The European Commission launches Amazon probe

John Naughton

The European commission has opened an antitrust investigation of Amazon, on the grounds that the company has breached EU antitrust rules against distorting competition in online retail markets. Amazon, says the commission, has been using its privileged access to non-public data of independent sellers who sell on its marketplace to benefit the parts of its own retail business that directly compete with those third-party sellers. The commission has also opened a second investigation into the possible preferential treatment of Amazon’s own retail offers compared with those of marketplace sellers that use Amazon’s logistics and delivery services.

The good news about this is not so much that the EU is taking action as that it is doing so in an intelligently targeted manner. Too much of the discourse about tech companies in the last two years has been about “breaking them up”. But “break ’em up” is a slogan, not a policy, and it has a kind of Trumpian ring to it. The commission is avoiding that.

It is also avoiding another trap – that of generally labelling Amazon as a “monopoly”. As the analyst Benedict Evans never tires of pointing out, a monopoly in what market, exactly? In the US, Amazon has about 40% of e-commerce. That looks like near dominance, in competitive terms. But e-commerce is only 16-20% of all retail. “So,” asks Evans, “does Amazon have 40% of e-commerce or 10% of retail? Amazon’s lawyers would argue, entirely reasonably, that Amazon competes with Walmart, Costco, Macy’s and Safeway – that it competes with other large retailers, not just ‘online’ retailers. On that basis, Amazon’s market is ‘retail’ and its market share in the US is between 5% and 10%.

On the other hand, if you’re a book publisher, then Amazon definitely looks like a monopoly with more than half of all book sales and probably three-quarters of all ebook sales. The moral for regulators, therefore, is that if you want to go after a monopolist then choose the market carefully. And this is what the commission has done, because in Amazon’s own online “marketplace”, where third parties sell stuff on its platform, it very definitely is a monopoly. And, according to the US House of Representatives recent inquiry, it is abusing its power in that particular marketplace. The EU inquiry will be into whether that is also happening in Europe.

The traditional response to such charges is that if people want to trade in Amazon’s hyper-efficient online marketplace then they have to play by Amazon’s rules. After all, nobody’s forcing them to be there. (The same argument is made about Apple’s app store.) That might work if there were dozens of alternative marketplaces, but network effects have led to a situation where a winner has taken all. In the online world, Amazon is a giant while all others are minnows. And the pandemic has further reinforced its dominance. So it really matters if the company is indeed abusing its monopoly in its own marketplace. What makes it worse is that Amazon is both a player in that marketplace and the adjudicator of complaints about its behaviour. Judge and jury and all that.

Breaking Amazon up is unlikely to be an effective remedy to this kind of problem. What is probably needed are laws that regulate behaviour in online marketplaces, which, for example, make it illegal both to run a market and trade in it on your own account. That’s not to say that break-up might not be appropriate in some cases. Maybe Facebook should be forced to disgorge Instagram and WhatsApp and Google to liberate YouTube. Even then, though, history provides some cautionary tales.

Take AT&T, for example, which for many decades was a lightly regulated monopoly with total control over the US telephone network. This had benefits, in the sense that the country had a pretty good analogue phone system. But it also had grievous downsides, because it meant that AT&T controlled the pace of innovation on communications technology, which effectively gave it the power to apply the brakes to the future. The company rejected the idea of packet-switching (the underpinning technology of the internet), for example, when it was first proposed in the early 1960s. Worse still, in the mid-1930s, after a researcher at Bell Labs invented a method of recording audio signals on to magnetised wire reels, he was forced to stop the research and lock away his notebooks because AT&T feared that it would damage the telephone business. So a technology that proved essential for the digital computing industry was hidden away for 20-plus years.

