Why tower portfolios change hands at multiples that look indefensible on year-1 EBITDA, and how much of that multiple rests on tenants not yet signed and on the exit multiple.
The multiple you can pay is the one at which your cash flows, debt and exit return your target, and for towers it rests mostly on tenants not yet signed. On Pellam Towers, the fictional 850-tower portfolio of The Digital Infrastructure Investor, year-1 EBITDA of 13.15 million supports up to 21.5 times for a 10 per cent equity IRR, with 7x debt and a 21x exit. The fund paid 21.44x and earns 10.06 per cent. With no new tenants, the ceiling falls to 16.6x.
Worked in full in The Digital Infrastructure Investor by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
A tower's costs are set by the site; its revenue is set by the number of operators on it. Each new tenant arrives on existing steel for a small strengthening cost and falls almost entirely to EBITDA. So tower EBITDA grows much faster than revenue, and a buyer pricing on year-1 EBITDA is really pricing years 5 to 10. Pellam's EBITDA grows from 13.15 to 27.79 million, 2.11 times, at 8.7 per cent a year, while revenue goes from 29.16 to 41.99. The mechanism is worked in how the tenancy ratio drives cell tower returns. This article asks the buyer's question: given that growth, what multiple is affordable?
The figures are the book's fictional case. Amounts in $ millions unless marked.
| Input | Value |
|---|---|
| Towers; colocation tenancies at purchase | 850; 382 |
| Anchor rent per tower; colocation rent, $ a month | 2,050; 1,750 |
| New tenancies a year, per tower; colocation churn | 3%; 1% |
| Sites on leased land; ground lease buyouts a year | 70%; 40 at 16x rent |
| Price; transaction costs | 282; 2% |
| Acquisition debt, interest only at 6.25% | 7x year-1 EBITDA |
| Refinancing in year 7 at 6.5% | 7x year-7 EBITDA |
| Exit, year 10, less 1.5% costs | 21x year-10 EBITDA |
Step 1: the operating cash flow. Revenue is the anchor's rent on all 850 towers plus colocation rent on the average tenancies. Deduct ground rent, site costs and overhead for EBITDA. Deduct maintenance, strengthening for each new tenancy and the ground lease buyouts for free cash flow.
Year 1: EBITDA 13.15, free cash flow 2.92, a free cash flow yield on the price of 1.04%
Year 10: EBITDA 27.79; exit = 27.79 × 21 × (1 − 1.5%) = 574.8
The year-1 cash yield is thin because the buyouts, 8.83 million in year 1 rising to 11.52 by year 10, convert ground rent into owned land. That spending lifts EBITDA later and is part of what the multiple pays for.
Step 2: the debt. 7 times 13.15 is 92.1 at closing, interest only; in year 7 it is refinanced at 7 times year-7 EBITDA, 155.6, and the excess comes back to equity. Equity put in at closing is 196.5.
Step 3: the returns at the price paid.
Entry multiple = 282 ÷ 13.15 = 21.44x year-1 EBITDA (19.45x year-2)
Unlevered IRR = 9.11%; equity IRR = 10.06%
Step 4: solve for the price. Hold every input and move the price until the equity IRR is exactly 10 per cent.
Maximum price = 282.9, or 21.5x year-1 EBITDA; headroom over the price paid: 0.9
For a 10% unlevered IRR: 261.3, or 19.9x
Excel: Goal Seek on the price cell, setting =IRR(EquityCashFlows) to 10%
| Entry multiple | Price | Unlevered IRR | Equity IRR |
|---|---|---|---|
| 17.0x | 223.6 | 11.86% | 14.27% |
| 19.0x | 249.9 | 10.53% | 12.16% |
| 21.0x | 276.2 | 9.35% | 10.40% |
| 21.44x (paid) | 282.0 | 9.11% | 10.06% |
| 23.0x | 302.6 | 8.29% | 8.90% |
| 25.0x | 328.9 | 7.34% | 7.59% |
Each turn of entry multiple costs from about two-thirds of a point of equity return at the top of this range to just over a point at the bottom. The fund paid almost exactly its ceiling.
| Exit multiple | With 3% new tenancies a year | With no new tenancies |
|---|---|---|
| 17x | 18.4x (241.5) | 14.2x (186.7) |
| 19x | 19.9x (262.2) | 15.4x (202.7) |
| 21x | 21.5x (282.9) | 16.6x (218.7) |
| 23x | 23.1x (303.6) | 17.8x (234.6) |
Two things set the price. Growth is worth about five turns: without a single new tenancy, the same portfolio supports 16.6x, not 21.5x, and the difference, 64.3 of price, is a bet on operators' future network plans that nothing in the master lease guarantees. The exit multiple is worth about 0.8 of a turn of entry for every turn of exit. A buyer who assumes the next owner pays 21x is pricing in a market that holds at today's level for ten years; at 17x the ceiling is 18.4x.
The common mistake is to read the entry multiple against comparable transactions and stop there. A 21.44x multiple can be made to look expensive or cheap depending on which comparables are chosen, and neither comparison says whether this portfolio, with this growth and this debt, returns the target. The second mistake is to set the exit multiple equal to the entry multiple by default. It is not neutral: here it adds 0.26 of a point, 10.32 per cent against 10.06, and it hides the fact that the return depends on the market as much as on the towers. The full model, with the price, the tenancy rate and the exit multiple in yellow input cells, is in the free workbook for Pellam Towers. What the buyout line itself returns is worked in what multiple to pay for a ground lease buyout.
Because EBITDA grows with each tenant added to existing steel at almost no cost, and long escalating leases make that growth unusually visible. On an illustrative 850-tower portfolio, year-1 EBITDA of 13.15 million reaches 27.79 million in year 10, 2.11 times, so a 21.44x entry multiple is 10.1x year-10 EBITDA. The multiple prices growth, not current earnings.
Build the free cash flow, debt and exit for ten years, then solve for the price at which the equity IRR equals the target. On an illustrative portfolio with 7x debt and a 21x exit, that price is 282.9 million, 21.5 times year-1 EBITDA. At a 10 per cent unlevered target it is 261.3 million, 19.9 times.
About as much as tenancy growth. On an illustrative 850-tower portfolio, each two turns of exit multiple move the price a fund can pay for a 10 per cent equity IRR by about 1.6 turns of year-1 EBITDA: 18.4x with a 17x exit, 21.5x with 21x, 23.1x with 23x. Assuming exit equals entry quietly assumes the market will pay today's price in ten years.
This article is one calculation from The Digital Infrastructure Investor. The book takes the same case from first principles to the decision, chapter by chapter, and every figure it prints is a live formula in the free companion workbooks.
Get the book on Amazon →Free companion files
Also on Amazon UK · Amazon Germany · Amazon France · Amazon Canada
Reading guide: infrastructure, data centres and energy → · All 1340 articles →
If this book helped, or didn’t, a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.