The state of telecoms in Africa
[By Russell Southwood]
Africa’s road to high-speed broadband is being achieved in leaps and bounds. Every week brings news of another piece of the jigsaw fitting into place. This week it’s the completion of the national fibre backbone in one of Africa’s larger markets. However, there’s still remains a yawning gap between the promise of ubiquitous, cheap bandwidth and current realities of the continent.
It’s hard not to be excited by the news that Angola (and its usually rather ponderous incumbent, Angola Telecom) has completed 10 000km of a national fibre backbone that connects every province in the country. It is a huge place and suffers from an enormous range of practical difficulties. Nevertheless, it has completed the telecommunications equivalent of building a motorway network across the country.
In this same week, one of my analysts brought to my attention an infographic from Google on download speeds in Africa. This claims to show the fastest download speeds in Africa, ranging from 10,1Mbit/s in Ghana to 1,38Mbit/s in Nigeria. It puts the world average download speed at 8,48Mbit/s.
The idea that Ghana has the best download speeds in Africa will cause a long, dry chuckle amongst my Ghanaian colleagues followed by them beating their heads against the wall slowly in frustration. If the average household download speed achieved is 10,1Mbit/s, I will (as a non-hat wearer) duly eat my hat: naturally, lightly sautéed with olive oil, garlic and red onions.
At this point, I can hear the siren voices saying: why does it matter? Things are getting better. It matters because if Africa is to do all the things that the Internet and data access promise, these have to happen at a speed that allows more or less instantaneous access rather than needing to make a cup of coffee while you wait for something to download. Demand for content and services is being throttled by the inability of operators to deliver reasonably priced, fast (10Mbit/s, I wish) and reliable bandwidth. There is significant evidence that Africans (particularly young ones) want Facebook, YouTube and other more local Internet services as much as any other group of citizens in the world. The key to them being able to get them is delivery on that promise of fast, affordable bandwidth.
The blockages are many and for those who follow these things, have a familiar ring to them.
— International bandwidth access: by the end of 2012, nearly every coastal African country will have a landing station. The only exception is likely to be Eritrea but its rulers take perhaps too great a pride in being exceptional so there’s not much can be done there. There’s some doubt about the southern reach of the Ace cable and others may join that list. In some coastal countries, like Ghana and Nigeria, there will be five landing stations and international bandwidth will sell in the lower hundreds of dollars per Mbit/s.
In many other countries, the World Bank has encouraged nationally-led operator consortia to run the single landing station and this should ensure open access and reasonable prices. But there are a number of countries (notably Cameroon and Gabon) where old-fashioned incumbents will sell a little of their huge fibre inventory at artificially high prices. In other places like Congo-Kinshasa, where the government insisted that the (almost non-existent) incumbent be the licence-holder, the jury is out.
— Pity the landlocked: there are 12 landlocked countries in Africa and as a number of studies have shown, they suffer multiple disadvantages because they lack access to the sea. One of the most notable is that the transit price for getting their data to all these new international landing stations often costs the same or more than it costs for their data to complete the journey from the landing station to London or New York.
One of the continent’s powerful mobile players was telling us recently that it was impossible to get to get reasonably priced transit bandwidth out of one of its West African landlocked country markets. This same operator is the cause of this problem in other territories rather than the victim of it. Right hand, meet the left hand.
Considerable effort has gone into creating equitable open access to international landing stations but rather less into tackling the problem for some of Africa’s more disadvantaged economies. Wioccs’s East African Backbone reserves capacity at reasonable prices for landlocked members of its consortium. But there is nothing similar elsewhere in Africa and the cross-border expansion of Africa’s carriers’ carriers (like Phase3 Telecom, Suburban and KDN) has hit a plateau from a combination of capital and licensing issues. Indeed, KDN is currently being sued by one of its suppliers in the Kenyan courts. A World Bank scheme to use fibre deployed by the members of the West African Power Pool has disappeared without trace.
— National backbones — the problems come home to roost: if there is a problem with transit pricing, the same issue is reflected at a national level. Bandwidth from Lagos to London is now down into the low hundreds per Mbit/s and will go lower as more cables arrive. However, the considerably shorter journey from Lagos to Abuja costs US$1 000 to $1 200 per Mbit/s. Nigeria is one of the most competitive countries and has historically led on the regulatory front, so why is this occurring?
