What is the average cost per revenue passenger mile (RPM) for airline operations in 2027?

Why the industry publishes CASM instead of cost per RPM

The first thing to understand is that "average cost per revenue passenger mile" is a derived number, not a reported one. Open any U.S. airline's 10-Q or investor deck and you will find CASM — cost per available seat mile — front and center, usually alongside CASM-ex (CASM excluding fuel and special items) and PRASM or TRASM on the revenue side. You will not find a line item labeled "cost per RPM," because airlines do not manage costs against filled seats. They manage costs against flown capacity.

The reason is mechanical. Nearly every dollar an airline spends is incurred the moment the aircraft is scheduled and dispatched, regardless of how many passengers are aboard. Fuel burn changes only marginally with payload — roughly 2–4% swing between an empty and full narrowbody on a typical stage length. Crew pay, gate fees, navigation charges, maintenance reserves, aircraft ownership, and station staffing are entirely capacity-driven. The genuinely passenger-variable costs — catering, in-flight service, credit card interchange on the ticket sale, baggage handling, and passenger liability insurance — usually total well under 10% of operating expense. So a metric denominated in available seat miles (ASMs) tracks what actually drives spending, and a metric denominated in revenue passenger miles (RPMs) does not.

That said, cost per RPM is a legitimate and useful analytical construct, and it is exactly the right denominator when you are asking a unit-economics question: what does it cost this airline to move one paying customer one mile? That is the number you compare against yield (revenue per RPM) to see whether the passenger economics work. Yield minus cost per RPM is the passenger-level margin, and it is the cleanest one-line summary of whether a route, a fleet type, or an entire carrier is making money on the flying itself.

The conversion is straightforward arithmetic. Load factor is defined as RPMs divided by ASMs. So:

Cost per RPM = CASM ÷ load factor

If a carrier reports CASM of 14.0 cents and a load factor of 84%, cost per RPM is 14.0 ÷ 0.84 = 16.7 cents. Every empty seat gets its costs redistributed onto the paying passengers. At a 100% load factor the two numbers converge; at a 70% load factor, cost per RPM runs roughly 43% above CASM. This is the whole reason load factor discipline dominates airline management thinking — a two-point load factor move shifts cost per RPM by several percent without a single dollar of cost being cut.

One caution before you use these numbers anywhere serious: mainline versus system reporting matters enormously. Network carriers publish both a "mainline" CASM (their own metal only) and a "consolidated" or "system" CASM that folds in purchased regional capacity flown by contract partners. Consolidated CASM runs materially higher — regional jets carry 50–76 passengers over shorter stages, so their unit costs are structurally worse — and the gap can be two to four cents. Comparing one airline's mainline figure against another's consolidated figure is one of the most common analytical errors in this space, and it will make a perfectly healthy carrier look bloated.

The two ways to build the number, and where each one breaks

There are exactly two credible paths to an average cost per revenue passenger mile, and they produce different answers for different reasons. Knowing which one you are looking at is more important than the number itself.

Path one: top-down from reported financials. Take total operating expense from the income statement, divide by RPMs from the traffic release. That is your cost per RPM. It is auditable, it ties to filed statements, and anyone can reproduce it. It is also the broadest possible definition — it includes cargo-attributable costs, loyalty program operating expense, maintenance-for-others, and every corporate overhead dollar, all loaded onto passenger miles that did not necessarily cause them. For a carrier with a large cargo operation or a big third-party MRO business, top-down cost per RPM is inflated by activity that has nothing to do with passengers.

Path two: bottom-up from a cost stack. Build the number from its components: fuel per ASM, labor per ASM, ownership per ASM, maintenance per ASM, distribution per ASM, and so on; sum them; divide by load factor. This is what airline planning departments actually do, because it lets you flex individual lines. Want to know what a 15% fuel price move does to unit economics? Only one line changes. Want to model a fleet swap? Ownership and maintenance move together while fuel improves and labor stays flat. The bottom-up build is far more useful for decision-making and far more prone to omission — people routinely forget aircraft rent, capitalized maintenance amortization, or the difference between mainline and regional cost structures.

The practical recommendation is to build bottom-up and reconcile to top-down. If your stack sums to 13 cents and reported operating expense divided by RPMs comes to 17 cents, you have four cents of something you have not accounted for — usually regional capacity purchase expense, cargo and other operations, or special items like fleet impairments and labor contract ratification bonuses. That reconciliation gap is where the actual learning is.

