Spectrum management and implications for Open Access fibre

An International Telecommunications Union forum on November 17 2007 reached a treaty aimed at meeting the global demand for radio-frequency spectrum, which has been fuelled by rapid technological developments and growth in the ICT sector. The treaty, reached at  the World Radiocommunication Conference 2007, aims to ensure “the most rational and efficient ways to exploit the limited resource of radio-frequency spectrum” and among others, stipulates that valuable radio spectrum now used mainly by broadcasters should be opened up to broadband services offered by mobile phone operators.
The new treaty came hot on the heels of the Connects Africa Summit, where there were calls for infrastructure-sharing and for effective spectrum management as some of the ways to improve connectivity in Africa.
This CIPESA/Fibre for Africa paper makes the case for reform in the allocation and management of spectrum for telephony, and argues that efficient spectrum management has a bearing on efforts to deliver affordable fibre in Africa.
Neotel’s quest for ‘cheap’ spectrum
South Africa’s Second National Operator Neotel last March reported a milestone when it was awarded a licence by the Independent Communications Authority of SA (ICASA) to operate on the 800 MHz frequency. Without access to that spectrum, it would have been forced to work through a sphere owned by a competitor – Telkom SA – which would have represented an expensive option for Neotel. Neotel said the 800MHz frequency band is critical for ensuring consumers benefit optimally from technologies whose handsets are available at low cost. ICASA said the allocation would “open up access to new, innovative and world-class technologies, and the provision of services to low-income consumers for whom telecoms has previously been unaffordable”.
This narrative about Neotel has a bearing on efforts to deliver affordable fibre in Africa. It shows that the market is the loser where few operators have access to a strategic and scarce resource such as spectrum (or international fibre). It follows that where such monopolies or duopolies exist, regulatory mechanisms are needed to ensure that they do not arbitrarily decide what to charge for their services. While this could call to mind the need for multi-national collaboration in developing regional fibre, it also raises questions as to whether the more favoured spectrum bands in Africa are optimally managed and utilised.
What ITU Says
Spectrum is the lifeblood of mobile communications, enabling operators to employ certain technologies, and to operate in certain localities. But its allocation has often been inequitable, offering incumbents and other early comers advantaged access, thereby raising the barriers to entry for new players. With open access premised on the need for numerous operators co-existing in a level playing field, it becomes evident that equitable spectrum allocation and management are necessary for open access to work in Africa.
ITU recommends that spectrum management should objective, timely, transparent and non-discriminatory. Indeed, a number of African countries have enshrined these principles in their national ICT policies. Kenya’s ICT policy says the communications commission shall manage radio frequency spectrum to achieve, inter alia, efficient and affordable telecommunication services. Zambia and Zimbabwe’s policies commit to establish equitable and cost-effective mechanisms to manage their spectrum in order to allow for the development of modern, effective and efficient communication systems.
The Ghanaian and Ugandan experiences
Despite the realisation that spectrum allocation needs to be equitable and aimed to ensure a level playing ground and affordable services for consumers, the reality on the ground does not always reflect this. Ghanaian telecom company WESTEL was fined US$25 m by the regulator National Communications Authority (NCA), for failing to roll out the number of lines required under its licence agreement. It had hooked up 3,000 lines rather than the obligatory 50,000 lines by 2002. The company attributed its slow progress to inability to get from the NCA the GSM band spectrum it needed, and the prohibitive interconnection arrangement it had with the incumbent, Ghana Telecom.  Consequently, WESTEL suspended its investments, and was in 2006 renationalised, then a majority stake in the company was sold to a new group. By then it had accumulated US$38.5m in outstanding fines and licence fees.
In Uganda, it emerged earlier in 2007 that the regulator nine years ago signed agreements with telecom operators stipulating that channels in the GSM 900 band would be shared equally among a maximum of three operators. An agreement with MTN, which got the Second National Operator (SNO) licence in 1998 stipulates that: “The Licensee (MTN) shall be informed of any applications for frequencies in the GSM 1880 – 2200 MHZ band. The licensee will be given opportunity to comment on the application prior to any [Communications] Commission’s action. Such comments shall be taken into account by the Commission.”
Edward Baliddawa, who chairs Uganda’s parliamentary Sessional committee on ICT, contends that this clause threatened to not only lock out permanently any operator from the GSM 900 band, but also went ahead to put unfair conditions on the other bands that any investor might have to be forced to resort to. This was because the two national operators (MTN and Uganda telecom) had also taken up chunks of channels within the 1800 MHZ band in addition to the GSM 900 which they already fully controlled.
