The Antitrust Treatment Of Intellectual Property

Antitrust policy towards intellectual property has traditionally focused on defining the permissible restrictions that the property holder could initially impose on its licensees.

In recent US Federal Trade Commission cases, however, issues relating to downstream renegotiation or reacquisition of licences by the property holder have been important. In this section, we explore the policy issues raised by these downstream considerations and suggest a criterion for analysing their antitrust implications. Reacquisitions of licenses or renegotiations of license terms may become desirable from a business perspective as new information about technology and/or the market becomes available.

New information may suggest that the initial licence agreement has deficiencies to be remedied. While any inter-company relationship has the potential to be improved through downstream renegotiation, arrangements involving intellectual property are particularly ripe for improvements.

 

This is because intellectual property licensing generally involves new technology and markets that are typically characterised by great uncertainty. These arrangements also often involve entrepreneurs or companies venturing into functions and markets that are new to them. Also, such arrangements often require the formation of untested partnerships that sometimes require adjustments over time.

What differentiates intellectual property from other forms of property?

Many argue that intellectual property should be treated the same as other property and not be given special status. For example, regulators should allow any type of licence or combination with another company except when that company is an actual or potential competitor.

However, identical treatment may only mean that the framework for assessing agency action should remain the same. It is possible that intellectual property as a class will generally present different issues and situations than other types of property, and these differences, in turn, will often lead to different antitrust decisions.

Uncertainty.

There are uncertainties regarding the value of intellectual property and how it can best be used and marketed. Furthermore, the inventor is likely to know more about the characteristics of intellectual property than either potential licensees or the antitrust agency. Uncertainty also exists because the novelty normally associated with intellectual property implies that the uses and future development potential of the intellectual property are often not well known.

In many cases a pure sale or simple transfer of the licence may be problematic because of the on-going development of the intellectual property. Innovations that are likely to be the base for future patentable and unpatentable small and large improvements pose additional problems for potential licensing or venturing partners.

Intellectual property is often invented by individuals or companies that are not best suited to exploiting the property. This feature makes some sort of sale, licence or joint venture appropriate both for private and social welfare reasons. The need to find the optimal, or at least an acceptable, partner exacerbates the problems of uncertainty discussed above because sale or contracts now must be negotiated by two or more groups with potentially widely differing views about the future.

While these uncertainties and information asymmetries also exist for other inputs, they exist to a greater degree with intellectual property. Consequently, licensing contracts are likely to be complicated and to contain special features for dealing with information problems.

Differences between intellectual property and ordinary property also exist regarding the need to renegotiate contracts. When there is significant uncertainty about the value and use of intellectual property, parties may renegotiate contracts more often. Renegotiations may arise because it is impossible to anticipate the types of changes in conditions necessitating contract adjustments in the future. Even where changes in conditions can be predicted, such conditions may be difficult to verify, thereby making contingency provisions based on these conditions ineffectual.

Public good.

The public good characteristic of intellectual property also distinguishes it from other types of property. Unlike private property, it is difficult to control the use and dissemination of intellectual property once it is released. The incentives to create intellectual property may be diminished unless there are contractual provisions and restrictions on its use that preserve the value of the property for the patent holder.

In the absence of these safeguards, innovators are more likely to suppress the introduction of intellectual property into the marketplace and to reserve it exclusively for their own use. This suppression not only reduces the consumption value of the innovation but it also retards the rate at which subsequent spin-offs from the original innovation are developed.

In addition, a major concern of parties to an intellectual property contract is whether improvements or blocking patents might be developed by one of the parties to the detriment of the other. These efforts are also likely to be guided by marketplace demands and competition.

Antitrust and property rights: welfare and policy considerations

Public policy with respect to innovation involves a trade-off between inducing investment in innovation and trying to deliver the full benefits of innovation - once it has occurred - to consumers as quickly as possible. Here we look at three welfare considerations: the overall incentive to innovate; the incentive to license and develop the innovation; and the case-specific welfare effect.

Overall incentive to innovate.

Many economists believe that innovation rather than competitive pricing is the source of the largest improvements in consumer welfare. In economics parlance, dynamic (innovative) efficiency is more important than static (price) efficiency.

A primary issue in inducing innovations is the problem associated with the appropriability of innovation rents and its effect on generating appropriate levels of investment in innovation. This problem can be greatly mitigated by an intellectual property policy such as patent protection.

Because a patent gives the holder exclusivity but not necessarily an economic monopoly, how the patent can be exploited in the marketplace will influence the expected rents to innovation and therefore the investment in innovation.

