On the Economics of the Google Android Case

A general challenge facing competition authorities in the digital era is learning how to apply the traditional tools of competition policy in multi-sided platform environments. This column argues that the Google Android case offers a great example of the need to consider the implications of the market’s two-sidedness. 

 

[Editors’ note: This post first appeared on VoxEU. Reprinted here with permission.] 

 

In July 2018, the European Commission fined Google €4.34 billion for illegal practices related to the Android mobile operating system. A key objection raised by the Commission revolves around licensing terms that prevent manufacturers from preinstalling Google’s app store (Google Play) unless they also agree to preinstall other Google applications, including Google Search and Google’s web browser (Chrome). The Commission argues that Google Play is a ‘must have’ feature for device manufacturers because it is by far the most important app marketplace for the Android ecosystem. Faced with the need to carry Google Play, manufacturers are left with little choice but to also preinstall the other applications in the bundle. This, argues the Commission, has significant potential to negatively affect the vitality of competition in the markets for mobile search and mobile web browsers. Indeed, evidence suggests that the default application configuration is a key determinant of consumer choice (over 95% of search queries on Android devices are made via the pre-installed Google service; for Windows Mobile, where Google Search is not preinstalled, the figure falls to less than 25%). In sum, concerns were raised that bundling Search with the Play store results in most manufacturers installing, and consumers using, Google Search by default, restricting the ability of rival search providers to compete on merits. 

 

Bundling and Theories of Harm

 

Several observers have drawn a parallel between the Android case and earlier cases in which Microsoft was found to have illegally bundled the Windows operating system with its own web browser (Internet Explorer). However, the key economic concerns at play in those cases were quite different. The consensus that emerged in the aftermath of the Microsoft cases is that bundling was motivated by a desire to foreclose competition in the web browser market in order to protect the primary monopoly, the Windows operating system (Carlton and Waldman, 2002). Indeed, Microsoft feared that a competing web browser might itself become a platform capable of running applications, and thus a substitute for Windows. In the Android case, by contrast, there is little fear that Google feels the need to protect the core monopoly (Google Play) from any particular threat. Rather, the primary concern is that Google is leveraging the dominance of Google Play to achieve an unfair advantage in the more competitive markets for mobile search and web browsing applications.

 

This distinction is significant because it raises the question of whether Google could really profit from this kind of leverage strategy. Indeed, if Google Play is so valuable to manufacturers, Google could simply license it at a high price and let them install the search engine and browser of their choice. This is in a nutshell the traditional ‘Chicago School’ line of reasoning, arguing that anticompetitive motives cannot explain the observed bundling. Later research (e.g. Whinston 1990) highlighted conditions under which anticompetitive bundling can be profitable, namely, when it deters the entry of potential rivals and allows the firm to charge monopoly prices. In the Android case, this theory is not really convincing – Google’s main competitors for search and web browsing are large, well-established firms (such as Microsoft) that are growing rather than reducing their investment in the market. 

 

Without any logically consistent theory for why bundling might lead to profitable leverage, one should lend more credence to the idea that bundling is motivated not by anticompetitive intent, but rather by the kinds of efficiency gains outlined in Google’s argument.

 

A New Theory of Harm

 

In a recent paper (de Cornière and Taylor 2018), we show that bundling can, in fact, be profitable by virtue of its effect on competition once one accounts for some of the key features of mobile app markets. 

 

These features are:

 

  1. the existence of revenues for developers when consumers use their applications (e.g. advertising revenues for search engine apps), which induces developers to offer ‘slotting fees’ to manufacturers to be installed as default – for instance, Google is said to pay Apple $3bn to be the default search app on the iPhone; and
  2. a form of complementarity between applications – the presence of an application such as Google Play on a device increases the demand for this device, which means that more consumers will also use this device’s default search engine, generating more revenues for its developer.

Given that devices featuring Google Play sell more units, competition between search engine apps to be installed as default should result in high fees paid to manufacturers. 

 

By bundling Google Play and Google Search, Google deprives its rivals from the potential complementarity, and reduces their willingness to offer payments to manufacturers. Indeed, search rivals would expect that any manufacturer who would install their application as default would come without Google Play, and would therefore sell fewer units.3 Facing less aggressive rivals, Google can offer smaller payments to manufacturers in exchange for being installed as default. 

 

The main effect of bundling is to shift the rent from the manufacturer to Google…. Even if a rival search engine was more efficient than Google (and thus willing to bid more if they were on an equal footing), in many cases this strategy would prevent it from being installed on most devices. 

Among its objections, Google argues that bundling allows royalty-free licensing – if applications were offered on a stand-alone basis, Google would charge a high fee to manufacturers to install Google Play, which would result in higher prices for consumers. 

 

We highlight a different effect, namely, that bundling reduces the price that Google has to offer to manufacturers for them to install its search engine as default.

 

The main effect of bundling is thus to shift the rent from the manufacturer to Google. This in itself is not necessarily problematic (if Google is more efficient, total welfare and consumer surplus are unaffected). The issue is that even if a rival search engine was more efficient than Google (and thus willing to bid more if they were on an equal footing), in many cases this strategy would prevent it from being installed on most devices. 

 

Would a ban on bundling restore efficiency in this market? Our results indicate that this is not necessary. Indeed, a multiproduct upstream firm like Google could achieve the same outcome by using a strongly asymmetric pricing structure – charging a high price for Google Play, and offering a large subsidy for Google Search. One way to prevent this would be for competition authorities to regulate application pricing, by imposing a price cap on Google Play and/or a price floor on Google Search or Chrome, a solution which has its own problems (what should be the regulated prices? How should they evolve over time?). 

 

Conclusion

 

A general challenge facing competition authorities in the digital era is learning how to apply the traditional tools of competition policy in multi-sided platform environments. Economists have consistently argued for a rational approach that acknowledges specific forces that arise in two-sided markets. The Android case offers a great example of the need to consider the implications of the market’s two-sidedness. A one-sided analysis might raise doubts about the potential for anti-competitive bundling in this environment. But acknowledging the two-sidedness leads us to a new theory of harm (Choi and Jeon 2016, Caffarra et al. 2018). Indeed, the hallmark of a platform is that one side of the market (e.g. app developers) benefit more (e.g. through app revenues) from a platform that has many members (e.g. consumers) on the other side. It is the existence of this benefit and its sensitivity to the popularity of the platform with consumers that is the key to the reasoning outlined above. 

 

References

Caffarra, C, O Latham, M Bennett, F Etro, P Régibeau, R Stillman (2018), “Google Android: European ‘techlash’ or milestone in antitrust enforcement?”, VoxEU.org, 27 July.

Carlton, D W and M Waldman (2002), “The Strategic Use of Tying to Preserve and Create Market Power in Evolving Industries”, RAND Journal of Economics 33(2): 194–220

Choi, J-P and D-S Jeon (2016), “A Leverage Theory of Tying in Two-Sided Markets”, Working Paper.

De Cornière, A and G Taylor (2018). “Upstream Bundling and Leverage of Market Power”, Working paper.

Whinston, M D (1990) “Tying, Foreclosure, and Exclusion”. American Economic Review 80(4): 837-859

 

Alexandre de Cornière is an Assistant Professor at the Toulouse School of Economics.
 
Greg Taylor is Senior Research Fellow and Associate Professor at the Oxford Internet Institute, University of Oxford.
 

Disclaimer: The ProMarket blog is dedicated to discussing how competition tends to be subverted by special interests. The posts represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty. For more information, please visit ProMarket Blog Policy.