We are pleased to bring you a report entitled "International Roundtable on Antitrust & Intellectual Property in Standard Setting" by Charles T. (Chris) Compton. The Intellectual Property Committee was one of the five committees that sponsored that June 18, 2002 program. Chris Compton is a partner with Wilson Sonsini Goodrich & Rosati in Palo Alto, California, where he heads the firm’s Antitrust Practice Group.

As this program’s flyer pointed out, standard-setting activities often occur among competitors, raising questions under the antitrust laws. Moreover, standard-setting increasingly "involves the creation and use of intellectual property, raising issues of protection of IP rights versus the needs of the standard, and the potential for anticompetitive behavior based on the use (or misuse) of IP rights." In addition to the kind of "submarine patent" claims raised by the FTC in both Dell and Rambus, compulsory licensing policy has come into play where IP rights may have been characterized as essential facilities.

CRA’s Stanley Besen opened the discussion with a hypothetical to illustrate how standard-setting organizations ("SSOs") should bargain for IP rights. In his hypothetical there are a number of competing technologies, each with a sponsor and all of which are equally capable of performing the same function. None of the sponsors produces the product for which the IP will be used. Each entails different manufacturing costs, but those costs will be the same for all producers of the final product. Finally, there is an industry standards body whose members are the firms that produce the final product, but which does not include the IP owners themselves.

Besen then posed a number of questions. First, should the standards body choose a standard? While fear of antitrust liability might be a deterrent, the answer is certainly yes, because a standard will be essential for developing a market for the final product. Second, and more difficult: Which technology should the standards body choose? The economist’s answer: "The technology with the lowest manufacturing costs." This means, of course, that the standard chosen will not necessarily be the one that is technologically "best."

Besen next asked what rights the SSO should obtain from the winning IP sponsor. The answer, in theory, is "the right to use the winning technology for the term of its IP protection at a license fee determined at the time the technology is chosen." Answering a final question, the license fee should ideally be "some amount between zero and the difference in manufacturing costs when using the lowest and second lowest cost technologies." By way of example, if the per-unit manufacturing cost using the lowest cost technology is $9, and the per-unit manufacturing cost using the second lowest cost technology is $10, the license fee cannot exceed $1 per unit. (Assuming that all users pay the same license fee).

These answers are the theoretical outcome of an auction held by the producers of the final product, in which they exploit the ex ante competition among sponsors to limit the magnitude of the license fee. In the perfect world, the winning technology should have the lowest sum of licensing fee and manufacturing cost. Of course, competing technologies normally are not all equally capable. How good the technology is will therefore be a factor, along with manufacturing costs for the ultimate product, in determining the appropriate license fee. But even the best technology should not be chosen as the standard if it does not offer a low enough license fee that, when coupled with manufacturing costs, results in the lowest cost alternative for producers of the ultimate product.

Varying the hypothetical further: if one technology sponsor is also a manufacturer of the final product, it has an inherent advantage over the others and should be prepared to accept a lower license fee. There may also be an incumbency advantage which should factor into the license fee auction. This will arise where further technology competition is anticipated down the road in which the winner of this round may be favored. A further variation: if the investments in R&D required to develop the competing technologies are not sunk, then the auction should take into account the effect of the license fee on the incentives of sponsors to invest in R&D, perhaps by accepting a higher license fee than could be obtained in an auction. Finally, conflicts of interest and antitrust issues arise if the hypothetical is changed to permit IP sponsors to also become members of the SSO – a not uncommon scenario in the real world.

Willard Tom of Morgan Lewis & Bockius led off this part of the discussion by pointing out the tension between the FTC’s decision in Dell and the European Commission’s 1994-1995 decisions concerning interim standards proposed by the European Telecommunications Standards Institute ("ETSI"), in connection with the private digital mobile radio communications. Dell involved the concealment of intellectual property rights which led to the creation of market power when JEDEC adopted a standard incorporating Dell’s undisclosed IP. In a network industry where compatibility is critical, once product manufacturing begins which relies upon the adopted standard, the holder of undisclosed IP can surface and demand royalties from manufacturers locked into the standard. Moreover, the assumption is that there are numerous comparable technical alternatives that might have been chosen. In that context, the FTC’s remedy in Dell was to mandate royalty-free licensing. That kind of remedy can discourage participation in SSOs if sought too quickly or broadly, and the FTC recognized that pitfall.