Eventually, though, the “break ’em up” mania took hold, and in the early 1980s AT&T was dismantled into seven companies – the “baby bells”. You can guess what happened: some of the babies grew and grew and swallowed up others, with the result that there are now two giant corporations – AT&T and Verizon. So even if WhatsApp, YouTube and Instagram were liberated from their existing parents, network effects and capitalist concentration will make them into a new generation of tech giants and we will be back here in 20 years wondering how to regulate them. The truth is that regulation is hard and focused and intelligent regulation is even harder. So maybe the way the EU is going about it is the path to follow.

[A version of this post appeared in The Observer, 15.11.2020]

Should you have a right to a Facebook account?

Alina Utrata

Now that the 2020 US presidential election has concluded, the post-mortem evaluation of how well social media platforms performed will begin. Since the content moderation debate has mostly focused on platforms’ willing or unwillingness to remove content or accounts, the post-election coverage will almost inevitably center around who and what was removed or labeled.

In 2019, the New York Times published a feature story about individuals who had had their Facebook accounts suspended—possibly because they had been misidentified as fake accounts in a general security sweep. However, these users do not know for certain. Individuals were only told that their accounts had been disabled because of “suspicious activity”—the appeal process to restore suspended Facebook accounts is not a transparent one, and cases frequently drag on for extended periods with no resolution. (Facebook, as the article documents, has quite sophisticated techniques for catching individuals attempting to make multiple accounts, foreclosing the solution of a “second Facebook account” for suspended users.)

Facebook CEO Mark Zuckerberg has frequently said that he does not want the platform to become the “arbiter of free speech.” Free speech, however, is a restriction that only applies to governments. As the existence of these suspended accounts show, in reality Facebook can limit speech or ban users for almost any reason it cares to put in its terms of service. It is a private corporation, not a government. 

The problem of the exclusionary policies of private corporations might be less acute in a competitive marketplace. For example, it could be inconvenient if a personal feud with your local corner store leads you to being banned from the shop; but it is always possible to buy milk from another store down the road. It is different, of course, if you happen to live in Lawton, Oklahoma—or one of the hundreds of communities across the US where Walmart is the dominant monopoly, capturing 70% or more of the grocery store market. Being banned from Walmart (either for using their electric scooters while intoxicated or violating the store’s policy on not carrying guns) might be far more significant for your life and livelihood.

Facebook’s form of monopoly power means that being banned from the platform can have significant consequences for individuals’ lives: loss of the data hosted on the platform (like photos or old messages), the ability to use messenger to connect to friends and family, or participate in professional or social groups only organized on Facebook. Some people depend on Facebook for their livelihoods, communicating with customers or selling on Facebook’s marketplace; or for political campaigns, reaching out to voters in a run for local city council, for example. The same dynamics are true for other digital monopolies, like Amazon. The recent House Judiciary report found that Amazon can, and often does, arbitrarily lower third-party sellers’ products in their search ranking, lengthen their shipping times, or kick them off the site entirely. About 850,000 businesses, or 37% of third-party sellers on Amazon, rely on Amazon as their sole source of income. Monopolies can be, as Zephyr Teachout argues, a form of economic tyranny.

There are two general approaches floated to remedying this monopoly power. The first is to “break them up.” Facebook or Amazon’s policies might be less important if there are many e-commerce or social networking sites in town—and perhaps their policies would improve if they had to compete with other platforms for users or sellers. On the other hand, Facebook might argue that the value of social networking sites are the fact that they are consolidated. As the sudden surge in popularity of the app Parler may soon demonstrate: there’s very little point in being on a social networking site if the people you want to reach aren’t there too. Alternative social networking sites may simply be complementary, rather than competitive. Similarly, Amazon might argue that it is convenient, and beneficial, to both consumers and sellers that e-commerce is located all in one place. Instead of searching online (by using another monopoly, Google) through hundreds of webpages with no guide as to quality, you can go to one portal at Amazon and find exactly what you want.

