How do you calculate cost per enplaned passenger for airline budgeting in 2027?

The scenario that makes the number bite

A regional carrier's finance team sits down in October 2026 to build the 2027 budget. Station costs at a mid-size hub have been drifting up — the airport authority signaled a rate increase, ground handling renewed at higher labor rates, and the carrier is adding a fourth daily frequency on one route while trimming another. The station director wants to know one thing: what will it cost us per boarded passenger next year, and is that number going the right direction?

This is the moment where cost per enplaned passenger stops being an accounting artifact and starts being a decision tool. The carrier has roughly $14 million in station-attributable costs at that airport and expects to board somewhere between 640,000 and 700,000 passengers. Divide the first by the second and the answer swings from about $20.00 to about $21.90 depending on which traffic case you believe. That $1.90 spread is not a rounding error — across the network it is the difference between a station that clears its contribution hurdle and one that gets flagged for restructuring.

The trap is that most teams calculate the metric backward. They take last year's actual cost, apply an inflation factor, take last year's enplanements, apply a growth factor, and divide. That produces a number that is arithmetically defensible and operationally useless, because cost and traffic do not move together. A large share of station cost is fixed or step-fixed — the gate lease does not care whether the aircraft is 60% or 90% full, and the station manager's salary does not vary with load factor. Meanwhile enplanements move with schedule, gauge, and demand. When you inflate a ratio, you implicitly assume the numerator scales with the denominator. It does not.

The same structural problem appears in the airport-side version of the metric, where an airport authority calculates CPE as the total airline payments (landing fees, terminal rents, apron fees, and any other aeronautical charges) divided by total enplaned passengers. Airports use it as the headline competitiveness number in bond documents and in negotiations with carriers. Airlines use it as an input to station economics. Both parties are computing something that looks like the same ratio but is built from different cost pools, and confusing the two is one of the most common analytical errors in the space.

There is a broader lesson here that applies well beyond aviation. Any per-unit cost metric — cost per ton-mile in trucking, cost per bed-day in hospitals, cost per subscriber in telecom, cost per order in a distribution network — has the same failure mode. The denominator is volatile, the numerator is mostly not, and averaging them together hides the operating leverage that actually drives the business. The discipline that fixes it is identical everywhere: build the numerator and the denominator as separate models, then divide only at the end.

How the calculation actually works, step by step

Start with definitions, because this is where most disagreements originate. An enplaned passenger is a revenue passenger boarding an aircraft at a given airport. It counts the boarding, not the person and not the trip. A round-trip passenger who departs from and returns to the same airport generates two enplanements there. A connecting passenger who arrives on one flight and departs on another generates one enplanement at the connecting airport (the outbound leg) — the arrival is a deplanement, which is counted separately and is not part of CPE. Total passengers handled at an airport is roughly twice enplanements, which is why quoting a CPE against total passengers understates it by about half. That single error is common enough that it is worth checking explicitly whenever you inherit someone else's model.

The airline-side calculation runs like this:

Step one — define the cost pool. Decide what belongs in the numerator and write it down. A station-level pool typically includes airport charges (landing fees, terminal rent, common-use gate and counter fees, apron and remote stand fees, passenger boarding bridge charges), ground handling (above-wing and below-wing, whether in-house payroll or contracted rates), station overhead (management salaries, training, uniforms, local IT), fueling throughput or into-plane fees, deicing where applicable, and any local security or facility charges the carrier pays directly. What does not belong: aircraft ownership, crew, fuel commodity cost, maintenance, and network overhead — those are system costs allocated on a different basis. If you fold them in you get "cost per passenger," which is a fine metric, but it is not CPE and it cannot be compared against anything published.

Step two — separate fixed from variable. Split each line into fixed (does not move with traffic within the year), step-fixed (jumps at a threshold — an extra gate, an extra shift, an additional ramp crew), and variable (moves proportionally with departures or passengers). This split is the single highest-value thing you can do to the model. A typical station will land somewhere in the region of 50–70% fixed-plus-step-fixed, with the remainder variable, though the mix depends heavily on whether handling is in-house or outsourced and whether the carrier holds preferential-use or common-use gates.

