Internet.org is misleadingly marketed as providing access to the full Internet, when in fact it only provides access to a limited number of Internet-connected services that are approved by Facebook and local ISPs. In its present conception, Internet.org thereby violates the principles of net neutrality, threatening freedom of expression, equality of opportunity, security, privacy and innovation.
This chapter critically examines the relatively few examples of regulatory implementation of network neutrality enforcement at national level outside the EEA examples of Norway, Slovenia and the Netherlands explored in Chapter 2. I draw on co-regulatory and self-regulatory theories of implementation and capture, and interdisciplinary studies into the real-world effect of regulatory threats to TMP. This involved appropriate fieldwork to assess the true scope of institutional policy transfer.2
The developed countries have recently legislated for or regulated for ‘net neutrality’, the principle that IAPs should not discriminate between different applications, services and content accessed by their users.3 This victory for net neutrality proponents came after 20 years of attempted discrimination between content streams within the walled gardens of both fixed and mobile IAPs, such as AOL in the 1990s, and Vodafone Live/360 in 2002–11, which was intended to challenge the Apple App Store and Android/Google Play.4 Alongside their walled gardens, these IAPs enforced monthly data caps preventing their customers having unlimited use of the Internet. Fixed-line walled gardens failed in view of the easy access to the Open Internet at increasingly low cost offered by broadband access. Continued attempts to maintain walled gardens since 2006 have focussed on both ‘negative’ and ‘positive’ net neutrality. I explain both in turn.
Negative neutrality is the blocking and throttling of content that threatens the business model of the IAP. This can be relatively benign when it is spam email and viruses that are blocked. It can also be self-serving and anti-competitive when it is unjustified and unreasonable restrictions on user’s preferred content that is affected – for instance P2P file sharing or video streaming. It is this ‘negative’ net neutrality which is the target of most legislation in the area, based on the generic regulatory principle of ‘first, do no harm’, in this case eliminating the harms caused by unreasonable negative blocking or discrimination. Cases in the US such as Madison River and Comcast were about blocking, and is it this that rouses much consumer anger and political action.5
Zero rating is ‘positive’ net neutrality violation involving not blocking, but treating some content better than general Internet traffic. As cable TV provides high definition and standard video and television channels at high fees in a separate logical pathway to the general Internet traffic on its cable, some telecoms companies hope to partition their Internet traffic to replicate this business model. Several IAPs attempted this practice over lengthy periods, notably by excluding television channels from monthly data caps for users, positively discriminating in favour of their affiliated content and against other video providers (such as YouTube).
‘Walled gardens’ thus reappear with much more specialised walls – restrictions that affect only certain non-affiliated types of Internet traffic, such as social networks or video. This exclusion of preferred content from data caps is described as ‘zero rating’ because all that downloading costs precisely zero in terms of counting towards monthly bills.6 Note that many fixed IAPs have virtually unlimited data use as part of their offer, made possible because maximum speeds and user profiles mean that the cumulative download burden does not overstrain the network.
Zero rating is only possible when users take an IAP subscription which has a data cap, which is generally at a much lower limit when imposed by mobile rather than fixed IAPs. Unlimited data plans mean users can download as much data as needed using the Open Internet pipe, whose speed is restricted only by the Internet itself, or the type of CDN used to supply media.7 When a cap applies to a monthly subscription (such as 1 Gigabyte a month8), that limits the amount of content that a user will choose to access. If data is as expensive as it can be in developing countries, any content can prove too expensive to access for the average user. Offering certain content on a ‘zero-rated’ basis means that content will not be included in the data cap – which is particularly useful if that content is streamed video, audio or an application used regularly, such as social network Facebook or messaging app WhatsApp. That content may be locally stored, relieving congestion in the network, as a result of partnership with the IAP, justifying in network engineering cost terms the decision to reduce the apparent end user cost, if not eliminating it completely.
