In attempting to address legitimate privacy concerns, Google could be trying to monopolize surveillance.
As Big Tech’s leviathan wraps its tentacles around online markets, regulators at both the state and national levels are scrambling for solutions. On October 17, a state senator from Ohio hosted a legislative hearing in Cleveland to examine whether state antitrust laws should be amended to tackle competitive issues presented by tech platforms. Earlier in the week, the Federal Trade Commission announced the conversion of the Technology Task Force, tasked with investigating tech platforms, into a permanent division in its Bureau of Competition. And last week, several state attorneys general gathered in New York to discuss how to style an antitrust complaint against Facebook.
Another investigation that lies at the intersection privacy and antitrust recently got underway in Congress: In late September, the Wall Street Journal reported that the House Judiciary Committee is studying whether Google seeks to throw its weight around the Internet’s naming system. Arcane stuff, for sure, but it’s not too hard to see what’s likely motivating the online advertising giant.
The Internet’s domain naming system (“DNS”) works much like a phone book, translating domain names such as promarket.org to Internet Protocol addresses. This function is needed because domain names are easier for people to remember than a long string of numbers. Currently, no single entity has control of the DNS system.
This openness could soon end. In June, Google announced plans to make encrypted DNS generally available in its Chrome browser. By encrypting DNS data, Google would prevent third parties, such as ad tech intermediaries or internet service providers, from tracking users while they explore the web.
What’s motivating Google? From Google’s point of view, the incremental profits to be made by potentially monopolizing DNS can’t be that attractive, particularly given the company’s complete control over the operating system (Android) and the browser (Chrome). Controlling DNS doesn’t given Google any significant additional visibility into users beyond what it soaks up from controlling the browser. Indeed, Google announced that it has “no plans to centralize or change people’s DNS providers to Google by default.”
Google’s plan to encrypt DNS is more likely driven by the search monopolist’s desire to preserve its market power in online search advertising. According to eMarketer, Google accounts for 37 percent of all digital advertising and a staggering 78 percent of search advertising in the United States. That’s a monopoly worth protecting from advertising intermediaries, even by anticompetitive means.
Why should a third party track users through their browsing history? None of these intermediaries, argue Google’s defenders, should be able to do so without a user’s permission. Fair enough. But Google also tracks users through their browsing history without permission, which renders this “only we should track” argument a touch ironic and self-serving.
|“Google’s latest encryption threat is not an isolated maneuver but instead represents part of a broader strategy aimed at closing its advertising ecosystem to third-party providers.”|
So long as an entire online ad industry is built around tracking users—and in an ideal world, the extent of surveillance would be severely curbed—it is still better for online advertising services to be competitively supplied than to be monopolized. When compared to the mammoth size of Google, advertisers and publishers are vulnerable customers on Google’s advertising platform. Too much selling-side power means Google can raise the price of ads to advertisers. And too much buying-side power means Google can suppress publishers’ compensation levels (or compensation shares).
Google’s latest encryption threat is not an isolated maneuver but instead represents part of a broader strategy aimed at closing its advertising ecosystem to third-party providers. To wit, in September, Google started to test a new feature that will block third-party tracking cookies within Google Chrome, which will inhibit retargeting by independent ad tech companies. And in 2017 Google ended keyword sharing for marketers on referring URLs. This exclusionary conduct can be understood as a means to prevent third-parties from getting a foothold in a small niche of the advertising space, with the prospect of someday competing with Google for the larger search advertising market.
When asked about Google’s new DNS policy at an October 18 House Judiciary Hearing on the Role of Data and Privacy in Competition, FTC Commissioner Rohit Chopra correctly expressed a concern that the market for internet browsers was highly concentrated, and that these types of market design matters should be decided by Congress, and not by a monopolist with a conflict of interest.
In its defense, Google claims it wants to encrypt DNS to help prevent hackers from spoofing or snooping on the websites that users visit. Enhanced privacy and security are certainly noble causes, and a potential benefit for end users. But the relevant question from an antitrust perspective is whether this conduct—or any of the exclusionary conduct mentioned above—generates anticompetitive effects for advertisers or publishers.
Even if encrypting DNS will generate a privacy benefit for end users, so long as there exists a less restrictive alternative to prevent spoofing or snooping by hackers, Google’s efficiency defense would crumble. Notwithstanding the Supreme Court’s ruling in Ohio v. American Express for two-sided “transaction” platforms such as credit cards—which overlooked a clear harm to merchants from Amex’s no-surcharge rule because of the possibility that some of the overcharges flowed back to cardholders—a speculative privacy benefit for end users cannot be used to offset a harm to a distinct customer group (here, publishers or advertisers).
Google’s defenders point out that Mozilla is also encrypting DNS, and because Mozilla lacks market power—it has a single-digit browsing market share—Mozilla must be encrypting DNS for pro-competitive reasons. Ignoring Mozilla’s financial dependency on Google, this “small-folks-are-doing-it” rationale is used repeatedly by antitrust defendants, but with limited success. Smaller rivals might be mimicking the exclusionary conduct of the dominant firms, in which case the conduct of a small firm cannot serve as a competitive benchmark.
Moreover, anticompetitive effects are the result of the interaction of market power and exclusionary conduct; both elements are necessary to generate a harm. This means that small firms are free to engage in all sorts of conduct that would be proscribed for a firm with monopoly power.
None of this is to say that Google’s exclusionary conduct described here has been shown to be anticompetitive. To do that, one would have to link any one of these restraints—encrypting DNS, blocking third-party cookies, or ending keyword sharing on referring URLs—to higher ad prices or lower payments to publishers. Because the first two strategies have not yet been implemented by Google, however, that traditional proof of anticompetitive effects is not available.
But perhaps investigators could build a case for intervention by showing that ad tech intermediaries or internet service providers constrain (albeit marginally) Google’s pricing power over advertisers or Google’s buying power over publishers. Or they could show that these players have the potential to evolve into something larger that eventually would constrain or topple Google’s advertising monopoly. With the right showing, regulators (or a court) could tap the brakes on Google’s exclusionary scheme before the leviathan extends its grip on search advertising for another decade.
Hal Singer is managing director at Econ One, a litigation consulting firm, and an adjunct professor at Georgetown’s McDonough School of Business, where he teaches advanced pricing. He has consulted to various internet service providers, including AT&T and Verizon.
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.