There are many competitors but only two of them (MTN and Nitel) have genuinely national networks. The long-standing problems with the incumbent Nitel and its multiple failed privatisations mean that MTN comes close to being a de facto monopoly operator at this level.
Other countries have chosen to make building a fibre backbone of this sort a national priority but these initiatives are not without issues. Uganda’s Chinese-built and financed backbone is widely acknowledged to have been over-costly and there are doubts about its operational effectiveness. In Tanzania, the government has made much play of separating out TTCL’s national fibre backbone (again Chinese-built but operating more effectively) within the company. But it has insisted that it can only sell a relatively high minimum amount of bandwidth to a limited group of operators and its pricing structure still produces artificially high prices. Contrast this with Ghana’s National Communications Backbone company that offers a flat rate per Mbit/s across the whole country.
There’s also a problem of investment displacement. Again, take the example of Tanzania. The private sector would have built 70-80% of the network that the government took a loan from China to build. So why didn’t it allow the private sector to build it (focusing on regulating price and access) and agree that its (lesser) financial contribution would build those parts the market wouldn’t?
— Mobile networks not fit for data purpose: the Irish are said to say: “If you want to go there, I wouldn’t start from here.” In a little over 10 years, Africa’s mobile operators have put up voice networks that cover anywhere between 30-80% of the continent’s population with voice coverage. In the last several years, they have been steadily upgrading these networks to handle data with the seemingly endless acronyms that promise high-speed data and only occasionally deliver it.
However, what started as a narrow pipe voice network with no Internet Protocol (IP) elements is now creaking at the seams as it seeks to go off and become an all-singing, all-dancing data network. It’s like the streets of Nairobi or Lagos or any other African city: the build up of traffic at different points during the day turns the road into a car park where nothing moves.
One major mobile operator told us that in one of its larger country markets, rural data demand using GPRS and Edge was doubling in volume every six months. Already at this 2-2.5G level, data traffic exceeded voice traffic by 60/40 and on 3G, the proportions are 90/10. The same pattern is apparent across all operators. The future is an IP-enabled data network that will carry the content and services that will replace some of the voice revenues as average revenue per user goes down. For better or for worse, mobile operators are central to the process of delivering affordable data to the widest number of people. However, they are currently struggling to transform what their networks can do in data terms with varying degrees of success.
— LTE for all — meet the future?: the mobile operators’ strongest card for continuing to be taken seriously in terms of data delivery is long-term evolution, or LTE. The Kenyan government decided that the quickest way to achieve this (and it has a good track record on speed of movement, see Teams) was to put out a tender for an open access, national network. In the absence of this, it may turn out that LTE and high-speed data delivery on it, will be the thing that further entrenches the market position of the new mobile incumbents.
In order to build an open access LTE network, you need access to the mobile operators’ tower network and so the arm-wrestling begins. New incumbent Safaricom and old incumbent Telkom Kenya have the power to negotiate a two week extension on the deadline. Since the winning bidder has to include an operator with an extensive tower network (Safaricom?), it will be necessary to negotiate with them placing the towers into the hands of a trusted third party operator, like Eaton, Helios, American Towers or another. The failure to get this kind of open-access structure right will put most other operators at a disadvantage against those who can make the investment.
The alternative is deep-pocket investment in fibre (to the home, office and cabinet) of the kind being carried out by insurgent challengers like 21st century, Jamii Telecom and Wananchi. But the skew to mobile use makes this a useful supplement rather than the central play. In this context, not enough African Governments have allowed their utilities to sell “dark fibre” as has happened in Uganda.
The strange case of technology as the game-changer: Another approach to breaking the back of this affordable access everywhere problem has been the argument that certain technologies would be “game-changers”. Over the last five years WiMax has had much airplay for this tune. It made early promises of both mobile data and voice but the latter was never delivered. Unfortunately, its base station technology, even when it was working at its best, was too expensive and had no customer device ecology at the right price. On that score, Wi-Fi wipes the floor with WiMax in terms of delivering bandwidth cheaply and reliably and has existing, cheap customer devices, not ones that will be ready “real-soon-now”.
The Holy Grail in technology terms is a low-cost, IP-enabled base station that operates on small amounts of satellite bandwidth to reach edge markets that need below an E1 of bandwidth. So far, everyone has delivered things that produce incremental cost changes but not the step down in costs that is needed. It’s a complicated bundle to get right involving renewable power, footprint and satellite optimisation. But this is the frontier that will begin to see changes in the core network over the next 10 years if it can be delivered. Why have an extremely cheap, IP-enabled base station at the edge of the network and not start replacing existing, end-of-life equipment with it in the core network?
