A new DOT rule will hide airline maintenance delays starting in 2026

Frontier Airlines, JetBlue, and Spirit sit at the bottom of U.S. delay rankings — and Southwest’s single-fleet Boeing 737 operation is more late-prone than its simplicity suggests. But here’s the uncomfortable truth: no standardized source publishes a carrier-by-aircraft-type delay rate.

What travelers actually have is carrier-level reporting, hub geography, utilization patterns, and regional-feed risk. And on October 19, 2026, a new DOT rule will shrink the “controllable delay” category, making the fleet-reliability question even harder to answer from public data.

The most delayed airlines, ranked

The worst-performing U.S. carriers cluster in three distinct camps. Frontier and Spirit run high-utilization models — aircraft that fly more turns per day, which means any single disruption ripples into two or three late departures. JetBlue carries the structural penalty of a Northeast-heavy network. But the single most delayed percentage in the retrieved carrier table belongs to Southwest at 27.5% of flights arriving more than 15 minutes late — a finding that runs against the operational simplicity analysts usually attribute to its all-737 fleet.

That mismatch is the first crack in the fleet-story narrative. Southwest’s single-fleet model should, in theory, simplify maintenance and crew allocation. Instead, turn-times and boarding congestion are the cited operational drags, not the aircraft itself. The fleet is the same; the operating tempo around it is the problem.

One caveat about Spirit: its 2026 ranking showed a 29.1% delay rate, but the carrier ceased operations on May 2, 2026. Ranking a dead airline is historically useful but booking-relevant only as a warning about its operational model surviving elsewhere.

What aircraft maintenance delay data actually shows — and where it stops

The Bureau of Transportation Statistics reports delays by carrier, not by aircraft type. The Airline On-Time Statistics dataset runs from January 1995 through June 2026, and it’s built from 17 U.S. carriers above a 0.5 percent revenue threshold — but the reporting structure never captured a field for the specific aircraft flying each leg. That’s the structural reason a fleet-type delay rate doesn’t exist in BTS data.

The FAA does publish Service Difficulty Reports — records of aircraft malfunctions and maintenance inspection activity, downloadable by year. But treating those raw totals as ready-made delay figures would be a category error. They lack fleet-size and utilization normalization, so an older, larger fleet will always generate more reports regardless of reliability. This piece deliberately sidesteps that flaw.

What we can verify is narrower. PSA operates roughly 150 CRJ-700/900 aircraft and Republic roughly 254 E175s in major-brand feed. Both regionals often cancel and delay more flights than mainline operators — but no primary-source maintenance or delay-driver figure was verified for either fleet in this session. The same gap applies to A320 versus 737 sub-variants: no retrieved material maps them to BTS delay outcomes for the named carriers.

This is the core investigative gap. Not that the data is hidden — that it was never assembled in normalized form. Until someone builds a carrier-by-aircraft-type delay table from raw schedule and tail-number data, travelers are left with proxies: hub geography, utilization intensity, and the operational patterns of the airlines that fly each fleet type hardest.

Why Delta and Alaska still lead on-time performance

Delta lands at 17.5% over-15-minute late flights with a 12.2-minute average delay; Alaska does better at 16.7% and 9.1 minutes. Both share a structural advantage the worst performers don’t: hubs outside the congested Northeast corridor. Delta’s Atlanta and Salt Lake City operations and Alaska’s Seattle-focused network avoid the airspace chokepoints that punish United at Chicago O’Hare and Newark.

But there’s a quieter inflation hiding inside these numbers. Between 2019 and 2025, airlines padded domestic scheduled block times by 8 to 12 minutes. That means today’s on-time arrival isn’t the same achievement it was six years ago — the schedule itself got slower, so more flights clear the 15-minute threshold without actually operating more efficiently. Airlines love to let that detail sit in the fare rules.

Alaska’s numbers also now include Hawaiian integration, noted in the underlying source material — a reminder that merger math can shift a carrier’s reported profile even when its core operation hasn’t changed.

The Section 511(b) delay rule change and what it hides

On September 3, 2026, the Federal Register published the final rule implementing Section 511(b) of the FAA Reauthorization Act of 2024. It takes effect October 19, 2026, and it creates a new causal category for ten specific events — events that “shall not be included within the Air Carrier code.”