Research into comparative net neutrality law in the Global South has recently been carried out by several NGOs and is well reported in the specialist media.9 Odlyzko notes that the zero-rating debate exists in one Asian country, but does not explore this in depth, while Marsden discussed monthly caps before zero rating had become commonly identified.10 Net neutrality dates to the 1990s, as does zero rating, even if the term of art was coined much later.11 There are ten times more mobile (5.6 billion) than fixed line connections (572 million) in developing countries, whereas the developed world ratio is 3:1. There are five times more mobile broadband subscriptions in the developing world with 2.37 billion to only 429 million fixed subscriptions (developed world 1.09 billion mobile to 365 million fixed at a ratio of 3:1). Seventy per cent of Internet users totalling over 2 billion people are outside the EU/US.
Social networking using Web2.0 software expanded from a very low base at the start of the smartphone era. Brown and Marsden explain that ‘Facebook grew from nothing in 2004 to become the second most popular destination Web site in the world by 2012.’12 It grew to 100 million users by autumn 2008, surpassed 1 billion monthly active users (MAUs) in 2012 and 1.5 billion in 2015, which included 210 million in the US/Canada and 112 million in its second largest market, India, in autumn 2014. Facebook was floated on the stock market in 2012, warning investors in its prospectus that ‘There is no guarantee that popular mobile devices will continue to feature Facebook, or that mobile device users will continue to use Facebook rather than competing products.’13
A particular business model for this practice is that of dominant social network Facebook, which from 2009 introduced Facebook Zero with mobile IAP partners, and in 2015 introduced a wider walled garden called ‘Internet.org’ (which despite its name is an Intranet for 30–40 affiliates), which was rapidly renamed Free Basics in late September 2015.14 In May 2015 opposition to the highly exclusive and non-transparent Internet.Org had led to content owners abandoning their previously negotiated tenancies, and mobile IAPs dropping the service.15 Free Basics has less powerful gatekeeper functions than Internet.Org and more content is permitted, with officially only technical grounds for refusal, but it is still only governed by a contract with Facebook, which Facebook can change unilaterally.
Politicians and telecoms executives who now claim to be in favour of net neutrality are in fact conceding that blocking and throttling users is no longer acceptable to politicians and therefore regulators. They largely only concede ‘negative’ net neutrality. ‘Positive’ net neutrality is a much more contested topic, and where download limits apply or ill-defined Specialised Services carry the zero-rated content, this concept of zero rating will be heavily contested. That is more the case with mobile than fixed networks, and more the case with developing nations’ mobile IAPs than developed.
The book thus far has relied to a large extent on the experiences of developed nations, especially the United States, EU and United Kingdom.16 I focus here on four case studies, beginning with the earliest effective regulation in Chile. This chapter summarises each nation’s development of net neutrality, and focuses on its implementation of regulation against zero rating since 2014. The methodology used both literature review and empirical interviews in the course of 2015.
Chile has the world’s earliest net neutrality law (from 18 August 2010)17 and an implementation of regulation permitting zero rating from 2014. Law 20.453 includes a provision which adds Article 24(h–j) to Law 18.168 ‘General de Telecomunicaciones’. Article 24H expressly forbids IAP practices that ‘arbitrarily distinguish content, applications or services based on the source or ownership thereof’. This would be relied upon by those opposed to zero rating. The original law required IAPs to self-report on any violations, resulting in infringement only for failure to report. Cerda reports that there were ‘allegations of negligent supervision of the law by public authority’ in failing to enforce consumer rights.18
In Chile,19 all four mobile IAPs (Claro, Entel, Telefónica and VTR) were notified to cease zero rating in 2014.20 The regulator’s (sub-secretary of communications: Subtel) conclusion was misreported in the developed nations’ media as banning all zero rating from 1 June 2014, when it applied to social networks, notably Facebook and therefore Internet.Org.
Subtel stated: ‘las empresas que entregan algunas redes sociales gratis, lo que hacen es privilegiar el uso de estos servicios, mediante el acceso a una Internet bloqueada, excluyendo las redes sociales privilegiadas’ – social networking apps received positive discrimination (‘privilegiadas’) when included in the zero-rated offer. The Chilean situation is complicated by Wikipedia Zero announcing on 22 September 2014 that it had negotiated an exemption from the rules, on the basis that it is neither a social network nor a commercial offer.21 As carriers have not asked Subtel to confirm this exemption, and Wikimedia does not have standing (as a non-carrier) to request that official explanation, the evidence for this is Wikimedia’s version of the exchange and its continued zero-rated offer in Chile.
