By Albert A. Foer and Sandeep Vaheesan
In early January, the Federal Trade Commission (FTC) closed its nearly two-year investigation into Google’s conduct. Unanimously, the Commissioners stated that Google’s alleged favoring of its own vertical search features in search results was not an antitrust violation. They found that changes to Google’s search algorithm were intended to offer more informative search results. The FTC acknowledged that modifications of Google’s algorithm deprived some vertical search sites of traffic, but stated that harm to competitors is a “common byproduct of ‘competition on the merits.’” Responding to other allegations, Google voluntarily agreed to stop appropriating content from vertical search engines and allow online advertisers greater flexibility in managing concurrent ad campaigns on multiple search engines. Investigations into Google’s practices continue in other jurisdictions, including the European Commission (EC) and the Korea Fair Trade Commission (KFTC). Given the high level of cooperation between these authorities in the Google matter, it seems unlikely that the pending investigations will reach significantly different results, though they may lead to legally binding commitments. Under a proposed settlement with the EC, Google has, among other things, offered to label its own vertical features and display three relevant non-Google vertical sites in search results. Reports indicate that many competitors of Google and some consumer groups are unhappy with the proposal, believing it to be inadequate. If adopted, this remedy would not differ dramatically from Google’s original, voluntary pledges to the FTC. (Given the scrutiny it has received and will continue to receive, Google would be foolish not to abide by these legally non-binding commitments to the FTC.)
We do not have access to the proprietary data and diversity of viewpoints available to these various antitrust authorities. But the question we would like to raise in this editorial is: “what if?” Suppose the FTC had found that Google’s search bias violated antitrust laws, or that such a finding could be made by another antitrust authority around the globe, concluding that the remedy proposed by Google to the EC was insufficient. Beyond the clear labeling of Google’s own services and the prominent displaying of non-Google sites, what remedy would be appropriate to counter the allegedly illegal conduct adopted by the firm?
Ongoing supervision of Google’s search algorithm is fraught with problems. Google constantly revises its algorithms to improve its search results. It is doubtful that any of the antitrust authorities mentioned above have the institutional capacity to review every significant or complaint-inducing change to Google’s search algorithm. And, in the event Google’s search formula could be monitored, this regulatory oversight would raise multiple concerns. First, detailed regulation of a dynamic company may deter desirable innovation. Second, regulation affecting the display of content in one of the primary gateways to the Internet would raise legitimate concerns of government control over access to information and speech. A structural remedy would avoid the pitfalls of a regulatory solution. For example, an authority could, in theory, direct Google to divest its vertical search sites and confine itself to general search. This would reduce Google’s incentive to modify its algorithm to favor its vertical search offerings. Yet, structural separation, by prohibiting vertical integration between different categories of search, could impede functional improvements. It is a conundrum.
This all should not eliminate concerns over Google’s market power. Google remains the dominant search engine in the United States and around much of the world. Google’s power should concern everyone who values an open Internet, not to mention the protection of privacy. At present, Microsoft’s Bing is the primary constraint on Google’s ability to manipulate search results. Users dissatisfied with Google have the option of using Microsoft’s search engine. Without Bing’s presence, a monopolistic Google would have the power to change its search algorithm with less regard for user preferences. The focus in military intelligence on capabilities rather than intent of a potential enemy is instructive. For example, consider the risks presented by Google under different, more ideological leadership. Could it rig its search results to give prominent rankings to only conservative (or only liberal) news sites and blogs, and even attempt to manipulate the preferences of undecided voters?
Despite the limitations of antitrust law, governments around the world are not powerless to preserve an Internet with multiple search engines. After all, they are among the largest purchasers of goods and services in the world. To promote the public interest in free and open access to information on the Internet, they should use their power to ensure the long-term survival of Bing (and other smaller non-Google search engines). They could, for example, require IT vendors to make Bing the default search engine on millions of government computers and move a significant fraction of advertising for government jobs and contracts from Google to Bing. Governments have other opportunities to tilt a playing field in the larger interest of maintaining a competitive industry structure. These policies could increase the number of visitors to Bing directly and also make Bing a more attractive resource for search relating to government.
Critics of this proposal might allege that it would be an unprecedented attempt by governments to “pick winners.” Ordinarily, the antitrust laws protect the competitive process against collusion and exclusion. In universal search, however, the antitrust laws alone likely cannot preserve a competitive market. Moreover, if Google turns out to be a natural monopoly, because of economies of scale and scope, public utility regulation or public ownership — the normal governmental responses — have significant shortcomings, including free speech objections. With the gatekeeper role that search engines play in the Internet, the existence of multiple engines is arguably even more important than having multiple producers in most markets. Furthermore, direct promotion of competition by national governments has famous precedents. During the Second World War, the US government constructed several aluminum smelters because Alcoa did not have sufficient capacity to meet wartime demand. Following the war, the government sold the smelters to Kaiser and Reynolds at discount prices to end Alcoa’s decades-long monopoly. More recently, the Brazilian government opted to purchase computers with open-source Linux instead of Microsoft Windows, to save money and develop an affordable operating system designed for local needs.
As a general principle, governments should not offer direct support to some companies, and not others. And, furthermore, the actual power of governments to promote competition through their purchasing decisions is far from certain. But sometimes neither antitrust nor regulation can adequately protect the public. The case of Google requires us to consider all alternatives, including the deliberate use of public resources to maintain effective choice in information sources on the Internet.
Albert A. Foer is President of the American Antitrust Institute and Sandeep Vaheesan is Special Counsel. A version of this article originally appeared as an editorial in the Journal of European Competition Law & Practice.
Journal of European Competition Law & Practice is a peer-reviewed journal dedicated to the practice of competition law in Europe. Primarily focused on EU Competition Law, the journal includes within its scope key developments at the international level and also at the national EU member state level, where they provide insight on EU Competition Law.