Therefore, one must design patent and antitrust policies to satisfy a participation constraint that the inventor must earn a sufficient return on his investment to induce the desired amount of innovative activity. Satisfying the participation constraint may necessitate some adjustments in the antitrust treatment of intellectual property.

Little empirical evidence exists on the influence of antitrust licensing restrictions on investment in innovation. Because use exclusivity is so closely related to market product exclusivity, it seems nearly impossible to assess the impact of an intellectual property policy on incentives to invent without considering antitrust policy as well. Perhaps the point is best understood as referring to the effects of small changes in antitrust policy.

Also, antitrust treatment of intellectual property, when the value of the intellectual property is closely intertwined with intellectual property held by one's competitors, would appear closely to affect innovation investment

Assuming that an antitrust agency's enforcement actions have a fairly significant impact on innovation incentives, the agency needs to assess whether existing intellectual property policies are over-encouraging or under-encouraging innovative activities before determining an appropriate antitrust policy. There appears to be little agreement on this question.

Thus, there is disagreement over whether minor changes in antitrust policy matter for inducing innovation and, if they do, there still remains the question of whether the joint effect of current patent and antitrust policies results in too little or too much innovation.

Incentives to license and develop the innovation.

In designing antitrust policy one must respect the self-selection principle, which asserts that the inventor may always suppress the invention and reserve it for his or her (or its) exclusive use. The investor will only make the innovation available to others if he can arrange sufficiently profitable licensing agreements with other companies. The profitability of licensing is affected by antitrust treatment of intellectual property.

While consumers benefit from the existence of a monopoly-supplied innovation, they benefit more if the same innovation is supplied more competitively. In addition, the full value of the innovation will often not be reached without involving additional companies that can provide capabilities for further development, access to other complementary intellectual property, new applications, efficient and available manufacturing, and effective marketing and distribution.

Through its choice of how to exploit the property (for example, patent, licence, joint development), the holder of the intellectual property determines the extent to which consumers will benefit from the property. When the holder of the intellectual property can contract in ways that will preserve a sufficient share of rents from the invention, other companies will often be enlisted to increase the total profit pie that can be earned through taking the invention to market.

Anticipated antitrust restrictions on current and future use affect these choices and therefore affect the degree to which consumers will benefit from the innovation. Thus, antitrust policy may encourage or discourage various forms of licensing and co-operative arrangements. The extent of licensing may affect the incentives for future innovation as well. Several analyses show that licensing to a rival reduces the incentives for it to innovate. Over the long run this may reduce the rate at which new products are brought to market. On the other hand, prohibitions on wider licensing may stifle competition in the short run while increasing the future rate of innovation.

In designing policy for intellectual property, one must respect existing information constraints. The inventor is likely to be privately informed about the value and novelty of his innovation.

Resolving problems associated with private information may necessitate complicated or unusual use restrictions in licensing agreements, which in turn may necessitate some adjustments to antitrust policies regarding intellectual property. The presence of asymmetric information (when one party has private information the other lacks) also increases the difficulty of regulating licensing arrangements and preventing anticompetitive behaviour.

Case-specific welfare effects.

Until now, we have been discussing how existing antitrust policy and law affect private choices about the development and use of intellectual property. When a particular case is presented, however, many will argue that the greatest weight should be placed on the impact of the decision on welfare relating directly to the product market under discussion.

This direct static welfare effect with respect to the particular piece of intellectual property in question involves conventional antitrust cost-benefit accounting oriented to determining if consumer welfare will increase or decrease as a result of the proposed transaction.

Downstream adjustments in contracting

How should an antitrust agency view proposed modifications to previously approved contracts and licenses between licensor, Company A, and licensee, Company B?

To think about this problem, consider the following hypothetical situation.

Company A holds the patent to a new process. The antitrust laws encourage A to license the process to other competing companies because such licensing has the potential for decreasing price. Imagine that the antitrust law would permit A to license to a rival, Company B, with a bail-out clause that would permit A to revoke the licence and to purchase the marketing assets of B at any time in the future.

Now suppose instead that A originally offers B a permanent licence without provisions for repurchasing the licence. At some later time A attempts to buy back the licence from B. Should the acquisition be permitted?

We propose the following contracting criterion to aid one in thinking about this situation: approval of the acquisition should be conditioned on whether an initial long-term contingent licence allowing A to repurchase (or adjust the terms of) the licence would have been legal. We believe that evaluating cases using this criterion may allow one to draw some useful inferences about the behaviour of A and B.