In fact, Claro (a subsidiary of Mexican operator America Movil, also active in Brazil, Columbia and other Latin American nations) was permitted by the Chilean regulator to continue zero rating as long as it formed part of a wider data plan that customers could choose.22 This was because data plans were included in the new zero-rating offer, removing the part of the complaint relating to ‘cuando los usuarios salen a través de un enlace externo, las empresas piden pagar’ – that non-zero-rated websites have to pay for users to exit zero rating onto the wider Internet. Zero rating would have to stop when users exhausted their data plan each month, in order that they were not left with only zero-rated content, which would be very explicit discrimination.
Brazil has had zero rating since prior to 2014, when it was a common practice by several mobile ISPs. Like Chile, Brazil has a bicameral constitution with a powerful directly elected executive president. Brazil had discussed net neutrality since the mid-2000s, with its formal advisory committee on Internet governance passing a resolution known as the Decalogue in 2009, which in part stated: ‘Filtering or traffic privileges must meet ethical and technical criteria only, excluding any political, commercial, religious and cultural factors or any other form of discrimination or preferential treatment.’23 This led to a period of public consultation led by the Ministry of Justice in 2009 (29 October–17 December) over a potential new legal framework. In 2011 the Chamber of Deputies (lower house of parliament) began negotiations on a law on privacy and net neutrality led by Deputy Alessandro Molon, which stalled in 2012/13.
In late 2013 the political process was accelerated due to President Roussef’s concerns over foreign surveillance of telecoms and Internet traffic (specifically her own communications), resulting in the Senate ratifying the Chamber of Deputies’ proposed law in a single month.24 Law No. 12/965 (the Marco Civil da Internet) was signed by the president at the opening ceremony of the Net Mundial conference in São Paolo in April 2014.25 The relevant section is Article 9, which states: ‘The party responsible for the transmission, switching or routing has the duty to process, on an isonomic [equality before the law] basis, any data packages, regardless of content, origin and destination, service, terminal or application.’ According to Article 9(3) ISPs must ‘act with proportionality, transparency and isonomy’ and ‘offer services in non-discriminatory commercial conditions and refrain from anti-competition practices’. The question for regulators implementing zero rating is whether it is proportional, transparent and non-discriminatory.
Unsurprisingly for such a rushed final law, the consequent implementation has proved controversial, not least because it is not clear which of two consultative bodies and the Ministry of Justice should be in charge of the drafting and enforcement of the subsequent rules.26 Article 9(1) states that it ‘shall be regulated in accordance with the private attributions granted to the President … upon consultation with the Internet Steering Committee [CGI] and the National Telecommunications Agency [Anatel]’. In 2015 both the regulator and the Ministry issued consultations, the latter organised together with the CGI in the period 28 January-30 April.27 The results of the consultation were made public in a Presidential Decree on 11 May 2016, which banned zero rating.28
It is unclear whether zero rating or Specialised Services will be effectively regulated at the time of writing, despite the 11 May 2016 Regulation. At the 2015 Summit of the Americas in Panama on 10 April, President Rousseff met Mark Zuckerberg and was photographed with him,29 he in a suit, she in a Facebook hoodie.30 Her pronouncements in favour of Facebook’s work in Brazil with poorer communities, and by inference Internet.Org, were a public scandal in view of the open consultations then ongoing. However, it is not clear what benefit such public lobbying achieved for Facebook/Internet.Org/Free Basics.