Why Low-end Spectrum is Critical
The catch is that GSM 900MHz transmits over longer distances than 1800MHz. This means that GSM 900 operators use less base stations to cover a similar area than do operators in the 1800MHz range. Additionally, GSM 900MHz has better penetration characteristics for buildings, and some of the handsets available in African markets are not GSM 1800 compliant. In many African countries, the number of operators in the GSM 900MHz zone has been two or three.
And as South African communications minister Ivy Matsepe-Casaburri noted during the October 29-30 Connects Africa summit, effective spectrum management should entail creating more licences in the highly sought-after low-end frequency spectrum, which she believes would encourage private investors to provide mobile wireless technologies and infrastructure in rural areas. Others have argued elsewhere that the “last mile” is where inadequate or inappropriate spectrum allocation could prove a stumbling block to improving affordable connectivity; and that this is especially so in countries where geographical distances dramatically raise the time and cost of making the last-mile connection.
The Case for Re-allocation
In Uganda, the Communications Commission says it has reviewed operators’ licenses as part of the implementation of the new licensing regime that came into effect on January 2 2007. At least two additional national telecoms operators have been licensed and the regulator has allocated them some of the GSM900 frequencies earlier allocated to the three existing operators.
Looking beyond Africa, in Finland the 900 MHz band was assigned unequally between the three nationally operating mobile operators. The two smaller mobile operators complained that they were less competitive as they did not have GSM 900 channels and their cost structures were resultantly inefficient. In the wake of regulatory revisions, towards the end of 2005, the regulator carried out a more equitable re-assignment of GSM 900 channels. And with more European countries aspiring for a more equitable allocation of spectrum, in France and Finland, all 2G/3G operators have the same amount of spectrum.
Secondary Trade
Some consider that the implementation of secondary market for frequencies would help to ease access limitations. This would work by having mechanisms which enable those companies that have big allocations to sell some of their surpluses. It has been argued also that the best way to improve competition especially in cases when one or few operators have rights of use for spectrum, but don’t use them effectively, is to soften or cancel restrictions related to technology use. And then the use of technology neutral licences may certainly be one step which may contribute to alleviate potential competition problems as this, at least in the long run, will increase the amount of spectrum which to a large extent may be used to offer competing services.
The GM Association says spectrum should be allocated “on the basis of achieving economically efficient, competitive and structurally desirable outcomes rather than to extract monopoly rents from the industry.”  It says if the market is the best allocator of scarce resources, as most economists would argue, it is important that countries should be able to develop their own spectrum trading arrangements.
A recent Infodev report notes that in the absence of strict regulations governing the use and non-use of frequencies, private operators may be tempted to ‘bank’ licences, being motivated by the prospect of a future sale, or simply by the desire to keep the frequency out of the hands of a competitor. It adds that the process of returning spectrum can be an awkward one unless it was written into the licence awarding the original frequency assignment.
Spectrum Commons?
A way forward is to advocate for a ‘spectrum commons’ model as an alternative to the market-based model. The spectrum commons would be administered by an independent organisation constituted of representatives from the government, the private sector and civil society. Such a model would be designed to produce a more democratic allocation of spectrum. It would start from the principle that the spectrum should be regulated in the public interest and for public benefit. Commercial use of the spectrum would need to demonstrate social and economic benefit and would be considered a form of “leasehold” of a portion of the spectrum commons. Spectrum “rental” charges would be levied and applied to the public good with a proportion being re-invested in the improvement of the communications environment through support for civil society communications initiatives and other communications services for public benefit.  – CRIS and APC, Involving Civil Society in ICT Policy
The GSM Association says the citizens of 19 African countries are still being forced to pay too much for international calls, due to the maintenance of a monopoly on international gateway services. It argues that introduction of competition into the international gateways market can reduce call prices by up to 90% and double call volumes, giving the example of Kenyan mobile operator Safaricom which received an international gateways license in 2006 and was able to cut international call prices by 70%. In similar vein, reducing the stranglehold some companies have on spectrum could translate into tangible benefits for users. And the process of unbundling would itself be an enabler of open access initiatives, such as ongoing backbone initiatives, and the upcoming marine fibre networks.

