What an elusive remedy in the Google Shopping case says about the future of regulating big tech

The absence of an effective remedy in Europe’s seminal Google Shopping case nearly three years since the decision reveals much about the limits of antitrust laws and the need for a new frontier in tech regulation.

While the European Commission and Google fight it out on appeal over whether the company violated antitrust laws by using Google Search to promote its own comparison shopping service, a debate outside the courtroom focuses on the appropriate remedy in the case. The relief to date is not working the way it was expected to, raising serious questions about the underlying theory of the case and the limits of competition laws. Much is at stake as the rest of the world watches to see whether Europe’s quagmire means that antitrust laws need to be pushed further in order to deal with digital platforms, or if their limits reveal the need for new forms of tech regulation.

European Commission vs. Google

Google Shopping was the first of Europe’s three big antitrust cases against Google (with total fines summing to € 8.2 bn), but it remains the most controversial and potentially precedent-setting for its use of competition laws to moderate how online platforms treat the businesses that use them.

The case involves comparison shopping services, which are websites that allow users to search and compare products sold by different sellers (“merchants”) on the internet. A user searches for a product and is shown a list of various merchants selling it (often at different prices and shipping rates). Once the user clicks through to make a purchase from a particular merchant, the comparison shopping service takes a cut of the proceeds. A user might get to a comparison shopping service in several ways, but often it is by searching for a product on a general search engine (such as Google Search) which generates search results that include links to the third party service showing the searched-for product.

In response to complaints from some comparison shopping services, the European Commission (the EU’s competition enforcer) opened an investigation of Google in November 2010. In a June 2017 decision,¹ the Commission found that Google had violated competition laws when it boosted its own comparison shopping service (now called Google Shopping) over those of competing services in the results shown to Google Search users. It did this in two ways. One, by creating a user-friendly shopping box (with product images and prices) that featured prominently on search results and contained links only to products listed on Google Shopping. Two, by tweaking Google Search so that third party comparison shopping services were demoted in search results, while Google Shopping was not.

As a result, the Commission concluded, visibility of third party comparison shopping services in Google’s search results dropped off, and with it corresponding user traffic. Meanwhile, visibility and traffic for Google Shopping increased. As the competing services had no other good way to attract users, the Commission concluded that Google’s conduct “risks undermining the competitive structure” of the comparison shopping services market. (Google disputes all of this.)

The Commission found that Google’s conduct was an unlawful “abuse of dominance” (known as “monopolization” under US antitrust laws). By giving “more favourable positioning and display” in search results to Google Shopping than to competing comparison shopping services, Google had “extended” its strong position in the less-competitive “general search market” to improve its smaller position in a more-competitive “comparison shopping services market.”

There is much debate in academic, policy, and legal circles about whether Google’s conduct should be unlawful under competition laws. One camp believes the Commission’s decision would have to be based on a legal theory that imposes a “duty to deal” on monopolists under certain limited circumstances which they believe are absent in the Google Shopping case. (Google makes this very argument in its appeal of the decision.²) The other camp says this case is not about a duty to deal, but rather the extension or “leveraging” of monopoly power in the market for search to the separate but related market for comparison shopping. (The Commission argues this in the appeal.) Commentators have dubbed this a “self-preferencing” theory of the case–a sort of favoritism by a digital platform for its own services over competing ones.

The Commission fined Google € 2.4 bn and ordered it to cease and desist its conduct by giving “equal treatment to rival comparison shopping services and its own service” in applying the “same processes and methods to position and display” rivals as to itself.³

Remedy proves elusive in practice

Google was left to decide how to comply with the remedy. It made two main changes to open up its general search to third party comparison shopping services–in particular the coveted shopping box that appears at the top of a search result. First, a competitive pay-for-placement auction allows anyone to bid for priority placement on Google Search results (including in the shopping box). Second, Google Shopping has been made into a separate business unit and it, too, has to bid for priority placement.

