This article, available here, analyses situations where platform operators design their platforms in a way that is liable to exclude intra-platform competitors. Exclusion in intra-platform markets require certain intricacies that existing theories of harm in antitrust law do not anticipate; thus, applying those theories unyieldingly is liable to cause confusion and result in judicial error. Authorities must formulate policies that detect anticompetitive exclusion without deterring innovation, and apply that policy consistently across comparable cases. Existing cases reveal that some authorities and courts have been taking a sensible approach to intra-platform exclusion, whereas others, especially in the EU, have shown a tendency to protect excluded intra-platform firms at the expense of consumer welfare.
The paper is structured as follows:
Section II defines software platforms and describes platform owners’ relationships with third-parties that distribute services through their platforms.
Software platforms are code-based infrastructures that facilitate exchanges and transactions through the creation of one or multiple downstream ‘intra-platform’ markets. Through a platform, users can transact with each other and create value that would that would otherwise not materialise due to prohibitive transaction costs. Due to network effects, platform markets are prone to tip in favour of monopoly. Whether tipping occurs in practice depends largely on inter-platform compatibility: generally, the less inter-platform compatibility, the more likely the platform market is to tip. Economies of scale in production reinforce the tendency of software platform markets towards concentration.
Due to the inevitability of concentration in platform markets, analysing competition within platforms is just as important as examining competition between platforms. Because platform operators are highly dependent on their intra-platform participants for success, they must manage their relationships carefully to maintain platforms that are attractive to both users and downstream firms. Additionally, platform owners must carefully consider the impact that entering into one of their intra-platform markets could have on their rivals’ incentive to innovate for the platform.
Section III analyses the circumstances in which intra-platform exclusion is likely to take place
It is generally in platform owners’ interests to maximize their platforms’ compatibility with third-party components to ensure that intra-platform firms can succeed on the merits of their products (which increases platform value). However, platform owners occasionally make design decisions that reduce certain intra-platform firms’ ability to compete within the platform. However, deciding to stifle intra-platform competition may deter innovation within platforms. One might wonder, therefore, why a platform owner would have the incentive to foreclose intra-platform firms.
Even though there are many pro-competitive factors that could motivate intra-platform exclusion, there are limited circumstances in which a platform operator could have the incentive to foreclose more efficient firms. For example, a platform operator may have the incentive to leverage its market power into a downstream market where that market is subject to scale economies. To illustrate, say a dominant platform operator bundles a complementary product like a mapping service with its platform, which users can access through the platform’s interface and by other means. The mapping service market is characterised by fixed production costs. The dominant firm’s mapping service competes with a more efficient mapping service, which is also available by other means. If the platform operator contractually obliges its users to use only its mapping service when they need directions, the rival mapping service may be unable to cover its fixed costs and may exit the market. With decreased competition in the mapping service market, consumers that only demand a mapping service turn to the monopolist’s offering. Thus, the platform operator extends its market power into the mapping service market and earns monopoly profits there too. Consumers are harmed because they are denied access to a superior mapping service.
Furthermore, a monopolist can tie an inferior good to the monopolised good in order to preserve profits in the monopolised market. This strategy would work as follows. During a first period, the monopolist in the primary market ties to deny a more efficient competitor in the tied product market sufficient scale to compete. In a second period, both firms can again produce the tied product, but the competitor can also enter the monopolised product market. By tying, the monopolist ensures that the competitor cannot profitably enter the market. The monopolist sacrifices some profits in the first period in order to preserve its monopoly over the primary product in the second period.
Section IV proposes a framework for analysing intra-platform exclusion.
According to this framework, intra-platform foreclosure may occur if the following are present: (i) a dominant platform which gives rise to at least one intra-platform market; (ii) an intra-platform market in which competition has allegedly been foreclosed; (iii) the presence of intra-platform switching costs that allegedly cause the exclusion; and (iv) an argument by the platform owner that its behaviour is justified by reference to some benefit of integration. This model can be broken down into three fundamental elements.
Firstly, it is important to consider the defendant’s market power over the distribution of the intra-platform product market that is allegedly foreclosed, as only this speaks to the ability of the platform operator to exclude. To possess the power to exclude, the platform owner must have the ability to deny its intra-platform competitors sufficient scale to profitably stay in the market. This is unlikely to occur if there are other platforms through which complementary products can be distributed. Authorities must therefore distinguish between pricing market power and distributional market power.
Secondly, authorities should consider the degree of costs that users face in switching to competing downstream options. If consumers face low intra-platform switching costs, they will thwart any attempt by the platform owner to impress upon them a feature they do not prefer by simply adding a different feature to their platform bundle. On the other hand, the larger switching costs become, the greater benefits a rival’s option must bring to encourage users to switch; otherwise, the platform design will exclude competing options. This is not a binary assessment, as has been the case in many traditional (especially tying) cases. For instance, a platform owner might close part of its platform system off from competition as a matter of platform design (e.g. Sire for Apple). Even if platform owners do allow users to switch to alternative products within the platform, certain features of platform markets can raise rivals’ costs of reaching users. Firstly, users may be reluctant to switch from a preinstalled option if she has to pay even a trivial amount for a rival option. Secondly, there are costs associated with seeking out and evaluating competing options. Thirdly, users may face learning or data-porting costs if they decide to switch. Fourthly, pre-installed components that are integrated with the underlying platform or other pre-installed components are likely to increase the cost of switching to competing options. Lastly, pre-installation can deter users from switching if the monopolist’s option itself constitutes a platform characterized by network effects.
