Government policy should as far as possible liberate the market to continue the success story of private ownership of radio spectrum and fixed infrastructure, instead of interposing well-intended but misguided interventions to correct perceived high pricing or underservicing. Such interventions can only be counterproductive, and should be resisted. A look at our telecommunications policy history explains why.
During the negotiations for a new dispensation in SA, it was recognised that telecommunications policy would be of major importance to a developing economy. Even before the 1994 elections several potential models were considered, but it was soon clear that the government was intent on maintaining a tight grip on what it saw as a strategic sector of the economy.
The only gesture towards market liberalisation was to issue two licences for the emerging technology of mobile telephones. They were issued in 1993, to Vodacom and MTN, on the premise that cellphones were toys for the rich and the profit motive would result in limited market penetration of perhaps 1-million people.
Fixed-line telephony was considered the infrastructure backbone on which to build access for the poor and previously disadvantaged. It was reserved for Telkom, which enjoyed a state-guaranteed monopoly from 1997 until 2002. This experiment was a disaster. A 30% stake was sold to Thintana, a "strategic equity partner" consortium led by SBC Communications of Texas, which specialises in exploiting telecoms networks in developing countries. A clause in its shareholder agreement, which was kept secret, explicitly exempted both Telkom and Thintana from parts of the then new Telecommunications Act of 1997. This permitted SBC to underinvest in infrastructure and milk Telkom for all the profit its monopoly was worth.
The nominal quid pro quo for these extraordinary monopoly rights was that Telkom had to install a paltry 1-million lines into underserviced areas. It duly did so, and the number of fixed lines peaked in 2000 at about 5.5-million. However, as soon as its legal obligation was discharged, Telkom disconnected many of these new lines for non-payment. Today, there are only 1.4-million fixed telephone lines left, which is the lowest total since the 1980s.
Second and third national operators were supposed to be licensed once Telkom's monopoly period expired in 2002, to create competition in the fixed-line market. The idea of a third operator was ditched early on, and the second operator, Neotel, didn’t get off the ground until 2006. Though handpicked for this purpose by wise government apparatchiks, it failed to spur much competition and made almost no impact at customer level.
The much-anticipated second national operator petered out into the disappointment of a failed acquisition bid by Vodacom for the paltry sum of R7bn, equivalent to just over a month's worth of the mobile operator's revenue at the time. In June 2016 it was announced that Liquid Telecom, a subsidiary of the Zimbabwean company Econet Wireless had bought Neotel for a mere R6bn.
When Neotel entered the market in 2006 it had to build infrastructure from scratch. This was widely seen as an obstacle to competition, so in 2007 a government committee recommended that the local loop be unbundled. The local loop is the cable that connects customer premises to the local exchange, and it was thought to be the most difficult and expensive segment of network infrastructure to duplicate. Local loop unbundling would give new market entrants the ability to use copper laid by Telkom, without having to worry that Telkom would interfere with its access to customers. The deadline for this process was 2011. It never happened.
The last that was heard of local loop unbundling was in 2019, when a National Treasury report recommended: "Immediate enforcement of local loop unbundling, as per the ICT policy, would enable multiple providers to have access to the local loop, enhance competition and reduce unnecessary infrastructure duplication." This, about copper infrastructure that Telkom is already in the process of decommissioning, because it is outdated and needs to be replaced by fibreoptic networks. Trust government to be 20 years behind the times. As Duncan McLeod argued in TechCentral in 2019, it's too late for local loop unbundling now.
Telkom resisted the effort to unbundle the local loop throughout. From a shareholders' point of view, of course, this resistance is entirely legitimate. After all, it isn't like Telkom is wholly government-owned anymore. A part of Telkom was listed on the stock exchange, and the local loop was among the assets that was sold to investors. Expropriating it would unjustly harm investors who bought shares in good faith.
Ironically, the failure to unbundle the local loop has been a blessing in disguise for SA consumers. Because Telkom's new competitors struggled to get access to customers' premises, they began to lay their own infrastructure. Multiple fibre network operators, including Vumatel, Octotel and Frogfoot, today compete vigorously against Telkom's OpenServe to roll out new fibre infrastructure to businesses and consumers.
The deployment of fibre would have been much less enthusiastic if Telkom's local loop assets had in effect been nationalised, as intended. Private ownership of fixed-line infrastructure is what ultimately spurred investment and improved services. With mobile operators, too, government regulation stifled the market for years. Even with only two licensed operators in the 1990s, however, the market grew quickly. Some subscribers preferred mobile phones because of convenience, but many chose to switch because you had to be creditworthy to get a fixed Telkom line. Profit-driven mobile operators pioneered prepaid services to overcome this problem.