In practice, in 2014 Anatel chose not to regulate zero rating. TIM (the Brazilian subsidiary of Telecom Italia Mobile), in partnership with WhatsApp, released a zero-rating plan that allowed subscribers to use the app in zero rating. Marcelo Bechara, counselor of Anatel, refused to regulate in the absence of specific prohibitions: ‘If there is no prioritized traffic, I do not see why it breaks the Marco Civil. This is the free market. It’s free business.’31
In 2015 Claro abandoned a previous offer that provided zero rating only, and adopted the Chilean approach with free WhatsApp, Facebook and Twitter offered only to users who also subscribed to data plans (pre- or post-pay).32 Claro CEO Carlos Zenteno had said in April that zero-rating plans were no longer part of the carrier’s strategy, as less than 1 per cent of customers used only Facebook or Twitter, and in June added: ‘It’s an evolution. We realized that it has no purpose only to offer zero-rating access to one site.’33 Claro argues that zero-rating on top of existing data plans represents a positive discrimination that the consumer chooses. Anatel’s decision on this issue will be critical to the future of Brazilian zero rating.
Ramos states that:
the gap between those who can pay for data caps and those who cannot afford them could lead to a two-tier internet: the ‘internet of the rich’, or those who are wealthy enough to pay for the unlimited access; and the ‘internet of the poor’, which would give access only to a few applications that would be affordable to poor people.34
In Brazil, such a digital divide has a potent political force, given that the policy of progressive governments, since Cardoso was elected in 1994, has been to narrow the inequalities that grew in the military dictatorship and before. Brazil was becoming a less unequal society until its recession, which began in 2015. But as Ramos explains:
the existence of two different ‘internets’ could distance the rich from the poor (with application providers creating services aimed for the rich and ‘light versions’ aimed for the poor). Ultimately, it could lead to a replica of the social apartheid currently perceived in many developing countries, where slums have limited access.35
It could lead to a perceived ‘gringo net’ where only the rich can afford to access the full Internet with its many foreign apps and services. That said, the ISPs in Brazil plead not to be made tools of social engineering, arguing that inequality is a matter for governments not companies, however integral their service to the socio-economic landscape.
Brazil has consulted on net neutrality in two phases, the first running in spring 2015 in which the zero-rating issue emerged as the most significant and commented-upon controversy, the second from 27 January 2016. The second phase resulted in the Ministry of Justice Regulations via Presidential Decree in 2016,36 and the eventual fate of zero rating remains uncertain. It remains legal in the absence of Anatel action.
On 8 February 2016 India banned zero rating.37 India has a population of 1.25 billion, with a billion mobile users or almost 80 per cent of all citizens, but low data use on smartphones, and only 26 million fixed telephone connections.38 Only 57 per cent of Indian (and 43 per cent of Brazilian) smartphone users actually use data plans at all, and the average amongst those Indians who do was 80MB a month in 2015 (3–5 per cent of developed national average usage).39 With a very low fixed Internet subscription rate, most Indian consumers primarily rely on the mobile Internet for data. The regulator is the Telecom Regulatory Authority of India (TRAI), which had consulted on net neutrality in 2006 when the issue first arose, with little public debate.40 By contrast its spring 2015 consultation produced over a million emails in reply, focused on zero rating.41
In India, three zero-rated options were offered in 2015, by both Internet.Org, owned by Facebook using the Reliance network, and Airtel (the largest mobile IAP in India with 226 million customers at April 2015). In summer 2015 an Indian government committee suggested that the locally based Airtel’s zero-rated option should be permitted but foreign-controlled Facebook’s Internet.Org prohibited.42 In response to concerns most vociferously raised in India but also in Brazil, the US, and other nations, Facebook made the terms of Internet.Org more transparent in May 2015, effectively opening access in principle to any app developer who could meet its terms.43 Nevertheless, Facebook’s privacy policies continue to apply and it is not possible to use Internet.Org without also being a Facebook user, while Facebook accesses all your tracking behaviour while logged in to any partner sites and can share that with mobile IAPs.
