Regulation in Sub-Saharan Africa has hit a brick wall, write RUSSELL SOUTHWOOD and STEVE SONG. What changes are needed for a data-centric future?

Sub-Saharan African countries have had over a decade of pioneering regulation that opened up competition and got communications services to a large number of people. But competition has turned to consolidation and the new data-driven future poses difficult challenges.

The period from 1994-2010 will probably be seen as the golden age of regulation in Sub-Saharan Africa. New regulatory bodies issued licences to eager investors willing to compete with each other to make money in this new industry. The sheer thirst for service from the new mobile network operators (MNOs) drove growth at a dizzying pace. It bought new wealth to the operators, to business and government (through new taxes) and it opened up a world of new social and economic opportunities for Africa’s citizens.

The achievements of these new markets gave African countries a justifiable pride in their success. Leapfrogging fixed technology and operating in the toughest of circumstances gave everyone a sense that the near impossible was actually being done. Innovations like mobile money showed that it was possible to create change at every level of the economy, even among the poorest of citizens.

The enemy of these changes was often the state-owned incumbents that resisted change at every turn. Their extensive monopoly privileges were not matched by an equivalent sense of responsibility for the public tasks they were meant to undertake.

But even before the move to all-data platforms became the next tectonic shift in telecoms, the dazzling spectacle of almost stratospheric growth hid several less welcome developments. Newly minted communications regulators were often far from independent and in many cases administered existing market realities rather than focusing on the bigger picture of what regulation was for: getting communications services to all citizens. Over time, as MNOs became more powerful in the market, one or more of them took over the incumbency role from the state-owned operator and sought to entrench their own newly acquired privileges.

Now the challenge of the irreversible shift from a number of highly controlled vertical markets including voice and SMS to an all-data platform has put the business model of all the MNOs under pressure. A data-centric future means that all services will be delivered via the same digital platform including services like financial payments.

This digital future entails substantial new investment in infrastructure that can carry large amounts of data yet with lower rewards than voice and SMS has given in the past. It puts the traditional MNO business model – historically based on voice and SMS revenues – under considerable strain.

As a result, along with other factors, consolidation has begun to happen in many markets. For example, the slow exit of Millicom’s Tigo from Sub-Saharan African markets is the most dramatic of these consolidations.

MARKET ANALYSIS

Sub-Saharan Africa is made up of 49 countries so it is often difficult to generalise beyond broad trends. To make some judgments about how competitive and easy to enter these markets are, we have assembled a table (see page 20) that looks at the major players in these 49 markets. The table analyses the number of national mobile and fixed operators in each market.

In the near total absence of large internet service providers (ISPs), the majority of revenues in each market are generated by, in many cases, a relatively small number of MNOs. Alongside this, there is the classic economic competition assumption that more operators in a market will mean that it is more competitive.

At a high level, the good news is that 53%3 of the 49 Sub-Saharan African countries have either some version of three or four mobile operators or in the case of Ghana, five operators. The different versions include a MNO created through the licensing of the former state-owned incumbent fixed-line operator. Ownership of an MNO by the state is not always a formula for commercial success as the case of Gamcel4 and many others shows.

The bad news is the flip side of this coin: 47% of countries still have markets where there is either only a monopoly operator (4 countries) or a mobile duopoly (20 countries). In other words, after over 20 years of regulation, just under half of the countries in the continent still do not have anything that could reasonably be described as a competitive market.

But even those with a wider range of operators still suffer from things that create less than optimal market conditions. In two of the countries – Angola and Namibia – government capital predominates. In two countries – Kenya and Senegal – there is a clear dominant operator and in both cases the government owns a significant share in these operators. Even where there is no clear cut case of there being a dominant operator, some MNOs have considerable market power through having the largest wholesale fibre optic network in a country.

In Cameroon, the state-owned incumbent Camtel has a monopoly of both wholesale networks and landing stations. Only 18 African countries have privatised their state-owned incumbents and many are inefficient, riddled with patronage and are sources of income for the politicians who control them.

Beyond these formalised state links, there are also networks of patronage through local shareholdings that are granted by government, regulators and operators. Angola and Cameroon are the most egregious examples5 although the former under its new president seems to want to change things.

Continuing commercial pressures are also likely to make these markets far less competitive. Seven of the more competitive markets have operators that are either likely to merge or close down in the short-term. Therefore, in many countries, the mobile space – which represents the majority of voice and data revenues – has become less competitive.

The unintended consequence of assigning spectrum licences to operators at high prices, whether through auctions or other means, has been to create a serious barrier to market entry. Cost of market entry combined with the dominance of existing MNOs discourages most attempts to create new business models. Only an operator with massive wealth and political backing, such as the recent example of Reliance Jio in India,6 can create a business model that is fit for a more data-centric market.