Step three — build the enplanement forecast bottom-up. Enplanements equal departures × average seats per departure × load factor. Do not forecast enplanements as a single growth rate. Forecast the schedule (departures by month, by route), the gauge (seats, which changes when equipment swaps happen), and the load factor (which has a ceiling and behaves differently by season and route). Multiply them. A carrier that adds a daily frequency on a 76-seat regional jet at 82% load factor adds roughly 62 enplanements per day, about 22,700 per year — the arithmetic is simple, but only if you keep the three drivers separate.

Step four — divide, then decompose. Compute CPE monthly rather than annually, because seasonality moves both terms and the annual average hides it. Then decompose the year-over-year change into a cost effect and a volume effect, so you know whether the number moved because costs went up or because traffic went down.

mermaid flowchart LR Q["Choosing a CPE model"] --> A1["Cost basis"] Q --> A2["Rate structure"] Q --> A3["Forecast cadence"] A1 --> B1["Fully allocated<br/>ties to P&L, poor for decisions"] A1 --> B2["Directly attributable<br/>decision-grade, no P&L tie"] A2 --> C1["Residual<br/>lower CPE, airline bears risk"] A2 --> C2["Compensatory<br/>higher, stable CPE"] A2 --> C3["Hybrid<br/>split by cost center"] A3 --> D1["Locked annual budget<br/>clean accountability"] A3 --> D2["Rolling reforecast<br/>tracks reality, weak commitment"] B2 --> E["Recommended core"] C3 --> E D1 --> E D2 --> E E --> F["Publish base + low + high cases"] </invoke>

There is a build-versus-buy dimension too. Small carriers can run this in a spreadsheet, and honestly should — the model is not complex, and the discipline of laying out the driver tree by hand is worth more than the tooling. Larger operators push it into a planning system where the schedule feeds the enplanement forecast automatically and rate cards live in a maintained table. The failure mode of the spreadsheet is version drift; the failure mode of the system is that nobody can explain where a number came from. Whichever you choose, insist that the driver tree — departures, seats, load factor, rate card — is visible and editable, not buried.

Pitfalls that quietly corrupt the number

Confusing enplanements with total passengers. The most common single error, and it produces a CPE that is roughly half the true value. Always confirm which the source used.

Inflating the ratio instead of modeling the terms. Taking last year's CPE and adding an inflation percentage assumes proportional cost and traffic movement, which is exactly the assumption that operating leverage violates. Model numerator and denominator separately and divide once.

Mixing airport-side and airline-side pools. An airport's published CPE covers what airlines pay the airport. An airline's station CPE covers everything the airline spends at the station, including handling the airport never sees. The second is typically several times the first. Comparing them, or averaging them, produces nonsense.

Ignoring the credit structure. At a residual airport, this year's low CPE may reflect a strong concession year, not efficient operations. If parking revenue softens, next year's CPE rises with no operational change. Always ask what share of the credit is demand-sensitive.

Averaging across seasons. A station with a 2.5× summer peak has a summer CPE far below its winter CPE, because fixed cost spreads over more boardings. An annual average describes neither month. Compute monthly, then weight.

Letting the load factor do the work. If CPE improved 6% and load factor rose 4 points, almost none of that was cost management. Decompose every variance into cost effect and volume effect before anyone declares a win. The mechanical decomposition: hold enplanements at prior-year level and revalue at current cost to get the cost effect; hold cost at prior-year level and revalue at current enplanements to get the volume effect; the small remainder is the interaction term.

Forgetting step costs when modeling growth. A model with only fixed and variable buckets will always say the next departure is cheap. It is cheap right up until it triggers a gate, a shift, or a ground support equipment purchase — and then it is not.

Budgeting rate increases from rumor. Airport rates are set through a published process. Get the document. A carrier that guessed at a rate increase and was wrong by two dollars per enplanement on 700,000 boardings missed by $1.4 million on one line.

Treating the metric as a target. CPE is diagnostic, not a goal. A station can improve CPE by cutting the ground handling that causes the delays that cost far more in downstream disruption. Pair it with on-time performance, mishandled-bag rate, and turn-time compliance so nobody optimizes the ratio into an operational hole.

Related questions

What is the difference between CPE and CASM?

CPE is a station or airport metric — cost per boarded passenger at a specific location. CASM is a system metric — operating cost per available seat mile across the network. CPE diagnoses local station economics; CASM measures overall unit cost competitiveness. They answer different questions and should both be tracked.

