forcing and bundling are pervasive in the computer industry. They are also plausibly represented as innovatively efficient. Such claims need to be subjected to close scrutiny. In such scrutiny issues of market definition and barriers to entry would be expected to loom large. Suppose, for example, a computer supplier A requires its customers to have the hardware it supplies maintained by B if they want to receive software maintenance and support services from A. Since these arrangements are vertical, the competition analysis will rest on market definition. If a separate market exists for the provision of maintenance services for A's hardware, A would inevitably have power in that market. Should no distinct market be able to be identified for A's hardware, A's power in it would be insubstantial or non existent. If enough users consider that their total outlay on a computer system spread over its expected lifespan includes hardware maintenance costs as part of their original decision to purchase the system then a single market could be identified. It would include not only the primary product (the computer hardware) but also the secondary service (maintenance and support). The difficulty with this kind of analysis is that it assumes that barriers to entry are structural rather than behavioural and ignores the inhibitory effect on entry or exit of the bundling scheme itself. To require software licensees to purchase or licence certain products or services from licensors or third parties where this cannot be justified on the grounds of quality control or technical efficiency and then argue that consumers have set the boundaries of the market with their wallets may be putting the analytical cart before the analytical horse. Since the problem in such cases is one of characterising or measuring barriers to entry, guidelines based on market share will be of less assistance here.

Not only may restrictions on unbundling suites or compilations of computer programs and reselling or relicensing individual programs be found to be anti-competitive on the facts, they may also lie outside the scope of the relevant grant as a matter of law. Under the two step test proposed by the Ergas Committee and retained in the Government Response for the per se provisions in ss 45, 45A and 47 this would introduce some highly artificial distinctions into Australian competition enforcement. Bundling one's own products or services must already surmount a substantial lessening of competition hurdle and would continue to have to do so under the government's proposals. Forcing licencees to consume the products of others, however, might (on one interpretation of the government's methodology) lose the benefit of the substantial lessening of competition test proposed for s 47(6) if it is outside the scope of grant. But not all forms of third line forcing will be outside the grant. [55] A further distinction can be made here between the sale of physical objects such as CD-ROMS through which software is loaded onto a computer and the licensing of that software. Thus, in one United States copyright infringement case such a provision in an end user licence agreement was found to be unenforceable since the intellectual property owner's right were exhausted under the first sale provisions in the copyright statute.[56] Were an Australian court to take a similar line (nothing turns for this purpose on the exhaustion of rights rule being judicially inferred rather than expressly legislated) the competition analysis would be unnecessarily complicated by a search for the boundaries between sale and licence.

4 Best endeavours clauses

A licensing arrangement may stipulate in relation to a copyrighted computer program or electronic database that the licensee employ its best endeavours to exploit the product and not use, sell or license competing programs or databases. Being on all fours with that considered in Transfield it is likely to be found within the scope of grant even if the clumsy 'relates to' were to survive a drafting cull. Post Ergas, liability would then depend on market definition and power.

5 Lump sum fees and output restrictions

A licensing arrangement which stipulates in relation to a copyrighted computer program or electronic database that the licensee pay a large upfront license fee which is deductible from later royalties earned by the licensee appears to be both within the scope of grant and competitively harmless unless the licensor has market power. On the other hand a requirement that the licensee produce a minimum quantity of the product made using the software coupled with a restriction on the use of other intellectual property rights to compete with the products made using the licensed intellectual property right may foreclose the market and certainly would be beyond the grant. Once again, it is difficult to see what guidelines have to offer here. If they are outcome oriented the nature of the practice will tell market players nothing that they need to know while to ban or bless either practice outright would be to impose a wholly spurious certainty on the debate.