Thirdly, it is important to consider the procompetitive rationales for increasing switching costs for users. Platform owners often have the incentive to bundle their own downstream products and services with their platform code (or otherwise increase intra-platform switching costs) because doing so renders the platform more valuable to consumers. It is important not to sacrifice these consumer benefits just because they incidentally make it harder for intra-platform firms to compete. A court or authority should therefore accept evidence that the platform design plausibly brings some consumer benefits in terms of functionality, productivity, or speed of service. These benefits should be specific to the platform owner’s design decision (i.e. not be replaceable by a third-party option that would bring greater benefits). In a similar vein, integration of complementary components can produce functionalities or quality of service for end-users that cannot prevail if the firm’s components are distributed on a standalone basis or substituted for third-party alternatives. Conversely, efficiencies that superficially support non-liability may be irrelevant if they support software bundling generally, but are not specific to the platform owner’s option.
Section V examines cases of intra-platform exclusion in software markets.
The author begins by reviewing the U.S. Microsoft/Netscape case. Microsoft was accused of protecting its then monopoly position in the market for Intel-compatible personal computer (PC) operating systems (OSs) by staving off the threat posed to it by Netscape’s web-browser, Navigator. Navigator was ‘middleware’ that allowed software developers to create cross-OS applications without incurring the cost of porting them to other OSs. Microsoft believed that developers would flock to write for Navigator instead of Windows, which threatened to ‘commoditise’ the underlying OS and therefore undermine the indirect network effects that entrenched Microsoft’s monopoly position. The DC Circuit Court’s decision in this case laudable: Microsoft instigated some intra-platform switching costs solely for exclusionary purposes, and was ordered to reduce those costs accordingly. For those practices where increased switching costs were justified, the court absolved Microsoft of liability.
The second case is the EU’s Microsoft / Windows Media Player case. The European Commission found that Microsoft illegally tied Windows Media Player (WMP) to its dominant Windows OS platform, which foreclosed competition in the market for media players. As with US v Microsoft, distributional market power was satisfied because Microsoft held a monopoly position in the market for operating systems, which were necessary to use media players. The author argues that the tying test applied in the EU took insufficient account for the efficiencies generated by software bundling. Such an approach deprives dominant firms of the opportunity to show that their product designs are beneficial to consumers and explicitly relieves authorities of the need to examine the direct effects of their allegedly exclusionary designs. Instead, a likelihood of foreclosure is presumed directly from the act of bundling complementary components with a platform.
Finally, the author discusses the various Google cases. These cases follow various allegations that Google has abused its dominant position in the market for general internet search services by favouring its own vertical search services, like Google Shopping, Maps and Flights, on its search results pages. Because users tend to click predominantly on prominently positioned results, the product design results in decreased traffic to Google’s vertical search rivals. Allegedly, Google’s intra-platform rivals depend on Google search to distribute their products and they must overcome the intra-platform switching cost that users face (i.e. their bias towards prominently displayed results). In the US, the FTC ended its enquiry once Google offered ‘ample evidence’ of ‘plausible procompetitive justifications’ for the bundling of its intra-platform offerings with its search results pages. Likewise, the English High Court considered that the bundling of search and mapping services were justified by efficiency justifications.
In Google Search, the European Commission found that Google’s prominent display of its shopping service had the effect of putting competing suppliers of comparison shopping services at a competitive disadvantage. The Commission also found that an update to Google’s relevancy algorithm had the effect of demoting links to rival comparison-shopping websites in Google’s natural search results, whereas Google Shopping was not prone to being demoted. The Commission’s analysis not only seems to conflict with the exclusion analysis undertaken by the English High Court as regards mapping services, but also failed to consider Google’s efficiencies properly. Ultimately, the author broadly considers that: ‘While US authorities and courts have largely taken a sensible, consumer centric approach to alleged intra-platform exclusion, their European counterparts have ignored vital, unique elements of intra-platform exclusion and thus have produced flawed results.’
This is an interesting piece, particularly in its attempt to develop a framework applicable to intra-platform exclusion in general. While I am broadly sympathetic to the argument that exclusion (or, more accurately, foreclosure) should be required for infringements to occur, I always found that this raises an important question: what do we mean by foreclosure, particularly in the absence of clear pricing and efficiency metrics? This fundamental issue cuts across the proposed approach, and would ultimately be a matter of evidence, thereby potentially undercutting the potential for certainty and coherence that the author is aiming for.