A Cancellation Wave Looks Like One Event. The Bill Splits Three Ways.

A cancellation wave arrives as a single number: six hundred flights, a thousand, a weekend of departures struck off the board. That number leads the coverage and is close to the least informative quantity in the event. Two waves of identical size — same carrier, same passenger count, same hours until the operation stabilises — can produce financial consequences that differ by an order of magnitude. The variable that separates them is not scale. It is the cause code attached to the cancellations, read against the passenger-rights regime that governs the departure.

The operational question — when does the airport clear — belongs to schedulers. The structural question asks who absorbs the cost, and the answer is written into statute long before the storm forms. Three regimes govern most of the traffic that matters: the European Union, the United States, and Canada. Each assigns the same physical event to a different balance sheet.

The Cause Code Is a Legal Trigger, Not a Description

Every large aviation market requires carriers to attribute disruptions to a category. The five collected by the US Bureau of Transportation Statistics were written for measurement and are now used for judgement.

• Air Carrier — "circumstances within the airline's control (e.g. maintenance or crew problems, aircraft cleaning, baggage loading, fueling, etc.)."
• Extreme Weather — "significant meteorological conditions (actual or forecasted) that, in the judgment of the carrier, delays or prevents the operation of a flight such as tornado, blizzard or hurricane."
• National Aviation System — "a broad set of conditions, such as non-extreme weather conditions, airport operations, heavy traffic volume, and air traffic control."
• Late-Arriving Aircraft — "a previous flight with same aircraft arrived late, causing the present flight to depart late."
• Security — terminal evacuation, re-boarding after a breach, inoperative screening equipment, or screening queues "in excess of 29 minutes."

Two features matter more than the list. The first is the clause "in the judgment of the carrier," sitting inside the weather definition: the classification is not an instrument reading but a call made by the party whose exposure depends on the result. That locates discretion in the reporting chain; it is not an allegation of bad faith.

The second is that weather is not confined to the weather bucket. Non-extreme weather falls under National Aviation System, and BTS notes that during 2020, 45.8% of NAS delays were weather-driven. Weather also propagates into Late-Arriving Aircraft, a category of transmission rather than origin. The Extreme Weather line alone captures perhaps a third of the weather share, so any claim that a wave was "mostly weather" or "mostly the airline" describes how overlapping buckets were filled, not a physical finding.

The European Union: The Cause Code Prices Itself

Regulation 261/2004 converts the classification into cash directly. Article 7 sets three fixed bands: EUR 250 for flights of 1,500 kilometres or less, EUR 400 for intra-Community flights of more than 1,500 kilometres and all other flights between 1,500 and 3,500 kilometres, and EUR 600 for everything above that. Where re-routing limits the arrival delay, the carrier "may reduce the compensation provided for in paragraph 1 by 50%."

Article 5 adds a notice ladder, the part most often skipped. Entitlement falls away if the passenger is informed "at least two weeks before the scheduled time of departure." Between two weeks and seven days, the carrier escapes by re-routing with departure within two hours and arrival within four hours of the original schedule. Under seven days, the tolerance narrows to one hour and two hours. A wave announced eight days out and one announced eight hours out are legally different events at an identical flight count.

The Gate, and What Survives It

Then comes the gate. Article 5(3): "An operating air carrier shall not be obliged to pay compensation in accordance with Article 7, if it can prove that the cancellation is caused by extraordinary circumstances which could not have been avoided even if all reasonable measures had been taken." Three conditions stack there — extraordinary, causal, unavoidable by reasonable measures — and the burden of proving all three sits with the carrier.

The missed detail is that the exemption is written against Article 7 alone. Article 8 — the choice between reimbursement of the full ticket cost and re-routing "to their final destination at the earliest opportunity" — is not gated on cause. Neither is Article 9: "meals and refreshments in a reasonable relation to the waiting time," hotel accommodation "where a stay of one or more nights becomes necessary," and ground transport between airport and hotel.

A carrier that successfully argues extraordinary circumstances has removed the fixed, per-passenger, capped liability and kept the open-ended, duration-linked one. Winning the argument is not the same as avoiding the cost.

That inverts the intuitive reading of a large weather wave. The highest care cost arises precisely where the compensation defence succeeds: an airport closed for two nights, thousands stranded at a hub with thin hotel inventory. Article 7 exposure is bounded at EUR 600 per passenger; Article 9 exposure is bounded only by duration and room rates.

