How do towerco business models affect tower valuations?

Why is an independent towerco’s tower often worth more than an operator-captive tower? How much value is added if an operator carves out their assets into their own towerco, and leases them up? How does one evaluate the investibility of new carve out towercos? India is a great example market to compare tower business models since all four aforementioned business model variants are represented, and since examples of all four business models could be coming to market in the near term.

How do you scale up a tower company? The build-to-suit model creates a lot of capital value and is a great way to startup, but building a portfolio site by site is a painstaking way to reach scale. If you’re lucky enough to have access to affordable capital, the path to achieving economies of scale can be shortcut through strategic acquisition. But what should you buy? Are the assets of another independent towerco always worth more than operator-captive towers?

Since the start of 2011, towercos have acquired 10,125 towers from other towercos (counting only the six transactions where the price was disclosed). The average value of a towerco tower in international markets (i.e. everywhere except the USA) was US$233,855.

Since the start of 2011, towercos have acquired 88,835 towers from MNOs (counting only the 44 transactions where the price was disclosed). The average value of an operator tower in international markets (i.e. everywhere except the USA) was US$147,820.

This simplistic deal comparison analysis of course neglects several critical factors – the lease rates are not in the public domain and neither, on many occasions, are the tenancy ratios, which means we cannot compare the acquisition of towerco’s towers with those of MNOs in terms of Tower Cash Flow (TCF) multiples. Where we have seen TCF side bay side, the multiples are generally similar, indicating that the main reason that a towerco tower is generally more valuable than an operator tower is that the tenancy ratio is generally higher. Nonetheless, the sample size in this analysis is just about big enough to suggest that an independent tower has generally been valued at a little over 1.5x that of an operator-captive tower over the last four years.

It remains a widely accepted truth that a portfolio of high quality independent towers built with multiple tenants in mind is typically worth more than an operator-captive portfolio, which is usually built to meet the needs of just the one tenant. But it’s not that black and white – there are high quality independent towerco portfolios and less good independent towerco portfolios. There are operator-captive portfolios built for a single tenant, with a tenancy ratio very close to one, and then there are operator-led towercos with high quality structures and tenancy ratios over two (although, usefully to keep things simple for the above analysis, no operator-led towerco portfolios have changed hands since 2011).

Let’s use India to further explore this discussion the point.

 

First some context on the Indian tower market. Recent spectrum auctions have stimulated increased appetite to acquire Indian towers. “We believe there could be acceleration in 3G deployments following the March 2015 low-and mid-band spectrum auctions,” said Jonathan Atkin, Managing Director, RBC Capital Markets. “Low 3G penetration nationally (8%) and in urban areas (25%) provide an attractive backdrop for leasing growth. 1,800 MHz is a plausible band for LTE deployment, while 800 MHz spectrum, currently utilized for 2G, is a logical candidate for continued band realignment to 3G.”

After a hiatus in tower transactions widely attributed to the restructuring of the cellular operator market in the wake of the cancellation of 122 operator licenses in 2012, India’s cellular operators’ new spectrum, and associated need for capital, has re-invigorated the tower transaction pipeline. Investment bankers are dusting off their tower industry rolodexes again in India – it’s tower sale season India!

It seems like everyone is either a buyer or a seller in the Indian tower industry today. In the past weeks and months we’ve seen reports that Reliance Infratel and their 52,000 towers could be sold, reports that IDEA Cellular has 11,000 towers for sale, reports that BSNL could finally carve out their 62,000 towers into a long-mooted towerco, while Viom Networks (42,200 towers), Tower Vision (~8,600 towers) and Ascend Telecom (~4,500 towers) have all been touted as potential acquisition targets in the recent past. And the tower business of GTL Infrastructure (29,432 towers) is probably acquirable by anyone with an appetite for a turnaround project.

Whilst “Thousands of Indian telecom towers for sale” may not be a new headline, neither is “American Tower seeks all-India footprint” – the world’s most acquisitive towerco having previously rolled up XCEL, Transcend Infrastructure, over 4,000 towers from Essar Group and, more recently, KEC International. American Tower’s organic growth rate in India has trebled in the last year; they added 1,351 Indian towers in the last 12 months, up from an increase of 409 in the preceding 12 month period. But organic growth alone is not going to get American Tower where they want to be in India. “We believe American Tower would consider further M&A in India in order to gain scale beyond its current 13,000 sites (e.g. Viom or Reliance Infratel),” said Jonathan Atkin, an analyst who follows American Tower for RBC Capital Markets. In fact, American Tower still has only 13,289 towers in India – a substantial portfolio in any other market, but less so in India where 40-45,000 towers are required to have a deep “all-India” footprint, enabling operators options not to be inhibited by a limited ownership pattern.