A third variant deserves a mention because it shows up constantly in commentary and is nearly always misused: CASM-ex. Airlines report it because fuel prices are exogenous and special items are non-recurring, so CASM-ex is the cleanest read on whether management is controlling what management can control. It is genuinely useful for that. It is useless for answering "what does it cost to fly a passenger a mile," because fuel is 20–30% of the answer. Never divide CASM-ex by load factor and call it cost per RPM.

How to decide which construction fits your question

The choice between top-down and bottom-up is not a matter of taste — it follows from what you are trying to decide. A route profitability question needs a bottom-up stack with a marginal cost view. A carrier comparison needs top-down consistency. An investment thesis needs both, reconciled.

The distinction that trips people up most is average versus marginal. Cost per RPM as normally computed is a fully allocated average — every overhead dollar spread across every passenger mile. But almost no real airline decision is an average-cost decision. Whether to sell the last twelve seats at 60 dollars is a marginal-cost question, and the marginal cost of an incremental passenger on an already-scheduled flight is small: incremental fuel for the added weight, catering, interchange fees, and a bit of handling. That is why revenue management systems will happily sell deep-discount seats that appear to price below fully allocated cost per RPM. They are not losing money; they are covering a marginal cost that is a fraction of the average.

Flip it around and the logic still holds. The decision to add a frequency, open a station, or take delivery of an aircraft is a long-run decision where nearly all costs become variable, and there the fully allocated average is the honest benchmark. The rule of thumb worth internalizing: use marginal cost for the last seat, average cost for the next airplane.

The concrete numbers behind each cost component

Here is where the abstraction has to turn into figures a practitioner can actually use. The ranges below reflect the structural economics of U.S. and European operations in recent years and the direction of travel going into 2027; treat them as planning bands, not as any specific carrier's reported results.

Fuel. Historically the single largest or second-largest line, and the most volatile. At jet fuel in the 2.20–3.00 dollars per gallon band, fuel typically runs 3.5–5.5 cents per ASM for a mainline narrowbody operation, more for older or smaller aircraft, less for a new-generation A320neo or 737 MAX flying long stages. Divide by an 84% load factor and fuel alone is roughly 4–6.5 cents per RPM. The single biggest lever here is not hedging — it is fleet age and gauge. Current-generation narrowbodies burn on the order of 15–20% less fuel per seat than the aircraft they replaced, and going up a size class spreads the same trip fuel over more seats.

Labor. For U.S. network carriers labor has become the largest single expense line, roughly 4.5–6.5 cents per ASM after the pilot and flight attendant contracts ratified across 2023–2024 reset pay scales upward by large double-digit percentages. That converts to something like 5.5–7.5 cents per RPM. Labor is the least flexible line in the stack: contract rates are fixed for years, minimum monthly guarantees floor the variable portion, and staffing ratios are set by regulation and by the schedule you published months ago.

Aircraft ownership. Depreciation plus rent, typically 1.3–2.5 cents per ASM depending on fleet age and whether the carrier owns or leases. An older owned fleet shows a very low ownership cost per ASM and a punishing maintenance and fuel cost; a new leased fleet shows the reverse. Never evaluate ownership cost without looking at fuel and maintenance in the same breath — the three trade against each other directly.

Maintenance. Roughly 0.9–1.8 cents per ASM for mainline operations, but lumpy and heavily dependent on where each airframe and engine sits in its check cycle. Engine shop visits are the dominant driver, and supply chain constraints on engine parts and turnaround times have kept this line elevated and unpredictable. Carriers with power-by-the-hour agreements smooth the line at the cost of a higher long-run average.

Landing fees, navigation, and station costs. Around 1.5–3.0 cents per ASM, driven overwhelmingly by stage length and gauge. These are largely per-departure charges, so a carrier flying 500-mile stages amortizes them over a fifth as many miles as one flying 2,500-mile stages. This is the single largest reason short-haul unit costs look terrible next to long-haul.

Distribution and selling. Roughly 0.7–1.5 cents per ASM: GDS fees, travel agency commissions, credit card interchange running about 2% of the fare, and reservations staffing. Direct-booking share is the lever, and it is one of the few cost lines where a carrier can move the number in quarters rather than years.