Facially, the pay-for-placement system adheres to the Commission’s “equal treatment” requirement: Google Shopping must bid for placement on Google’s search results just like any of its competitors. A representative for Google has said the remedy is a success, with 600 comparison shopping services participating in auctions.⁴ More recently, Google announced that it is reverting to a more relevance-based approach for searches done on its Google Shopping service.⁵ But there’s no indication of any change to the pay-for-placement auction system that governs Google Search results (which is what is at issue in the Google Shopping case).

The Commission has not said publicly whether Google’s pay-for-placement system complies with the remedy, or whether it is having the intended effect of restoring competition in the market. However, public reports indicate that Google made the changes in consultation with government officials. And since then, the Commission has repeatedly said that it is monitoring Google’s compliance with the ordered remedy.⁶ But by November 2019, a top Commission official had acknowledged that traffic to third party comparison shopping services has not increased.⁷ This suggests the remedy is not having the intended effect, since diversion of user traffic from competing services was at the core of the Google Shopping decision.

The third party comparison shopping services have repeatedly said that the remedy is not working, most recently in a November 2019 letter to the Commission.⁸ They say that the shopping box appearing at the top of the results on Google Search continues to drive nearly all user traffic (95%) to services other than the third party comparison shopping websites. And although the Commission says the remedy means that now “81% of Shopping Units include at least one offer of a competitor, and 45.9% of clicks go to these competitors,”⁹ the third party comparison shopping services say that it is not them, but Google, who is the beneficiary of that traffic. (The European Consumer Organization agrees.¹⁰) That is because the Commission has attributed to “competitors” any traffic going to merchants even if it is routed through (and generating revenue for) Google. In their view, Google’s compliance with the remedy is continuing to divert traffic to its own services in a continuation of the very conduct deemed unlawful in the Commission decision.

As for why the third party comparison shopping services don’t steer the traffic back to themselves by bidding in the auctions for higher placement in Google Search results, they say that it would be economically futile to do so. Because Google’s separation of its Google Shopping business is just “meaningless internal accounting,” its bids “aren’t real” but rather internal payments that “cost Google nothing” as a whole.¹¹ In their view, all bidders are not treated alike on the auction because the economic reality is that they could never (while maintaining a profit) “buy back” the traffic lost to Google Shopping.¹²

So the competing services believe that the steps taken by Google to try to comply with the remedy have not cured the problems identified in the Commission’s decision, mainly, the diversion of traffic from their services to Google’s. And while the Commission has not made public any analysis of how consumers have been impacted by the remedy, the third party comparison shopping services say that product listings prioritized on Google Search are being offered at higher prices than on their own websites.¹³ If true, this would pose yet another problem for the Commission, whose remedy under prevailing competition law orthodoxy must–first and foremost–benefit consumers.

So although the Commission has not been very forthcoming about the results of its efforts to monitor Google’s compliance with the Google Shopping remedy, there are many signs that competition (at least as it is defined in the prohibition decision) has not yet been restored.

The root cause of the troubled remedy: a novel and unclear theory of the case

Antitrust remedies—especially in cases like Google Shopping that deal with “single-firm conduct” by big companies—have a controversial history. A famous example is the US Department of Justice’s case against Microsoft, which went to the brink of a historic break-up of the company before an appeals court reversed course. But remedies short of divestiture can also be controversial. When the European Commission brought its own antitrust case against Microsoft, its remedy required the company to un-bundle its operating system from its own video application and to ensure the system’s interoperability for certain third party developers. This drew the ire of a school of legal and economic commentators who dislike imposing a “duty to deal” on companies (the doctrine is very limited under US law, for example.)

But more often, remedies are not even talked about—an after-thought in cases brought by enforcers more focused on proving liability (how the law was violated). Government retrospectives on the effectiveness of remedies imposed in past cases—especially non-merger cases—are exceedingly rare.

So, too, in the Google Shopping case it does not seem that the remedy received enough attention. The decision spends a mere nine paragraphs (out of over 200) on the remedy, with the ultimate position of the Commission being that it “was not required to specify further ‘how its requested remedy is meant to work.’ It is for Google and Alphabet, and not the Commission, to make a choice.”