American Tower has reportedly been in protracted negotiations to acquire some or all of the equity in Viom Networks. But our sources suggest those negotiations have reached an impasse, both on the basis of a divergence in strategies for diversification of the business model beyond macro towers, and on the basis of a seemingly insurmountable gap between valuations. Unsourced rumours in the press suggested American Tower bid a little over US$3bn for Viom Networks, while Viom’s CEO has publicly stated a valuation of US$4bn.

One of the obstacles to the fully-fledged resumption of tower transaction deal flow in India is the lack of a meaningful benchmark for the valuation of an Indian tower since the 2012 market restructuring. Between 2008 and 2012, 42,007 Indian towers changed hands at an average cost per tower of US$115,028. But since the cancellation of 122 MNO licenses there has been only one small Indian tower deal (earlier in 2015, American Tower acquired KEC International, including 381 towers, for US$13mn – just US$34,121 per tower). The 2012 market restructuring saw Indian towerco’s lose thousands of tenancies almost overnight, taking a big bite out of operating profits. But Indian tower companies are managed by astute, experienced leaders, and the 2012 setback inaugurated a new level of focus on site level efficiency and profitability. For example, Viom Networks delivered their maiden profit in FY2012/13.

So what better indicates the value of an Indian telecom tower; Indian towercos ability to cost effectively erect a new site for as little as US$25,000, the acquisition of a few hundred KEC International towers for less than US$35,000 per tower, or the tremendous potential for India’s towercos to harvest tenancy and tower portfolio growth from India’s transition to mobile broadband, fuelled by spectrum acquisitions?

The potential of Indian towercos to create capital value is illustrated by the soaring share price of Bharti Infratel, up over 350% since a low of Rs 126 in 2013 to Rs 443 (at time of writing). Bharti Infratel’s performance is also fuelled by generous dividend payouts and by the towerco’s recent inclusion in the MSCI (Morgan Stanley Capital International) Global Index. Bharti Infratel’s market cap at time of writing had risen to ~Rs 850bn or just over US$13bn, a market cap which, combined with a tower count of 85,892 (including the 42% equity holding in Indus which sits on the Bharti Infratel balance sheet) suggests a valuation of a little over US$150,000 per tower.

Bharti Infratel’s soaring share price also introduces another critical factor into the Indian tower market: a second credible trade buyer. Could a seller ever feel they would attract a “full value” bid if American Tower (an infamously canny buyer) were the only credible counterparty bidding at auction? Bharti Infratel may now have a US$1bn acquisition and improvement budget, with an appetite for domestic sale and leaseback opportunities, as well as diversification into Wi-Fi, DAS and fibre, and/or an expansion into Bangladesh, should the Bangladeshi regulator create favorable market conditions.

As well as two strategic buyers with appetite and low cost capital (and we wouldn’t rule out the possibility of edotco being a third major towerco bidder), there are also prospective FII buyers of Indian towers like Macquarie, Blackstone, Carlyle Group and Providence Equity. And of course India has no shortage of prospective sellers. Let’s climb back on the rumor mill, recycle some of the stories, categorise each opportunity by business model, and share our take on the reality of which towers could be acquired in India, and how their business model affects valuation.

A hypothetical buyer of Indian towers may be confronted with two similar sized potential portfolios they could acquire: Viom Networks has 42,200 towers, Reliance Infratel (RITL) has ~52,000 (although RITL has been inconsistent in defining their tower count – in their FY12/13 annual report they said they had “nearly 50,000” towers, in FY13/14 it was “43,379” – we await the FY14/15 report with interest)! Tenancy ratios show a wider differential: Viom Networks claims to have a market leading tenancy ratio of 2.4, with an order book that could propel that to 2.7 by year end 2015, compared with around 1.6 at Reliance Infratel (Reliance Infratel’s tenancy ratio is not in the public domain – we’ve received estimates ranging from 1.41 to 1.84 from various sources, thus quote 1.6 as a weighted midpoint).

Let’s compare the tenancy mix.

We would estimate that ~85% of RITL’s tenancies come from either Reliance Communications (RCOM) or Reliance Jio (JIO). RCOM is a solid anchor tenant, JIO is a solid second tenant, and there are currently no ownership linkages between RCOM and JIO. However, one cannot escape the feeling that while Mukesh Ambani’s JIO needs voice services and towers, while his brother Anil Ambani’s RCOM needs spectrum, some form of future asset integration seems possible. RITL’s TCF doesn’t all derive from RCOM and JIO: the company recently signed co-location agreements with Aircel and Tata Teleservices, for example.

