South Africa – #DataMustFall – the unaffordability of mobile Internet access
The issue of the high costs of mobile data or mobile Internet access in South Africa has been kicked around for several years, with little success either in terms of producing a convincing analysis or, more importantly, of designing an effective intervention. Work on reducing mobile termination rates over many years seems to have achieved little in this regard.
Among the more obvious suggested causes of the high prices have been:
· An uncompetitive market structure (dominated by two players);
· Ineffective legislation; and
· A weak regulator.
A central problem is legislative complexity, that regulating high prices and market structures falls between two stools, the Competition Commission and the Independent Communications Authority of South Africa (ICASA), and that the Electronic Communications Act (as amended) is not fit for purpose. Consequently, many options cannot be implemented quickly and would require months, if not years, to draft new laws, push them through a new parliament, that might have other priorities, then have independent bodies implement them correctly, following consultations, impact assessments and the inevitable delays from appeals, almost certainly to the Constitutional Court. Consequently, there can be neither a quick fix nor a silver bullet.
At the request of the ANC government an inquiry was launched in 2017 by the Competition Commission, under Section 43B(2) of the Competition Act 1998 (as amended):
… to understand what factors or features of the market(s) and value chain may cause or lead to high prices for data services, and to make recommendations that would result in lower prices for data services.
It has now published its interim report, just in time for the election.
Benchmarking
The Competition Commission has noted the complexities of comparing the prices for mobile data with pre-paid and post-paid, with ‘free’ Megabytes and the rolling over of unused data. It identified work by International Telecommunication Union (ITU), Tarifica, ICASA, and Research ICT Africa (RIA) in which South Africa featured.
RIA reported a quarter-by-quarter deterioration of the ranking of the data prices charged in South Africa. While, ICASA found:
Disturbingly … that Vodacom [Vodafone Group] and MTN [Group] prices in South Africa are considerably higher than the prices they charge in other countries in which they operate.
It is worth looking at data from the Rewheel Digital Fuel Monitor, which shows some very much lower prices for the first half of this year in Europe and OECD countries. At the extreme is Finland with a median price of €0.10 or ZAR 1.60 for Gigabyte, while the EU average was €1.90 or ZAR 30.40.
In the SA Telco Sentiment Index, Brandeye found very little sympathy with the operators:
SA consumers are unhappy with telco providers’ data pricing and poor customer service as 1 in 10 threatened to switch networks and 44.4% of complaints on social media went unanswered by providers.
The Business Day cartoon (above) captures some of the feeling about the two biggest operators.
In terms of competition analysis, one material finding was that South Africa performed markedly worse for pre-paid than for post-paid customers, which:
… indicates a potential structural problem with retail prices in South Africa, whereby poorer, prepaid consumers are exploited with relatively higher prices than the wealthier post-paid consumers.
Consequently, use amongst post-paid subscribers was growing, while it was flat amongst pre-paid customers. This limits or precludes the regular use of apps for education, health, job searches, shopping and the like. Given poorer customers purchase smaller bundles of data they are most heavily hit by out of bundle (OOB) payments, up to half their spending, compared to negligible levels for subscribers buying larger bundles.
The Competition Commission argues that the mechanism was that poorer consumers were limited to mobile operators, having far fewer opportunities to use alternative networks. Whereas, wealthier consumers were likely to have Wi-Fi on their fixed broadband service at home, at work or university, and in shopping malls. A drop in mobile data price for poorer customers was unlikely to result in a surge in data usage leaving operators no better off.
The benchmarking revealed little that ICASA did not or should not have known for years, but set it out more formally and in terms of competition law. Importantly, it described, perhaps tentatively, why mobile operators can and are able to discriminate against poorer customers. Some data was not provided, notably comparisons of the amounts of spectrum available to operators, the levels of capital expenditure, Herfindahl–Hirschman Index (HHI), and the rules for sharing masts and towers. A survey of consumers would also have been helpful, to obtain insights into the willingness to spend money on data and the reasons for holding multiple SIM-cards and devices.
A shortage of spectrum
One cause of the high prices has been attributed to the failure of government and ICASA to release yet more spectrum to mobile operators. This has been the subject of extensive debate and featured in the 2019 party manifestos. It is also said to be the subject of a dispute between the current minister and ICASA, with the former allegedly interfering in matters that the latter ought to determine independently.
A significant part of that spectrum has been tied up in a gross policy failure, the migration of broadcasters from analogue to digital technologies, which was supposed to have released frequencies for mobile broadband. Amongst the causes of delays were the decisions by Zuma to keep changing ministers and then to split the ministry, followed by the decision of Ramaphosa to reverse the split, with yet another change in ministers.