Although the Commission may have been within its right to keep mum on the details of a remedy, some second-guessing of the approach is justified given its apparent failure to bring about the intended outcome in the market. And the exercise proves useful to exploring some flaws in the underlying decision that—through no fault of those bringing the case—flow from limitations in the current competition laws.

A novel theory of liability turns out to be a shaky foundation for effective relief

Where a spotlight is shone on a remedy, the focus is as much as anything on the case’s underlying theory of liability (Was the law broken?). The reason should be obvious: it is tough to cure a violation of the law if you don’t know how it was broken in the first place. In Google Shopping, a novel and unclear theory of the case contributed to the challenges in achieving an effective remedy.

The Commission decision says that competition laws have been violated when Google “diverts traffic in the sense that it decreases traffic from Google’s general search results pages to competing comparison shopping services and increases traffic from Google’s general search results pages to Google’s own comparison shopping service.”

Many antitrust experts have struggled to understand the theory of the case in a way that would reveal a clear set of factors for telling what Google did that was unlawful apart from what it did that was lawful. Opponents of the decision argued on appeal that the decision “undermines legal certainty” because a violation “is not based on any coherent legal standard” but rather on a “vague claim of ‘more favourable positioning and display’ that has no apparent limiting factors” and “makes it impossible to ascertain the additional factors or legal principles that render such favouritism.”¹⁴ In other words, future lawful and pro-competitive conduct could get caught up in what the Commission considered to be unlawful and anti-competitive in the Google Shopping case.

The inexact theory of the case goes a long way to explain the confusion in the ongoing debate over an appropriate remedy. The Commission avoided bringing this case under a more restrictive and difficult-to-prove theory about a monopolist’s “duty to deal” with its competitors, instead opting for a more flexible “monopoly leveraging” theory adapted to the novel claim that Google “self-preferenced” its own services. In so doing, it renounced any remedy that for Google would “require it either to transfer any asset or to enter into agreements with one or more competing comparison shopping services.” But this resulted in a dilemma for devising effective relief in the case: how is the Commission to remedy the self-preferencing conduct if not by imposing some sort of duty to deal on Google?

Because simply ordering Google to treat third parties “equally”–without any further guidance–has apparently not worked. Google’s compliance with the remedy to date has not restored user traffic to competing comparison shopping services. So although requiring Google to treat the third party services equally may have better matched the government’s preferred theory of the case, that doesn’t change the fact that it has failed to remedy the self-preferencing conduct. On the flip side, the most plainly obvious remedy for that conduct–requiring Google to reinstate a purely relevance-based algorithm for its search results–became unavailable to the Commission once it decided not to bring the case under a duty to deal theory.

This “damned if you do, damned if you don’t” scenario that the Commission finds itself in with imposing a remedy in the Google Shopping case is a direct consequence of the theory of the case. The fact that an ineffective pay-for-placement auction remains the only option on the table for complying with the remedy—and nearly three years in, there’s no sign that the Commission is considering any alternatives—calls into question the legal basis (at least under current antitrust laws) for bringing the case.

Lack of guidance on market impact makes it difficult to know what an effective remedy is

The absence in the Google Shopping decision of a meaningful analysis of the actual impact of Google’s conduct on the market (what antitrusters call the “anti-competitive” and “pro-competitive” effects) makes it difficult to know what a working remedy would even look like.

In covering the potential negative (anti-competitive) effects on the market, the Commission’s decision says that Google’s “[c]onduct has the potential to foreclose competing comparison shopping services, which may lead to higher fees for merchants, higher prices for consumers, and less innovation.” But despite a nearly seven-year investigation, the decision provides little more than a few conclusory paragraphs on these three points. Similarly, on the other side of the ledger, the decision provides a bare-bones response to the pro-competitive “justifications” for Google’s conduct–mainly, that the changes Google made to its search service improved the quality of its search results and so were justifiably driven by what online shoppers wanted.