ETSI members had also been concerned about "submarine patents," and so adopted a draft IPR undertaking requiring its members to license all technology necessary for a standard unless the IP holder expressly reserved its right not to license identified IP within 180 days from the date standard development was begun. Acceptance of this policy was a condition for membership in ETSI. Numerous ETSI members complained that it was wholly unrealistic to identify, in the first 180 days of a long-term development, what technology they possessed that might be encompassed by that standard. They argued that, in effect, ETSI was making a condition of membership a commitment broadly to license unknown intellectual property. With this advance knowledge that the SSO would "appropriate" technology, there was an incentive not to invest in the IP to begin with. In its preliminary position, the European Commission objected to the ETSI "license by default," which was thereafter abandoned.

ETSI’s rejected policy represents the converse of the Dell threat: possible over-deterrence of the "submarine patent" threat through a policy that effectively appropriates R&D investments made by the SSO’s members – thereby discouraging participation in pro-competitive SSOs. Dell and ETSI therefore frame the issues: what are appropriate licensing policies by the SSO, and what should be the government policy in its enforcement role?

Brian Siritzky of Pillsbury Winthrop then spoke from the perspective of one counseling a client on whether or not to join an SSO, thereby giving up some control over its IP. Siritzky drew a sharp distinction between small, emerging companies with novel IP and the larger, well-established corporations with hundreds or even thousands of patents and patent applications. There is little downside for the larger company in joining an SSO. It can extract licenses as well as information from its smaller competitor members. Further, it will likely enjoy relatively low license fees in any patent pooling arrangement thanks to its large patent portfolio. The small start-up, by contrast, may suffer relatively high costs of participation, and have far less influence or control over the SSO operations and outcome. With only one IP arrow in its quiver, the smaller company has to make a difficult though important choice: if it joins the SSO, it will have to disclose its IP and perhaps provide details or even copies of pending patent applications. And if its technology is adopted and all the relevant technologies are licensed in a patent pool, its share of royalties may be based on the number of patents it contributed and may therefore be quite small. If its technology is not adopted as an integral part of any standard, the start-up has lost much of its opportunity fully to exploit its novel technology. By not participating in the SSO, however, the start-up virtually assures that its technology will not be adopted, and may be swept aside by the competition utilizing the SSO’s standard.

Siritzky therefore asks a number of questions of particular importance to the small start-up: Is the SSO open to all? What is the cost of entry? What are the IP disclosure requirements? What are the consequences for non-disclosure, deliberate or inadvertent? Has it already been determined what the license terms and royalty rates will be? Is there a patent pool arrangement?

Maurits Dolmans then discussed the application of European competition law to standardization activities and associated IP policies and licensing arrangements.

The Internet SSO, "Worldwide Web Consortium," is currently playing out some of these issues. For example, it is considering to exclude from the organization those not willing to offer royalty-free licensing of technology encompassed by the standards. The ETSI IPR Undertaking case illustrate the danger of SSO rules excluding those not willing to accept free or overbroad compulsory licensing as a condition of membership. Such rules can be considered collective boycotts where competitors are excluded from necessary participation. In light of this case and the ETSI/DVSI case, Dolmans suggests the need for the SSO to do an objective analysis of the value and the cost of competing technologies when selecting a standard – "costs" being not only the license royalties, but possibly also the disadvantages of eliminating the open source community as a technology supplier as soon as Internet standards include IP, and the disadvantage of excluding potentially superior IP-based technologies. It is prudent for the SSO to have verifiable data showing the cost-benefit analysis in its choice of a standard. The necessary outcome of the SSO’s terms of participation may be the exclusion of certain patented IP from the standard; but in the absence of an objective and verifiable evaluation, there should be no exclusionary rule at the outset which might later be characterized as an impermissible boycott.