Internet.Org’s policies were carefully analysed by the Centre for Internet Studies in India.44 It was a matter of great priority for Facebook to expand its mobile network partnerships rapidly internationally, especially in India, in the face of a decline in youth MAUs in its home US market from 2013. The prize for Free Basics was to grow the number of subscribers in the Indian market more effectively. Zuckerberg stated:
[through] Internet.org in India now, there are already more than a million people who now have access to the internet who didn’t otherwise … in terms of DAU (Daily Accessing User) growth, the three largest countries were India, the US and Brazil.45
The threat of regulatory action was expressed in July 2015 by the Joint Secretary of the Department of Telecommunications, V. Umashankar:
[I]f the need arises, the government and the regulator may step in to restore balance to ensure that the internet continues to remain an open and neutral platform for expression and innovation with no [IAP], or for that matter any content or application provider, having the potential or exercising the ability to determine user choice, distort consumer markets or significantly controlling preferences based on either market dominance or gatekeeping roles.46
He explained that the Telecoms Committee report delivered in July 2015 proposed ex ante regulation: ‘a licensee has to file the tariff plan with TRAI prior to the launch. TRAI would examine each such tariff filing carefully to see if it conforms to the principles of net neutrality and that it is not anti-competitive by distorting consumer markets.’47 Should zero-rating have already begun, as with Internet.Org and Airtel, ‘penalties will be levied if there is a violation’.48
Facebook’s partnership with third largest mobile operator Reliance Communications (RCom) to deliver Internet.Org was suspended on 24 December 2015 by Reliance, based on a request from the regulator TRAI.49 The sequence of events was apparently that RCom informed the regulator on 23 November that it offered Free Basics, to which the regulator replied on 21 December, and asked the carrier not to deploy before submitting the terms and conditions, which included tariff plans. This led Facebook CEO Zuckerberg to interrupt his paternity leave to write an extremely aggressive statement in a major Indian newspaper on 28 December, accusing critics of misrepresenting Facebook’s plans.50 This backfired spectacularly by raising the spectre of economic colonialism, which is a very emotive issue for India, even 70 years after gaining independence from the UK. Guha and Aulakh explain that:
On December 9, Facebook started a mass campaign on its platform asking users to support Free Basics and urged them to email Trai declaring their support of ‘digital equality’. Free Basics was sought to be conflated with digital equality, with Facebook pitching the product as a solution to connect the unconnected billions. [TRAI] had called Facebook’s Save Free Basics campaign a ‘crudely majoritarian and orchestrated opinion poll’. It also pulled up Facebook for the responses, which the regulator said didn’t address any of the questions posed in the consultation paper. On January 1, Trai asked the company to alert its users to send revised responses to the questions on the consultation paper as a vote for Free Basics did not hold up as a valid response.51
The Prime Minister, who had been a supporter of Free Basics less than four months earlier, advised Facebook to behave less aggressively: ‘government must not allow any platform, no matter how popular, to monopolise any information system in the country as it can have far-reaching social, political and economic ramifications.’52 This was the clearest indication of political pressure on the regulator to find against Facebook, which it did four days later.
The resulting regulations ban zero rating by both Free Basics via its Indian partner mobile network RCom, and domestic network Airtel’s own zero-rated offer. Those offers that subscribers have already received were permitted to continue for six months (to August–September 2016), but any breach of that or any zero-rated (called ‘differential pricing’ in the Regulations) offer to new subscribers would make the licensed network operator liable to daily fines of 50,000 Indian Rupee (about US$700–750). Licensing is permitted and controlled by the Indian Telegraphy Act 1885. Though these fines are low, the context of the regulator’s power over other licence conditions makes it unlikely that a network operator would not comply.
India’s road to a zero-rating ban has been unusual: the regulator in spring 2015, and the Prime Minister in September 2015, appeared minded to support differential pricing, but the strength of public opinion and lobbying directed by civil society coalition SaveTheInternet.in, compounded by Facebook’s culturally insensitive aggressive lobbying, led to a complete reversal within months.53 Whether that decision leads other (post-colonial or otherwise) regulators into similar bans remains to be seen.
Canada had a chequered record on net neutrality until 2015, with rules proclaimed by the regulator in 2009 but not enforced until 2015. In 2011 the regulator explicitly supported capacity-based billing (rate caps),54 which led the main IAPs to stop throttling video and other high bandwidth content as they had admitted to doing since 2008. It then adopted greater enforcement practices for net neutrality in 2014.