The Commission may be within its legal rights to not go into the details on these issues. But the absence of adequate guidance on how Google’s conduct (then) may have actually impacted the market makes it difficult to tell whether the steps it has taken (now) to comply with the remedy are restoring the market to a more competitive state. And these omissions start to look tone deaf when considering the Commission’s view that its decision imposes no “duty to deal” on Google because it could comply with the remedy by removing altogether the shopping box from its search results.¹⁵

The risks of failing to adequately consider the actual impact on the market are especially strong in a case like Google Shopping, where a plausible argument can be made that some of the conduct deemed unlawful might have benefited consumers. Google’s addition of a user-friendly shopping box to its general search gave online shoppers easily navigable and instant results that took them directly to the point of purchase.

Third party comparison shopping services no doubt believe that they have a great offering–with innovative value-add features like product filters and reviews–that they would like to see visibly promoted on Google Search. But it may be that the users simply prefer a streamlined shopping box that directs them straight to a merchant’s website. Even the comparison shopping services have had to admit they lost traffic “because Google’s [shopping box] is increasingly satisfying user demand for comparison shopping services directly.”¹⁶ Or that when Google did a test run that let users toggle their Google Search results to show links to comparison shopping service websites instead of merchant websites, users still didn’t take.¹⁷

The Commission never meaningfully engages on how Google’s design decisions were influenced by what consumers want, brushing it aside by saying that it “does not object to the … way that Google displays or organises its search result pages” but rather to the unequal way in which it chooses what appears in those results. But the decision walks a very fine line under the competition laws in trying to discern where design decisions stopped and self-preferencing began. This is made evident by the fact that–nearly three years in to the Commission’s prohibition–there is still no sign of an antitrust remedy that works and keeps the shopping box intact.

If the struggles experienced by comparison shopping services result from consumers voting with their clicks, it is difficult to imagine any working remedy that is within the confines of the Commission’s theory of the case–and, as is discussed next, within the limits of current antitrust laws.

Alternatives for relief proposed by third parties are not workable under modern antitrust laws

Compounding the Google Shopping remedy dilemma is that alternatives for how Google could be made to comply with the decision’s remedy–proposed by third party comparison shopping services–are not workable under prevailing antitrust orthodoxy.

Calls to impose rate caps on the fees charged by Google Search in its auctions for priority placement are an antitrust non-starter.¹⁸ There is perhaps no idea more renounced in modern competition law and policy than telling companies what prices to charge for their products. Protecting the competitive conditions of the market–not direct rate-setting by the government–is currently the preferred mechanism for ensuring competitive outcomes.

Characterization of the Commission’s decision as a “call for equal access to relevance-based search results” that requires Google Search to “return to the relevance-based search results” would be almost as unpopular under the competition laws.¹⁹ (In addition, as discussed above, to being a poor match for the Commission’s chosen theory of liability in the case.) There is a strong reluctance in mainstream antitrust to having governments or courts get involved in micro-managing how a business designs its products. The Google Shopping case–where the simplest remedy of requiring a return to relevance-based search results has been avoided so as to “not interfere with either the algorithms Google applies or how it designs its search results pages”²⁰—shows the lengths to which antitrust enforcers will go to avoid being put in this position. (This reluctance was likely compounded by fears of opening a Pandora’s box. Already, the Commission is hearing calls to force Google to change how its search results look and work in other vertical markets, such as Google Local Search.²¹)

Whether or not it’s a realistic prospect, competition policy is instead more apt to count on the market to correct itself. Here–the theory goes–by counting on consumers to rebel against Google if they believe its gatekeeper status calls for a return to more relevance-based search results.

Forcing a divestiture of the Google Shopping business so that its bids are more fully independent of the rest of the Google organization is probably the most antitrust-friendly of the proposed alternatives, but it is still not a likely outcome.²² Almost unthinkable just a few years ago, talks of “breaking up big tech” are now more common in policy-making circles. But the concept remains outside the antitrust mainstream. There is little precedent for it in non-merger cases, with the most infamous and recent example in tech being the failed attempt by the US DOJ to break-up Microsoft in the early 2000s. For its part, the European Commission has not even tried to force a divestiture as a remedy in a prohibition decision involving single-firm conduct during that time or ever since.

So even if given the benefit of the doubt about their effectiveness, the proposed alternative remedies are unworkable because they fall outside of mainstream legal and economic antitrust orthodoxy.

