Almost no consumers have the resources to assess the quality of information online. Search engines seek to remedy asymmetries in information, effectively providing a quality certification service to consumers. Google claims to rank organic results ‘‘scientifically’,’ based strictly on relevance and the quality of the listings. Ninety two percent of all Google search traffic occurs on the first page, encompassing the top ten organic results and paid ads, which reflects high levels of consumer trust.
This paper, available here, argues that Google’s search engine is indispensable for innumerable companies, which cannot compete effectively when Google fails to rank organic results according to relevance. However, Google’s ad-based business model creates incentives for it to promote paying advertisers or its own business, in particular by lowering the rank of more relevant results. This leads to lower quality in the search market, to lower output in downstream markets and, ultimately, to lower consumer welfare – independently of whether Google operates downstream or not. As a result, Google should owe a duty to promote economic activity by ranking organic results according to relevance and quality.
This argument is developed as follows:
Section 2 details how Google and Amazon promote economic activity.
Before the internet, no accessible source offered consumers ostensibly unbiased information concerning multiple sellers at no cost. Nowadays, consumers can obtain such information instantaneously by typing a few words into Google or Amazon. Search engines supply consumers with the information necessary to find lower prices and better quality. In addition to facilitating product comparison, search engines also lower transaction costs associated with purchases. Reflecting the trust in which Google is held, consumers do not expend resources reviewing all the search results produced by its search engine: a recent study found that a super-majority of consumers – 92% of all search traffic on Google – check only the first page of Google search results.
In addition to reducing transaction and search costs, the Internet has also led to lower prices across a variety of products. By increasing variety and enabling transactions that otherwise would not occur, holding price constant, Google and Amazon enhance total welfare by a multiple of the consumer welfare generated solely by lower prices. Search engines have created the ‘‘long tail’’ of Internet purchases, as consumers can now find less popular books, music, and movies that collectively can rival bestsellers and blockbusters in terms of sales.
Section 3 explains why Google’s incentives may lead it not to maximise economic activity.
Google purports to rank its searches by reference to the quality and content of the relevant webpage. Businesses seek the best organic rankings because consumers value the editorial integrity on which Google trades. The service that Google provides inherently facilitates economic activity.
On the other hand, Google is also able to favour paid over organic results – in effect, something its business model incentivises it to do. To monetise searches, Google intermingles paid and organic results, with paid advertisements generally appearing at the top and bottom of organic results. Google charges per click, and also when clicks on paid ads produce sales. This paid search corresponds to 70% of Google’s revenue, USD $90.27 billion in 2016. Google’s financial incentives are thus to favour paid listings, because this generates revenues it would otherwise not earn.
This leads to harm to consumers by comparison to a search engine that earned profits differently. By moving a business that otherwise would appear on the first page to another page, or by artificially lowering a result in the first page from, say, first to ninth, Google degrades the quality of its organic search results by promoting the rank of less relevant links in order to make users click on paid results. Alternatively, Google may lower the rank of potential advertisers in organic results to create a stronger incentive to purchase ads. In any event, Google does not enhance quality, competitiveness, or in any way benefit consumers by maximising its ad revenue to the detriment of organic search results that maximise further economic activity.
Section 4 discusses how the refusal to deal doctrine could apply to a dominant online search engine.
Traditional refusal to deal claims involve a monopolist that both controls an essential facility upstream and operates in a downstream market where rivals need access to the input to compete. The antitrust concern here is that denying access to the essential facility will exclude rivals and allow the monopolist to extend market power downstream, resulting in higher prices and curtailed quality and follow-on innovation. For refusals to deal by search engines to fit into this narrow conceptual category, the monopolist would have to compete and threaten monopolisation downstream. While Google does not technically refuse to deal with businesses that appear somewhere other than in its first page of search results, Google’s finagling of rankings, which causes these businesses to appear beyond the first page of results, can amount to a constructive refusal to deal.