The United States: The Cause Code Prices Almost Nothing

The Department of Transportation states it plainly: "Contrary to popular belief, for domestic itineraries airlines are not required to compensate passengers whose flights are delayed or canceled." There is no domestic analogue to Article 7. On international itineraries the department points to Article 19 of the Montreal Convention — a fault-based route requiring a claim filed with the carrier, not an automatic entitlement triggered by distance.

What the United States regulates tightly is the refund. Under the 2024 final rule, refunds become automatic rather than request-driven. A refund is owed when a flight is cancelled or significantly changed, where a significant change means "departure or arrival times that are more than 3 hours domestically and 6 hours internationally; departures or arrivals from a different airport; increases in the number of connections." Refunds must be issued "within seven business days of refunds becoming due for credit card purchases and 20 calendar days for other payment methods," and must be paid "in cash or whatever original payment method the individual used," not in vouchers or credit. Parallel thresholds apply to delayed baggage: 12 hours after a domestic flight arrives at the gate, and 15 to 30 hours for international arrivals.

That rule is cause-blind: a hurricane and a crew-scheduling collapse generate the same refund obligation on the same clock. The cause code still drives regulatory attention, performance statistics and reputational cost, but moves no cash by itself. The downside from a mishandled wave sits in rebooking, in the working-capital effect of returning fares within seven business days, and in voluntarily published service commitments.

Canada: A Three-Way Split Instead of a Two-Way One

Canada's Air Passenger Protection Regulations cut the middle out of the binary, sorting disruptions into three categories — "within the airline's control," covering scheduled maintenance, aircraft choice and staffing; "within the airline's control but required for safety"; and "outside the airline's control," covering extreme weather, emergencies, security issues, labour disputes, airport problems and government travel bans.

Compensation attaches only to the first. Where a disruption is within the carrier's control and not safety-related, notice was 14 days or less before departure, and arrival was at least three hours late, the schedule is fixed by carrier size.

Arrival delay	Large airlines (CAD)	Small airlines (CAD)
3 to 6 hours	400	125
6 to 9 hours	700	250
9 hours or more	1,000	500

Passengers have "within 1 year of the delay or cancellation" to file. Two consequences follow. The safety carve-out offers a classification route that never requires arguing the event was external: an unplanned mechanical issue can be conceded as internal and still fall outside the schedule. And the size split makes the same nine-hour disruption a CAD 1,000 liability at one operator and CAD 500 at another — incidence depends on who was flying the route, not only on what happened to it.

Winners and Losers Inside a Single Wave

Side by side, the picture stops being airlines versus passengers. It becomes which passengers, holding which itineraries, at which carriers.

Position	Outcome when cause is external	Outcome when cause is internal
EU short-haul passenger	Care and re-routing only; EUR 250 not payable	EUR 250 plus care and re-routing
EU long-haul passenger	Care and re-routing only; EUR 600 not payable	Up to EUR 600, halved if re-routing is tight
US domestic passenger	Refund on the statutory clock	Refund on the same clock
Large Canadian carrier's passenger	No cash; treatment obligations apply	CAD 400 to 1,000 by delay length
Small Canadian carrier's passenger	No cash; treatment obligations apply	CAD 125 to 500 by delay length
EU carrier, multi-night closure	Article 7 avoided; Article 9 rises with duration	Both obligations run together
US carrier, same closure	Refund liability only; no care mandate	Refund liability plus reputational exposure

The sharpest asymmetry there is between regimes: the same aircraft on the same day meets a capped cash obligation on one side of a single legal test, an uncapped care obligation on the other, and in the United States neither — only a working-capital event on a seven-business-day fuse.

The Slot Channel: Where Scale Does Matter

One mechanism restores the raw count, and it runs through airport capacity rather than cash. Under the European slot framework, Council Regulation (EEC) No 95/93, historic precedence is conditional. Article 8(1)(a) provides that "a slot that has been operated by an air carrier as cleared by the coordinator shall entitle that air carrier to claim the same slot in the next equivalent scheduling period." Article 10(3) sets the condition: a series does not carry that entitlement "unless the air carrier can demonstrate to the satisfaction of the coordinator that they have been operated, as cleared by the coordinator, by that air carrier for at least 80 % of the time during the period for which they have been allocated."