Everything else. Passenger service, insurance, IT, and corporate overhead round out roughly 1.0–2.0 cents per ASM.

Sum the mainline bands and you land at approximately 13.5–18.5 cents per ASM for a full-service U.S. carrier, which at an 84% load factor is roughly 16–22 cents per revenue passenger mile. A low-cost carrier flying a single fleet type at high density and high utilization on longer stages compresses that materially — CASM in the 8–11 cent range, cost per RPM in the 9.5–13 cent range at comparable load factors. Ultra-low-cost carriers push lower still on paper, though the gap has narrowed as their labor costs converged toward the majors and as they shifted toward less dense, less utilized flying.

Regional operations sit at the opposite extreme. A 76-seat regional jet on a 400-mile stage can show CASM north of 20 cents and cost per RPM above 25 cents. That is not mismanagement; it is arithmetic. Fixed per-departure costs divided by few seats and few miles produce large unit numbers. It is exactly why network carriers have spent years upgauging regional flying to larger aircraft wherever scope clauses and pilot supply allow.

Three normalizations are mandatory before comparing any two of these numbers. First, stage length — unit costs fall roughly with the square root of stage length, so a carrier averaging 1,400-mile stages will look better than one averaging 700 miles even if they are equally efficient. Second, gauge — seats per departure spreads per-departure costs. Third, utilization — daily block hours per aircraft spreads ownership cost. Analysts adjust CASM to a standard stage length precisely so that these structural differences do not masquerade as operating performance.

What is actually moving the number into 2027

Several forces are pulling in opposite directions, and the net effect on average cost per revenue passenger mile depends heavily on carrier type.

Pushing costs up: labor rates locked in by contracts ratified in 2023–2024 continue to step up annually, and those escalators are contractual, not discretionary. Airport costs are rising as terminal redevelopment programs at major hubs are financed through landing fees and rental rates. Maintenance costs remain elevated because engine shop visit turnaround times have not fully normalized and parts availability constrains the market. Aircraft delivery delays have forced carriers to keep older, thirstier airframes flying longer than planned, which raises both fuel and maintenance per ASM.

Pushing costs down: continued fleet renewal toward new-generation narrowbodies delivers a genuine per-seat fuel improvement; upgauging spreads fixed costs over more seats; and load factors have held at historically high levels, which mechanically lowers cost per RPM without changing CASM at all. Direct-booking share continues to grind distribution costs down.

The wildcard is sustainable aviation fuel. SAF currently costs a multiple of conventional jet fuel, and mandates in the EU and UK are ramping through the late 2020s. The blend percentages are still small enough that the effect on total fuel cost per ASM is modest today, but it is directionally upward and it lands unevenly — carriers with heavy European exposure absorb more of it than domestic U.S. operators.

Netting it out: the realistic expectation is that mainline unit costs drift upward in nominal terms while low-cost carrier and mainline cost gaps continue to narrow. Cost per RPM in the high teens for U.S. network carriers and the low teens for low-cost operators is the reasonable planning assumption, with fuel price the dominant source of variance around it.

Implementation: building and sequencing the analysis

If you are actually constructing this metric — for a route model, a competitive benchmark, or an investment case — sequence matters. Doing steps out of order produces numbers that look precise and are wrong.

A few sequencing rules worth stating plainly.

Pull traffic data before financials, not after. The traffic release defines the denominators, and if you build the cost stack first you will unconsciously fit it to a number you already have in your head. Getting ASMs, RPMs, and average stage length locked down first also forces you to decide up front whether you are working mainline or consolidated — a decision that changes every subsequent line.

Separate regional capacity purchase expense explicitly. On a network carrier's income statement this appears as a single large line covering everything paid to contract regional operators. It is not decomposable into fuel, labor, and ownership from the outside, so any bottom-up mainline stack must exclude it and any consolidated comparison must include it. Mixing the two is the single most common error.

Reconcile before you divide. Dividing an unreconciled stack by load factor just propagates the error and makes it harder to find, because now the discrepancy is scaled by a factor you introduced. Get top-down and bottom-up within a few percent, then convert.

Run sensitivities as ranges, not points. Fuel at plus or minus 50 cents per gallon moves cost per RPM by roughly 1.0–1.5 cents on a mainline narrowbody operation. A two-point load factor swing moves it by 2–3%. A labor contract step-up of 4% moves it by roughly 0.25 cents. Presenting a single point estimate for a metric with this much natural variance overstates precision and invites bad decisions.