What if an IP sponsor offers to license on reasonable and non-discriminatory terms, but refuses to disclose to the SSO its specific IP, which may be justified in the case of non-public patent applications? If the SSO permits this, no competition law issue need arise so long as the IP owner does not seek injunctive relief preventing implementers from making, using and selling compliant products. The offer to license on FRAND terms constitutes a waiver of injunctive relief, and conditions for exemption under article 81(3) EC would not permit such relief either. The IP owner is therefore limited to a request for damages – equal to the amount of FRAND royalties. This means that he question of what constitutes a fair license term or royalty will have to be decided at the end of the process, by a court if the parties do not agree.

If some judicial body must ultimately determine a fair and reasonable license fee, there are various methodologies available. These include the weighted average cost of development, production and marketing of the technology; same firm price comparisons; market-based pricing; profit comparisons; and the "efficient component pricing rule," also known as the "Baumol/Willig Rule" after the celebrated economists. In sum, incomplete disclosure of IP by an SSO member, if combined with the promise to license on fair, reasonable and non-discriminatory ("FRAND") terms, should normally not give rise to liability in Europe – unless the member knows the patent to be invalid or non-essential and uses its disclosure successfully to achieve adoption of its preferred technology or to torpedo the standard.

As for compulsory licensing, Dolmans opined that this remedy should only arise if the owner refuses to contribute its IP to the potential standard, having actively supported the IP technology as the basis for the standard. Otherwise, compulsory licensing should be an option only where the IP owner is dominant and: (1) the refusal to license is discriminatory; (2) there are acts equivalent to patent abuse, including "excessive" pricing; (3) where the refusal prevents the emergence of a product with new functionalities not otherwise available through the use of technology outside the standard; or (4) where there is a deliberate withholding of IP information with the intention of misleading the SSO, as the FTC alleged in Dell. In sum, under Article 82 EC a compulsory license might be imposed on members or non-members of an SSO if a refusal to license results in a prohibited "abuse of dominance, and only in "exceptional circumstances."

The FTC’s Susan Creighton next offered some comments from the perspective of a U.S. enforcer (with the necessary disclaimer that she did not speak for the Commission or any individual Commissioner). Ms. Creighton’s focus was on unilateral conduct, such as that alleged in Dell (or now in Rambus). She reminded us that standard setting can occur not just in SSOs, but even through government sponsorship or through successful efforts by a single firm in achieving a de facto standard. Ms. Creighton then emphasized that the teaching of Dell is not limited to conduct in violation of the formal rules of an SSO. Antitrust liability may arise under Section 2 of the Sherman Act, for acts of monopoly or attempted monopolization, or under the FTC Act, on a much broader canvas. Dell is not so broad, however, as to suggest that the FTC is contemplating a new set of rules or guidelines for SSOs.

We must recall that antitrust liability for a member of an SSO, as elsewhere, arises with exclusionary conduct having the requisite anticompetitive effects. That conduct may include silence by the member of an SSO for the purposes of deceiving other members of the organization in their adoption of a standard; it may include as well affirmative misrepresentations for the same purpose or even silence in the face of a duty to speak, with the intent to deceive. By definition, there can be no competitive or business justification for deceptions or half-truths. Any of these conduct variations may or not be violations of a given SSO’s rules.

As for Section 2’s requirement of anticompetitive effect, there must be actual or threatened monopolization of a relevant economic market. Also required is a causal connection with that anticompetitive effect – for example, the showing of a large difference between the value of IP in an ex ante auction versus the ex post reality. There is, of course, no causation if the IP holder did not even participate in the SSO’s decision adopting a standard. In commenting upon the global nature of these issues, Ms. Creighton made the provocative observation that there might occur in future a situation in which U.S. authorities might seek to block efforts to enforce patents held by a European SSO member whose conduct as such violated the standards suggested in Dell (and Rambus).

