Goodbye Burning Man, Hello Brussels

Europe will regulate the internet giants

With its historic antitrust actions against AT&T in 1974 and Microsoft in 1998, the Department is again enforcing the Sherman Act to restore the role of competition and open the door to the next wave of innovation – this time in the vital digital markets”. DoJ press release, Oct 2020.

Market processes are often incapable of ensuring fair economic outcomes with regard to core platform services. Moreover, existing Union law does not address, or does not address effectively, the identified challenges to the well-functioning of the internal market posed by the conduct of gatekeepers which are not necessarily dominant in competition law terms.” EU proposed Digital Markets Act, December 2020.

2020 was the year five internet giants – Amazon, Apple, Facebook, Google and Microsoft – became indispensable to daily life, 25% of the S&P500 and the new number one target of regulatory enforcement (replacing banks).  This paradox of success is not new and has similarities with the multi-decade regulation of telco networks. One aspect in common is the difference in approach between the US and Europe. Brussels and its recently proposed Digital Markets Act (DMA) is likely to prove more influential, fundamentally challenging the modus operandi of the internet giants. Ahead of the DMA being signed off later in 2021, expect severe cognitive dissonance at Mountain View, Menlo Park etc, followed by vigorous lobbying, which will fail to have any effect. As the DMA is then implemented from 2022, expect an intense multi-year legal battle at the ECJ. Key lesson from the telcos: Brussels usually wins.

The US anti-trust approach disrupted telco corporate structure, the EU approach disrupted telco market cap

The US administration’s main thrust against the perceived dominance of the internet giants is the launching of anti-trust lawsuits (DoJ – Google, FTC – Facebook). US law requires the DoJ to show actual monopolistic behaviour and then applies remedies to try and control that specific behaviour. In the Google suit for example, the DoJ suit is alleging Google has made agreements with third parties that restrict user’s choice over search engines and seeks to remedy these agreements. The EU has similar anti-trust powers but also an alternative mechanism that can regulate companies if the very nature of a market structure might encourage monopolistic behaviour in the future, regulating “ex ante” (or before the fact). The point is that the European ex ante regulation can be much more wide ranging and pre-emptive, targeting any current or future potential competitive bottlenecks, whilst the DoJ ex post approach is limited to targeting areas where a court judge has agreed that monopolistic behaviour has actually happened. DoJ ex post tends to target actual executive behaviour, Europe ex ante tends to target and then change fundamental market structure.

This ex post / ex ante divide between US and Europe has already been seen in telco regulation. The US engaged in a lengthy anti-trust pursuit of American Bell (AT&T). The first DoJ call to regulate American Bell under the Sherman Act was in 1913; at the time it was refusing to interconnect with rival telephone networks. A lengthy DoJ legal campaign led to AT&T’s eventual break-up in 1984. In Europe, dominant incumbent telcos were left intact but in 2002 forced by new EU ex ante regulation to physically open up (unbundle) access, allowing third parties to resell the telco’s local loop.

The break-up of AT&T was traumatic to the corporate entity and also its internal monopoly equipment supplier Western Electric, but it didn’t tackle the key competitive bottleneck. The key local loop bottleneck was left unharmed, broken up into regional fiefdoms (RBOCs), with little if any increase in local loop competition, stable or rising prices and leading ultimately to the local pieces coalescing back together (into Verizon and AT&T). US law didn’t tackle the natural monopoly, where the bottleneck is not the result of corporate behaviour but inherent to the business itself. The break-up of AT&T created a lot of fireworks but was not particularly disruptive to end-user telecom prices or ultimately to shareholder value.

In contrast, the European ex ante regulation of its incumbent telcos completely changed the telco market and seriously undermined the incumbents’ profitability. The EU regulation forced the incumbents to allow third parties to unbundle their network on a cost-plus basis, which meant installing their own switching equipment in the incumbent local switches and then renting the local loop (last mile) for around eu10/m. The European incumbents at the time were charging their end customers around eu80/m for the same thing. The first company to really capitalise on this new arbitrage opportunity was Iliad which launched in France in 2002 offering fixed line, broadband and voice calls all-in for below eu30/m, reselling the France Telecom network. Similar unbundlers gradually emerged in many other EU markets. More than half of European incumbent lines ended up being unbundled in this way and incumbent revenues, profits and market cap duly tumbled.

