A tower’s site costs are paid by its first tenant, so every later tenant is close to pure margin, and the price usually assumes they arrive.
The tenancy ratio (tenancies divided by towers) drives cell tower returns because a second tenant costs about $15,000 to accommodate, earns $21,000 in its first year and pays back in 0.71 years, with almost no added running cost. On an 850-tower portfolio, lifting the ratio from 1.45 to 1.69 takes EBITDA from $13.2 million to $27.8 million and the margin from 45.1 to 66.2 per cent. With no new tenants, equity IRR falls from 10.06 to 5.72 per cent.
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 →
The tenancy ratio is the number of tenancies on a tower portfolio divided by the number of towers. It is the single figure that organises tower economics, because a tower’s costs are set by the site and its revenue is set by the number of operators on the steel.
In the book’s worked case, the Pellam portfolio has 850 towers carved out of a mobile operator. That operator, the anchor, stays on all 850 under one master lease at $2,050 per tower per month, with a fixed escalator of 2.5 per cent a year. A second operator rents space on 280 towers and a third on 102, each at $1,750 per tenancy per month with a 3 per cent escalator. The colocation tenancies number 382. The ratio is (850 + 382) / 850, or 1.45. A ratio of 1.00 would mean every tower carries only its anchor.
Start with one tower on leased ground. Seven towers in ten stand on land the fund does not own: 595 of the 850 sites pay a landowner $1,150 a month, rising 3 per cent a year. With only the anchor on the steel, ground rent alone takes 56.1 per cent of the anchor’s rent. Upkeep, insurance and overhead still have to be paid from what is left. A one-tenant tower on leased ground is a thin business.
Now add a second operator. The tower may need strengthening first: a heavier mount, reinforced members, sometimes a larger foundation. The model assumes $15,000 of capital expenditure per new tenancy. The tenancy brings $21,000 of rent in its first year. The payback is 0.71 years. After that, the landowner is already paid, the compound already fenced, the permits and power feed already in place. The new rent falls almost entirely to EBITDA.
That is the whole economic idea of a tower. The fixed costs of the site are paid once, by the first tenant. Every later tenant is close to pure margin.
In year 1 the portfolio collects $20.91 million of anchor rent and $8.25 million of colocation rent. Ground rent of $7.93 million is the largest single cost, against $3.57 million of site costs and $4.50 million of overhead. EBITDA is $13.15 million on revenue of $29.2 million, a margin of 45.1 per cent. The anchor pays 71.7 per cent of revenue.
The base case assumes new tenancies arrive at 3 per cent of the tower count each year, which on 850 towers is 25.5 a year, and that 1 per cent of colocation tenancies leave each year. The path follows, with the ratio measured at each year end.
| Year | Tenancy ratio | Revenue $m | EBITDA $m |
|---|---|---|---|
| Year 1 | 1.47 | 29.2 | 13.2 |
| Year 3 | 1.53 | 31.7 | 15.9 |
| Year 5 | 1.57 | 34.4 | 18.9 |
| Year 7 | 1.62 | 37.3 | 22.2 |
| Year 10 | 1.69 | 42.0 | 27.8 |
Revenue grows from $29.2 million to $42.0 million. EBITDA grows from $13.2 million to $27.8 million, and the margin widens from 45.1 to 66.2 per cent. Most of the widening is operating leverage from new tenancies. Part of it is not: the portfolio also buys land under some leased sites each year, and ground rent falls as it does. An honest reading credits the margin to both.
Nothing in the master lease obliges any operator to add a single antenna. So the case worth running first is no growth at all.
| Case | New tenancies a year | Equity IRR % |
|---|---|---|
| Base case | 25.5 | 10.06 |
| Break-even to target | 25.1 | 10.00 |
| No new tenants | 0 | 5.72 |
With no growth, the equity IRR falls to 5.72 per cent and the unlevered IRR to 5.96 per cent, against 10.06 and 9.11 per cent in the base case. More than four points of equity return depend on tenancies that do not yet exist.
The threshold matters more than the scenario. The equity IRR reaches 10 per cent at 0.0296 new tenancies per tower a year, or 25.1 a year across the portfolio. The base case assumes 25.5. The cushion is 0.4 of a tenancy a year across 850 towers. A portfolio that keeps adding tenants, and keeps growing EBITDA every year, can still miss its target if it adds them slightly more slowly than modelled.
Treat the colocation rate as the price, not as an input beneath it. The growth assumption is a statement about the operators’ future network plans. Diligence should test it against evidence:
Then set the rate of new tenancies to zero in the model and walk it up until the equity return crosses the target. The gap between that number and the base case is the real margin for error. The full tower case, with ground leases, the merger risk and the exit, is worked through in The Digital Infrastructure Investor, and every figure above is reproduced by a live formula in the companion workbook.
A ratio of 1.00 means every tower carries only its anchor tenant. In the book’s 850-tower case the ratio is 1.45 at purchase and 1.69 by year 10, which lifts the EBITDA margin from 45.1 to 66.2 per cent. Each tenancy above one adds rent with almost no added site cost.
In the worked case a new colocation tenancy needs $15,000 of strengthening capital expenditure and brings $21,000 of rent in its first year. The payback is 0.71 years, less than nine months. After that the rent falls almost entirely to EBITDA, because ground rent, permits and the power feed are already paid.
With no new tenants over ten years, the equity IRR on the 850-tower portfolio falls to 5.72 per cent and the unlevered IRR to 5.96 per cent, against 10.06 and 9.11 per cent in the base case. The target is reached at 25.1 new tenancies a year; the base case assumes 25.5.
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 105 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.