While it doesn’t make the headlines, an incipient trend is afoot. The traditional operator model of owning and operating towers (vertical integration) is changing to shared and leased access.
The reason is simple: spending money on network investment in an environment where revenues don’t grow to pay back those investments is not sustainable. Do not be fooled by the fact capex intensity ratios (as a share of revenue) have remained stable at 15 per cent to 20 per cent over the past ten years. The amount of money left after capex (free cash flow) is in structural decline. Add to the mix impending demands of new investment for 5G, coverage obligations attached to spectrum licences and persistent debt loads, and the decision to slim down is less a choice than an imperative. Tower sales mitigate capex demands by shifting a portion of network spend to opex.
Bharti Group’s spin-off of its Indian towers into Infratel in 2007 was, at the time, an isolated case which has since re-emerged following the creation of China Tower in 2014. Third parties now control and operate 60 per cent of mobile base stations in China, a striking figure given their presence was essentially zero in 2014. While China has the highest outsourced ratio in the world, in its wake have followed a wave of infrastructure spin-offs in partnership with private equity groups (for example Telefonica, Altice) or sale and leaseback deals with tower companies (for example Verizon with American Tower in 2016, and Zain with IHS in 2017).
One could be forgiven for regarding network economics as a dry subject, but this matters. The basic telecoms network operating model is undergoing a permanent change, albeit in slow motion. The prevailing asset-heavy model of the past will be replaced by a more diverse and changeable range of network options, with a host of new competitors on the block.
It is a common penchant among high-flying tech industry bellwethers and start-ups to decry the utility business model as boring, inefficient, hamstrung by labour unions and delivering low margins: in short, something that should be operated by the public sector given the lack of a profit motive.
The rise of tower companies and their insertion into the telecom sector value chain offers a sharp tonic to this position.
A short primer. Mobile networks have broadly two levels in the value chain: passive and active. The passive level includes real estate (sites), towers (masts) and ancillary equipment such as power and housing. The active level includes the radio access network (RAN), backhaul links and core.
Tower companies target the passive level by taking on ownership of the assets and then wholesaling access back to mobile operators. Operational efficiency in the passive level of the network value chain is significantly higher when tower companies control it. Why? Because tower companies operate multi-tenancy arrangements, with the major companies running ratios of between 1.8 and 3.0 customers per tower.
Operators have not historically provided access to competitors on a wholesale basis, although that is now changing in some countries. Because the infrastructure cost base is largely fixed following the initial investment to lay the towers, substantially all incremental revenue derived from an additional operator siting their equipment on a given mast flows through to profit. This is buttressed by rent escalators of 2 per cent to 3 per cent built into the annual leasing contracts. Coupled with a demand stream driven by an inexorably rising tide of data traffic, the net effect is a near bullet-proof business model with EBTIDA margins ranging between 30 per cent and 60 per cent. Who said you can’t make money from being a utility?
Future point 1: pooling costs need not mean a loss of control
Tower leasing, as opposed to direct ownership, does two things for operators. First, it reduces the size of the balance sheet. Second, it shifts costs from capex to opex, which is typically between 50 per cent and 70 per cent of network spend. Estimates from the field suggest cost savings for an operator of 70 per cent over a five-year term and 50 per cent over ten years under a leasing agreement with a tower company compared to direct build outs (price premiums are applied for long term security).
But what about 5G, which necessitates a different network planning strategy to anything that has come before it?
The interesting potential change involves small cells. Operators could deploy their own small cell networks, but there would be heavy overlap in city centres given cramped cell sizes. A tower company may choose to deploy its own small cell network and sell capacity to multiple operators on a wholesale basis.
Crown Castle and American Tower in the US have positioned small cells as a priority investment area. The advantage of this approach is a faster time to market by pooling costs. Recent modelling in the UK reported in the Telecommunications Policy journal, for example, suggests that under a shared small cell approach, 5G population coverage would reach 67 per cent some 18 months faster than by continuing the existing two-by-two national network approach. In this scenario, operators would still be able to differentiate on network quality based on their spectrum holdings.
Future point 2: sweat the asset
Beyond cost savings, there is also the option to sell access to a spun off tower portfolio on a wholesale basis to competitors (excluding the RAN). Altice explicitly cited the opportunity to wholesale capacity to other operators in its rationale for spinning off towers into a JV with KKR in France (SFR) and Portugal (PT).
In the latter, Altice claims “it will be the first independent tower company…allowing other operators to access towers and expand their 4G/5G networks, in line with the Portuguese regulator’s latest recommendations and best market practices about sharing of infrastructure.”
While these operator companies are at the vanguard, our expectation is that such co-opetition will grow in popularity if for no other reason than the tower assets would otherwise be underutilised.
Operators without towers? Not in full but even in part is a sensible shift and one that will continue into 5G.
– Tim Hatt, head of research, GSMA Intelligence
The editorial views expressed in this article are solely those of the author and will not necessarily reflect the views of the GSMA, its Members or Associate Members.