Infrastructure Concession Transaction Multiples: Why the Middle Half Spans 8.1x to 16.5x

Last updated: August 5, 2026

A single EV/EBITDA headline for "infrastructure concessions" obscures more than it reveals. In Concessio's matched sample of 210 transactions, enterprise value to EBITDA multiples run from 8.1x at the 25th percentile to 16.5x at the 75th percentile, with a median of 12.2x — and where a given transaction lands inside that range depends on structural features of the individual concession, not on the sector label alone. This page sets out the distribution itself, the structural features that move a transaction within it, and why Concessio does not yet publish per-unit multiples (EV per passenger, EV per TEU) as a range — only as individually named, fully cited comps.

What does Concessio's EV/EBITDA distribution look like?

n = 210 transactions.

The interquartile spread — 8.1x to 16.5x — is wide enough that "the multiple" is not a single number an analyst can anchor to without knowing which quartile a given asset belongs in.

Why does the middle half span 8.1x to 16.5x?

Four structural features move a transaction within that range:

• Remaining concession term. A shorter remaining term compresses the cash-flow runway an acquirer is buying, which pressures the multiple downward relative to a freshly awarded or recently extended concession. Remaining term is not a fixed number for a given concession over its life — extensions lengthen it, early terminations or hand-backs end it — so two transactions in the same sector, priced years apart, can be buying very different lengths of remaining cash flow even off the same original contract.
• Tariff-setting mechanism. Concessions with indexed or formula-based tariff escalation carry more predictable revenue than those where tariffs are renegotiated periodically at the grantor's discretion, and that predictability is priced.
• Demand risk versus availability risk. An asset paid on traffic, volume, or throughput (demand risk) carries a different risk profile than one paid on availability regardless of usage — the two structures are not interchangeable inputs to a multiple even within the same sector.
• Proximity to a regulatory reset. Where a concession's revenue is set under an allowed-return, allowed-revenue, regulatory-asset-base, or regulated-tariff framework, timing relative to the next scheduled reset matters: a transaction priced immediately before a reset carries more revenue uncertainty than one priced mid-cycle.

None of these four variables is visible from EV/EBITDA alone — which is why the multiple by itself is not enough to benchmark a specific transaction.

Why aren't per-unit multiples published as a range?

Per-unit multiples — EV per passenger for an airport, EV per TEU for a port, EV per kilometer for a toll road — are, in principle, a cleaner comparable than EV/EBITDA because they strip out margin differences between operators. In practice they require the enterprise value and the volume denominator to come from the same disclosure, at the same transaction date — and in Concessio's experience building this dataset, most disclosures give one or the other, not both matched. That scarcity shows up directly in the per-unit metrics Concessio currently computes:

Concessio does not publish a per-unit multiple range for any of these categories.

What do individual per-unit comps look like in practice?

Where a matched EV-and-denominator pair exists, the individual comp is fully citable — it just isn't a market benchmark on its own.

Toulouse-Blagnac Airport, France — enterprise value of US$770 million, equivalent to US$100.5 per passenger. Source: legifrance.gouv.fr.

Aéroports de Lyon, France — enterprise value of US$1,083 million, equivalent to US$113.4 per passenger. Source: vinci.com.

Port of Brisbane 99-year lease, Australia — enterprise value of US$2,096 million, equivalent to US$2,141 per TEU. Source: statements.qld.gov.au.

Toulouse-Blagnac, Lyon and Port of Brisbane are individually named, fully cited transactions — not a distribution large enough to publish as a range, and not one Concessio will manufacture by averaging them.

Where does this data come from?

Concessio's sourcing standard for transaction records is the same standard applied across every dataset on the platform: primary government and regulatory filings, stock-exchange disclosures (SEC EDGAR, HKEX, and peer exchanges), gazette publications, and multilateral development bank project databases. Sources that fail that bar — Wikipedia, paywalled secondary aggregators, fund marketing pages — are excluded from the pipeline by policy.

Concessio does not currently publish a citation-rate statistic for the transactions dataset the way it does for the clause library. The description above is the acceptance standard the pipeline applies, not a verified compliance figure for this specific 210-transaction sample. Full detail on sourcing rules and how records are screened: /docs/methodology.

To query the underlying transaction-level records — including the individual deals behind this distribution — Request Demo.

Figures as of August 2026. Published August 5, 2026.