The Department of Telecommunications and Postal Service (DTPS) made serious errors in drafting the Electronic Communications Amendment Bill it presented to Parliament, containing important provisions on spectrum. This resulted in its withdrawal until after the election.
The Competition Commission proposes that spectrum assignment be designed to be pro-competitive. Effectively this means caps on the spectrum held by MTN and Vodacom, but not Cell C, Rain and Telkom Mobile. Such a decision might well end up in the courts and might anyway prove counterproductive.
Facilities sharing
The Electronic Communications Act 2005 (as amended) contained overly complex provisions, inspired by European Union legislation, for network sharing that proved extremely difficult or even impossible to implement.
Nonetheless, the Competition Commission:
… is of the view that efforts to enhance facilities access and sharing can substantially reduce operating costs and ensure the rapid deployment of competing infrastructure, to the potential benefit of lower prices eventually.
It is very doubtful whether proposals for network sharing can be implemented under the existing legislation, with any attempt to do so risking delays and failure. It almost certainly needs new legislation and the strengthening of the regulator.
Alternatively, measures could be taken under competition legislation before the Competition Tribunal, based on the provisional findings of dominance against Vodacom and MTN.
Improving price-based competition
The Commission notes that the retail mobile market “has remained stubbornly concentrated”.
It is worth recalling that the decision to introduce competition was taken in 1993 and that the ANC government botched the subsequent entry of the third operator and abandoned plans for a fourth. The entry of the fourth operator was largely accidental, with Telkom creating 8ta only after it was bought out of Vodacom, going on to perform very poorly, despite access to substantial financial resources and infrastructure.
It is best not to disinter the appalling efforts of government with the under-services areas licences (USALs), the two Sentech fixed wireless network initiativess or Broadband Infraco, other than to note the complete absence of post mortem reports. Policy failures are rarely subject to governmental or parliamentary inquiries.
Setting aside how South Africa got where it is and any associated blame, the Commission finds that:
Vodacom has a share in mobile services more generally, and data services specifically, that exceeds the thresholds used in the Competition Act for a conclusive determination of dominance. MTN has constantly skirted around the threshold level where there is a rebuttable presumption of dominance. These shares have barely changed over time.
It found that the two operators were largely able to ignore the other operators, what the Commission rather quaintly calls “challenger networks”. Cell-C was licensed in 2000, in somewhat peculiar circumstances, while Telkom launched 8ta in 2010, and Rain launched a data only service in 2018.
The Commission notes that the first two “challengers” have failed to make progress by offering lower prices. The explanation offered is that MTN and Vodacom have significant first-mover advantages in terms of network quality and coverage. Customers will not readily switch away from the big two players, given the markedly poorer networks of their rivals.
The problem with this analysis is that the first-mover advantages date back a quarter of a century. If they still exist today, as the Commission argues, then the explanation cannot lie with the firms, but with government and the regulator. In any politico-regulatory system some decisions create path dependencies, but subsequent analyses, legislative changes and market interventions provide opportunities to remedy serious failings. Clearly, these interventions have been ineffective in respect of allowing first-mover advantages to become entrenched. The question is why after so many years and so many interventions do first-mover advantages still exist and why have the government and regulator failed to facilitate challenges?
However commendable it would be remedy these deeply entrenched advantages, it would not bring short term benefits.
Virtual operators
There are some who believe mobile virtual network operators (MVNOs) can bring more competition and shake up the market. This requires regulated wholesale access to the networks of MTN and Vodacom that present legal and regulatory challenges and consequently create uncertainties and delays.
The former use of service providers did little good for consumers and was never properly analysed.
Forcing the introduction of more MVNOs might help, but it is far from certain and unlikely to be quick.
Fixed lines
A case can be made for improving fixed line broadband, if only to facilitate offloading from cellular networks and to provide Wi-Fi for public use.
The saga of the provision of broadband over the Telkom PSTN points to the serial failures of the state, a state-owned enterprise and the regulator.
Today, the concern is to drive the provision of more fibre to the home (FTTH) by the likes of Liquid Telecom and Vumatel. Any encouragement is to be welcomed.
Self-criticism and price reductions
The Commission has suggested the operators confess their sins and make commitments to reduce prices, especially the differential against poorer users. However, the weakness of this request is shown by its alternative:
Absent such commitments, regulators should coordinate around a legislative or regulatory means to achieve such outcomes which may include amendments to the ECA, additions to ICASA’s End-User and Subscriber Service Charter Regulations, obligations or an investigation of excessive pricing to lower income consumers by the Commission.
