At the moment, the strategy of the major tech platforms seems to be to deny any bias and malicious self-favoring. The facts, however, do not speak in their favor. The question, therefore, isn’t whether Big Tech is guilty of self-favoring—it is—but whether self-favoring comes at the expense of consumers and, if so, how to deal with it.
Without resorting to an overly simplistic statement, it remains difficult to evade the perception that the digital economy is prone to self-favoring and other forms of anticompetitive behavior.
Several allegations of thwarting competition have been thrown against tech companies such as Amazon, Facebook, Apple, and Google in recent years, with self-favoring being one of the most prominent. These allegations are partly the reason why the House antitrust subcommittee launched an investigation into the market power of the four major digital platforms and confronted the CEOs of these companies last month.
Except for Amazon, which acknowledged during the Congressional hearing to be investigating the matter internally, it appears that the strategy of tech platforms is to deny any bias and malicious self-favoring. However, the facts do not speak in their favor.
Google is perhaps the best example. When navigating the internet using Google, in addition to the organic blue links, the search engine yields a number of Google offerings at premium places unavailable to competitors, making the self-favoring rather obvious.
In light of this, the relevant question should not be whether Big Tech is guilty of self-favoring. It is. Rather, the question should be whether self-favoring comes at the expense of consumers and —if so—how to deal with it.
While authorities in Europe appear to have embarked on a combination of regulatory amendments and antitrust investigations into the conduct of digital platforms, their US counterparts have so far confined themselves to the latter. This, of course, raises the question of antitrust’s ability to police against self-favoring, which was the very purpose of the Congressional hearings.
This question is particularly pertinent, given that the many allegations could easily give the impression that Big Tech is Bad Tech, fundamentally at odds with consumers’ interests, and that antitrust intervention is easily available and only a matter of priorities.
Unfortunately, both of these assumptions rest on a false narrative. Not only is it too simplistic to label Big Tech firms as bad, given that billions around the world frequently use their products and services, it would also be false to consider self-favoring as per se detrimental to the interest of the consumers and clear antitrust infringement.
The antitrust investigations into Google in the US and the EU are a case in point, providing a basis for a critical evaluation of self-favoring in the digital economy and of antitrust as a possible remedy.
Self-Favoring as an Antitrust Infringement
The scope of antitrust provisions differs across jurisdictions. Generally, it reserves condemnation for actions that are not only anticompetitive—e.g., by foreclosing competitors—but also outside the scope of competition on the merits. Even companies with an almost-100 percent market share are allowed to offer better or cheaper products at the expense of competitors, potentially eliminating the competitors.
Consequently, there is no theory of harm than can accommodate self-favoring, unless coupled with something else. This can be illustrated by the cases of Google Shopping and Google Android, delivered in 2017 and 2018 (respectively) by DG COMP, the premier antitrust enforcer in the European Union.
Regardless of being on appeal, the Google Shopping and Google Android cases represent the only examples of self-favoring elevated to actual antitrust infringements, beyond mere allegations and investigations, and can thus provide some guidance on the matter.
The Google Shopping Case: More About Discrimination and Search Bias
In Google Shopping, DG COMP took issue with the way Google ranked results following a search, as this involved an (illegal) advantage to its own comparison shopping service. For this, DG COMP doled out a €2.42 billion fine against Google in 2017, accompanied by an order to stop the practice within 90 days.
The theory of harm advanced by DG COMP in the Google Shopping case did not involve self-favoring but rather search bias and discrimination. Namely, that Google’s offerings were allotted premium placements on the top of the yielded search page and came enriched with pictures and other information, while competitors were referred to the organic blue links at the bottom.
This was deemed anticompetitive and illegal, not by virtue of the algorithm used to rank possible search results placing some before others, but because the algorithm was not applied to Google’s offerings. This preferential treatment translated into visibility, as end-users were unlikely to click on anything below the first few hits, securing Google a competitive advantage detrimental to the consumer’s long-term interests.
However, because the Google search engine is free to use, Google’s business model partly relies on pushing other Google products to allow recoupment. The decision therefore stands for an obligation to secure ranking neutrality, even when ranking is unprofitable and recoupment must be secured from other offerings.
A rather expansive reading of the above concepts explains why the US Federal Trade Commission (FTC) in 2013 had decided not to pursue the matter further. It also explains why the outcome of Google’s appeal cannot be considered a foregone conclusion, as the consumer harm is not obvious.
Google Android Involves Tying
In Google Android, DG COMP intervened against Google’s policy for licensing the Android smartphone operating system (OS), including levying a fine of €4.34 billion on the company in 2018.
An OS is required to operate smartphones and is installed by manufacturers on their devices before they are sold to end-users. While some manufacturers like Apple use their own OS, most prefer to source from a third party. In contrast to other companies’ software—e.g., Microsoft—the Android OS is available for free, provided that the manufacturer agrees to pre-install certain Google apps, including Google Search and Google Chrome. This policy secured Google’s offerings a competitive advantage over possible alternatives by the virtue of coming preinstalled.
Once again, the theory of harm advanced was not self-favoring but another form of abusive practice: in this case, tying. However, it wasn’t tying in a traditional manner. Normally, tying strategies are deployed for the purpose of leveraging market power between adjunct markets, gradually monopolizing these markets. DG COMP did not refer to this motive but rather to a strategy directed at countering the effect of the gradual migration of internet searches from (stationary) PCs to mobile devices.