Howard Morse next spoke, noting that adherence to the RAND (reasonable and nondiscriminatory license terms) standard alone was insufficient to stop all anticompetitive conduct by members of an SSO. RAND, moreover, is often meaningless to clients, though there might be normal or customary royalty rates in certain well-established industries. Morse urged that it should be generally lawful for SSO members to negotiate and determine licensing terms before the standard is fixed. Ex ante negotiations should not be a per se exercise of monopsony power, though the rules of many SSOs prohibit the practice. BMI, Northwest Wholesalers and other cases suggest that the rule of reason should apply given that the purpose of the SSO and its rules is to facilitate the creation of a new product, which is procompetitive. We need a clear government statement of policy, however, to encourage or at least provide latitude for ex ante determinations of reasonable licensing terms, rather than forcing reliance on the vague RAND standard. Further, the patent ambush issue as raised by Dell and Rambus should only be a concern if there are in fact alternatives to the technology chosen for the standard.

Morse then spoke about patent pools. These can certainly be an efficient vehicle for promoting the development and dissemination of new IP-based products, but they can facilitate collusion as well, particularly in foreclosing downstream competition. In the DOJ’s Business Review Letters on MPEG and DVDs, the agency emphasized that those patent pools mandated licenses to all, based on the RAND standard, and that the royalties would be a "tiny fraction" of downstream prices and "small relative to the total costs of manufacture." With respect to what constitutes permissible discrimination in royalties, we should note that the DOJ in those Business Review Letters did permit different royalty rates for different categories of licensees. In Morse’s view, a licensing term should only be discriminatory in violation of RAND if it permits a competitor to foreclose downstream competition.

R. Hewitt Pate, Deputy Assistant A.G. at the Antitrust Division, spoke next. He commented that the agency cannot say much at a substantive level in this debate, but will continue to focus on process and procedure rather than, for example, what the proper royalty amount should be in a given situation. The DOJ will ask, for example, whether participants with market power have the ability or intention to "capture" the SSO for their own purposes, either through their ownership of IP needed by the SSO or in more traditional ways.

Further, the DOJ is skeptical about any legislated exemption from the antitrust laws for SSOs and their members. Even the issuance of new guidelines in this area seems problematic, given the wide variance in the number of different factual settings and very complex issues dealt with by SSOs. Instead, the Business Review Letter process has proved a good approach, as reflected in the MPEG and DVD consortia. The DOJ is committed to making this process both faster and more useful to corporations engaged in standard setting. (In this connection, Pate recommended Mark Lemley’s paper at the recent George Washington University Conference.) Pate did not disagree with Howard Morse’ view that application of the per se rule is inappropriate in testing ex ante decisions by SSO members on reasonable licensing terms. Even in a rule of reason analysis, it is possible to have a large number of major players with collective market power raising serious antitrust issues. The DOJ is going to want to know under what circumstances and by what terms are the parties collectively setting, up front, RAND license terms.

Mark Powell of White & Case next spoke on the European perspective of RAND. With fairness the E.C.’s touchstone, the proper acronym is actually "FRAND". At the end of the day, there will be a great deal of uncertainty as to appropriate terms, since royalties must properly balance the interest of the licensor and licensee. After ETSI, SSOs in Europe are understandably reluctant to get into ex ante determinations of FRAND terms. Moreover, the addition of such ex ante license negotiations may substantially prolong the already extended standard setting process. The European Commission, for its part, wants to leave to IP experts the determination of what constitutes FRAND in a given case. Powell sees no need for the E.C. to draw up additional guidelines. Rather, the EC must continue to respond to up-front licensing requirements on a case-by-case basis, as more is learned in this delicate balancing of interests between licensee and licensor.

Further reflection on the European perspective came from Luc Peeperkorn, of the Commission’s Competition Directorate, who stressed that this is an area where policy is not well developed yet. The views that he and his colleagues Magdalena Brenning and Nick Banasevic expressed were therefore very tentative and only theirs, not those of the Commission.