Finally, always pair the cost figure with yield. Cost per RPM in isolation says nothing about whether the airline makes money. A carrier at 18 cents cost per RPM earning 21 cents yield is healthier than one at 12 cents cost earning 12.5 cents. The spread is the business; the cost number alone is trivia.

Adjacent metrics that make cost per RPM interpretable

Cost per RPM belongs to a family of unit metrics, and it is nearly useless outside that family. The full working set:

PRASM (passenger revenue per available seat mile) sits on the revenue side of the same denominator as CASM, which makes PRASM minus CASM the cleanest margin-per-unit-capacity read available. TRASM adds cargo and ancillary revenue and is the better comparison for carriers with meaningful non-ticket revenue — which now means essentially all of them, given how large bag fees, seat assignment fees, and loyalty program revenue have become.

Yield is revenue per RPM — the direct revenue-side counterpart to cost per RPM. Yield is far more sensitive to stage length than PRASM is, because longer flights command higher absolute fares but lower per-mile fares. Comparing yields between a short-haul and long-haul carrier without normalizing is meaningless.

Break-even load factor is the load factor at which yield exactly covers CASM. It falls directly out of the same inputs: CASM divided by yield. If CASM is 14 cents and yield is 18 cents, break-even load factor is 78%, and every point above that is contribution. This is often the most communicable version of the entire analysis for a non-specialist audience, because it collapses cost, price, and volume into a single threshold.

Cost per enplanement matters for airport and ground operations, where the cost driver is passengers handled rather than miles flown. Station staffing, ticketing, and baggage systems scale with enplanements, not RPMs, which is another reason short-haul networks carry structural cost disadvantages — the same passenger generates a station cost event more often per mile flown.

The broader lesson generalizes well beyond aviation. Any capacity business — hotels, freight, cloud infrastructure, telecom, live venues — faces the same fundamental split between a capacity-denominated cost metric and a consumption-denominated one, and the ratio between them is always some form of utilization. Hotels run cost per available room night against occupancy. Trucking runs cost per mile against deadhead percentage. Cloud providers run cost per provisioned core-hour against utilization. In every case the capacity metric drives operational decisions and the consumption metric drives pricing decisions, and confusing the two produces the same category of error: pricing off a fully allocated average when the decision was marginal, or building capacity off a marginal cost when the decision was structural.

Sanity checks before you trust your own number

A handful of quick tests will catch most errors before they reach a slide.

Check the ratio between cost per RPM and CASM. It must equal one divided by load factor. If your load factor is 84% and your cost per RPM is not about 19% above your CASM, one of the three numbers is wrong. This is the fastest possible arithmetic check and it catches an astonishing share of mistakes.

Check cost per RPM against yield. If your computed cost exceeds yield by a wide margin for a carrier that reported an operating profit, you have almost certainly loaded consolidated or total costs against mainline RPMs, or omitted ancillary and cargo revenue from the comparison.

Check against stage length. If a carrier's average stage is under 800 miles and you calculated a cost per RPM below 12 cents, be skeptical — per-departure costs alone make that hard without extreme density and utilization.

Check the fuel line independently. Fuel cost per ASM should approximately equal (gallons per ASM) times (price per gallon). Gallons per ASM for a modern narrowbody operation runs roughly 0.012–0.016. If your fuel line implies something far outside that band, you have a units error, and units errors in this space are common because ASMs run into the hundreds of billions and small mistakes hide well at that scale.

Check that special items are handled consistently. Fleet impairments, contract ratification bonuses, and one-time settlements can swing a quarter's CASM by a cent or more. Either include them everywhere or exclude them everywhere, and say which you did.

Finally, state the period. Airline unit costs are seasonal — first quarter CASM runs materially above the annual average because capacity is lower and fixed costs are spread thinner, while third quarter runs below it. A quarterly figure compared against an annual figure will mislead by a meaningful margin every single time. When someone quotes an average cost per revenue passenger mile without naming the period, the carrier, and whether it is mainline or consolidated, the number is not yet an answer to anything.

Related questions

What is the difference between CASM and cost per RPM?

CASM divides operating expense by available seat miles — all flown capacity, filled or empty. Cost per RPM divides by revenue passenger miles — only occupied seats. Cost per RPM equals CASM divided by load factor, so it is always the larger number whenever load factor is below 100%.

