FlyerIntel Research

Cancel Early or Run Late

In bad weather an airline can cancel more and run closer to schedule, or cancel less and land late. The data shows which airlines did which.

Published October 12, 2026 · Source period July 2023 – July 2026

Summary

Republic Airways arrived on time 69.0% of the time in bad weather against 64.7% expected, and cancelled 8.7% against 5.7%. PSA Airlines arrived on time 59.1% against 65.0% and cancelled 10.9% against 6.2%. Republic Airways and Endeavor Air cancelled more than the others and still arrived on time more often.

Key findings

  • 1,428,157 bad-weather departures from 13 airlines were compared with the other airlines at the same airports and months, on cancellations and on the arrivals of the flights that operated.
  • Delta Air Lines cancelled 2.7% against 3.9% expected and arrived on time 69.3% against 65.8%.
  • Southwest Airlines cancelled 2.8% against 3.7% expected and arrived on time 65.2% against 67.2%.
  • Republic Airways cancelled 1.5 times the expected number and arrived on time 69.0% against 64.7% expected.
  • Across airlines the correlation between cancellation points above expected and on-time points above expected was -0.27.
  • PSA Airlines had 20.6% of completed bad-weather flights arrive 60 or more minutes late against 15.9% expected.

Two ways to handle a storm

An airline facing a storm can cancel flights early and fly the rest close to schedule, or keep most of them going and let them arrive late. Cancellation rates alone cannot tell which an airline chose, and neither can arrival rates, because each leaves out what the other counts. Here both are measured for the same 1,428,157 bad-weather departures from 13 airlines, each against what the other airlines did at the same airports in the same months.

In bad weather: cancellations and arrivals, against the other airlines at the same airports

In bad weather: cancellations and arrivals, against the other airlines at the same airports
AirlineBad-weather departures comparedCancelledExpectedArrived on timeExpectedPoints against expectedArrived 60 or more minutes lateExpected
Republic Airways98,2338.7%5.7%69.0%64.7%+4.213.5%15.9%
Delta Air Lines261,7132.7%3.9%69.3%65.8%+3.512.4%14.7%
United Airlines200,9002.9%5.0%66.5%63.1%+3.514.7%16.3%
Endeavor Air21,4355.2%3.3%71.4%69.6%+1.812.2%12.3%
Envoy Air44,5398.8%7.2%58.5%58.7%-0.219.0%19.9%
Spirit Airlines47,5755.1%4.2%64.0%65.2%-1.216.1%15.5%
American Airlines216,2745.5%6.1%63.3%64.5%-1.317.2%16.2%
Southwest Airlines215,6742.8%3.7%65.2%67.2%-2.013.1%14.3%
Alaska Airlines48,6522.4%2.4%71.1%73.9%-2.98.7%10.1%
SkyWest Airlines126,5915.4%3.1%61.9%65.9%-4.017.5%13.4%
Frontier Airlines45,0054.6%4.3%58.3%63.1%-4.819.3%16.1%
JetBlue61,5025.2%5.1%60.5%66.2%-5.718.9%15.1%
PSA Airlines40,06410.9%6.2%59.1%65.0%-5.920.6%15.9%

Departures in bad weather at covered airports, matched to the other airlines by airport, by month and by weather. Expected is what the other airlines’ rates would give this airline’s flights. On-time and late shares are of completed flights; cancelled and diverted flights are not in them. Sorted by points of on-time arrival above expected.

Where each airline lands

Delta Air Lines and United Airlines cancelled fewer flights than the others and also arrived on time more often than the others. Southwest Airlines cancelled fewer flights than the others and arrived on time less often than the others (Southwest Airlines: 2.8% cancelled against 3.7%, 65.2% on time against 67.2%). Republic Airways and Endeavor Air cancelled more flights than the others and arrived on time more often (Republic Airways: 8.7% cancelled against 5.7%, 69.0% on time against 64.7%). SkyWest Airlines and PSA Airlines cancelled more flights than the others and arrived on time less often, so the extra cancellations bought no better arrivals. Alaska Airlines, Frontier Airlines, and JetBlue cancelled about what the others did and arrived on time less often. Republic Airways arrived 4.2 points above the expected on-time rate; PSA Airlines arrived 5.9 points below it.

Does cancelling more buy better arrivals

Across the 13 airlines the two gaps showed no clear relationship, with a correlation of -0.27 between points of cancellation above expected and points of on-time arrival above expected. Republic Airways and Endeavor Air are the exceptions that fit the idea. With 13 airlines this is a rough guide and not a measurement of a policy.

The long delays

PSA Airlines had the largest excess of arrivals 60 or more minutes late: 20.6% of its completed bad-weather flights against 15.9% expected, 4.7 points above the expected share. Republic Airways, Delta Air Lines, United Airlines, Envoy Air, Southwest Airlines, and Alaska Airlines had fewer long delays than expected. For 10 of 13 airlines the on-time gap and the long-delay gap point the same way: more arrivals on time goes with fewer arrivals that late.

What this does not show

The comparison is by airport and month and does not hold the hour, the length of the flight or the aircraft. A cancellation can be a decision made the day before or an aircraft that never arrived, and the data separates neither. A cancelled passenger and a passenger on a flight that lands late have different problems, and a ranking by on-time arrival does not weigh one against the other. Standard errors treat flights as independent, which a storm over one airport is not, so the screen for a difference of 3 standard errors is looser than it sounds.

Methodology

Calculated by FlyerIntel from 21,708,321 individual flight records in the US Department of Transportation Bureau of Transportation Statistics Reporting Carrier On-Time Performance dataset, covering July 2023 through July 2026. Following BTS methodology, a flight counts as delayed when it arrives 15 or more minutes after its scheduled arrival time. Cancelled and diverted flights have no arrival time and are excluded from on-time rates rather than counted as late; cancellation rates are reported separately. Full definitions are published at /methodology/. Bad weather means IFR or LIFR conditions, or a thunderstorm, snow or ice, fog, heavy rain or high wind in the hourly observation at the origin airport for the scheduled departure hour. For each airline and each stratum (one airport, one month of one year, one kind of weather) the other airlines’ cancellation rate, on-time arrival rate and share arriving 60 or more minutes late were applied to the airline’s own flights, and the results summed; a stratum counts only when the others have at least 120 departures and 120 completed flights in it. Better or worse on time needs 1.5 points and 3 standard errors; Fewer or More cancellations needs 3 standard errors and a ratio at least 10% from one. Airlines with fewer than 10,000 matched bad-weather departures are not shown.

Limitations

The source dataset covers scheduled domestic flights operated by carriers above the DOT reporting threshold; smaller carriers and international flights are not included. Flights sold under a mainline brand but flown by a regional carrier are attributed to the operating carrier, which is how the DOT records them. Past performance describes what has happened and is not a prediction about any individual flight. The correlation is over a handful of airlines and is a rough guide. The standard errors assume independent flights and are smaller than the real uncertainty. The data does not say when a cancellation was decided.

Sources

  • FlyerIntel analysis — FlyerIntel · source
  • Iowa Environmental Mesonet ASOS/METAR archive — Iowa Environmental Mesonet, Iowa State University · source
  • Reporting Carrier On-Time Performance — US Department of Transportation, Bureau of Transportation Statistics · source

Full FlyerIntel methodology · Report a problem with this analysis