Regulated Airports vs Demand-Risk Toll Roads

Who Carries the Traffic Risk Decides Everything Else About the Asset

A demand-risk toll road and a regulated airport can sit in the same infrastructure portfolio and still be valued on completely different logic. The split starts with one question: who is exposed if traffic disappoints?

On a toll road, revenue is volume times tariff, and the operator absorbs the shortfall if fewer vehicles show up than planned. On a regulated airport, the regulator approves an asset base and an allowed rate of return, and sets aeronautical charges to recover that return. Passenger numbers still matter to retail and some commercial revenue lines, but the core tariff is a regulatory calculation. That difference in who bears the risk also explains why the two assets reset and end differently and need different valuation tools.

Telling the Two Apart in a Filing

The two models describe themselves differently on the page, and the tell is usually in the first paragraph of the relevant note.

Ferrovial’s 407 ETR toll road runs on a 99-year term that ends in 2098; Vinci’s Escota motorway ends in February 2032. Both are disclosed as concessions with a stated end date. Aena’s Spanish airport network discloses no single expiry for the core business: it operates under a regulatory framework, DORA II, that runs 2022 to 2026 and then resets. A missing expiry date in the concession table is itself the sign of a regulated asset.

Not every airport in a portfolio is regulated this way. Vinci’s own Lyon airports sit inside an ordinary finite-life concession that ends in 2047, the same structural shape as a toll road, just with a passenger terminal instead of a carriageway. The regulatory, no-expiry model applies to networks such as Aena’s; most other airports still run under an ordinary concession.

How Revenue Resets

A demand-risk toll road’s revenue moves on two levers set out in the concession deed itself: traffic growth and a tariff escalator, often tied to inflation or a fixed annual step. Those formulas are fixed for the life of the contract; the operator cannot ask the grantor to revisit them because a particular year’s traffic came in soft.

A regulated airport’s revenue resets on a different clock. At the end of each regulatory period, the regulator re-approves the asset base and recalculates the allowed return, and the new figures apply until the next reset. Aena’s per-passenger charge for 2025 carried no increase versus 2024. The DORA II formula produced that result, just as a toll road’s deed produces its escalator; management chose neither.

A Worked Illustration: HarbourLink and AeroGate

The primer works both models through hypothetical archetypes on round numbers, HarbourLink for the toll road and AeroGate for the regulated airport, so the mechanics are visible without a real company’s noise sitting on top of them.

HarbourLink is a 30-year toll road with A$500m EBITDA, a 75% margin, 2.0% annual traffic growth, a 2.5% CPI-linked toll escalator, 25% tax, 8% of EBITDA set aside for maintenance capex, and net debt at 4.0 times EBITDA. Discounted to its expiry at a built 8.21% nominal post-tax WACC, with nothing assumed to survive beyond year 30, the after-tax, after-maintenance cash flows are worth approximately A$6.16bn.

AeroGate is a regulated airport with a EUR 6,000m asset base and a 7.50% allowed pre-tax return, giving a return on RAB of EUR 450m (the asset base times the allowed return). Allowed revenue is a bigger number than that return alone: it also carries the regulatory depreciation that returns the capital over the asset’s regulatory life, plus the operating expenditure the regulator lets the airport recover. The primer quantifies those two extra legs on Aena’s own filed asset base: a return on RAB of EUR 720.9m, EUR 312.9m of regulatory depreciation over an illustrative 30-year life, and EUR 600m of illustrative allowed opex, adding to EUR 1,633.8m of allowed revenue. There is no expiry to discount to, so AeroGate is valued as a multiple of its own asset base: 1.0 times that base, the level at which the asset earns exactly its allowed return, gives an enterprise value of EUR 6,000m, less net debt of EUR 1,000m and divided across 500m shares, EUR 10.00 a share. A 1.15 times premium case, allowing for traffic running ahead of the regulator’s assumptions, lifts enterprise value to EUR 6,900m and equity to EUR 11.80 a share.

Neither figure reads a real company as cheap or dear. They are teaching numbers, run through the arithmetic a reader would apply to any comparable asset.

What Happens at the End of the Term

HarbourLink’s present value has nothing left in it after year 30 because the asset itself has nothing left: it reverts to the grantor, and the accounting treatment for these arrangements requires that assumption. A model that assumes any value survives past expiry is assuming away the defining feature of a demand-risk concession.

AeroGate has no such cliff. When DORA II ends in 2026, the next period starts on new terms. For DORA III, Spain’s Council of Ministers set an 8.32% pre-tax WACC for 2027-2031 on 15 September 2026. The airport’s value sits in the asset base and the return the regulator allows on it.

Which Valuation Tool Follows Each Model

The horizon difference means each asset needs its own tool.

A demand-risk toll road is valued on a cash-flow projection to its expiry, with the asset’s value at the end of that projection set to nothing, because nothing is left to value. A regulated airport is valued off the asset base itself, typically as a multiple of that base, with a separate multiple of its regulated earnings used as a cross-check. Running a toll-road-style projection on a regulated airport invents an expiry date that does not exist in the filing. Valuing a toll road as a multiple of an asset base it does not have skips over the one fact that determines its worth: how many years are left.

Where This Goes Wrong

Treating a regulatory period as though it were a concession expiry is the most common mistake here. DORA II ending in 2026 is the point at which Aena’s charges reset for the next period. A model that zeroes the Spanish network out at that date has invented an ending the filing never describes.

The second assumes every airport in a portfolio uses the same model. A group can hold a regulated, no-expiry network alongside ordinary finite-life airport concessions, and each needs its own treatment, so check what the filing calls the arrangement before assuming which one applies.