Highlights‎ > ‎

Intel to live "the quiet life of a monopolist"? (Back on the AMD v. Intel lawsuit)

posted Mar 23, 2012, 11:44 AM by Julien Pillot   [ updated Mar 24, 2012, 7:52 AM ]

On March 4th, AMD (Advanced Micro Device) completely ceased its activity as a CPU producer (Central Processing Unit – that is to say microprocessors for Personal Computers), putting an end to a divestiture process started in March 2009 when AMD decided to spin off manufacturing business to GlobalFoundries (a joint-venture with ATIC, an investment company formed by the government of Abu Dhabi). Although AMD now plans to focus on the sole chip design, it is very likely that most of its R&D efforts will now concern highly profitable technologies rather than commodities such as CPUs. Such a move would conclude an almost forty years-competition between AMD and Intel (which respectively hold 18.8% and 80.3% of the 2011 CPU global market). In a nutshell, the risk of switching from an oligopoly market structure to a monopoly one cannot be ruled out, at least on a medium-term basis.  

Basically, the competition among the two most important CPU manufacturers largely exceeds the sole market frontiers. Indeed, AMD has a long history of litigation with Intel. During the 2000’s, AMD sued Intel in many commercial areas for unfair competition. According to AMD, as the single operator technically able to serve all the market, Intel took advantage of higher production capacities to offer PC manufacturers retroactive fidelity rebates that could not be turned out. While Intel settled legal disputes in the US through a private transaction with AMD (2009, see the humoristic cartoon at the bottom of that bill), in Japan (2005) or in the EU (2009), Intel was found to have abused of its dominant position by engaging in naked restraints likely to leverage its market power from the uncontestable side of the market to the contestable one. Interestingly, but perhaps ironically too, in sanctioning such an exclusion strategy, both the Japanese antitrust authority and the European Commission aimed at preventing a potential market foreclosure…

The vigorous competition in which AMD and Intel are engaged has been the topic of a substantial academic literature. Amid all these articles, a very recent work by R. Goettler and B. Gordon in the field of New Industrial Economics particularly caught my attention (R. Goettler and B.Gordon, “Does AMD Spur Intel to Innovate More”, Journal of Political Economy, 119(6), December 2011, 1141-1200).  In this paper, in which the authors’ purpose is to estimate an equilibrium model able to explain the relationship between market structures and innovation, we can find the very interesting scheme replicated here (click to enlarge). As shown in this figure, a certain degree of market foreclosure would be profitable both in terms of innovation rate and of consumer welfare. Monopolizing a significant share of the CPU market allows Intel to invest in R&D without taking advantage of what Adam Smith used to phrase “the quiet life of a monopolist”. From the consumer perspective, according to this scheme, such a situation is also profitable, not because clients benefits from a price competition, but rather because final users can enjoy continuous technological progress in CPUs.

At this stage of the analysis, and assuming that AMD plans to exit the CPU market on the long term, a series of remarks has to be formulated. First of all, if the rationale of the 2009 European Commission decision was to fine Intel for market foreclosure then, on a static-basis, the Commission has to be blamed for the lack of a concrete effect-based analysis on the consumer welfare. At the opposite, if the very essence of the Commission decision in Intel was to prevent predictable market monopolization – and that is in what I believe – then the Commission was probably right in pronouncing sanctions against Intel…

Nevertheless, even in this second scenario (less critical vis-à-vis the Intel decision), the European Commission and the quoted American authors' expertises are partly misguided because they both fail to consider two essential characteristics of the CPU market in particular, and most generally of the whole computer industry. In the one hand, CPUs are not a stand-alone product but are part of an entire computer architecture made of several interconnected and interdependent components bound to evolve in a symbiotic way (the ecosystem metaphor is not overrated here). As the heart of a composite system, CPUs are at the Operating System disposal to achieve more or less complex tasks, appealing other devices (motherboards, graphics and audio cards…) in the most efficient way to maximize the final user experience. Consequently, approximating consumer welfare and innovation rate on the sole CPU market-basis is complete nonsense, both from economic and technological perspectives. In the other hand, the computer industry changes continuously, as well as consumers’ habits. While experts worry about the future of the traditional PC industry (+ 0.5% global growth in 2011), the “nomad” side of the market (especially tablet PCs and smartphones) knows an exponential growth drove by tough technological competition and rapid product obsolescence. Due to numerical convergence, the frontiers between the two sides of the market have never been as poorly-defined. Such a trend could be likely to make a dominant position in the nomad CPU market much more profitable than in the PC market ; now Intel is only the fourth global player in this very strategic market (just behind Qualcomm, Samsung and Texas Instrument). As far as one can judge beforehand the likelihood of such scenario, a shrewd industrial economist (and a meticulous competition authority) would be well-advised to monitor Intel’s next (pricing) moves on the PC CPU market in order to prevent a potential leveraging strategy.