A second approach to tackling monopolies is regulation. For example, the state can and does get involved if a private corporation excludes you on the basis of a protected identity, such a race or sexual orientation. US Senator Elizabeth Warren’s calls for Amazon, Apple or Google to choose whether they want to be “market platforms” or “market participators” is another example of the state’s attempt to impose regulations in order to make sure that these monopolies are more fair. The government also gets involved when it involves product safety. For example, the Forum on Information and Democracy just published a report outlining recommending principles for regulating quality standards for digital platforms, in the same way that governments might require standards for food or medicine sold on the market. In this approach, the state imposes limits or controls on corporations to try and curb or reform their power over consumers. However, this approach requires active government enforcement and involvement. As the House Judiciary report documented, even though they are equipped with anti-trust laws, many US regulatory agencies have been slow or unwilling to take on the Big Tech monopolies. Corporations also point out that government involvement can stifle innovative and entrepreneurship.

However, there might be a third approach: democratization. Mark Zuckerberg has said that, “in a lot of ways Facebook is more like a government than a traditional company.” If that is the case, then it has been a long time since the United States tolerated a government with the kind of absolute power Mark Zuckerberg exerts over Facebook (as CEO and founder, Zuckerberg retains majority voting shares). So could we democratize Facebook, and make it a company ruled by consent of the governed rather than fiat of the king? Could Facebook users appoint representatives to a “Constitutional Convention” to draft Facebook’s terms of service, or adopt a Bill of Rights to guide design and algorithmic principles? Facebook’s Oversight Board has already been compared to a Supreme Court, so why not add a legislative branch too? Could we have elections on representatives to a Facebook legislature, which would pass “laws” about how the online community should be governed? (A Facebook legislature would arguably be more effective than the referendum process Facebook tried last time it experimented with democratization.) 

Crucially, however, any democratization process would have to be coupled with genuine democratic reform of Facebook’s corporate governance: a Facebook Parliament in name only wouldn’t achieve much if Mark Zuckerberg retained absolute control of the company. True democratization would require in a change not just in who we think represents Facebook, but who owns Facebook—or, rather, who ought to own Facebook. Mark Zuckerberg? Or we, its users? If the answer is Mark Zuckerberg, a Facebook account will always be a privilege, not a right.

Democratizing digital sovereignty: an impossible task?

Julia Rone

The concept of digital sovereignty has increasingly gained traction in the last decade. A study by the Canadian scholars Stephan Couture and Sophie Toupin in the ProQuest database has shown that while the term appeared only 6 times in general publications before 2008, it was used almost 240 times between 2015-2018. As every new trendy term, “digital sovereignty” has been used in a variety of fields in multiple often conflicting ways. It has been “mobilized by a diversity of actors, from heads of states to indigenous scholars, to grassroots movements, and anarchist-oriented “tech collectives,” with very diverse conceptualizations, to promote goals as diverse as state protectionism, multistakeholder Internet governance or protection against state surveillance”.

Within the EU, Germany has been a champion of “digital sovereignty” — promoted in domestic discourse as a panacea, a magic solution that can at the same time increase the competitiveness of German digital industries, allow individuals to control their data and give power to the state to manage vulnerabilities in critical infrastructures. As Daniel Lambach and Kai Opperman have found, German domestic players have used the term in very vague ways, which has made it easier to organize coalitions around it to apply for funding or push for particular policies. Furthermore, the German Federal Foreign Office has made considerable efforts to promote the term in European policy debates. It has been more cautious at the international scene, where the US has promoted an open Internet (which completely suits its economic and geopolitical interests, one must add) and has been very suspicious of notions of digital sovereignty, associated with Chinese and Russian doctrines above all. Attempting to avoid qualifications of sovereignty as necessarily authoritarian, French President Emmanuel Macron proposed in a 2018 speech at the Internet Governance Forum a vision of the return of the democratic state in Internet governance, as different from both the Chinese model of control and the Californian model of private self-regulation. This unfortunately turned out to be easier said than done.