The most consequential item on that list for fleet-reliability watchers is unscheduled maintenance that cannot be deferred. Today, a non-deferrable maintenance write-up is counted as a controllable air carrier delay. After October 19, it won’t be. The reclassification doesn’t change whether the aircraft is actually broken — it changes whether the airline gets blamed for it in the rankings.

For travelers trying to assess airline reliability, the effect is subtle but real. Post-rule data will sit on a different footing from earlier years, which means year-over-year comparisons will become less meaningful. A carrier that looks improved in 2027 may simply be reporting fewer controllable maintenance events because the definition narrowed, not because its jets are actually running better.

How BTS actually counts a delay minute

The delay rankings everyone quotes rest on a thinner methodology than most travelers assume. BTS treats a flight as delayed only once it arrives 15 or more minutes after schedule. Then, and only then, does the agency begin counting delay minutes. If a single delayed flight is assigned multiple causes — say, a late inbound aircraft and a maintenance write-up — the minutes are prorated across those causes rather than counted as one undivided total.

That proration is quietly significant for the fleet question. A maintenance event contributing 48 minutes to a flight that’s also 72 minutes late due to a late-arriving aircraft gets diluted in the public record. Maintenance delay minutes never appear as a clean, standalone total for a given aircraft type. Between 35 and 58 percent of total carrier delays trace to late-arriving aircraft — the largest cause category BTS tracks — which systematically masks the fleet-specific contribution underneath.

The reporting structure reinforces the gap. ASQP 234 filings cover nonstop scheduled-service flights from the 17 carriers above the revenue threshold, with airlines filing delay causes monthly. Aircraft type was never part of that filing. So the granular fleet delay rate travelers want isn’t just unpublished — it’s structurally absent from the dataset the rankings are built on.

What this means for travelers booking through U.S. hubs

If you’re connecting through a U.S. gateway to Asia-Pacific, the practical read is straightforward: prefer mainline Delta or Alaska for tight first legs, where the late-arrival percentages run 17.5% and 16.7%. That’s not carrier loyalty logic — it’s the clearest verified signal in the data.

Be equally deliberate about regional feed. If your itinerary’s first leg is operated by PSA or Republic under a major-brand code, treat the connection with extra margin. The retrieved material consistently shows regionals canceling and delaying more than mainline operators, even though no normalized fleet-type driver is verified.

Don’t bother trying to filter by aircraft type yet — the standardized delay data doesn’t exist. What you can do instead is track the carriers and hub patterns above, because those are the only published signals with real comparative value.

One thing to watch: after October 19, 2026, maintenance-driven delays will partly disappear from the “controllable” category. If a carrier’s delay numbers improve suddenly in the next reporting cycle, check the rule change before giving the airline credit. The jets didn’t change — the accounting did.

Key terms

Air Carrier code

The Air Carrier code is the DOT delay category that holds airlines responsible for disruptions within their operational control. It covers events like crew scheduling, maintenance, and turnaround delays that the carrier could have prevented. Under the Section 511(b) rule taking effect October 19, 2026, ten events — including non-deferrable unscheduled maintenance — move out of this code, shrinking what counts as a controllable airline delay.

Service Difficulty Reports

Service Difficulty Reports are FAA records of aircraft malfunctions and maintenance inspection activity, available for download by year. They capture raw incident counts without normalizing for fleet size or utilization, so larger fleets naturally generate more reports. This article sidesteps them as delay proxies because no ready-made carrier-by-aircraft-type normalization exists in the retrieved material.

ASQP 234

ASQP 234 is the Airline Service Quality Performance dataset that feeds BTS on-time rankings. It covers nonstop scheduled-service flights from 17 U.S. carriers above a 0.5 percent domestic revenue threshold, with airlines filing delay causes monthly. Aircraft type was never part of the filing, which is why a fleet-specific delay rate is structurally absent from the rankings.

Section 511(b)

Section 511(b) is the provision of the FAA Reauthorization Act of 2024 that reshapes DOT delay-cause categories. The final rule implementing it was published September 3, 2026, and takes effect October 19, 2026. It creates a new category for ten specified events declared outside the Air Carrier code, putting post-rule delay data on a different footing from earlier years.

Questions? Answers.