The U.S. Justice Department has presented its evidence in the antitrust case alleging that Google unlawfully maintained a monopoly over “general search services” by “lock[ing] up distribution channels” through “exclusionary agreements” with makers and marketers of devices. Google’s agreements with Apple, for example, have made its search engine the default in Apple’s Safari browser.
The government contends that such agreements serve little purpose other than to suppress competition, but they look rather different in light of economics. The agreements at issue were the only means available to distributors for securing a share of the billions of dollars that search engines receive annually from advertisers.
Search engines do not charge users for searching, but some queries return ads along with the organic search results, and advertisers pay the search engines when users click on ads. These payments are what funds the search engines’ operation and improvement.
Distributors get a share of the search-ad revenue in exchange for boosting a particular search engine’s usage and hence its ad revenue. A simple and obvious way to do that is to make the search engine a default (in a browser or other means of accessing the internet).
Users find it convenient for a device to be fully functional right out of the box. A significant fraction of users stick with a default search engine, although they can change the search engine of a browser or download a different browser with a different default search engine. In its case against Google, the government has stressed user stickiness and often referred to the “power of defaults.”
The power of defaults would be at its zenith in a world of undifferentiated search engines, because every user would stick with every default. In such a world, the owner of a distribution channel could capture all of the profit from search advertising by auctioning off default status. To claw back some search-ad profit, a search engine would have to differentiate itself.
With differentiated search engines, the power of defaults is limited by “leakage,” when some users override the defaults and switch to preferred search engines. And if one search engine does much better than its rivals in providing what users want, leakage will be highly asymmetric.
Suppose the leakage from search engine S1 is 30%, while the leakage from any other search engine is 70%, with 50% of users leaking to S1 and 10% leaking to each of two other search engines. S1 then gets 70% of a channel’s usage if it is the default and 50% otherwise. Any other search engine gets 30% of the usage if it is the default and 10% otherwise.
Under the forgoing assumptions (which might not resemble the real world), default status has the same value to every search engine. A channel owner still can extract all of that value by auctioning off default status, but a channel owner now gets just 20% of the ad revenue, because default status delivers just 20% of users.
If, instead, the leakage from S1 is only 10%, default status is worth more to S1 than to the other search engines, so S1 would bargain with the owner of a distribution channel over the division of the ad revenue. If they split the S1’s gain from default status (as economic theory predicts), S1 gets 70% of channel ad revenue, and the channel owner again gets 20%.
The government would contend that a search engine like S1 should not enter into any agreements that confer default status. But S1’s rivals then would enter into such agreements, which would harm the majority of users preferring S1. Getting nothing in return, S1 would not share its ad revenue with distributors.
The government makes much of the importance of “scale” to the quality of a search engine. It argues that greater usage could lead to quality improvements for S1’s smaller rivals, and that such quality improvements would place competitive pressure on S1. But the government can only speculate about what might have been.
The legal roadmap for the case is the 2001 Microsoft decision. The court in that case held that a claim of monopoly maintenance could be sustained only through proof of conduct that was “not a form of competition on the merits” and was “reasonably capable of contributing significantly to a defendant’s continued monopoly power.” The government directed its proof to the latter but slighted the former.
If U.S. District Court Judge Amit Mehta finds that Google has a monopoly and that the agreements at issue were reasonably capable of contributing to its maintenance, he will turn to a legal question: Did antitrust law oblige Google to forgo usage-promoting distribution agreements when its rivals otherwise would enter into them, using the power of defaults to take users from their most preferred search engine? I think not.