The argument for this criterion is that if the law were going to permit repurchasing of the licence under an ex ante arrangement between A and B, it should not capriciously block the transaction later simply because it was not planned. Allowing more of these transactions would likely increase the amount of pro-competitive licensing that will occur and the use of more innovative forms of contracting and joint development of intellectual property. But one might ask why A did not write a complete contract with B allowing for these actions in the first place. Why should the antitrust authority need to solve problems between A and B that the parties should have anticipated and resolved initially?

This counter argument to the criterion overlooks the fact that there are several reasons why a complete contract specifying provisions for licence termination or contract adjustment may not have been written.

It may have been too costly for A and B to specify all the possible future contingencies and actions appropriate for each contingency. Further, some contingent contractual provisions may not be enforceable if the court cannot verify ex post whether a particular contingency has actually occurred.

While we do not find the counter argument to be particularly persuasive, there is a related argument that is an important caveat to the use of the criterion. The type of contract that A signs with B may reveal information to the antitrust agency about the properties of the patented process A owns.

As noted above, it is costly to write and enforce complicated contingent contracts with bail-out clauses. If A insists on the contingent contract with the repurchase clause perhaps it indicates that the process is easily imitated and transferable to other applications. The self-selection principle suggests that to deny such a contract clause may cause A to suppress the introduction of the process. Thus, in this instance, the enforcement agency may be compelled to approve the contract.

On the other hand, suppose A sells a permanent licence to B without any repurchase provisions. This may reveal that there is less need for A to protect the value of its patent against imitations and expropriations by other companies. Alternatively it may reveal that B has duly compensated A for the increased risk of expropriation in return for guaranteed access to A's process.

Here, the argument for allowing repurchase of the licence is somewhat less compelling. It is not necessary for ensuring that the innovation is disseminated in the market. The argument against repurchase is as follows. The fact that A did not initially contract for the termination of the licence reveals that A did not believe that this provision was necessary to ensure the success of the licensing agreement. Consequently, there is no reason to allow the licence to be terminated now since the parties did not feel it necessary to plan for the termination originally. We believe that there is a danger in carrying this line of reasoning too far, especially with respect to contract adjustments. An inference that unless a provision was included in the original licensing agreement it must not have been necessary to the success of the operation is not always valid. For example, it may be impractical to plan for all contingencies in the initial contract and, further, some provisions may be excluded because of

disagreements over them that might have caused the negotiations to break down.

Would open-ended contract provisions, for example provisions that allowed renegotiation but specify little else, circumvent the inability to plan for all contingencies problems? Such provisions might appear to solve the ex ante incentives problem posed were antitrust authorities only to use a case-specific welfare analysis. However, we do not think that the authorities would treat this provision as a get out of antitrust scrutiny free card, so this solution seems unlikely to solve the unpredictable contingencies problem.

Fixed-term licences present other related and interesting issues. When a licence terminates (and we see no obvious reason why such provisions would be illegal under antitrust law), the licensor and former licensee may negotiate a new licence.

Antitrust scrutiny would than be directed to the new licence agreement in much the same way as antitrust scrutiny of renegotiation using the contracting criterion.

However, there is one significant difference: the information that should be used to assess the welfare effects in the renegotiation is what exists at the time of the renegotiation instead of what existed at the time of the initial licence.

The contracting criterion is intended to help ensure that proper incentives for licensing and post-licensing actions are maintained in situations where uncertainty (especially concerning possibilities about which no one is aware), verification problems and differences of opinion prevent such incentives from being directly specified in the contract.

From the incentives point of view, using only information (and lack of information) available at the time of the original licence agreement seems appropriate because the incentives for licensing and anticipated post-licence actions depend on that information. This would not be the case in the new licence setting. Indeed, from the antitrust evidence point of view, restricting one's analysis to using prior information could allow some dubious agreements or activities to pass antitrust scrutiny because of lack of information rather than because of the substance of the information. In such cases, post information would be useful.

Conclusion

Downstream adjustments to licences or buy-backs of property rights present antitrust policy makers with interesting problems. One approach would be to consider only the case-specific effects on welfare, taking the existing situation as given.

For example, a producing licensor could not buy back its licence from a licensee or renegotiate the licence unless consumer welfare is improved because the efficiencies that result (and are passed on to the consumer) outweigh the negative effects of reducing competition.

Such an approach is too narrow. Antitrust policy should also consider the indirect effects of the policy on incentives to license or otherwise exploit the intellectual property. While a case could also be made to consider the indirect effect of the policy on incentives to innovate, consideration of such an effect would require an assessment of the value of antitrust policy and intellectual property rights policy taken together. A more instructive approach for analysing downstream situations is to consider whether the same action as a contingent part of an original contract would have been acceptable to the antitrust agency. If so, then, permitting the downstream transaction should be seriously considered.