These look to be medium-term efforts that could easily be delayed or rendered ineffective. They are unconvincing threats, unless they can be turned into something that can be implemented quickly and decisively.
One lever that might work in the short term is the allocation of so-called high demand spectrum, where the Commission suggests:
Pro-competitive assignment may include spectrum caps on larger operators, asymmetric assignments and set asides for new entrants such as the WOAN, in a manner that ensures a prospect of commercial success.
Again these are medium- to long-term solutions. They presuppose that the Treasury could be persuaded to forego large sums of money from the two dominant operators to allow changes to the market structure.
None of the ‘challenger’ operators, Cell-C, Rain or Telkom has the money to buy expensive spectrum and to deploy a major network. Assigning the spectrum to Infraco or Sentech would be to ignore their inexcusable past errors. The WOAN is an untried model that has the look of a Gupta-inspired scheme, with the requirement for a remarkable amount of capital to build out a national network.
Given the dangerous fascination of the government with the fourth industrial revolution (4IR), there will be pressure to deliver 5G networks in short order, something only MTN and Vodacom can do at any scale. Indeed, there is already some sort of dispute between the Minister and ICASA over 5G spectrum plans. Path dependencies leave MTN and Vodacom in the key position, having much of the necessary infrastructure, even though 5G needs many new masts and antennas.
A crude but potentially effective option would be to attach price controls to any new spectrum licences.
MTN and Vodacom could be allowed spectrum provided they adhered to price caps based on international benchmarks and specifically limit the differential between post-paid and pre-paid charges. The cheapest data unit prices in Gigabytes per Rand would be subject to an international benchmark with a specified methodology and a list of comparator countries. Additionally, the most expensive prices could be limited to no more than, say, a 25 per cent more than the cheapest prices, a factor that might be reduced over time using a glide path. The operators could also be obliged to publish the cheapest, median, mean and most expensive prices on a monthly or weekly basis.
Separation
The Commission opens the way to litigation, legal uncertainty and delay by suggesting the use of separation:
In both these cases, some form of functional and/or accounting separation may be required of the larger networks if there is to be greater transparency as to the costs of the radio access network (RAN) and core network relative to the retail services.
It would be very risky to set ICASA the tasks of determining costs and regulating wholesale access, since it does not have the track record to suggest it has a reasonable chance to delivering. Moreover, it could take years to sort this out.
These are certainly some lessons from the Telkom settlement agreement with the Competition Commission that helped transform wholesale access to fixed infrastructure:
In addition, the history of failure to engage in necessary wholesale regulation, not just of mobile but also fixed line markets, which has resulted in entrenched concentration strongly suggests that reform to the legislative and/or regulatory framework is most likely required if the institutions are to deliver on this type of regulatory action going forward.
Again, this is not for the short term.
Conclusions
A new parliament will be convened in a matter of days, presumably with yet another ANC government that must face up to a range of challenges, one of which is improving mobile Internet access by addressing #DataMustFall. Much will depend on who is appointed Communications Minister and whether they take up the existing draft legislation or begin a new round of policy formulation to create some different proposal that might take many months, followed by many more months to pass the necessary law. In an ideal world, they would also seek to strengthen ICASA to be able to regulate dominant players more effectively, though the evidence is that the ANC is seeking to reduce its independence. Bringing ICASA under the Competition Tribunal would be one option.
Given the likely cost of migration to 5G and the absence of prospective market entrants, government is limited to the existing players. It can, in the medium term, try to boost alternative or ‘challenger’ operators with pro-competitive policies, but that will not solve the immediate problem.
Since the two major players are also dominant there are risks of making matters worse, of further entrenching their positions.
A price cap could be imposed as a condition for immediate release of spectrum to the existing operators. The cap would have to take the form of a maximum differential between the prices for post-paid and pre-paid users.
The operators make the usual complaints, that they have reduced their prices since the benchmarking exercises, that the circumstances in South Africa are not comparable, that investments necessitate higher rates, and so on. These echo mobile operators down the decades, having been heard on mobile termination rates (MTRs) and roaming charges with tedious regularity in the 1990s, 2000s and 2010s.
It would be helpful to conduct post-legislative scrutiny of the ECA and a peer review of ICASA, which could feed into future legislation and policies for the longer term.
Further comments
The Commission has invited stakeholders to make further submissions to [email protected], including comments on the findings and the recommendations by 14 June 2019.