Viom Networks’ tenancy mix has greater diversity. Tata Teleservices represent 31% of Viom’s customer base. However, Tata Teleservices’ dispute with joint venture partner NTT DoCoMo may compromise investors’ valuation of those tenancies. A further 31% of Viom Networks’ customer base comes from India’s other incumbent cellular operators: Vodafone India, Airtel and IDEA Cellular, plus JIO. Another 25% of tenancies come from Aircel, Videocon, BSNL, MTS Sistema and various others, including a small proportion from non-traditional MNOs such as ISPs. Viom’s diverse tenancy mix is rounded off by Telenor’s Uninor: “by sharing with each other, Tata Teleservices and Uninor were able to become all-India operators,” said Viom Networks’ Umang Das in a recent TowerXchange interview. “Our work with Uninor gave (Viom Networks) the opportunity to rollout 16,000 towers in a single year.”

A comparison of the debt positions of RITL parent RCOM and Viom Networks is also revealing. According to Moody’s, “while RCOM has begun to reduce its debt level – most recently with the US$1bn equity raised via the issue of warrants and a qualified institutional placement completed in June 2014 -- its leverage, as measured by adjusted debt/EBITDA, remained high at 5.2x for the 12 months ending 31 December 2014”. Meanwhile, in a recent interview in the Economic Times, Viom Networks CEO Syed Safawi said: “we are actually one of the very few companies to have reduced 15% of our net debt, well over Rs 1,000 crore, in the last 18 months in India... At the moment, our debt stands below Rs 7,000 crore (around US$1bn).”

In summary, RITL has the larger portfolio with a healthy tenancy ratio and good tenants. Having been built greenfield, rather than assembled through various mergers and acquisitions, RITL has minimal overlapping towers. RITL’s is a relatively new portfolio of towers, with an average age around five years, reflected in the fact that RITL towers are designed for India’s culture of infrastructure sharing: almost all their sites have a capacity to co-locate four tenants. “Understandably, Reliance is keen to proceed with a sale before the market is flooded with assets,” said BMI in a recent analysis. “We expect the sale of Reliance Infratel to attract considerable interest from local and international towers specialists alike… The company’s strategic presence in key metropolitan markets and its location in spectrum-rich circles mean demand for its towers is high; Reliance Infratel’s pan-India fibre backbone would be useful in reducing costs associated with routing traffic between its towers.”

Viom Networks’ portfolio may be smaller, but organic growth and tenancy ratio growth are both higher, indeed Viom claim to have secured a massive 25% share of incremental tenancies. Viom’s tenancy ratio is a significantly higher, and their tenancy mix is more diverse than that of RITL. There is a little risk in both companies’ tenancy mixes, but no more than is normal for a prospective tower deal. “Viom Networks is perceived as being better positioned and better run than Reliance Infratel,” summed up one analyst. Indeed, Viom Networks has the necessary headroom to be a buyer rather than a seller, although management seem less inclined to buy more Indian towers, and more inclined to invest internationally in an end to end “AssetCo+ServCo” model, which they call ‘Viom Next’.

All of this means Viom Networks is in less of a hurry to sell and, with their peer Bharti Infatel’s valuation soaring, they may feel they can realise a better valuation through IPO than through trade sale. TowerXchange therefore feel Viom Networks’ towers should attract a higher valuation, but Reliance Infratel may be more motivated to sell, thus are more to be likely acquirable by buyers seeking volume at an more easily justifiable price.

In a March 2015 interview with TeleAnalysis, BSNL Chairman and Managing Director Anupam Shrivastava said that his company’s “tower leasing proposal has gone to the ministry for its consideration. The proposal focuses on creating a separate business unit, creating a separate subsidiary company within BSNL which is fully owned by BSNL and creating a joint venture.” However the carve out of a BSNL towerco is structured, the prospect of some of India’s “old growth” towers coming to market is an exciting opportunity for network planners.

“BSNL being a Government owned company have not adopted the sharing model aggressively, and their tenancy ratio remains very close to one. Discussions about creating a separate BSNL towerco have been going on for a long time and the general feeling is that they will have to do this one day due to their financial issues and to generate funds. Their single tenancy ratio means their assets currently don’t have maximum value,” said TR Dua, DG of TAIPA in an interview in TowerXchange.