Peeperkorn thought that the area is also unlikely to develop quickly, as competition authorities have a dislike to get involved in pricing issues. There are some good reasons for this: the complexity in calculating what a reasonable or excessive price is; the asymmetrical information between company and authority; the competition authority not being equipped for price regulation in the same way specialised regulators are for their particular sector; and the on-going demands of price monitoring.

The Commission has therefore in general shied away from such exercises, instead expressing the principle of RAND and concentrating on procedure, such as asking for the possibility for arbitration to be foreseen and the need to keep separate cost accounting per activity. The Commission also has been involved in settlements where all parties are satisfied with the resulting price.

Peeperkorn thought a useful distinction could be made between two scenarios: Firstly, the situation where before the standard is set different technologies compete and are more or less able to become the standard. In this Article 81 situation the competition between the technologies should make it possible for a competition authority to avoid having to set the price. With the proper standard setting procedures in place it should be possible for the parties to arrive at a competitive price that can be regarded as reasonable. He stressed that where possible the price should be set before the standard is chosen, so as to avoid any increase in price as a result of the (extra) market power obtained by becoming the standard technology. He added that in such a situation it may be less clear to what extent or how the principle of non-discrimination should apply. It is clear that within the same market different users of the technology should pay the same price. But what about Ramsey pricing between different markets, asking a higher price/royalty to those companies operating in a market with an inelastic demand as compared to those operating in a market with an elastic demand? As it would be agreed before the standard is set and in competition with other technologies, could this not be pro-competitive?

Secondly, Peeperkorn described a scenario where there is only one technology available—i.e., there are no alternatives, or they are so inferior as not to be in the same market. It is this Article 82 situation in which it is really difficult to apply RAND. The IPR holder is dominant, at least in the technology market, even before the standard is agreed upon. Having to agree on the price/royalties before the standard is set may not be sufficient, as more or less the same high price will result. Becoming the standard will, in addition, increase the entry barriers to the technology market.

The question of what is reasonable in such a case is the same as asking what price is not excessive. Under Article 82 the Commission has the possibility, yes even the obligation, to prohibit excessive prices charged by dominant companies. For the reasons mentioned above, however, the Commission has never been too keen to apply this prohibition.

In the excessive pricing cases and court judgments of the past, which in general did not relate to IPRs or standard setting, the benchmark against which the excessive character was measured has always been cost. Most experience gained so far by DG Competition relates to access pricing for essential facility type investments like telecom networks and ports. The Commission is generally intervening in those cases on the basis of a complaint—usually because the owner of the facility is also active downstream and is allegedly using the pricing for its facility to push out downstream competitors. In such cases the facility owner is requested to apply RAND and base the price on cost. It is usual to request a certain separation of accounting between different activities so as to enable cost assessments. However, in these port cases it is generally left to the parties to find a mutually agreeable settlement. In the telecom sector the Commission has pushed for application of what is called the Long Run Average Incremental Cost, a kind of forward looking replacement cost principle. It is forward looking to disallow the old state monopolies to charge entrants for the incumbents’ inefficient investments of the past. Here the application is left to the national regulators.

The application of a cost-based rule may, however, be particularly difficult in an IPR related case. In addition to the normal data problems there may be the extra problem of how to account for R&D projects that were not successful but that need to be sponsored out of the successful projects’ returns. It seems clear that a marginal cost pricing rule will therefore not be sufficient, especially as the marginal cost of providing a license may be close to zero. A fully allocated cost rule, which allows coverage of all costs—including overhead costs, joint costs and failed projects in the same area—is the minimum that needs to be employed.

Speaking next from the European Commission was Magdalena Brenning, also from the Commission’s Competition Directorate. Within the European Commission, there are no precedents on late or no IPR disclosure in an SSO context. Instead Ms. Brenning stressed the importance of SSOs taking the EC antitrust rules into consideration in their internal rules; only in specific instances when these rules fail, should parties turn to the Commission. There are two particular reasons for this.