Saturating the African Marketplace With Fibre

By Wairagala Wakabi
Fibre optic cables could flood the Eastern coast of Africa, if plans by the South African and Kenyan governments, as well as various independent investors, come to fruition. Some of the plans are in fairly advanced stages, and seem likely to materialise faster than the inter-governmental East African Submarine Cable System (EASSy), whose take-off has been dogged by wrangles among stakeholders.
Efforts to construct fibre optic cables to link this part [Eastern] of Africa to the international marine cable system have gained momentum over the last year, and are partly attributable to failure by EASSy to take off as planned. Supported by up to 16 regional governments and, 30 telecom companies, and promoted by the New Partnership for Africa’s Development (NEPAD), EASSy was conceived in 2002, and as recently as January last year, its construction was anticipated to have begun by May 2006. Up to now, the ground-breaking ceremony is yet to take place.
Kenya and South Africa (SA), the two countries that have had the most vocal (and often conflicting) opinions about EASSy, have each made individual plans to connect to the international cable, as the misunderstandings over EASSy persist. What is not clear is whether these cables will be complementary to EASSy, or will offer competition that will render the premier regional infrastructure project less viable.
In South Africa, the South East Africa Telecoms (Seat) has been suggested, and the weekly magazine Financial Mail reported in January that the venture enjoyed the financial backing of giant US private equity firm Blackstone. Seat promoters are reportedly talking to South African entities they want to join in, and the magazine reported that local businessman John Mathwasa was among the brains behind this venture.
While details about Seat may still be cloudy, the position is a lot clearer regarding the ongoing expansion of Flag, which is touted as the world’s largest private undersea cable system. Owned by Indian telecom operator Reliance Communications, Flag Telecom owns and manages an extensive optic fibre network spanning Asia, Europe, the Middle East and USA. In the second half of this year, Flag hopes to connect the Kenyan port city of Mombasa to the network it is building in the Gulf region.
Flag Telecom has said the cable will further link SA to Kenya via Mozambique, Tanzania, Madagascar and Mauritius as part of a plan to revamp its global network by the end of 2009.
In mid 2006, Kenya Data Networks, a signatory to the EASSy memorandum of understanding, clinched a deal with Flag Telecom to construct a $115 million link between Mombasa and Yemen. Kenya says it wants to have multiple links to the international fibre optic network, to spur competition, enable affordable services and more reliable communications.
On January 31 2007, Kenyan Information and Communications minister Mutahi Kagwe signed a $2.7 million contract with Tyco Telecommunications to undertake a survey on the construction of The East African Marine Systems (Teams), an $80 million link between Mombasa and Fujairah in the United Arab Emirates. It is a venture where Telkom Kenya (majority owned by the government) will hold a 40% stake and Etisalat of Dubai a 20% shareholding. Private investors, who are likely to buy shares off the Nairobi Stock Exchange, will hold the remaining 40%. The plan is to complete the cable by November this year, though some independent analysts say it is not likely that the cable would come on stream before the second quarter of 2008.
Having alterative fibre optic providers should ideally have the effect of slashing international bandwidth prices which, until now, have been kept artificially high by monopolistic providers, in countries where they exist. In SA, Telkom runs the only two cables linking to the international fibre system, and has been accused of keeping the tariffs artificially high. In much of eastern and southern Africa, the new cables would be able to deliver cheaper bandwidth compared to that currently supplied by satellite.
But will this be case once these cables are in the hands of private cartels over which regulators have no authority? Or should governments take an interest in these planned cables embracing Open Access principles, just as EASSy has?
Telkom has said it is concerned about the number of proposed cables along Africa’s east coast, arguing that deployment of two or more cables within the same region would affect the commercial viability of all of them. But the NEPAD eAfrica Commission’s Policy and Regulatory Advisor, Dr Edmund Katiti, has told Reuters that since EASSy will be operated on a cost-recovery basis, anyone wanting to compete with that will have to operate on the same basis. The catch is that by the time EASSy comes on board, it is likely to find the competitors already entrenched in the market.
For its part, Kenya has argued that Bangladesh with a population of 20 million people has three cables that are fully utilised. It would follow therefore that for countries such as Kenya to become competitive in the global outsourcing business, they need access to more than one cable.
Separately, SA’s department of public enterprises is planning to use a newly created state enterprise, InfraCo (or Infrastructure Company), to build a submarine cable to link SA with other international cables in the British Virgin Islands. InfraCo has been capitalised with R647 m ($89.8 m) to enable it to provide wholesale international bandwidth to telecom operators and Internet service providers, among others.