What all of this comes to show is that beyond the fact that more and more political and economic players talk about “digital sovereignty”, the term itself is up for grabs and there is no single accepted meaning for it. This might seem confusing but I argue it is liberating since it allows us to imagine digital sovereignty as how we want it to be rather than encountering it as a stable, ossified reality. Drawing on a recent discussion on conflicts of sovereignty in the European Union, I claim that discussions about digital sovereignty have been dominated by the same tension as more general discussions on sovereignty – namely the tension between national and supranational sovereignty. Yet, as Brack, Crespy and Coman convincingly argue, the more important sovereignty conflicts in recent European Union politics have in fact been between the people and parliaments, as bearers of democratic sovereignty, on the one hand, and executives at both the national and supranational level, on the other. The demand for “real democracy now” that informed the Spanish Indignados protests reverberated strongly across Europe and in a decade of protests against both austerity and free trade protesters and civil society alike made strong claims for democratic deepening. Sovereignty is ultimately bound with the question of “who rules” and since the French Revolution in Europe the answer to this question at least normatively has been “the people”. Of course, how do “the people” rule and who constitutes “the people” are questions that have sparked both theoretical and practical, sometimes extremely violent, debates over centuries. Yet, the democratic impulse behind the contemporary notion of sovereignty remains there and has become increasingly prominent in the aftermath of the 2008 financial crisis in which the insulation of markets from democratic control has become painfully visible.

What is remarkable is that none of these debates on sovereignty as, ultimately, democratic sovereignty has reached the field of digital policy. Talk about digital sovereignty in policy circles has often presupposed either an authoritarian omnipotent state — as evidenced in Russian and Chinese doctrines of digital sovereignty — or a democratic state but where all decisions are made by the executive, as in Macron’s vision of the ‘return of the state’ in Internet policy. Yet, almost all interesting issues of Internet regulation are issues that deserve a proper democratic debate and participation. States such as France attempting to regulate disinformation without even a basic consultation with citizens have rightly been accused of censorship and stifling political speech.

Who can decide what constitutes disinformation, hate speech or online harms? There is no easy answer to this question but certainly greater democratic involvement and discussion in decisions about silencing political messages would be appropriate. This democratic involvement can take the form of parliamentary debates, hearing and resolutions. But it can also take the form of debates at democratic neighbourhood assemblies or organized mini-publics events. It can take place at the European level with more involvement of the European Parliament and innovative uses of so-far ‘blunt’ instruments such as online public consultations or the European Citizen Initiative. Or it can take place at the national level, with parliaments even of small EU member states building up their capacity to monitor and debate Internet policy proposals. National citizens can also get involved in debates on Internet policy through petitions, referenda, and public consultations. Such type of initiatives will not only promote awareness about specific digital policies but will also increase their legitimacy and potentially their effectiveness if citizens have a sense of “ownership” with regard to new laws and regulations and have taken place in coining them.

Some of this might sound utopian. Some of it might sound painstakingly banal and obvious. But the truth is that while our democracies are struggling with the challenges posed by big tech, a lot of proposals for regulation have been shaped by the presence and power of private companies themselves or have been put forward by illiberal leaders with authoritarian tendencies. In such a context, demands for more digital democratic sovereignty could emancipate us from excessive private and executive power and allow us to reimagine digital content, data and infrastructures as something that is collectively owned and governed.

The early years of the Internet were marked by the techno-deterministic promise that digital tech would democratise politics. What happened instead was the immense concentration of power and influence in the hands of a few tech giants. The solution to this is not to take power from the private companies and give it back to powerful states acting as Big Brothers but instead to democratize both. We can use democracy as a technology, or what the ancient Greeks would call techne, to make both private corporations and states more open, participative and accountable. This is certainly not what Putin, Macron or Merkel would mean when they talk about digital sovereignty. But it is something that we as citizens should push for. Is it possible to democratize digital sovereignty? Or is such a vision bound to end up as the toothless reality of an occasional public consultation whose results decision makers ignore? This is ultimately a political question not a conceptual one. The notion of “digital sovereignty” is up for grabs. So is our democratic future.