The mobile operators that often had to fight tooth and nail for fair treatment against the state-owned incumbents have now effectively drawn up the ladder behind them. Under pressure from over the top (OTT) players, they have fought hard to firewall traditional profit centres. As an increasing number of Africans become daily internet users, it is reasonable to ask whether the mobile operators will allow enough “air” into the new digital ecosystem (through revenue sharing) or whether they will seek to dominate these new content and service revenues as well.

Independent ISPs that might provide some alternative have struggled to find capital and have often lacked any market vision beyond servicing a relatively small number of corporate clients. For those unable to raise capital on stock markets, the interest costs of domestic loan capital remain in double digits in a period where the cost of international capital from Europe and the US has never been lower.

Large scale, internet backhaul companies with a multi-country presence are small in number (see last column in the table). There is very little progression from starting as a small network communications company and becoming much larger. There is also no route to creating a large number of diverse companies of this kind of any size. For as one of the managers from one of the larger ISPs recently put it:7 “The regulator looks after the larger operators rather than competition through local entrepreneurs. However, these would provide more jobs and choice for local users.”

Globally, in 2018 67% of people were unique mobile phone subscribers and 47% were mobile internet users. By contrast, only 45% of people in Sub-Saharan Africa were unique mobile

OVERVIEW OF SUB-SAHARAN AFRICAN COUNTRY MARKETS

 subscribers and only 24% were mobile internet users. But even the latter figure overstates internet use: based on data from operator, MTN, with the exception of Ghana and South Africa, only 30-40% of its users are active, defined as using more than

5 MB per month. The scale of the challenge to close this digital divide is considerable and not much will happen quickly unless more enabling regulatory environments are created.8

The GSMA graphic (see inset figure) highlights two major issues that require urgent attention. The first issue is that 30% of people in Sub-Saharan Africa are not using mobile broadband in spite of being within range of a mobile broadband signal. There are many barriers to use including gender relations, education, literacy and languages but the simplest one to address is the cost of data itself. Data prices have fallen in many countries across the continent but all too often they remain above what is affordable, especially in rural areas where incomes are lower.

The second issue is harder to solve: 47% of Africans live outside of mobile broadband coverage. Under market pressure, it is unlikely that most mobile operators will seek speedy rollout to these non “market addressable” areas. With a few notable exceptions, state owned telcos and universal service schemes have very limited success in closing the divide in terms of voice so it is unlikely they will be more successful with data.

Given these challenges, where should African regulators place their focus? If everything (including voice through things like WhatsApp) is increasingly just a part of a generic data service and ever larger parts of African business, government and civil society are dependent on access to data to do what they do, what is needed?

The objective must surely be to establish data networks that can deliver a high quality of connectivity to the maximum number of people in the country at the lowest possible price point. The income levels of users in Africa are not the same as countries in Europe and the US and therefore data prices for users need to be lower than at present in almost all Sub-Saharan African countries.

REGULATION OF FIBRE NETWORKS

At the heart of Africa’s data future are the fibre optic arteries traversing the continent. Because of their high capital cost, fibre networks often tend towards monopoly. If there is only one effective backbone network, then the price and quality of service needs to be regulated to meet the pricing objective outlined above.

Fibre infrastructure is a key national asset and should be treated like other key infrastructures like energy and water. Public ownership of fibre networks through government is not necessarily the best and only way of enabling digital transformation. It is a curious fact that state-owned fibre networks nearly always charge more for capacity than similar competitive networks.

Regulators need to focus on driving down prices at the wholesale level in order to enable competition in the last mile and to reduce overall retail prices. At the light touch end, this would be through insisting that there was a publicly available wholesale rate-card (with terms and costs of access) available from any dominant wholesale network operator, public or private. At the tougher end, it would involve agreeing a cap on wholesale prices through a clearly calculated formula.

Governments (at both local and national level) and regulators needs to ensure that the costs and ease of building fibre are reduced to their lowest practical level. Historically, rights of way (permission given by property owners to allow fibre to traverse their land) in Sub-Saharan Africa have been a form of tax-raising for local government. Often, the process of obtaining these permissions is both lengthy and bureaucratic. Once obtained, there is typically no legal or operational framework to protect the fibre optic cables from other civil works like road improvements. Rights of way should be coordinated through an agency capable of offering fast response times and administering the protection of subsequent works. Where appropriate, municipal authorities should actively invest in the building of “open access” fibre ducts to dramatically lower the cost for operators to deploy fibre networks in their cities.

INCENTIVISING MARKET ENTRY FOR NEW OPERATORS

To create new competition, there needs to be significant incentives that will help tip the balance against market advantages of the larger market entities like mobile operators. A sympathetic regulatory framework is needed that does not treat new, smaller operators on the same terms as one of the larger mobile operators.