Antitrust enforcement at a crossroads after Google Shopping

Nearly three years in to its finding that Google violated the competition laws, the European Commission finds itself in a quagmire. The existing remedy has not brought the relief that the complaining third parties–or, for that matter, the Commission itself–had intended for the Google Shopping case. And a novel self-preferencing theory of the case brought under a constraining body of antitrust law provides no adequate alternative relief.

What it means is that either the self-preferencing theory of the case strayed too far from modern competition laws and the inability to fashion effective relief simply reflects those boundaries, or the theory of the case is sound but an effective remedy within the bounds of mainstream antitrust orthodoxy does not exist. In either scenario, the competition laws as currently conceived seem to be the wrong tool for what the Commission is trying to do. And the fact that the whole process has–in apparent accordance with European antitrust laws and procedures–shrugged off what online shoppers might actually want is yet another sign that existing antitrust laws are not rising to the occasion.

It is telling that these are some of the very issues that the US Federal Trade Commission grappled with during a similar antitrust investigation of Google in 2013. The FTC decided not to bring an enforcement action because there was no viable antitrust theory for the case since “Google adopted the design changes that the Commission investigated to improve the quality of its search results, and that any negative impact on actual or potential competitors was incidental to that purpose.”²³ And as if previewing the limitations of the antitrust remedies available in the case, the FTC noted that it did not want “to second-guess a firm’s product design decisions.”

For certain, the FTC’s decision to close its investigation of Google brought its own wave of criticism, and with it much attention to the limitations of US antitrust laws in dealing with online markets. But what’s noteworthy is that the EU’s antitrust enforcer continued on an increasingly divergent path from the US that led it to its own controversial Google Shopping decision in 2017, the height of its efforts to rein in digital platforms with abuse of dominance antitrust laws. In so doing, it appears to have found the messy end that the FTC had steered clear of–reaching an internal limit to current competition laws.

The future of enforcement in tech: antitrust and beyond

The Google Shopping case may still yet turn out to be the seminal antitrust case in tech that its originators had hoped, if not for the reasons they expected. Even if not a reliable antitrust theory for tackling unfair treatment by digital platforms of third parties, the decision may prove useful as an intermediate step in the evolution towards more rational and effective regulation of the tech sector. Some of the lessons learned might be specific to antitrust, but the more important ones probably extend into new paradigms such as industry regulation.

The devil you know: tinkering with the competition laws

One effect of the challenges encountered in the Google Shopping case may be creative reinterpretations of existing antitrust laws. Even US antitrust enforcers are doing a double-take on Google (among other big tech players) in sweeping probes that could lead to novel antitrust claims aided by the tailwinds of the aggressive and forward-thinking enforcement in Europe. But antitrust is a heavily doctrinal field with incredible inertia. It is a slow-moving, methodical body of law not suited for making quick changes through the administrative procedures and court systems that render its decisions. Nowhere is this more apparent than in the Google Shopping case, which–like the EC’s earlier saga in the Microsoft case–will exceed a decade from inception through appeal.

Lawmakers could speed things up by making direct changes to the antitrust laws that specify additional conduct that is unlawful. For example, proposed amendments to Germany’s competition law expressly add self-preferencing to the list of potential abuses of dominance.²⁴ Some in the US are calling for lawmakers to nullify court decisions (including ones restricting the duty to deal) that have hampered antitrust enforcement in the tech sector.²⁵ In the EU, enforcers are more confident that existing antitrust laws can reach novel theories of liability (subject to the pending appeals in the Google cases), though many would agree that the development of the law has been too slow.

And even if expanded antitrust laws more clearly reached conduct like the self-preferencing at issue in the Google Shopping case, they would still need to empower its enforcers with effective remedies. So far, there has been no big push for greatly expanding antitrust remedies. The little attention they have received has been focused on the use of so-called “interim measures.” These stop-gap remedies–imposed while an antitrust proceeding is ongoing–are meant to protect competitive conditions until there is a final decision on whether a law was broken. Recently, the Commission revived their use in a case against Broadcom, France imposed them on Google,²⁶ and Germany’s competition law amendments propose to lower the bar for them. Yet interim measures are a limited solution. Putting in place a remedy sooner in the investigation does not mean the remedy will be any better at restoring competition.