BSNL is already making progress: JIO recently signed a deal to lease 4,000 BSNL towers at a typical market rate of Rs 38,000 (~US$600), hence TowerXchange’s estimate their prospective towerco would today have a tenancy ratio of 1.1.

Sometimes a low tenancy ratio is indicative of a longer, faster runway for growth by tapping pent-up demand. The BSNL towers have not been proactively marketed, and many are in unique locations where it would be very difficult to permit another site nearby, locations which other Indian MNOs have long coveted. Should they choose a joint venture structure which enables third party participation in the venture, a BSNL towerco would be a unique and attractive investment opportunity. If the value of BSNL towers is suppressed by lower initial tower cash flow (TCF) growth compared to their peers, it is somewhat compensated by the wider between current TCF and a theoretical ceiling of TCF.

The tower company business model creates long term recurring revenues and efficiencies that generate significant capital value. Having looked at Indian examples, TowerXchange understands why independent towercos are generally valued higher than operator-captive towers, but not all tower companies are created equal. There are independent tower companies with stagnant organic and tenancy ratio growth, and/or with mountains of debt, and there are independent tower companies built to create ‘annuity revenues’ – where owners have built nice solid cash flow – thank you very much – and haven’t leased them up as aggressively nor optimised operational efficiencies. Then there are thoroughbred tower companies, companies obsessed by service level improvements, site level profitability, and who drive market leading tenancy ratios. Many thoroughbreds are pure-bred, independent towercos like Viom Networks and American Tower, but some of the most respected, highest valued towercos in the world are carve-outs which still have MNOs as majority shareholders – Bharti Infratel being an excellent case in point.

Similarly, not all operator-captive towers are created equal. Some were rushed to market, taking shortcuts on permitting, and built as economically as possible with the minimum capacity to meet only the needs of their owners. Such towers often come with a tenancy ratio very close to one. They require significant improvement capex, and the investment of significant overhead to legalise the sites. But then we see operator-captive towers built in markets with a strong infrastructure sharing culture, such as India, Sri Lanka or the U.S., towers with additional tenants on them as a function of bi-lateral swaps or commercial co-location sales. MNOs who build and lease up robust towers, in good locations, with a full set of permits and a long, favorable and transferrable lease can attract towerco-like valuations.

Ultimately the most valuable tower in terms of current and prospective future TCF may not be the same as the most acquirable tower. The motivation of the seller is critical, as of course is the fit with the investment acquirer: how many towers overlap with their existing sites? Does the prospective acquisition create new or deepen relationships with anchor or substantial second tenants? Do third party towercos have a significant competitive footprint? The buyer’s investment thesis is critical: is the buyer seeking growth potential – are they looking for a platform into which they can invest tower marketing savvy to drive tenancy ratio growth? Or is their investment thesis more predicated on recurring TCF: towers can provide a solid “pensionable” income, and thus attract infrastructure funds to their proven long term cash flow generation potential.

Terms and lease rates in India are remarkably consistent in India at around Rs 38,000 (~US$600) per month for a ground based tower (GBT), discounted by around a third for rooftops. This consistency is important for this analysis as it means the main differentiator in tower valuation in India is derived from tenancy ratio (and operating efficiency).

The addition of second and third tenants means a discount is shared with all tenants on the tower, a model which we haven’t seen replicated in many other markets, and which creates a financial incentive to drive India’s deep-rooted culture of infrastructure sharing.

While TowerXchange has spoken to some stakeholders who feel pricing transparency is key to inspiring tenant’s confidence in the towerco model, other stakeholders feel pricing should be a ‘black box’ and should be more flexible. “Transparent, inflexible pricing makes the drive for tenancy ratios a matter of life and death”, said one towerco executive. “We need to offer very different terms to get the single tenant towers built that enable us to connect smaller Indian villages. This market for single tenant, rural towers has driven Indian towercos to be able to deliver towers for as little as US$25,000.”

1.    Lease rate

2.    Tenancy ratio

3.    Potential future tenancy ratio

4.    Amendment revenue (additional space leased by anchor tenant)

5.    Potential for additional amendment revenue as 3G and 4G rollout

6.    Desirability and uniqueness of locations

7.    Credit worthiness of tenants

8.    Cancellation clauses

9.    Completeness of permits

10.  Transferability of leases / remaining lease duration

11.  Ownership or ROFR on land under tower

12.  Operating efficiency

13.  Age and performance of any energy equipment

14.  Quality and capacity of structures and implications for improvement capex

15.  Integrity of the asset register

Whilst these are all tangible, quantifiable measures of value, intangible measures are often of equal if not greater importance, the foremost of which is the experience of the management team.