For example, is it really necessary to licence ISPs? A simple registration process with a low administrative fee (as is found in countries such as

PRINCIPLES FOR A MORE DIVERSE DIGITAL ECOSYSTEM

Open access: No single operator or cartel can use control of aparticular network layer to derive an unreasonable return. Networks are a national asset and should be available to all who participate in the market.

Diversity: Large oligopolistic organisations do not breed innovationand find it hard to change their business models. Nobody knows yet the best business models for data so a diversity of approaches and technology is essential.

Local: Getting local involvement from national to local level isimportant. Having a local minority shareholder is a start but is not enough; there needs to be steadily accumulating capital of both money and expertise locally. The new digital ecosystem needs to embrace a wide range of local actors. The voice of users/consumers should be represented at board level in large organisations.

Innovation: Enabling new market entrants to nip at the heels of theincumbents is the only way to keep the licensed “profit takers” from getting complacent. There needs to be innovation that will address service delivery to low-income and rural customers.

Private sector or state-owned are not the only alternatives: Neither the market nor the state alone will deliver coverage and services to low-income communities. Local structures are often best placed to respond to demand in rural areas.

Canada, New Zealand and the US) would be perfectly adequate. If there are a large number of ISPs and one fails, then it doesn’t affect the whole country.

Licensing costs are but one small part of the barriers to market entry experienced by smaller operators. The cost of spectrum – which is set by the price paid by larger companies – now acts as a means of excluding others. Each new iteration of mobile technologies – 3G, 4G, 5G – creates yet another price hurdle for smaller operators. Thus, as we’ll outline below, there needs to be several different approaches to offering cheaper spectrum access from “free” spectrum (as in the existing ISM bands) to spectrum sharing.

Lastly, there needs to be low-cost access to capital to allow smaller market entrants to grow over time. A revolving loan fund could be set up that allowed companies with below a certain level of turnover to apply for expansion loans. Compared with MNO investments, these sums would be modest but they would help nurture a diverse range of data providers with a far wider range of user business models.

These kinds of adjustments to favour smaller operators are eminently possible. For example, the Kenyan communications regulator will allow local company JTL to spread out its payments for its US$25 million spectrum licence over the 10 year period of the licence. This is a step in the right direction although the cost of spectrum access remains high. The need to service spectrum licence debt may still incentivise operators to build infrastructure in the most lucrative, densely populated urban areas rather than in sparsely populated, lower-income rural regions.

 TURBOCHARGING THE USE OF SPECTRUM

Spectrum pricing acts as a disincentive to opening up new markets, particularly for those on low incomes. Common sense tells you that a person in say Lagos, Nairobi or Dakar can afford a greater average cost of spectrum per head than a person living in a rural village: one size does not fit all. Worse still, the shape of what is possible – both in the market and outside of the market – is then distorted by a spectrum price set by a very small number of buyers. Governments and regulators giving in to the temptation to maximise revenues from the sale of spectrum licences, create the unintended consequence of making it hard to deliver a cost-effective service to low-income users.

A big advantage of the licence-exempt ISM bands is that a whole data ecosystem has evolved using relatively cheap, plug and play Wi-Fi technology. Indeed the majority of data traffic originates from these networks rather than the traditionally more expensive MNO networks. Regulators wanting to promote cheaper data prices should pay more attention to innovations in charging for spectrum elsewhere in the world.

The US is exploring expanding free access to spectrum in the ISM bands from 5 to 6 GHz.9 Also, it is opening up the Citizen Band Radio Spectrum (CBRS) at 3.5 GHz for shared LTE use, something that has gone almost unnoticed elsewhere. There are also existing ISM bands in the 1.7 and 2.4 GHz frequencies, ideal for broadband backhaul, that many African countries make it difficult to access.

Dynamic spectrum technology (formerly known as TV white spaces) offers a proven route to maximising the use of spectrum and sharing it between different operators. In many parts of Sub-Saharan Africa, spectrum is almost unused in uncovered rural areas but the MNOs protest loudly whenever any attempt is made to argue for its use. One approach to breaking this logjam is to create secondary use regulation for spectrum that would allow smaller companies to lease the spectrum in uncovered rural areas at a much lower price or perhaps in exchange for reduced universal service fund contributions from the primary license holder.

A key to creating a greater degree of innovation and competition is the need to have free spectrum that will allow small rural operators to connect back into the core network. Often a single operator will be able to lock out new and potential small entrants through the wholesale terms it offers. As an example, the 11 GHz frequency offers a lot of throughput over long distances and there are a number of innovative technology vendors selling cost-effective equipment that can make this possible.

