The European Commission's new draft guidelines on horizontal cooperation1 (to be finalised by the end of 2010) have set-up the legal framework to analyse different agreements between firms that can be exempt from the application of Article 101 against collusive agreements. Besides R&D, production, purchasing and commercialisation agreements, a crucial category analysed in the Guidelines is given by standardisation agreements, which are aimed at defining technical and quality requirements with which certain products or processes may comply. The role of these agreements is quite important in the private sector because of their impact on the innovative activity of high-tech firms (think of the crucial role of standards in Information and Communication Technologies (ICT)) and in the public sector for public procurement (think of the debate on the European Interoperability Framework, currently under discussion at the EU level to set common rules on the procurement of e-Government services by public authorities). Yet while the draft guidelines acknowledge the importance of Standard Setting Organisations (SSOs) in promoting innovation, competition, technology adoption and interoperability, they also stress that the discussions among competitors in SSOs may result in anticompetitive practices.2
How standard setting organisations work?
Representatives of firms participating in SSOs repeatedly meet to reach consensus on the rules that define a standard. In the ICT sector, such workshops are necessary to select the best technologies to include into a technology platform, and they contribute to promote R&D investment, organise the innovation activity in a common direction, and develop new technologies typically protected by patents. The selection of the most efficient standards and the incentive to R&D are therefore the main benefits from standardisation agreements.
After determining the composition of a standard, bilateral negotiations take place between the holders of patents that are essential to the standard and final adopters. Consensus is sometimes reached only several years after the workshops begin, and in the meanwhile firms may start the production phase by following the indications emerging from the conversations at the certification consortium. This implies that patent holders have a strong bargaining position at the negotiation stage and in extreme cases there can be the risk of the so-called hold-up. That is, investment by others is limited ex ante because it could be exploited by the patent holders ex post.
The standards certified by SSOs in the ICT industry include the Wi-Fi protocol, the ADSL and the SDRAM technologies. Although there are many instances of peaceful resolution of issues between parts, a few antitrust cases on both sides of the Atlantic testify that harsh competitive tensions can influence the process of ICT standards’ definition (FTC v. Rambus, FTC v. N-Data, EC v. Rambus and EC v. Qualcomm). As acknowledged by the Guidelines, these tensions arise from the conflicting interests driving firms with different business models, with companies slotting in along a spectrum going from pure innovators to pure manufacturers, with many falling somewhere in-between the two poles.
Pure innovators raise most of their revenue from the technology market; they are interested in having their patented technologies included in a new standard and in gaining the associated licensing revenue. Manufacturers in SSOs are often vertically integrated3; they join certification bodies to achieve coordination among industry participants and have a clear interest in paying low rates for standard technologies.
The different incentives that characterise the licensing stage can alter the outcome of the technology adoption phase and give rise to inefficient outcomes. This happens, for example, when integrated manufacturers sign cross-licensing contracts and exclude from a standard the technology of a pure innovator even though its inclusion would increase social welfare. Exclusion can arise because, from the point of view of manufacturers, cross-licensing preserves product-market and licensing rents, while contracting with pure developers is efficient from the technological point of view but leads to rents’ dissipation because of hold-up. This creates a bias in favour of cross-licensing that is welfare detrimental when the pure developer’s input is very efficient.4
Fixing the patent hold-up and promoting innovation
The normative implication is that standard setting consortia should adopt a policy of early licensing commitments for the essential patents to fix the hold-up problem and let integrated companies design the standard more efficiently. These commitments should not be anchored to vague definitions of fair, reasonable, and non-discriminatory (Frand) terms, which do not reflect the nature of existing SSOs and contradict the same nature of the intellectual property rights (IPRs) (which deliver market power by definition). The market should be allowed to set whatever licensing terms arise out of the early negotiations.
However, so far a policy of early-licensing commitment has received only timid support from SSOs, mainly because of firms' fear of antitrust authorities' intervention. The Guidelines seem to support the employment of an ex ante licensing policy by SSOs when they state that “if standard-setting organisations provide for ex ante disclosures of most restrictive licensing terms, this will not lead to a restriction of competition within Article 101(1)” (# 287). However, they also specify that “[p]rior to the adoption of the standard, agreements by IPR holders on the licensing terms they will disclose also constitute restrictions of competition by object within the meaning of Article 101(1)” (#267). The difference between the two expressions, and in particular between “licensing terms” and “most restrictive licensing terms” (emphasis added), is small but substantial and calls for an interpretative effort that clearly introduces uncertainty on the enforcement of the law by the commission.
The ambiguity of the guidelines generates two risks. First, the potential beneficial effects of ex ante disclosure of licensing terms would likely not be realised because of SSOs’ member firms fear of being punished for anticompetitive coordinated practices.
Finally, and this is a general problem of the part of the draft guidelines concerning SSOs, a dangerous bias against proprietary technologies could implicitly emerge even if it is clear that this would be highly detrimental to innovation and technological progress in ICT sectors and to efficient decisions in public procurement.5
Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreement
Tarantino, Emanuele, 2010, “Technology Adoption in Standard Setting Organizations: a Model of Exclusion with Complementary Inputs and Hold-up”
Sebastian v. Engelhardt, Andreas Freytag, Stephen M Maurer (2010), “Should governments promote open source software?”, VoxEU.org, 29 October.
1 See the “Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreement”.
2 The focus of the Guidelines should be only on Art. 101, but the text often leaves the impression that the Commission is trying to develop a tool to investigate SSOs both under Art. 101 and 102.
3 For instance, the evidence gathered by the FTC in the Rambus case shows a strong presence of integrated companies in JEDEC (the SSO of the case). Moreover, in ETSI, the SSO of the Qualcomm case, the voting rule allowed even a small minority to impose their standard configuration by employing a vote weighting system based on European turnover. This favoured European producers, many of which are vertically integrated (like Nokia and Sony-Ericsson).
4 A policy implication is that, if the technology of an excluded pure innovator is ascertained to be superior, antitrust authorities should closely scrutinise a defense argument based on the pro-efficient effects of cross licensing (see Tarantino, 2010, for a wider discussion). The Guidelines’ propose a number of methods to evaluate the intrinsic technological value of an input (# 284-285), but some of them appear highly inappropriate: an evaluation by independent experts of the value of proprietary technologies is the last step on the “slippery slope” toward price regulation (and thus not very desirable from a policy or technology point of view).
5 For a similar point see also the recent Vox column by Engelhardt et al. (2010), who support the idea that a bias of public procurement toward “too much open source could actually be a bad thing” for the incentives to innovate and allocate resources efficiently in the IT sector.