— Ubiquitous Wi-Fi access — giving local access: one of the remaining blockages is that access at the local level is fairly restricted. If you’re not a corporate customer paying premium prices, it’s difficult to get cheap and reliable household bandwidth or to find its equivalent through public, Wi-Fi hotspots.
At an early stage, some of Africa’s mobile operators (notably MTN) started experimenting with separating out their data traffic from their voice traffic at base station level. This practice is now widely described with the rather ugly phrase “Wi-Fi offload”.
As the number of smart and feature-rich handsets in Africa increases, customers will increasingly be encouraged by mobile operators — before the LTE nirvana arrives — to switch over to a Wi-Fi hot-spot or Wi-Fi mesh network. Google has been experimenting with this approach in Nairobi’s The Junction shopping mall and other operators are trialing a similar approaches.
As ever, the issue in competition terms is whoever entrenches themselves at this level could turn out to be the price “gatekeeper”. For mobile operators, the recurring question remains: is this core to our business? So far, they have defensively played every hand that looks threatening to them but the tide may turn.
— Fighting for the unconnected: there’s a lot of rhetoric around reaching the rural populations of Africa but not a great deal of action. When one large operator tells us that 10% of its base stations are commercially marginal, the scale of the challenge is apparent. Yet there is a clear interest in the Internet in rural areas shown by that stat quoted earlier of rural data use doubling and by national surveys carried out in places like Kenya.
Many regulators in Africa have collected large amounts of money from operators but this has largely stayed in their bank accounts because they have taken forever to set up universal service agency functions or separate organisations.
Where they have spent the money, it has tended to go back to the “usual suspects” (incumbent and mobile operators). In the main they have tended to extend their voice networks, leaving Internet/data the poor relation. (the exceptions include places like Uganda). The argument against these structures is that if you are relying on “the usual suspects” to do the work, it is an expensive financial structure that strips out a significant percentage for overhead costs before returning the money to the same operators. Therefore why not simply write universal service clauses into their licences that translate into the kinds of sums being extracted?
But the issue is perhaps one that requires closer attention of a different kind. Government policy makers need to say to the operators, either you go to these areas or we will give licences and spectrum to others who will do so on a local basis. This leads to three broad potential options:
1. The mobile operators (going the low-cost base station route) do their own coverage in these areas and the cost is deducted in whole or in part from their universal service obligations.
2. An independent, infrastructure sharing company offers operators the ability to connect to these areas at an agreed price per minute. This was what Ericsson was promoting two or three years ago in Tanzania, with optimised base stations that had larger coverage areas but very little has been heard of it recently.
3. You set up independent, small-scale operators and they get an interconnection agreement that might be asymmetrical to give them sufficient financial means to survive and again you could deduct an initial “market-generating” subsidy from the universal service funding obligation.
— It’s not the digital divide, it’s the electricity divide: it doesn’t matter whether it’s a mobile phone, a PC or a TV, they all require electricity. So the real divide will increasingly be between those who have access to reliable electricity to power these devices and those who don’t. There are two broad categories: firstly, those who already supposedly have access to electricity who would like reliable power that didn’t go down regularly and spike in ways that damaged their devices; and secondly, those with no electricity or struggling with occasional power, largely but not exclusively in rural areas.
For all the energy that goes into promoting universal access, not enough goes into addressing these power problems. At a recent broadcast conference, one broadcaster was speaking optimistically about the impact rural electrification would have on increasing TV audiences in Uganda. But for every Uganda, there are two or three African countries where addressing electricity supply seems to be in stasis.
What is harder to understand is why the kind of power roll-out scheme that operators came together to achieve in Uganda cannot be generalised across other countries? Also, why are the infrastructure sharing companies not addressing power issues? Why can’t there be small-scale, local power providers? You cannot separate out the achievement of a digital dividend from the provision of reliable power supply. The two silos of communications and power are related.
The arrival of the international fibre cables has provided a warm glow of achievement to many of Africa’s politicians but unless they focus on the remaining problems outlined above, the promise will always fall short of the potential.
- Russell Southwood is head of Balancing Act Africa