Do connecting passengers count as enplanements?

Yes, once. A passenger connecting through an airport enplanes on the outbound leg and deplanes on the inbound. That is one enplanement at the connecting airport. Hubs therefore show enplanement counts inflated relative to origin-destination demand, which is why hub CPE looks lower than the local market alone would support.

How often should CPE be reforecast?

Monthly for tracking, quarterly for formal reforecast, annually for the locked budget. Monthly is necessary because seasonality moves both terms; anything less frequent lets a traffic shortfall compound for a full quarter before anyone sees the cost-per-passenger effect.

Can a rising CPE be a good sign?

Sometimes. A new terminal entering rate base raises CPE immediately while the capacity it adds takes years to fill. If the facility unlocks growth, CPE declines from the new higher base as traffic builds. Judge it against the capital cycle, not against last year.

FAQ

What exactly counts as an enplaned passenger?

A revenue passenger boarding an aircraft at that airport. It counts boardings, not people or trips. Round trips generate two enplanements at the origin airport. Connecting passengers generate one enplanement at the connecting point. Deplanements are counted separately and never belong in the CPE denominator. Non-revenue passengers are typically excluded, though the treatment of employee travel varies by carrier — pick a convention and document it.

Should I use budgeted or actual enplanements in the denominator?

Budgeted for the budget, actual for reporting, and always show both. The gap between budgeted CPE and actual CPE is where the volume effect lives, and it is often the largest single driver of variance. Reporting only actual CPE hides whether the number moved because you spent differently or because fewer people flew.

How do I handle a mid-year schedule change?

Rebuild the enplanement forecast from the revised schedule rather than adjusting the prior forecast by a percentage. Then re-test every step cost threshold — a schedule change that adds a departure in a new time band can trip a shift or gate cost that a percentage adjustment would never surface.

Why does the airport's published CPE differ so much from my station cost per passenger?

They are different cost pools. The airport's figure covers only what airlines pay the airport — landing fees, rents, apron charges. Your station figure adds ground handling, station overhead, fueling fees, deicing, and local staffing, none of which flow through the airport. Expect your number to be substantially higher, and never present them as comparable.

What sensitivity range should I publish?

At minimum a base case plus a downside and an upside on the traffic side, since the denominator is the volatile term. A common framing is base, base minus three to five load-factor points, and base plus two to three. Present the resulting CPE as a range. If the range is uncomfortably wide, that is information about the station's operating leverage, not a flaw in the model.

Is CPE useful outside aviation?

The structure is. Any fixed-cost-heavy operation with a volatile volume denominator — hospitals, distribution centers, telecom networks, transit systems — faces the same modeling problem and benefits from the same discipline: build numerator and denominator separately, split costs by behavior, model step thresholds explicitly, and decompose every variance into a cost effect and a volume effect.

Sources

• https://www.faa.gov/airports/planning_capacity/passenger_allcargo_stats/passenger — FAA passenger boarding (enplanement) data and definitions
• https://www.bts.gov/topics/airlines-and-airports — Bureau of Transportation Statistics airline and airport data resources
• https://www.transtats.bts.gov/ — BTS TranStats database, including Form 41 airline financial and traffic schedules
• https://www.gao.gov/products/gao-15-107 — U.S. Government Accountability Office report on airport funding and rate-setting
• https://www.iata.org/en/programs/ops-infra/airport-infrastructure/ — IATA airport infrastructure and charges resources
• https://www.icao.int/publications/pages/doc9082.aspx — ICAO Doc 9082, Policies on Charges for Airports and Air Navigation Services
• https://airportscouncil.org/ — Airports Council International – North America, industry policy and economics resources
• https://www.faa.gov/airports/resources/publications/orders — FAA airport compliance and rates-and-charges policy publications
• https://www.transportation.gov/policy/aviation-policy — U.S. DOT aviation policy, including airport rates and charges policy statements

Related on PULSE

• [How do you build a bottom-up operating budget when volume is the volatile term?](/knowledge.html)
• [How do you separate fixed, step-fixed, and variable costs in a station model?](/knowledge.html)
• [How do you decompose a unit-cost variance into price and volume effects?](/knowledge.html)
• [How do you choose between fully allocated and directly attributable cost views?](/knowledge.html)
• [How do you build sensitivity cases instead of a single-point forecast?](/knowledge.html)