Market definitions and tech monopolies

John Naughton

The tech analyst Ben Evans has an interesting essay on his Blog about the difficulties of defining ‘monopoly’ in the context of the industry. This is one of his hobby-horses, but he’s always provocative. For example:

One of the basic building blocks of any competition case is market definition. If you’re claiming that a company has market dominance, and that it’s abusing that dominance, what market are we talking about? Very obviously, the company being prosecuted tries to draw the definition as widely as possible – ‘we compete with the entire planet!’ – and the prosecutor tries to draw it as narrowly as possible – ‘Ferrari has a monopoly of rear-engined Italian sport cars with horse logos!’

The fun part of this is that both of these definitions are true, and so you have dig rather deeper and work out what problem you’re trying to solve to work out what definition to use, because very often, picking the definition decides the outcome of the case, before it’s even started.

These questions come up a lot in talking about Amazon. If you read the accounts and do the numbers, you can work out that it has about 40% of US ecommerce (I wrote about this here). But US ecommerce is only part of total US retail – it was about 16% in 2019 and this year, with lockdown, it’s spiked to a bit over 20%.

So, does Amazon have ~40% of ecommerce or ~10% of retail? Amazon’s lawyers would argue, entirely reasonably, that Amazon competes with Walmart, Costco, Macy’s and Safeway – that it competes with other large retailers, not just ‘online’ retailers. Indeed, many people who argue most strongly for antitrust intervention against Amazon do so because it competes with physical retail – because they worry what Amazon will do not just to Costco but to their neighbourhood stores. On that basis, Amazon’s market is ‘retail’ and its market share in the USA is between 5% and 10%. 

On the other hand, if you’re a book publisher, you don’t care what Amazon’s share of shoe sales is. Amazon has well over half of US book sales, and probably three quarters of ebook sales. So if we’re arguing about how Amazon runs its books business, it unquestionably has market dominance. You have to pull out a segment, not the whole company. Of course, this works both ways – if you’re pulling out segments, then Amazon has 1% of US grocery sales (even Walmart only has 20%), and you can’t complain about it buying Whole Foods. 

You get the picture: he’s great at asking questions. But he doesn’t have the answers. And, at the moment, does anyone?


About this Blog



Our Aim

To create a common space for debating internet policy and regulation that goes beyond Silicon Valley exceptionalism and the Brussels bubble. It will…

… have an international perspective

The debate about democracy and technology is dominated by the US and Europe, with the rest of the world relegated to the side-lines. Now more than ever it matters to understand shared challenges posed by power, technology, and democracy, and to wrestle with how political regimes matter for the governance of technology, and how different ways of building technology matter for politics, government, and states. We strive to expand horizons, to explore what Washington and Brussels can learn from Delhi, Bejing, Seoul, analyse shared themes and notable points of divergence, and provide a platform for voices that are current ignored or under-appreciated in the public discourse about tech.

…assert the primacy of basic democratic principles

The meaning of democracy has never been stable. We believe that debates about technology and regulation are often debates about how to understand, interpret, and apply the ideal of democracy and what it means — in complex, modern states, which rely on technologies that most people have neither the interest nor expertise to understand — for citizens to govern themselves and forge their collective state. We aim to provide a forum for voices from across the world wrestling with that challenge, to share their own reflections on power, technology, and the changing meaning and experience of democracy.

… challenge the implicit determinism in traditional approaches to tech power

This blog starts from the presumption that technology is not an agent – people are. Instead of approaching technology as the agent of creative destruction, we invite contributions that question and probe the possibilities of human control, rather than assume its limits. Our contributors will explore: how the interaction between law, regulation, the administrative state, and the institutions of representative democracy affects what it means for a society to exert control over technology; and what we can learn about the underlying health of democracy in the present and its possibilities in the future. Rather than adding one more online bubble, we want to see our blog as a hammer, as a pen that changes hands, as a page of a story that might not be perfect but keeps us interested in what will happen next.