18 years later, European consumers pay around eu30/m for fixed (including broadband) and eu20/m for mobile, which cost around 2-4x as much in the US. Verizon and AT&T’s combined market cap of usd450bn is 1.7x the size of the combined European incumbent telcos’ market cap, despite the latter also having extensive holdings outside their domestic market (DT - Europe & US, Telefonica - Europe & Latam, Telenor - Nordics & Asia, Orange - Europe & Africa, etc).  

The lesson is clear – European ex ante regulation was much more disruptive to telco business models and shareholder value than US ex post. The same will likely be true for the internet giants as well.  

The devil was in the detail and the stock-market didn’t see it coming

This EU proposal to unbundle telco networks was written in the fine print of an EU regulation published in 2002. The regulation was arcane, most of the incumbent telcos downplayed it and most European telco analysts at the time were sanguine about it, focused on other issues (the aftermath of the 3G spectrum and TMT busts). After Iliad IPO’d in 2004, bringing its disruptive ULL business model to much wider attention, this view changed dramatically.

If we read the fine print of the proposed EU DMA regulation (google “Brussels 15.12.2020 2020/0374”, turn to page 39) there are similar clauses that seem quite bland but which have potential to wreak the same havoc on existing business models if applied comprehensively. At this point there is the likelihood of some cognitive dissonance back in California.  “Surely they don’t really mean that?”  The telco analyst at this point replies: “no, really, they probably mean every word”.

Take the following clauses (which I have summarised) from the proposed DMA. A gatekeeper (read internet giant) shall

·         Apply fair and non-discriminatory conditions of access for business uses to its software application store.

·         Allow third party applications and stores to access and inter-operate with the gatekeeper’s operating systems

·         Provide to business users free of charge access to aggregated and non-aggregated data that is generated in the use of the core platform service by the business user and the end-users engaging with the services of the business user, when the end-user opts-in to such sharing.

·         Provide to any third-party search engines access to ranking, query, click and view data generated by end-users of online services of the gatekeeper, subject to anonymisation of personal data.

·         Refrain from using data generated by business users (and the end-users of the services of that business) of its platform to compete with those business users

·         Refrain from treating more favourable in ranking services products offered by the gatekeeper

·         Provide advertisers and publishers access to the performance measuring tools of the gatekeeper and the information necessary for advertisers and publishers to carry out their own independent verification of ad inventory.

Taken literally these clauses threaten to disrupt several trillion-dollar business models, at least within Europe.

The EU goalposts will shift

The proposed EU DMA regulation already envisages extending the regulation to new types of internet services as they emerge, and to even regulate internet platforms that are not currently gatekeepers but are expected to become so in the near term. 

Longer term, there is likely to be a moderate waterbed effect, with internet giants charging higher rents in areas of their business that remain untouched by EU regulation, to offset revenue lost in the regulated part. This would probably draw a response from the EU in time. Regulation led European telcos to end- up halving their access charges, charging nothing for local and long-distance calls but charging eu1.00+ per minute (and eu10c per MB) for using their phone when travelling abroad. The EU ultimately reacted, creating its first regulation of EU roaming tariffs in 2007, and finally eliminating them inside the EU in 2017. 

The network effects are in any case so strong at the internet giants that the EU regulation will probably fail to stop the big getting bigger to some extent anyway. This is likely to lead to a further response from the EU.

Lobbying isn’t going to work. Legal principle might, to some extent.

The penny will have already dropped back in California and execs and owners will probably be quietly outraged. What about property rights, free markets, rewards for investment and risk taking? Many European telco execs felt the same when ULL took off in 2004. This time around the feeling is exacerbated by the apparent “them v us” nature of the European regulation, with the European rules seemingly targeting largely American enterprises. Surely the weight of US influence can be brought to bear?

From a telco analyst’s experience, this lobbying won’t work. Once a draft regulation has been published by the EC it tends to stick. European telco regulation proved completely impervious to lobbying, despite many European telco CEO trips to Brussels. This failure was despite most EU Member States (including Germany and France) owning a large equity stake in these incumbent telcos and sharing in the considerable downside of the market cap lost. This failure of lobbying reflected the institutional structure of the EU, with the European Commission (EC) able to push its own pro-regulation agenda and, as an unelected body, be relatively immune to outside forces. Similar institutional forces are at play in this new EU internet regulation.

The tendency for the EC to resist outside influence is strengthened in the internet case by some of the wider EU forces at play. The European Commission has long been arguing for the EU to have its own tax-raising powers and balance sheet. This Rubicon was half-crossed in the first covid crisis in 2020 – the EU now has its own eu750bn balance sheet but still no direct tax raising powers. The EC has the power under the DMA to fine internet gatekeepers (for infringement of the DMA) up to 10% of their global revenue on a one-off basis and up to 5% of daily global revenue on an ongoing basis. The fact these gatekeepers have tended to pay low amounts of corporation tax to most Member States – a vexed issue in the EU - will not have been missed. In the eyes of the EU large DMA-related fines could redress this perceived fiscal imbalance and also help fund the EU’s new balance sheet.