Lyndall Shope-Mafole, the SA director general for communications, told the SA parliament in January that InfraCo, which was formed by the communications arms of Eskom (a power company) and Transnet (a transport infrastructure company), would be available for any company – including Telkom – to use at cost. With www.fibreforafrica.net

Q&A with SEACOM President On Fibre Rollout

Construction of the 13,700 km Sea Cable System (SEACOM, www.seacom.mu) is underway and expected to reach completion by June 2009. The cable will comprise two fibre pairs, connecting South Africa, Mozambique, Madagascar, Tanzania and Kenya, to India and Egypt, with an option for connectivity into the UAE and Djibouti. CIPESA/Fibre-for-Africa (www.fibreforafrica.net) spoke to SEACOM Ltd. president Brian Herlihy.
Q. Who are the partners in SEACOM and how much are they investing in the network?
SEACOM has publicly announced its investors, it is 75% African owned with Agha Khan Economic Development Group (IPS) out of East Africa, Venfin, Convergence Partners and Shanduka group from South Africa. The remaining 25% is owned by Herakles Telecom, our New York based development company. Herakles management is also the management of Sithe Global (developer and investor of the Bujagali Hydro in Uganda) and Global Alumina, a $4.5 billion alumina refinery in Guinea.
Q. What do you envisage will be the prices for SEACOM’s bandwidth?
SEACOM’s pricing is the equivalent of $100 to $170 per Mb/s per month.
Q. In what ways will SEACOM be competitive compared to other fibre initiatives in the region, such as TEAMS and EASSy?
SEACOM is the only cable offering a PoP (Point of Presence) to PoP solution for Europe and Asia. I believe this is a large advantage and the purchase of onward capacity is a difficult process. SEACOM understands that TEAMS has very competitive pricing to Fujairah. SEACOM believes that its pricing to Europe is more competitive than EASSy’s pricing to Sudan.
Q. How will SEACOM assure affordability of services, and how has it been responsive to NEPAD’s calls for Open Access?
SEACOM has structured each landing point as an open access unit. This has been accomplished in two formats. First, the capital cost of the landing stations is a sunk cost for SEACOM, in other words SEACOM does not seek to recover the investment costs of the landing station through co-location fees.
Secondly, each landing station is built with an additional building whereby customers can put their own equipment at the cable station. SEACOM has published its prices through many different forums. The pricing is an 80% discount over current satellite charges and is the only cable offering capacity directly from a PoP in Nairobi to a PoP in Europe or India without the requirement to purchase onward capacity.
Q. Would you say the fact that SEACOM is not a local company in the countries where it is going to land fibre places it at a disadvantage compared to other cable systems such as EASSy and TEAMS?
Depending on each country’s regulatory regime, SEACOM has either established a local entity to operate the cable or partnered with an existing cable. SEACOM will bring experienced operators to the cable to ensure that the local entities maintain world class quality service.
Q. What are your comments to the assertion that it is not viable for the eastern coast of Africa to have three competing cables? Do you see a need for the cables to cooperate in some areas rather than duplicate everything?
The current capacity demand on the East Coast of Africa is very small. However, we believe that the future demand will experience exponential growth.  Having said that, the three competing systems would create a large over-supply, which could create a short-term glut.
Q. What is Herakles Telecom and where does it operate from?
Herakles Telecom is a development company set in New York. The management is the same management that works with Sithe Global (www.sitheglobal.com) and Global Alumina (www.globalalumina.com)
Q. There has been talk that Herakles staff were involved in the Africa One project which did not materialise, and that they collected money from African telecom, which money has purportedly not been refunded. How true is this, and how could it affect SEACOM’s operations?
I have noted the slander towards myself and JP (Jean Pierre de Leu) out of Kenya.  I can confirm that both of us worked and spent many years with Africa ONE for the hope that this project would go forward. However, neither of us were principals of this project and left the project from 2002. It was our understanding that the only countries that made a deposit were Eritrea and Mauritania.
We understand that the sales person for Eritrea was able to help that money be returned.  It should be noted that the Africa ONE name was sold by AT&T to a private investment group in 1998 and it was this group who was responsible for that money. ¬JP and myself have nothing to do with this entity.  Since Africa (2002) I have worked as a developer on over $5 billion of projects in Africa, each of which have excellent reputations and large impacts in their respective countries, including Bujagali in Uganda.
–    January 2008