Why is CASM-ex the wrong input for cost per RPM?

CASM-ex strips fuel and special items to isolate controllable costs. Since fuel is roughly 20–30% of total operating expense, dividing CASM-ex by load factor understates true cost per passenger mile substantially. Use full CASM when the question is what it actually costs to carry someone.

How much does stage length change cost per RPM?

Substantially. Per-departure costs — landing fees, ground handling, navigation charges, taxi fuel — are fixed per flight and spread over more miles on longer stages. Unit costs fall roughly with the square root of stage length, which is why analysts adjust CASM to a standard stage before comparing carriers.

What is break-even load factor and how does it relate?

Break-even load factor is CASM divided by yield: the occupancy at which revenue per mile covers cost per mile. At 14 cents CASM and 18 cents yield, break-even is about 78%. Every point of load factor above it converts directly to contribution margin.

Why are regional carrier unit costs so much higher?

Regional jets seat 50–76 passengers and fly short stages, so per-departure fixed costs divide across few seats and few miles. Cost per RPM above 25 cents is normal, which is why network carriers upgauge regional flying toward larger aircraft whenever pilot scope clauses permit.

FAQ

Do airlines publish an official average cost per revenue passenger mile?

No. Carriers report CASM, CASM-ex, PRASM, TRASM, and yield. Cost per RPM is a derived figure you compute by dividing CASM by load factor, or by dividing total operating expense by reported RPMs. Because it is derived, always state which construction you used and whether the inputs were mainline or consolidated.

What is a reasonable planning range for 2027?

For U.S. mainline network carriers, roughly 16–22 cents per revenue passenger mile at load factors in the low-to-mid 80s. Low-cost carriers sit closer to 9.5–13 cents. Regional operations run above 25 cents. Fuel price is the dominant source of variance — a 50 cent per gallon move shifts the figure by about a cent.

Does a higher load factor always lower cost per RPM?

Mechanically yes, because the same costs spread across more revenue passenger miles. But chasing load factor by discounting can lower yield faster than it lowers unit cost, which is why revenue management optimizes total revenue rather than load factor. A full airplane at the wrong fares loses money just as reliably as an empty one.

Which cost line is the largest for a network carrier?

Labor, following the pay scale resets ratified across 2023 and 2024, typically running 4.5–6.5 cents per ASM. Fuel is second at 3.5–5.5 cents per ASM at recent price levels, though it can retake the top spot when crude spikes. Together the two usually account for roughly half of total operating expense.

Why does my bottom-up stack not match reported financials?

The gap is almost always regional capacity purchase expense, cargo and other operations, or special items. Network carriers report a single consolidated line for regional flying that cannot be decomposed externally. Reconcile before converting to cost per RPM, because dividing an unreconciled stack by load factor just scales the error.

How do these unit economics compare to other capacity businesses?

The structure is identical. Hotels measure cost per available room night against occupancy; trucking measures cost per mile against deadhead; cloud providers measure cost per provisioned core-hour against utilization. In every case a capacity-denominated metric drives operations while a consumption-denominated one drives pricing, and confusing the two produces the same errors.

Sources

• https://www.bts.gov/topics/airlines-and-airports — U.S. Bureau of Transportation Statistics airline data programs
• https://www.transtats.bts.gov/ — BTS TranStats database (Form 41 financial and traffic schedules)
• https://www.iata.org/en/publications/economics/ — IATA Economics reports and industry cost analysis
• https://www.airlines.org/dataset/ — Airlines for America industry datasets
• https://www.eia.gov/dnav/pet/pet_pri_spt_s1_d.htm — U.S. EIA jet fuel and crude spot price series
• https://www.sec.gov/edgar/search/ — SEC EDGAR full-text search for airline 10-K and 10-Q filings
• https://www.icao.int/sustainability/Pages/Economic-Analyses.aspx — ICAO economic analyses of air transport
• https://www.faa.gov/data_research/aviation — FAA aviation data and forecasts

Related on PULSE

• How do you calculate break-even load factor for a route?
• What drives the gap between mainline and consolidated unit costs?
• How should stage length adjustments be applied when benchmarking carriers?
• What does a fuel price shock do to airline unit economics?
• How do capacity businesses outside aviation structure their unit metrics?