The first reason is of a practical and policy nature. Indeed, the Commission does not have the technical expertise to determine whether any chosen technology is the best one. It is therefore likely to focus on procedure instead of substance. Further, in light of its limited resources, it is not appropriate for the Commission to be drawn into a large number of standardization disputes on a case by case basis. Finally, due to the long lead-time of the Commission's procedures, the standard resulting from a patent ambush may be firmly entrenched before the Commission can take a position.

The second, and potentially more important reason relates to the nature of the legal instruments available to the Commission. The Commission can always intervene to put an end to an antitrust infringement, but it might not be in a position to grant the remedies sought by a complaint. When faced with a "patent ambush" problem in an SSO context, complainants may therefore be better off in the national courts under theories of breach of contract or unfair trading statutes.

Article 81 applies to collusive behavior, i.e., at least two firms must be involved. In a patent ambush situation, the problem arises of imputing one member's action to the whole organization. There may, however, be Article 81(1) liability arising out of an SSO agreement that excludes competitors from the technology market as a result of a non-transparent procedure. Article 81(3) may also be a basis for ordering compulsory licenses on FRAND terms (but probably not free of charge, as requested by some complainants).

As for Article 82, one must recall that unlike U.S. law, liability arises only for abuse of dominance, not anticompetitive creation thereof. Showing abuse may be problematic in a patent ambush context. The EC, moreover, has no equivalent to the Federal Trade Commission Act, which was the statutory basis for liability in Dell. To demonstrate this point: where a non-dominant SSO member intentionally conceals a patent that reads on the ultimate standard, and thereby becomes dominant as a result, is difficult to say liability arises under Article 82. Similarly, the subsequent assertion of IP rights against other members of the SSO may not constitute abuse of dominance, since the patent itself was properly granted in the first place. The only apparent area for Article 82 liability might arise if the IP holder applies unfair license terms, engages in excessive pricing or refuses to license in order to monopolize a downstream market.

The SSO rules may nevertheless incorporate certain safeguards. A mandatory arbitration clause would seem prudent for breaches of internal rules, since that would allow quick action before the standard based upon the patent ambush has been established. The SSO could also set functionality standards rather than those based upon specific IPRs, since several patents may be able to meet the same function, thereby making ambush more difficult.

A brief question and answer period followed these outstanding presentations by this international panel of standard-setting experts. The questions, the presentations, and the next-day’s announced action by the FTC against Rambus, all made clear that the standard-setting organizations, which generally serve an important and pro-competitive function, are increasingly raising troublesome antitrust issues. Those attending or listening in no doubt found this session extraordinarily useful.

1. The sponsoring committees are the International Antitrust Law Committee of the ABA Section of International Law & Practice, together with the Intellectual Property Law Committee and the International Antitrust & Foreign Competition Law Committee of the Section of Antitrust Law, the Committee on Antitrust Matters of the Section of Intellectual Property Law, and the Committee on Antitrust Law of the Section of Business Law.

***********************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************

Chris Compton is a partner with Wilson Sonsini Goodrich & Rosati in Palo Alto, California, where he heads the firm’s Antitrust Practice Group.

H e may be contacted at 650-320-4882,ccompton@wsgr.com.

This article is intended for informational purposes only and is not intended as legal advice or as a substitute for legal consultation in a particular case or circumstance.

***********************************************************************************************************************************************************************************************************************************************************************************************

AT-IP Report is the electronic Newsletter of the Intellectual Property Committee of the American Bar Association, Section of Antitrust Law. The views expressed in the AT-IP Report are the authors' only and not necessarily those of the ABA, the Section of Antitrust Law or the Intellectual Property Committee. If you wish to comment on the content of this Electronic Newsletter, please write to the ABA, Section of Antitrust Law, Attention: Intellectual Property Committee, 750 North Lake Shore Drive, Chicago, Illinois 60611. Copyright 2002 American Bar Association.