The gist of the argument in this paper – which can be found here – is intriguing, and plausible: tech giants do not compete within itemized relevant markets where they are monopolists. Instead, they are conglomerates that compete three-dimensionally as oligopolists across industries, which is what the author meant by a moligopoly. This blindness of antitrust to competition across markets is likely to lead to mistakes, and should be rectified.
The paper is structured as follows:
- Section I sets out the moligopoly hypothesis and tests it by reference to empirical data. The author begins by reviewing how the competition law literature’s default position is to characterise the tech giants as dominant firms. Competition law focuses on one industry segment – i.e. a “relevant market” – where the investigated tech giants often enjoy unassailable clout, and where substitution by actual or potential rivals is unlikely. For example, Google’s competitive stronghold is search, Apple’s core is its unique ecosystem, Facebook’s moat is its social networking platform, Amazon’s toehold is in online retailing, and Microsoft’s anchor is its operating systems for personal computers.
On the other hand, the technological literature constantly speaks of the intense degree of competition amongst tech giants. Financial data providers often categorise “monopolistic” tech giants as each other’s main rivals; these giants seem to compete with each other outside of the relevant market where antitrust authorities concentrate their dominance investigations; and investigated firms are often subject to competition from other tech giants in those markets where they are dominant. In their self-reporting, tech firms consistently single out “competition” as a risk factor, sometimes the main risk factor they face. In short, the analysis of non-antitrust literature tends to: (i) characterise each technology firm as a conglomerate that offers a mix of products and services; (ii) recognise the distinctly superior position held by each technology firm in one or more core businesses; and (iii) balance thus superior position with a variety of competitive pressures exerted “across industries” by other technology and non-technology firms in actual or future, non-core markets.
- Section II describes how moligopoly competition is multi-dimensional. Technology firms compete as rival sellers of substitute products or services in one or more product or service markets. Besides, technology firms compete to serve market segments that do not exist or to unearth new market footholds. And technology firms also jockey upstream to gain control of entrepreneurial assets, skills and capital, which hold the key to industry disruption. This show how, in contrast to the standard antitrust perspective, moligopolists are conglomerates with a core business which have entered into adjacent businesses where they compete against one another. The reason for this is that moligopolists are terrified of being disrupted, and hence are always in search of new market footholds. This inevitably leads them to veer away from their core business, and experiment in a varying array of fields, sectors and industries, sometimes adjacent, often peripheral – either through R&D or the acquisition of emerging start-ups.
The author calls this behaviour, following a Harvard Business Review article, “competition against non-consumption” – i.e. moligopolists try to provide goods and services to customers who are not now purchasing because existing products or services are “too expensive or too complicated”; or to develop “new market applications for entirely new customers”. This process of conglomerate expansion follows two directions. First, moligopolists engage in entrepreneurial ventures in frontier technological areas close to their core activities. Second, moligopolists also attack the lower end of peripheral markets with the hope of disrupting incumbent players. This process often makes moligopolists mimic each other. When one moligopolist discovers a potential new market foothold outside of its core, other moligopolists tend to follow suit. However, and importantly, this type of head to head, disruptive competition almost never targets the core markets where the technology firms are dominant. Third, technology giants compete over the creation, cultivation and appropriation of units or groups of entrepreneurs – i.e. competition over human resources.
- Section III describes why modern antitrust theory and its applications fail to capture moligopoly competition. From the author’s point of view, mainstream theories of competition law turn a blind eye to rivalry that arises outside of specific product or service markets. Instead, the focus is on supply and demand/prices and quantities in a market, and the operating assumption is that all things outside that market to remain constant. Besides, descriptive models of competition represent rivalry in mono-parametrical terms: firms compete either by setting prices or quantities.
The limitations of the mainstream, neo-classical framework adopted by antitrust are said to be nakedly exposed by competition between tech giants, where competition occurs across markets. These limitations trickle down into applied antitrust theory and lead to the adoption of crude analytical tools, rules and proxies. These unsuitable analytical tools are said to be “prone to generate decisional errors”: “The challenge for antitrust policy is thus to design tools that grasp the competition, or lack thereof, that occurs outside of antitrust relevant markets. Failure to do this will give way to the proliferation of flawed theories or the overgeneralization of convenient models, like for example, multi-sided theory which, albeit a sophisticated and elegant theory, gives rise to broad exoneration claims from antitrust practitioners.”
Ultimately, this points not only to competition law misunderstanding the competitive effects of business practices as a result of its reliance on mainstream economic theory, but also to a need to adopt a humble, evidenced-based approach to novel theories of harm. There are a number of “dystopian” theories of harm doing the rounds, including: (i) infanticide by acquisition (i.e. the acquisition of new, potentially threatening start-ups); (ii) theories of cyclical monopoly in new technological markets; (iii) virtual competition through algorithms and big data. A responsible approach commands the verification of each theory of harm on the basis of the facts and theoretical knowledge available in the light of the state of the art, and a confrontation other potential explanations.
- Section IV proposes specific measures to reroute antitrust policy in moligopoly markets towards the identification of barriers to entrepreneurial resources. This would be pursued as follows. To start, concerns with accumulation of market power in one relevant product or service market should be systematically filtered through a prior examination of the two dimensions of technology competition, in order to establish whether the firm under scrutiny is a moligopolist. If the firm under scrutiny is a moligopolist subject to fierce multi-dimensional rivalry, this should be the end of the antitrust inquiry, even in core-markets, because we are unable to measure cross-market competition but we can be reasonably sure that such competition exists. If the technology company is not subject to multi-dimension rivalry, then moligopoly competition is insufficient and the antitrust inquiry should focus in priority on conduct that elevates barriers on entrepreneurial resources, which are the engine of competition in digital markets.
Comment: I think this is a very good piece. I know the author is using this in a forthcoming book, so I’m not going to dissect it too much. I’ll just point out that it chimes with my previous emails (and the OECD’s work) discussing the limitations of competition law theory – and, by implication, of the underlying economic theories – as regards the new economy.
Naturally, many of these limitations applied to competition analysis of previous industries in the past, which is why I am somewhat disappointed by ‘wait and see’ attitudes to these developments. It is precisely because our available (theoretical) tools seem to lack explanatory power that alternative theories of harm have developed of late. What an antitrust approach that recognises this should do is to evaluate the risks of doing nothing against the costs of over-intervention, and this can only be done if there are theories of how antitrust harm can occur (assuming that such an estimation is even possible, which I doubt given the measurement issues that plague non-price parameters of competition).
In effect, this is what the author tries to do in the last section when he develops his antitrust policy towards moligopolies. However, because he does not elaborate on that theory of harm, I find the proposal to be the least convincing part of the paper. Nonetheless, this is a very good attempt at identifying weaknesses in competition law’s armoury when dealing with the new technological giants, and a valiant attempt at repairing those weaknesses. I look forward to the book.