Yet even where Google does not compete in the downstream market, attempts to maximise ad revenue adversely could affect price and quality due to Google’s market power upstream. When search engines rank better search matches (e.g. in terms of price-quality ratio) below inferior substitutes, they obstruct economic activity. This happens by preventing purchases that consumers otherwise might have made, by transferring surplus from consumers to producers, or simply by reducing both consumer and producer surpluses. Manipulated rankings could produce: (1) price discrimination between consumers for a given level of quality; (2) greater costs for the same level of quality; (3) a reduction on the quantity of goods or services that the consumer would consider purchasing; or (4) a reduction in the variety of goods consumers had access to.
As a dominant creator of markets linked to search queries, Google should have a duty to attempt to maximize consumer surplus in terms of the price-quality ratio that its organic listings present, incorporating pragmatism and judgement. As an alternative to requiring monopolisation by Google downstream, antitrust authorities could impose a duty on Google, to rank organic search results strictly according to relevance and quality. The objective of compelling access would not be to prevent a monopoly downstream, but to prevent a monopoly from utilising market power to inhibit economic activity in possible downstream markets. As an alternative to imposing a general duty, antitrust authorities could demand that the effective refusal to rank according to relevance must not perpetuate or facilitate the dominance of any undertaking in a downstream product or service market. This would require a novel approach to refusals to deal, but would also ensure that consumer welfare and economic activity downstream would be maximised.
Section 5 considers what tests of refusal to deal against Google could apply in practice.
The EU will compel access to essential facilities when exceptional circumstances exist. In the context of IP rights, these occur: (1) when access to the input is indispensable to operating on a neighbouring market; (2) the refusal excludes effective competition on that market; (3) the refusal precludes the appearance of a new product for which consumer demand exists; and (4) a dominant undertaking cannot objectively justify the refusal to license.
The author submits that these conditions are met by Google. Regarding indispensability, there are no alternatives in the market and the amount of investment in other distribution channels that a rival need to make in order to attain equivalent exposure as a high ranking in organic Google search would be immense. The refusal excludes market competition because it removes competitors from the first page, and thereby from the awareness of most consumers. The author seems to presume that the third requirement has been changed/lightened in Microsoft – from preventing new products from coming into the market, to jeopardising the incentive to invest in novel technologies or otherwise damage competition in the downstream market, e.g. by intentionally listing products with inferior price-quality. In this respect, Google would lead the relevant downstream market not to develop efficiently. Regarding objective justifications, business demands may support granting access on reasonable terms, rather than equal access. In granting certain businesses access to the first page of organic results, Google can comply with this condition only by ranking according to relevance and quality.
This is yet another paper that tries to assign a specific theory of harm to the Google Shopping decision – and it seems to interpret the decision, plausibly, as being concerned with business practices that foreclose competition in a market downstream. However, the author goes much further and seems to argue for an overall duty for Google not to rank its searches for anything other than relevance. For competition law to require this strikes me as being problematic – unless one concludes that self-preferencing is a problem that should be addressed by a per se rule, in line with recent German and Japanese proposals.
However, the remedy the author is proposing seems to go beyond self-preferencing. The argument is that any change to Google’s ranking practices that depart from relevance could, theoretically, reduce consumer harm. Such practices seems to include to introduction of paid (advertising) results, which provides the basis for Google’s business, or the mere identification of a Google service in the market downstream in the search page. The theory of harm underpinning such a remedy must be that Google should not be an advertisement-led business – or, as the author puts it, a court might find that a viable business model could allow entities that rank near the first page to bid for a boost to the top of the first page, for other businesses to pay more for placement at the bottom of the first page, and for all entities to pay for the chance to compete in Google Shopping. This is a rather bold theory, as one would think that the relevant competition counterfactual would be limited solely to the manipulation of search results. Oddly, this approach could logically be extended to the prohibition of advertising in all its forms as an abusive practice, or to finding agreements between advertisers and broadcasters to be anticompetitive merely because the advertisement is potentially misleading or less relevant – which is a very hard conclusion to countenance.