Combined with heavy competition in the handset market, which drove down prices, the poor, who had hitherto been forced to use public telephones, were soon taking to cellphones in their droves. Rapid advances in handsets led to high turnover, with many customers buying a new device every two or three years. This flooded the market for used handsets, which made mobile telephony even more accessible to the poor. The uptake among consumers surprised even the mobile operators. By 2000 there were 8-million mobile subscriptions.
A third mobile operator, Cell C, was granted a licence in 2001, seven years after Vodacom and MTN were established. It struggled to make a real impact in the market and always remained a bit player, but it did offer significant price competition at times. By 2010, there were 50-million mobile subscriptions. That year Telkom entered the competitive landscape for mobile telephony, offering significant competition for the major incumbents. Cell C fell into deep and sustained financial difficulty, and after being recapitalised in 2017 it has started to move its customers onto the networks of "partners" MTN and Vodacom, in preparation for decommissioning its own network and a fresh recapitalisation. ("Chasing the network, Cell C aims to deliver quality over quantity", February 3).
In 2018, Wireless Business Solutions, which had until then had operated its own fixed wireless network, expanded its services to offer 4G mobile data under the brand Rain. It offers unlimited data for phones at a remarkably low R299, or for any device at R479, with fairly widespread network coverage. It has started 5G rollout in Johannesburg and Cape Town. This offering alone undermines any argument that mobile data prices are too high, and that #DataMustFall.
Today there are almost 100-million active mobile subscriptions for a population of fewer than 60-million people. The vast majority of South Africans now have access to mobile telephony. According to the March 2020 State of the ICT Sector Report by the industry's regulator, the Independent Communications Authority of SA (Icasa), 89.5% of households have access to only a mobile phone, 7.1% have a mobile phone and a landline, 0.1% have only a landline, and 3.4% have no telephone at all. So the effective coverage of mobile telephony is 96.5% of all households. About 75% of SA households now have access to the internet. Four out of five of them can access it anywhere they are via their mobile device.
The foundation of the success of mobile telephony in SA lay in privately owned radio spectrum. Because it is a limited resource, the market ensured the best-qualified operators with the deepest pockets got to own it. This created an incentive to invest billions in infrastructure and develop technologies that made more efficient use of this spectrum. The government has been threatening to intervene to bring down supposedly high data prices and supposedly low internet penetration by rolling out its own national broadband data network for 20 years. It established Broadband Infraco as a state-owned enterprise in 2007 for this purpose. Despite significant taxpayer funding, the company has never made a profit, and its liabilities are almost twice its rather modest assets of only R1.35bn. Although it does offer long-distance backhaul capacity to other network operators, there is little evidence that it has had any significant impact on the market.
The latest attempt by government to muscle into the telecoms sector is the proposed establishment of a national wireless open-access network (Woan), which would be allocated some of the spectrum the regulator has for years failed to auction off to private operators. Even the total nationalisation of radio spectrum has been mooted, although in five years no progress has been made on either front. To anyone who has followed the fortunes of telecommunications policy, or indeed state-controlled enterprises of any kind, the apparent impotence of government-backed broadband policy will come as no surprise. Over-budget under-delivery is the hallmark of enterprises owned, controlled or driven by government.
That said, the policy proposals of the last few years are supremely misguided. The ostensible goal of spectrum nationalisation would be to allocate it to people and companies who otherwise wouldn't be able to afford it. But the price of spectrum is predicated on the expected future profit it can make, which, in turn, is based on the capacity for network infrastructure investment and the ability to provide services for which consumers are prepared to pay. Any organisation with a plan to do so will be able to raise the funding for it. If they can't, they don’t deserve to have valuable spectrum wasted on them.
To break the price mechanism is a basic mistake. It cannot result in "equitable allocation" of spectrum. It can only result in less infrastructure investment, lower efficiency, higher prices, and worse service than there otherwise would have been. Private ownership of radio spectrum has been extraordinarily successful at placing phones into the hands of millions of people, including the poor, while government control over fixed lines has proven to be an extraordinary failure, with fewer lines in operation than at any time since SA's transition to democracy. Had the local loop been unbundled as intended, there would have been much less competition in the infrastructure market, and we probably would not have seen the extensive fibre rollouts of today.
Let's keep private innovation and competition alive.
This article was first published on BusinessDay on 5 February 2021.