EMBRACING TECHNOLOGICAL DIVERSITY

Existing MNOs use identical technologies to operate their networks. Key parts of their business are often outsourced to the same vendors, from tower ownership and operation to billing systems to customer management. The key technologies are bought in large packages from a very small number of vendors. Not surprisingly there is not a great deal

• InterMEDIA | July 2019 Vol 47 Issue 2

of competitive difference in technology between operators and despite their best efforts, often not much difference in the services provided and the quality of them.

Data networks allow a much wider range of equipment to be used to deliver connectivity across the network and to customers. Much of what used to be built into hardware is now a software option, offering more opportunities to upgrade and change existing networks than existing analogue equivalents. Smaller companies are often in a better position to take advantage of these kinds of technologies, which can give them a competitive edge over their larger rivals on the all important customer needs of speed and price.

GETTING TO THE HARD-TO-REACH IN RURAL AREAS

MNOs have to produce a return for their shareholders and are therefore not always best placed to address markets with scattered, low-income potential users. Their level of overheads means that their minimum paying customer base per base station is in the low thousands. While network infrastructure costs in general are coming down for all players including MNOs, they remain relatively high-cost organisations trying to address a group of people who require a low-cost solution. The same is true of state-owned incumbents who carry many of the same kind of overheads and as already observed, find it difficult to focus on this kind of task.

There are two kinds of organisation that might fill this service delivery gap: small-scale ISPs and community cooperative networks. Regulators should encourage local entrepreneurs to set up data services in poorly served or unserved areas. These types of local ISPs can begin to address edge markets where there are pockets of potential revenues in small communities.

Rural community cooperatives can take many forms. They can be tiny as the one in Lawrencetown, Nova Scotia, Canada. They can address access challenges in regions that aren’t viable for large operators such as Rhizomatica is doing in rural Oaxaca, Mexico,10 or Zenzeleni in the Eastern Cape of South Africa.11 They can embody commons principles as Guifi.net is doing in Catalunya, Spain.12 They can enable farmers to dig their own fibre trenches to deliver 1 Gbps (symmetric) internet service as Broadband for the Rural North (B4RN) is doing in the UK.13 All of these structures and approaches have relevance in the African context.

Local organisations will help keep income in local communities. It is interesting to note that affordable access to broadband in rural areas is by no means limited to African countries or even low-income countries. The need for regulatory reform to address inclusive access is a global one.

Without change, regulators have hit a brick wall. A small number of mobile operators will not necessarily create a diverse and locally based digital ecosystem for Africa’s citizens. Regulators need to act directly to open their markets to a more diverse set of affordable access operators and solutions.

RUSSELL SOUTHWOOD is CEO of Balancing Act (balancingact-africa.com), a specialist in ICT in Africa; STEVE SONG is also a specialist in ICT in Africa and currently a shared spectrum advocate at the Network Startup Resource Center and a fellow-in-residence at the Mozilla Foundation.

REFERENCES 1 Southwood R (2018). Follow the money. Intermedia 46 2. 2 Millicom to offload remaining African units; Econet linked with deal. TeleGeograpy, 20 October 2017. bit.ly/2iFuzhH 3 Rounding means that these figures do not add up to 100%. 4 Gamcel calls for bail out or risk halting operation. Foroyaa, 29 May 2019. bit.ly/2KZ7jYS 5 Cameroon: Former Camtel D-G David Nkoto Emmane is under investigation for two corruption allegations, forbidden to leave the country. Balancing Act, 24 May 2019. bit.ly/2JmTKze / Angolan president cancels fourth mobile licence after shareholding controversy, Balancing Act, 24 May 2019. bit.ly/2FSCnFh 6 OECD (2016). How Mukesh Ambani shook up the phone industry, in charts. Economic Times, 17 October 2018. bit.ly/2XoYzBL 7 The challenges of being a small ISP in Ghana. Balancing Act, 26 April 2019. bit.ly/2Jp8cGK 8 The data in this section is taken from: The mobile economy – Sub-Saharan Africa 2019 GSMA. 9 US regulator proposes rules for 6 GHz Wi-Fi . Wi-Fi Now, 18 December 2018. bit.ly/329KQxu 10 ‘It feels like a gift’: mobile phone co-op transforms rural Mexican community. The Guardian, 15 August 2016. bit.ly/2bmzRGJ 11 The village that built its own Wi-Fi network. BBC video, 28 March 2019. bbc.in/2COBmMX 12 Discussed in: Navarro L (2018). Network Infrastructures: The common model for local participation, governance and sustainability. APC issue paper. bit.ly/2XqQs2P 13 ‘Broadband for the Rural North’ (B4RN)

– best practice in community-driven broadband roll-out. European Commisson, video interview. bit.ly/2Jst7IV

www.iicom.org