The EC seems to be backed up by most Member States, with the Euroactiv website reporting support for the recently published DMA by all Member States except Ireland so far. Most crucially, France and Germany have come out in favour with Germany even suggesting the regulation be extended to more internet companies.   

The EC does however have to act within basic principles set out in European law and adjudicated by the ECJ. Challenging the DMA regulation on legal principle will have more success than lobbying.

There do indeed seem to be a few potential weaknesses in the DMA, for instance regarding how to define what a gatekeeper is. The EU telco ex ante regulation of 2002 applied a concept of dominance already well established in European anti-trust law. The EU DMA strays onto more novel ground, proposing to regulate an internet company as a gatekeeper if it has a significant impact on the internal market, operates a core platform service and enjoys an entrenched position. 

The problem with concepts like “significant impact” is that they then have to be defined. The DMA proposes, as a first test, arbitrary quantitative criteria– to be a gatekeeper the company should 1) have more than 45m monthly active users (MAUs) in Europe and more than 10,000 active business users in Europe, and 2) should have European revenues of eu6.5bn per annum or be part of an entity with a market cap greater than eu75bn. Companies below these size thresholds are to be judged on qualitative criteria, and particularly on the extent of barriers to entry.

Ultimately these concepts - influence on the internal market, core service, entrenched position - are broad and subjective, and potentially represent new legal precedent in the EU.  Why apply these terms just to internet giants but not companies of influence with entrenched positions in any other EU industry (like Airbus, or VW, or SAP), particularly as more industries digitise and organise around internet-enabled platforms they are gatekeepers of. And what are the wider consequences of establishing “influence” as a reason to regulate? And why stop at eu75bn market cap? Whilst the EC itself may not be too concerned with creating new legal norms, the ECJ could be, in some areas.

There is also a question of proportionality, a key EU legal principle. EU law tends to focus on consumer harm. In the telco case, the harm avoided by intrusive regulation was evidenced in the prices charged to end-users, a clear consumer benefit to weigh against the regulatory intrusion. In the internet case, users already get service for free and the perceived harm is evidenced most directly by the frustration of other businesses in trying to gain access to them and their usage patterns. Most end-users are distinctly unbothered by this issue and also really don’t want their data shared around. The benefits of ex ante regulatory intrusion are much more abstract in this case and flow directly to businesses rather than consumers, whilst costs are partly borne by consumers (risk of loss of data privacy and data security). This may cause an ECJ judge to question whether a particular intrusive DMA regulation is proportional.

One could also legally challenge whether a particular internet giant really does have an entrenched position in a particular segment. Whilst telco local loop networks have proven to have strong and enduring monopoly characteristics, internet businesses can be more perishable, particularly in social media. Even Google’s search business is now being challenged by Amazon, now that it has hundreds of millions of Prime customers (one-day delivery is becoming the search killer app).  Some seemingly enduring entrenched positions are likely to prove perishable over time, questioning an underlying precept of the EU regulation.

Finally, the EC seems completely under-resourced to handle any legal challenge to its process, with just 80 full-time equivalent staff to be dedicated to implementing the DMA. The EC team will likely be beefed up in time, and Member States could second staff in the meantime, but the allocated resource does suggest that only a handful of companies will be targeted in the first few years of the DMA and probably in selected market segments.

So there are several potential chinks in the DMA armour. However, legal appeal is more likely to water down the effect of the DMA for some companies in some segments rather than altogether stop the new regulation. Ultimately Europe is exercising its right to construct its own laws and generally takes a more interventionist approach to regulating companies. Having initially fostered freedom for internet companies e.g. through net neutrality laws, the mood in the EU has gradually shifted and the internet is going to get regulated, period. Life is going to permanently change for the internet giants, at least as far as their European business is concerned.


European regulation has scope to be much more disruptive to internet business models and market cap than US regulation, once it is implemented (2022+). This might get overlooked by the financial markets, with EU regulation a dry, arcane issue, and Brussels far away from where most capital allocation is going on. The network effects at the internet giants are so strong that EU regulation will probably fail to stop the big getting bigger (which will lock EU regulation in place and lead to its extension over time). What will matter for internet-giant stock prices is more whether EU regulation slows the rate of growth down. Time will tell.