In 2008 the dominant incumbent Bell Canada was not ordered to stop throttling smaller IAPs to whom it provided wholesale connectivity, CRTC instead launching a wider inquiry into Internet Traffic Management Practices (‘ITMP’ was the acronym used).55 In October 2009 Canada’s regulator CRTC announced that it would in future examine infringements of net neutrality on a case-by-case basis,56 using existing powers under Section 36 of the Telecommunications Act 1993, which states: ‘Except where the Commission approves otherwise, a Canadian carrier shall not control the content or influence the meaning or purpose of telecommunications carried by it for the public.’57 Thus the regulator chose not to act on any individual complaints until 2011. In 2011 Geist then documented failures to investigate, let alone act.58 A much-heralded 2011 ruling on ITMP and data management caps was little enforced.59 Until 2013, Canada’s regulator claimed the power to regulate net neutrality, but chose to forebear, claiming no evidence of problems that would justify action by the regulator.60 Even in mid-2015, research by Geist reveals that CRTC emphasises transparency over fining miscreants where IAPs are shown to have misled consumers over net neutrality violations.61 The main form of Canada’s net neutrality rules is not the ITMP decision itself, but rather provisions in the Telecommunications Act that predate the Internet (section 27(2) (no unjust discrimination) and section 36 (no interference with content).62
Jean-Pierre Blais became chair of CRTC in 2012 on the standard five-year term, announcing his arrival with the intention to properly regulate the sector in which the regulator ‘has reputational baggage, I want to build it back up’.63 This was in contrast with his laissez-faire business-friendly predecessors who ‘would rubber-stamp almost anything they [the corporates] proposed’,64 including the net neutrality issue. In his first year, Blaise carried out four major interventions: he rejected former incumbent Bell Canada’s initial takeover of Astral Media, until conditions were imposed that later revised and approved the merger; he limited mobile phone contract durations to two years; he pressured mobile IAPs into halving international roaming fees with the United States; and he investigated unbundling television channels, leading to a decision to force unbundling in March 2015. Bell Canada’s president was also rebuked officially by CRTC for trying to interfere in editorial decisions to ban its TV station’s coverage of CRTC, Blais stating ‘An informed citizenry cannot be sacrificed for a company’s commercial interests … corporate interests may have been placed ahead of fair and balanced news reporting’.65 The president of Bell Canada was immediately replaced on 9 April 2015.
Note that Bell has cross-media ownership of CTV, Canada’s most popular TV channel, and until 2005 also owned the largest circulation newspaper, The Globe and Mail. The Bell Canada attempt to purchase Astral Media (the owner of TV channels HBO Canada and The Movie Network) was announced in March 2012, but regulatory clearance was only given when the majority of English-language programming was divested, alongside local programming and unbundling requirements which would dilute any perceived threat to the public interest posed by dominance of Bell’s programming in English-speaking Canada (note that Quebec, which has a quarter of Canada’s population, is officially Francophone with only 7.7 per cent native Anglophones, mainly around the city of Montreal66).
