Telecommunications and technology advisory firm BMIT has warned that South Africa faces complex choices and trade-offs in reaching decisions regarding the awarding of new spectrum licences for mobile broadband.
In a new paper, called Spectrum Policy for the Gs, the firm said that South Africa will have to be smart in the way it awards spectrum to a new wholesale open-access network (Woan) and to existing and potentially new network operators.
Government wants to create the Woan — to be owned by a consortium of private-sector players — to infuse more competition into the South African telecommunications industry and to challenge the dominance of market leaders Vodacom and MTN.
The plan has drawn criticism, though has been welcomed in principle by smaller operators, including Telkom and Cell C. The latter, which has a distressed balance sheet, said last week that it simply cannot afford to participate in a spectrum auction and so will likely lean on the Woan — whatever form it takes — for access to future 4G and 5G spectrum.
BMIT has done extensive modelling around the proposed Woan, including looking at whether there is a business case for it, what sort of investment it will need to make and what its drivers of profitability will be.
“It will be critical to coordinate policy goals with how spectrum is allocated, forging a sensible balance between interests that are complementary in some respects while they conflict in others,” said the paper’s author, Martyn Roetter. “Otherwise, the dissension and contradictions between the hopes for the economy and goals for society that are riding upon wireless networks, and the ability to deploy these networks efficiently, economically and productively to fulfil these hopes and achieve these goals, will persist.
“In the worst case, no one will be a winner in the medium and long term, even if there may be one or two financial winners among special or narrowly defined interests in the short term,” he said.
Roetter said South Africa should emulate Europe and Asia in placing an emphasis on licensing so-called mid-band frequencies for 5G deployment (typically 1GHz to 6GHz), while decisions about high bands (typically 24GHz to 48GHz) “should reflect the outcome of the imminent World Radiocommunication Conference (WRC-19) scheduled to start in late October, at which, among other items, global and regional allocations to terrestrial mobile services in these bands will be made”.
Whatever is decided in the coming months will have a “decisive impact on the valuations and hence costs and uses of spectrum in South Africa,” Roetter said in the paper.
“The outcomes will determine the capabilities, coverage and timing of the networks that are subsequently deployed to deliver new and improved services. The delay in awarding high demand spectrum licences, and communications regulator Icasa’s decision to include the 3.5GHz band in the future invitation to apply, open the opportunity for the country to coordinate the award not only of this high-demand spectrum in the 700MHz, 800MHz and 2.6GHz bands but also of other bands envisaged in many countries for 5G deployments (in other mid bands as well as the high-millimetre-wave bands) in a coherent and productive framework.”
Spectrum valuation, he said, is a complex process. The outcomes for prices of the same spectrum awarded even in comparable countries and market circumstances vary enormously, depending on the process that is employed and the conditions of the licences.
“It makes more sense to structure and adopt an awards process designed as best as possible to produce results that fulfil well-articulated purposes for networks in supporting economic and social objectives, rather than trying to determine the value of spectrum as an exogenous variable to be taken into consideration as if it is independent of the process itself.”
The long delay in allocating high-demand spectrum in South Africa creates an opportunity for the country to construct its spectrum policy within a longer-term perspective that anticipates the characteristics of next-generation technologies, as well as future and established uses of wireless networks, Roetter said.
“This situation could allow the country to build, at least regionally, a late-mover advantage based on spectrum, despite having fallen behind in exploiting fully the benefits offered by the deployment of 4G technology.”
South Africa has never formally licensed spectrum for 4G deployments, though operators have improvised by reallocating (or “re-farming”) 2G and 3G assignments to roll-out 4G infrastructure.
BMIT’s paper said several other countries are planning to, or have already held, multi-band 5G spectrum auctions, including frequencies within the low bands, mid bands and high bands.
“South Africa, like other countries, confronts the challenge of how to set or achieve prices and other conditions applied to spectrum licences in ways, and with costs for the licence holders, that will enable the services delivered over wireless networks to generate the greatest benefits for the South African economy and all the country’s residents.
“It is very complicated to determine the best strategy for spectrum allocations because multiple, sometimes conflicting, interests are necessarily engaged, both within the public and the private sector, in trying to influence decisions… Choices that are made to achieve a balance between these interests will exert substantial effects on the valuation of spectrum, and then on the distribution over time of the network costs borne by and the benefits network services will generate for various constituencies.
“These benefits may be spread widely over all the different segments or categories of the population and user groups, as defined by geography, or by sector in the business community, or by groups of consumers or residents with distinctive needs or expectations for services, or by different public services competing for attention and funding from the public purse. Or they may be concentrated on a subset of these constituencies contributing to unhealthy inequality in South African society.”
BMIT said two examples of the balance that must be struck in relation to spectrum are:
- The extent to which new market entrants are to be encouraged — for example, through limiting existing large operators by putting caps on the amounts of new spectrum they can acquire; and
- The trade-off between fiscal requirements (the desire of government to generate revenue directly, often in the short term, from the awards of spectrum licences) with the need for investment to foster long-term economic development. This investment may be hampered or unreasonably delayed if the costs of spectrum absorb a significant proportion of the funding available to licence holders to deploy networks.
BMIT said so-called dynamic (time and location) sharing or allocation of bandwidth is an idea and capability that could play a significant role in the operation of future networks, whether a Woan or one controlled by an existing operator.
“In principle, these techniques would help maximise network usage and revenues to offset high costs. One approach particularly applicable to a Woan is based on dynamic spectrum allocation, which provides network capacity to those who need it, when and where they need it, in real time,” it said.
“Ideally, bidders will be able to acquire bandwidth immediately, or for some defined time in the future, in response to projected demand and market conditions. Also, in principle, supply and demand for capacity or spectrum can thereby be continuously matched as closely as possible on every cell site on a network, helping to achieve a viable financial model for a Woan that supports not only existing commercial mobile services but also machine-to-machine, Internet of things and other applications so as to reduce the risk the network will be underutilised and fail to generate a return for its investors.
“The second approach is a solution that enables an operator’s underutilised capacity to move from areas where average network load percentages are low to locations where the percentage load is high and additional capacity is needed. High-demand growth leads to very uneven and generally low utilisation of operators’ existing network capacity. This solution may involve two overlapping networks or independent network layers. The benefit of moving mobile devices from a heavily loaded cell to an alternative network improves connection speeds and reduces capex and opex. It may also provide opportunities for an operator to provide wholesale excess capacity in specific network locations.” — © 2019 NewsCentral Media