Here cancellations behave like a stock, not a flow. Compensation settles event by event; slot utilisation accumulates across a season and is tested as a ratio. A wave large enough to push a series below 80% turns a disruption into a question about an intangible asset — an order of consequence no cause-code table shows.

Conditions That Undo the Conclusion

Several conditions collapse this frame, and they are not edge cases.

• Entitlement is not expenditure. Article 7 fixes an amount per passenger, not how many passengers claim it. Canada makes the point explicit with a one-year filing window rather than automatic payment. Realised cost is entitlement multiplied by take-up, and take-up is not in the statute.
• Small waves are dominated by ordinary rebooking. Where the cancelled block is small relative to fleet and network slack, the legal channel is a rounding error beside re-accommodation cost. The frame is built for the tail of the distribution.
• The classification is provisional. Article 5(3) requires proof of three separate elements. A cause code filed in an operations centre on the night of an event is a starting position in a dispute that may resolve much later, and differently.
• Delays and cancellations are not interchangeable. The Article 5 notice ladder, the Canadian three-hour arrival threshold and the US three- and six-hour significant-change definitions draw the line in different places. A wave of long delays produces a different liability profile from one of outright cancellations, at identical passenger disruption.
• Coverage gaps are large. Substantial traffic departs from jurisdictions with none of these frameworks. There the cause code carries no statutory weight, and incidence falls back to contract terms and the Montreal Convention's fault-based route.
• The opposite reading is defensible. Scale is measurable now while cause is contested for months, and a large enough wave produces re-accommodation, crew-repositioning and slot consequences that swamp any per-passenger schedule. Strongest at coordinated hubs, weakest for short contained disruptions, and not obviously wrong.

What to Watch Next Week

• The cause-code mix, not the cancellation count. Read Air Carrier against Extreme Weather and National Aviation System together, and treat Late-Arriving Aircraft as propagation rather than origin.
• Language, not just numbers. A carrier describing an event as extraordinary or exceptional signals the position it intends to defend under Article 5(3).
• Duration of stranding, not headcount. Under an EU-governed disruption, Article 9 care scales with nights, not with passengers denied boarding on day one.
• Refund timing complaints. The seven-business-day and 20-calendar-day clocks are objective and observable; slippage indicates back-office rather than front-line strain.
• Slot coordination notices at capacity-constrained airports. These show whether a wave has crossed from a cash question into a utilisation question under the 80% rule.

Concrete Framework — The Sequence That Works

1. Fix the jurisdiction before touching the number. The departure point determines whether the cause code is a cash trigger, a reporting label, or a three-way sort.
2. Pull the cause distribution, not the headline. In US data, record all five BTS categories separately; non-extreme weather sits inside National Aviation System and propagates through Late-Arriving Aircraft.
3. Locate the notice interval. Under Regulation 261/2004, two weeks or more ahead carries no Article 7 entitlement. Between two weeks and seven days, test the two-hour departure and four-hour arrival tolerances; inside seven days, one hour and two hours.
4. Separate the capped obligation from the uncapped one. Article 7 is bounded at EUR 250, 400 or 600 per passenger and reducible by 50% under tight re-routing. Articles 8 and 9 run regardless of cause and scale with nights.
5. For Canadian exposure, ask which of three boxes, then which carrier size. Only "within control and not safety-related" reaches the schedule, which differs by a factor of two to three between large and small carriers at every band.
6. For US exposure, model the refund as a timing event. Seven business days for card purchases, 20 calendar days otherwise, original form of payment. The question is cash conversion, not compensation.
7. Check the slot series last. At a coordinated airport, estimate the share of the season's series already operated against the 80% condition — the only channel in which the count itself binds.
8. Re-run the sequence once the classification is contested. The initial cause code is a position, not a finding, and any incidence estimate built on it should be stated as a range.

The cancellation count measures how many people were inconvenienced. It says almost nothing about who pays. That is answered by a cause code assigned under time pressure by an interested party, read against a statute written years earlier, in a jurisdiction chosen by an aircraft's position when the operation broke. Scale is the visible variable. It is rarely the decisive one.

This article is analysis of regulatory structure and is not legal, financial or investment advice. Regulatory provisions are summarised from published primary texts and may be amended; passengers with a specific claim should consult the current instrument and the relevant national enforcement body.