Zero rating is not common practice, and has not been definitively banned. However, a new CRTC case may lead to a definitive ruling: the Videotron ‘Unlimited Music’ case.67 The net neutrality regulatory battle in Canada played out as a broadcasting ownership battle, in which programme unbundling had as an integral part the decision to regulate zero rating in 2015. It was to be expected that net neutrality violations favouring the company’s preferred content would also form part of broadcasting regulation, specifically oversight of channel diversity. Unbundling of TV channels would not be appropriate alongside increased bundling of Internet distributed channels. CRTC ruled in February 2015 that Bell had been ‘unlawfully’ setting a double standard by exempting its $5-a-month Bell Mobile TV app from download limits it places on subscribers to its mobile network, giving it until 25 April to correct its pricing.68 It also ruled against Quebec rival Videotron. Both are required to change to per-Gigabyte pricing. Bell had argued that the Mobile TV service provides 43 channels, only 12 of which are owned by Bell, the remainder being owned by other Canadian channel operators. The action was based on a 22 November 2013 complaint by student Ben Klass, supported by Telus, who argued that Bell in effect was marking up prices for competing streaming services by as much as 800 per cent.69
On losing the action in 2015, Bell immediately filed a lawsuit in the Federal Court of Appeal, whose hearing is pending, arguing that the CRTC was wrong to issue its decision under the authority of the Telecommunications Act, because Bell Mobile TV app is a broadcasting service, but it acts solely as an IAP for other parties’ video. Bell says that broadcasting rules should therefore not apply, an argument approximating to that of the IAPs in the US who claim Title II telecoms regulation should not apply to their IAP activities. Moreover, given that Mobile TV was providing Canadian content in competition to over the top player NetFlix, nearly all of whose content is from the United States, the Mobile TV decision may be portrayed as opposed to Canada’s national content policies. It illustrates that not all zero-pricing plans may be opposed on the same grounds and potential public interest at stake.
The issue of zero rating is highly contentious – a ‘bad case’ on which to make net neutrality law, as van Eijk describes it. I suggest two regulatory actions to encourage the correct use of zero rating:
- treating zero rating as a short-term exception to net neutrality, and
- ensuring any such short-term exception is not exclusive, by subjecting such contracts to FRAND conditions.
These conditions are not dissimilar to the principles by which the Wikimedia Foundation permits Wikipedia Zero to be offered by mobile IAPs, in that it: ‘allows other public interest websites to ride onto its own scheme, eschews any exclusive rights or exchange of payment between itself and mobile carriers, and forbids carriers from selling the service as part of a limited bundle’.70 I consider exceptions, non-exclusivity and FRAND in turn.
Short-term exceptions to net neutrality are likely given the post hoc nature of regulation: regulators lay out ground rules then respond to complaints regarding infringing practices. Difficult marginal cases can require extensive investigation. Such processes can take several months in the case of effective regulators, requiring both technical and economic analysis, a call for evidence, hearings and enforcement notices. In the case of litigious market actors, appeals against decisions can take months, years or longer to reach constitutional courts as the final appeal court. There is nothing in zero rating to suggest it is anything but a straightforward case of discrimination, which should not be subject to such long appeal processes. As explained earlier, walled gardens are nothing new, represent obvious discrimination and have been outlawed by those countries with effective net neutrality regulation. Any attempt to offer a time-limited zero-rated offer as an introduction to mobile data use could be flagged as such and limited by regulation to perhaps 3–6 months. This would be subject to FRAND conditions and regulatory enforcement.
FRAND conditions could be applied to:
- mobile IAP contracts with Free Basics and other affiliated content providers, including the IAPs’ own subsidiaries, and
- conditions under which the content providers offer access to their own portals;
- however, if zero rating is not taken up by a significant part of the subscriber base (e.g. 10 per cent of each operator’s users), there may be a case for a de minimis exception from FRAND/non-exclusivity. It would be difficult to argue in practice that such a small number on a short-term basis distorts innovation significantly.
The first condition is relatively straightforward to implement in theory but difficult in practice, as it is basically vertical unbundling of the mobile IAP’s business unit arrangements. One could also compare it to the regulatory treatment under EU antitrust law of competitors to Microsoft’s applications interoperating with their dominant Windows operating system.71 However, not all regulators are capable of equal treatment of subsidiaries with competitors, especially in the resource-challenged developing world where independence and regulatory commitment are less easily maintained.
An alternative form of FRAND may therefore be to regulate de facto at a regional or global level, in establishing the ground rules for access to the zero-rated platform which mobile IAPs will offer. In this case, the regulated actor is the ‘host’ platform for those applications that will be offered. If applications to join such a platform offer – such as Free Basics or Wikipedia Zero’s offer – are established under FRAND terms that can be examined and monitored independently, then the platform which is established for one developing market may, with few modifications, prove to be that offered in many others.