Tinkering with antitrust law–whether expanding its reach into new types of conduct or broadening its remedy toolkit—will no doubt give enforcers some more flexibility in dealing with the online economy. And Europe’s experience with the Google Shopping case will have played some role in that evolution. But the changes will face doctrinal resistance, the process will take much time, and the ultimate results may be limited. That is why policymakers are already looking for a different path.

Recasting tech enforcement through industry regulation

Given the limits to what can be done with competition laws, the bigger impact of the Google Shopping case may be in pushing the policy mindset beyond antitrust enforcement and into industry-specific regulation.

Until now, the digital economy has managed to avoid intense sector regulation. But when it comes to dealing with issues not well-suited for antitrust laws–like those arising in the Google Shopping case–industry regulation may offer some advantages. For one thing, it could short-circuit the drawn-out proceedings and slow-developing case law that hamper antitrust. Compare, for example, the nearly decade-long (and running) saga in Google Shopping to the two-and-a-half years that will have passed between Australia launching a sector inquiry and putting in place digital platform regulations later this summer.²⁷ Quick action and nimble tools will be especially important in dealing with fast-changing digital markets.

In addition, sector-specific regulation may be well-suited for dealing with the sort of conduct at issue in the Google Shopping case, and in other cases that antitrust enforcers are pursuing against online platforms (for example, Germany’s leveraging-type case against Facebook and the EU’s self-preferencing investigation of Amazon). These cases display three key characteristics that have been identified as being better suited for sector regulation than for antitrust enforcement: conduct involving vertical chains of production, industries driven more by innovation than price, and two-sided markets.²⁸

Moreover, limitations of modern antitrust laws in fashioning behavioral remedies–like the dead-on-arrival proposals from the third parties in the Google Shopping case–need not constrain sector regulations. Decades of antitrust law and economics have erected a barrier–perhaps put there for reasons perfectly sensible in other applications–to telling businesses how to design or develop their products or what to charge for them. But industry regulations need not inherit that legacy. They can be drawn up from scratch with a narrow sector-specific focus. And unlike antitrust authorities bogged down by sweeping enforcement missions, regulators can singularly focus on monitoring and modifying remedial measures as circumstances change in a dynamic digital economy.

There is already some indication that governments are pivoting from antitrust enforcement to industry regulation. The UK’s competition authority has declined so far to recommend a market investigation of the online advertising sector because it would prefer for the government to pursue a “comprehensive regulatory framework to govern the behaviour of online platforms.”²⁹ The Australian competition authority’s comprehensive sector inquiry of digital platforms in 2019 led to recommendations for “holistic reform” including regulatory codes of conduct on dealings between platforms and third parties.³⁰ The EU passed regulations affecting self-preferencing conduct on certain business platforms,³¹ and is now working on broader online platform regulations that are expected to tackle broader issues of market leveraging and self-preferencing.³² And Japan is working on regulations that will control how certain categories of online platform businesses treat third parties competing in a related market (including limits on self-preferencing), as previously covered here.

These are all signs of what the future looks like in digital tech regulation.

Conclusion

When it is all said and done, more than a decade will have passed between the time that several competitors complained to EU antitrust enforcers about Google’s conduct in comparison shopping and a final (appellate) ruling on whether Google violated the law. And even if that ruling upholds the Commission’s decision in Google Shopping, a viable remedy for those competitors may never come to fruition. Whether one thinks that Google’s conduct should be stopped or not, clearly the current competition laws and procedures have proven not to be the best way to tackle the question.

The antitrust laws are not–and never will be–the cure for all that ails the digital economy. Given the limits to what can be achieved by changing the existing laws, other solutions such as industry regulation may prove to be more suitable for dealing with certain aspects of the tech sector. Pushing policymakers around the world–and perhaps also the tech businesses themselves–to recognize that may ultimately prove to be the legacy of the Google Shopping case.