This development potentially challenged Google’s control over internet searches, making it pivotal for Google to secure a place on mobile-based internet searches. According to DG COMP, Google’s initial investments—and its later promotion of Android—could be attributed to this. DG COMP was even inclined to attribute Google’s willingness to invest millions into the Android system to this fear. Evaluating the alleged motive is difficult but it does appear plausible. Mobile-based internet searches already account for more than half of all internet searches.
While providing a plausible motive for acting in an anticompetitive manner, DG COMP failed when it came to explaining how Google’s pre-installation policy is detrimental to consumers’ interests. Not only were the alternatives not licensed to third-parties (Apple) or only available against a fee (Microsoft), but as explained by DG COMP itself, Google invested very substantial resources to secure broad penetration of Android.
While acting out of self-interest, Google’s actions and policies had also secured cheaper smartphones, helping to transform smartphones from a luxury product for the few to a cheap commodity for all. This does not challenge the legal merits of the decision but does question its wisdom and reasoning, giving much to discuss on appeal, as DG COMP’s notion of consumer welfare appears very narrow.
Neither of the Google Cases Condemn Self-Favoring Directly
As can be seen from Google Shopping and Google Android, the cases were built on traditional forms of anticompetitive behavior—discrimination and tying, respectfully. This does not make the rendered conclusions invalid but does indicate that either self-favoring must be accommodated to a traditional theory of harm, or the possibility of expanding the concept of abuse must be explored.
In particular, the Google Shopping case appears to be based on an expanded reading of the concept of abuse and what can be demanded of companies with very strong market positions. This gives some ground to Google’s claim of being held accountable not only beyond what has been previously considered illegal by case law but also in a manner that fails to balance the different interests, including those of consumers. The FTC’s 2013 decision to not pursue charges against Google was largely rationalized by these concerns, combined with a recognition that any decision could be overturned on appeal had the case moved forward.
Naturally, other examples of self-favoring might be more straightforward compared to the Google experience. DG COMP’s new investigation into Amazon appears to involve potential abuse of sensitive data collected from independent retailers selling products on Amazon’s marketplace. Allegedly, Amazon uses this data to push its own products at the expense of alternatives offered by competitors.
Very little is publicly known about the Amazon investigation, making it difficult to offer comments on its legal merits. Amazon CEO Jeff Bezos did acknowledge during last month’s Congressional hearing that the company was investigating the matter internally, giving some support to DG COMP’s interest. Further, it would not be entirely unfair to claim that Amazon, by accepting a commission for the privilege of being offered on its platform, also has a fiduciary obligation to refrain from self-favoring. This fact not only sets the Amazon case apart from Google, but also makes the abusive practice more apparent and straightforward.
Self-Favoring as Detrimental to the Consumer Interests
As detailed above, the acceptance of self-favoring as an antitrust infringement is subject to some qualifications and only possible within a narrow set of circumstances—typically, if coupled with something malicious or unfair. Even then, it is not evident that self-favoring itself is detrimental to the interest of consumers.
Google, for instance, is not an altruistic company offering products for free. On the contrary, consumers pay for the rendered search services by giving up their personal information, which is then monetized by Google.
Google’s business model rests on recoupment against monetization of data blended with the selling of adjunct services. Deprived of access to this, Google might not have offered these services, nor maintained them, or even developed them in the first place. Essentially, the same story could be presented in defense of Facebook, which offers a platform for communication between billions of people in exchange for monetizing personal data.
A counterfactual analysis, often useful in antitrust investigations, could challenge the consumer benefits of intervention. It might even indicate that this is predominately beneficial to competitors, ultimately shielding these from competition.
Granted, Google does direct consumers in the direction of Google offerings, but this follows from the manner in which Google chose to present results: rather than just providing a list of organic blue links, Google blends in pictures and information on things like prices and availability, enhancing the consumer experience. So once again, what looks like a solid case from a legal perspective could, after closer inspection, be debated.
Should Self-Favoring be Policed by Antitrust?
While self-favoring can be anticompetitive, and antitrust can be used to police against this within a narrow set of circumstances, intervention can also be anticompetitive if it is made against a company that has successfully developed superior products and services that benefit consumers. If that intervention is done for the benefit of companies that have not delivered on the latter, that would give the appearance of punishment for being successful.
Applying antitrust to the digital economy, particularly to cases of self-favoring, commands some diligence or finesse, contributing to antitrust enforcers’ traditionally limited interest in pursuing cases.
The FTC’s 2013 decision to not pursue charges against Google is a prime example of this. In a lengthy memo, accidentally released to the public, the agency explained that while it was inclined to see problems associated with Google’s self-favoring, perhaps even (illegal) anticompetitive behavior, it had mixed feelings about the prospect of winning the case if pursued, as consumers did benefit from Google policies. On the balance, the FTC therefore preferred to close the case, subject to different commitments from Google.
In contrast to its US counterpart, DG COMP has decided to test the limits of antitrust and render two decisions on the matter, with additional decisions on the horizon. It also appears that the FTC and DOJ are currently contemplating revisiting Google’s practices, either for the purpose of investigating new potential infringements or as a way to rectify the FTC’s 2013 decision.
In light of this, it becomes apparent that antitrust can be used to police for self-favoring and other forms of anticompetitive behavior in the digital economy. However, it can only do so within a narrow set of circumstances, and only if the conduct is genuinely anticompetitive. The latter should not be accepted per se, but discussed much more substantially.
The essential question is therefore not if antitrust can be used to police the tech sector for self-favoring, but whether this would be prudent, as not all forms of self-favoring are detrimental to the consumer interest.