Mobile operators would like as much content delivered onto their networks as possible, including zero-rated and directly peered CDNs, such as Akamai or Level3. The appeal of Free Basics is the low bandwidth demand of its apps (no graphics, flash video). Some suggest directly peered CDNs should also be zero rated. It should be much cheaper (though not cost-free) to deliver content from a locally peered source. That should be passed on to the consumer, and zero rating is as good a way as any. Actual costs may be nearer zero than full price in any case. Note that without a data package alongside free content, content providers would be obliged to contract with a directly peered CDN – unless the zero-rating offer is very short term (e.g. three months maximum) to let new users ‘taste’ the edge of the Internet. I argue that FRAND and non-exclusivity should always be applied to zero-rated offers, short term or long.
I argued that zero rating is a relatively minor short-term problem, not technologically but price determinist as I now explain. The majority of ‘mobile’ data traffic is actually downloaded to devices via Wifi in the home, office or a hotspot location. It is not the cost of mobile data plans that is the dominant price driver, but that of hardware and prevalence of Wifi. There can never be as much Wifi in developing as developed countries, but open Wifi can be accessed relatively widely in countries where Internet policy is not dominated by the copyright maximalist lobby and morality (anti-pornography) cybercrime lobby. Hardware for mobile data is much cheaper than at its introduction a decade or more ago in the developed world, whether that be smartphones, laptops or tablets.72 Combining the huge advances in technology pricing/performance with the prevalence of Wifi hotspots in 2015, it is clear that the environment for rapid adoption of mobile Internet access is far better than for fixed access in 2000. This applies despite the extremely high prices for mobile IAP data, which only forms a small part of the adoptive environment required to access the mobile Internet (arguably, no mobile IAP access is required at all given that schools, cafes, universities and other public areas offer free Wifi). Only 43 per cent of Brazilian smartphone users even use their data plans.73
Jurisdiction will be the greatest challenge to any attempt to regulate the platform rather than the mobile IAP offering zero rating. There are three obvious routes to enforcement:
- via the telecoms regulator’s enforcement of platform neutrality on the mobile IAP, and therefore into the contractual terms of its agreement with the platform;
- via antitrust as a merger condition for any platform that chooses to expand into this area; or
- by a considered coordinated response by a network of net neutrality enforcement agencies at regional level, such as in BEREC.
The first has resource constraints, except that the better-resourced early mover regulators may establish ground rules that can be ‘copy and pasted’ by later-acting less-motivated regulators. The second is the type of net neutrality regulation that was adopted in the United States from 2005 onwards as an antitrust ‘default’ rule against large IAPs that wished to merge. In the global view of such mergers, a net neutrality undertaking for a limited time period was considered by the merger partners to be a small price to pay. The third is also difficult in practice to implement, though larger, well-resourced regulators (e.g. Germany/BEREC) advising their smaller cousins (e.g. Cyprus or Malta) can issue a decision or opinion that will help other regulators to take similar or identical action to enforce neutrality. Given the networks of regulators, consultants, civil society actors, academics and law firms that have exported and shared ‘best’practice in telecom regulation since the first liberalisations in the 1980s (in Japan, the US, Sweden and the UK), such networks can be expected to actively engage in spreading such practices internationally.
I considered whether zero rating poses a serious challenge to Open Internet use, examined the country case studies that demonstrate its regulation, and suggested areas for further independent research into the effectiveness of net neutrality regulation. I argued that zero rating is a relatively minor if highly controversial short-term problem as compared to Specialised Services, not technologically, but from a price-determinist view, as I now explain. Next to such a pervasive Internet policy problem versus privacy or free speech, is net neutrality an over-inflated sideshow, or a necessary precondition? Examination of national case studies helps to shed light on the extent to which net neutrality proves an essential pre-condition for solving other less technical, more politically accessible communications policy problems. More research is needed in this field as implementation of national and regional net neutrality legislation increases, especially in Europe; but this examination has shown that the roles of regulatory commitment, civil society activism and national political and market conditions are critical to the resolution of hard cases in net neutrality, specifically zero rating.