6 Negating Infringement Defences

Usually it will be the right owner who seeks to come within the four walls of the intellectual property right so as to avoid the application of purpose tests and per se prohibitions. Sometimes, however, it will be competitors or users who complain that the condition or arrangement in question seeks to cut back rights guaranteed to them under the relevant intellectual property statute. In such a case the restriction is at risk of falling outside the scope of the grant because it impinges on a defence against infringement which would otherwise be available to the licensee. Into this category may fall copyright licences prohibiting reverse engineering, disassembly or decompilation of software. These may arguably fall outside the scope of the grant because of the new division 4A of the Copyright Act 1968 (Cth)[57]. This allows a software licensee to reproduce (including by decompilation) a legitimate copy of a computer program (not a pirated one) in certain circumstances, namely:

• As part of the normal running of the program
• In order to study how it functions
• To make a back up copy
• To make an interoperable product
• To correct errors in the product
• To test its security

Software licensors who attempt to pre-empt some or all or part of these fair use provisions by contract are placing themselves beyond the intellectual property statute's protective umbrella.[58] Thus if an end user license stipulates that there is to be no reverse engineering, modification or testing the security of the program, not only will such provisions be unenforceable, they will also be, if the Government Response is enacted intact, beyond the protection of s 51(3). Since their purpose is arguably to substantially lessen competition they will be at risk under s 45 whether or not their effect is proven to be (or to be likely to be) anti-competitive on the facts. On the other hand where a computer program is patented as well as copyrighted there is no scope of grant problem on the patent side because no comparable statutory fair use defence exists for patent infringement in Australia. This threatens to introduce a wholly illogical distinction into the enforcement equation, one which it is impossible to justify on policy grounds and which demonstrates the inherent difficulties in bringing both parts of the Ergas Committee's two step test together in a coherent statutory scheme.

7 Grant back provisions and continuing obligations

Clauses which require licensees to surrender to the licensor their own intellectual property rights are often regarded as competitively suspect. Whether this is in fact so will largely depend on the degree to which the rights surrendered cover products for which there are no clear substitutes. If no such substitutes are discernible on the technological horizon the grant back will be at risk even if it is not complete. To take but one example, suppose the licensor of a computer assisted design system imposes the condition that it is to be a joint copyright owner of any works the licensed user produces using the expert system (whether it be the design of a new building, boat or nose). These run the risk of being construed as attempts to leverage the market power conferred by the copyright into other markets and as such are both beyond the grant and potentially anti-competitive. Equally problematic are attempts to project the licensor's market power (where it exists) beyond the statutory life of the intellectual property right in question such as provisions that the licensee not exploit the intellectual property right once the term of protection for that right has expired or post expiry obligations to make royalty payments.

VII Conclusion

This paper was researched and written without the benefit of sighting either the Dawson Report or the actual wording of an amended s 51(3). Nevertheless it now seems unlikely that there will be any legislative retreat from the position adopted in the Government Response. If enacted in their present form the government's proposals will only marginally tilt the competition balance against intellectual property owners and licensees. Most licensing arrangements will survive a substantial lessening of competition test if it is applied to them. The real objection to the government's proposals is not that they disturb existing expectations (these being no more relevant here than in any regulatory context) but that they do so in ways which are both economically illiterate and unnecessarily complex. If the government were, however, to have last minute second thoughts we would suggest that they might usefully note the following observations.

First, it is probably unwise to enact a special competition regime for intellectual property rights before reforming Part IV as a whole. Secondly, when and if Part IV is overhauled, the starting assumption should be that its provisions fall equally on all types of property. From this it would follow that purpose tests and per se liability should either apply to all market players or none. It would also mean there would be no need to distinguish between different forms of per se liability. Section 51(3) would then be redundant. Should this view be rejected and the section retained in some form, the scope of grant issue would have to be tidied up by finding a more appropriate form of words than 'relates to' as the connecting link between right and exemption. The question of non statutory rights also needs to be re-examined. It cannot simply be sidelined. Finally, drafting infelicities and conceptual lacunae cannot be cured by enforcement guidelines, especially if those guidelines are pulled in different directions by unclear and incompatible objectives.

© Queensland University of Technology Law & Justice Journal 2003
This article may be cited as: (2003) 3 (1) QUTLJJ 28