FlyerIntel Research
The Holiday Travel Penalty Is Smaller Than It Looks
Thanksgiving, the Christmas-to-New-Year run, the Fourth of July, and the Memorial Day and Labor Day weekends, measured against every other day in FlyerIntel’s data.
Published September 5, 2026 · Source period July 2023 – July 2026
Summary
Across the five biggest US travel weeks, covered airports cancelled 1.15% of scheduled departures against 1.51% on an ordinary day. Holiday periods cancel less, not more. On-time performance moves the other way, by less: 77.3% of holiday-period arrivals were on time against 78.2% otherwise, a gap of 0.9 points. The five weeks do not move together: Labor Day weekend is the most reliable period measured, at 83.9% on time, while Memorial Day weekend is the least, at 74.7% with an average arrival delay of 12.0 min.
Key findings
- Cancellation rate is lower than on an ordinary day in four of the five travel weeks, from 0.69% at Labor Day weekend, against 1.51% network-wide. Independence Day is the exception, at 1.59%.
- Labor Day weekend and Thanksgiving week post the best on-time rates of any period measured, both above the 78.2% ordinary-day rate.
- Memorial Day weekend is the least reliable of the five: 74.7% on time and an average arrival delay of 12.0 min, against 7.5 min on an ordinary day.
- Weather’s share of coded cancellations is above its ordinary-day 59.1% in all five weeks with enough coded cancellations to split, as high as 78.4%; the carrier-attributed share is below its ordinary-day 25.9% in four of them.
- FLL (Fort Lauderdale) shows the sharpest holiday-period drop in on-time performance: 67.4% against 73.0% on an ordinary day, a fall of 5.6 points. ANC (Anchorage) improves the most, gaining 3.0 points.
The warning that is only half right
Every year brings the same coverage: pack patience along with your bags, the busiest travel days of the year are coming, expect chaos. The data behind this report agrees with half of that and contradicts the other half. Cancellations, the outcome that actually strands a traveler, are lower than on an ordinary day in four of the five weeks measured here; Independence Day is not. What the warning gets right is delay: Memorial Day weekend, Independence Day and Christmas through New Year’s are measurably worse than an ordinary day, while Labor Day weekend and Thanksgiving week are measurably better. There is no single "holiday effect" in this data. There are five different weeks, and they do not behave alike.
On-time arrival rate, by period
Cancellation rate, by period
Why the cancellations that do happen, happen
| Period | Weather | Carrier | National Airspace System | Security |
|---|---|---|---|---|
| Thanksgiving week | 78.4% | 17.2% | 4.4% | 0.00% |
| Christmas through New Year’s | 62.4% | 33.0% | 4.5% | 0.11% |
| Independence Day | 63.6% | 18.9% | 17.6% | 0.00% |
| Memorial Day weekend | 71.3% | 18.6% | 10.1% | 0.00% |
| Labor Day weekend | 76.0% | 17.9% | 6.1% | 0.11% |
| An ordinary day | 59.1% | 25.9% | 14.9% | 0.04% |
The cancellation reason is the single code the operating carrier files with the DOT. A period is shown only where it recorded at least 500 coded cancellations.
Every covered airport, holiday weeks against an ordinary day
| Airport | City | Holiday-period on time | Ordinary-day on time | Difference |
|---|---|---|---|---|
| FLL | Fort Lauderdale | 67.4% | 73.0% | -5.6 pts |
| RSW | Fort Myers | 73.9% | 79.2% | -5.3 pts |
| DTW | Detroit | 76.3% | 79.8% | -3.5 pts |
| MKE | Milwaukee | 77.5% | 81.0% | -3.5 pts |
| STL | St. Louis | 75.8% | 79.3% | -3.5 pts |
| OMA | Omaha | 78.6% | 82.0% | -3.4 pts |
| ORD | Chicago | 71.5% | 74.7% | -3.3 pts |
| EWR | Newark | 75.3% | 77.9% | -2.6 pts |
| BUF | Buffalo | 77.1% | 79.6% | -2.5 pts |
| MCO | Orlando | 72.6% | 75.0% | -2.4 pts |
| MIA | Miami | 71.0% | 73.3% | -2.4 pts |
| MSP | Minneapolis | 78.9% | 81.3% | -2.4 pts |
| SJU | San Juan | 74.1% | 76.4% | -2.3 pts |
| BWI | Baltimore | 73.9% | 76.1% | -2.2 pts |
| DFW | Dallas–Fort Worth | 71.0% | 73.2% | -2.2 pts |
| JFK | New York | 75.0% | 77.2% | -2.2 pts |
| ABE | Allentown | 81.0% | 83.0% | -2.0 pts |
| MDW | Chicago | 74.7% | 76.6% | -1.9 pts |
| DAL | Dallas | 74.4% | 76.1% | -1.7 pts |
| IAH | Houston | 75.8% | 77.5% | -1.7 pts |
| MCI | Kansas City | 79.5% | 81.1% | -1.7 pts |
| MEM | Memphis | 79.6% | 81.3% | -1.7 pts |
| TPA | Tampa | 76.7% | 78.3% | -1.7 pts |
| PHL | Philadelphia | 74.8% | 76.3% | -1.5 pts |
| IND | Indianapolis | 80.5% | 82.0% | -1.4 pts |
| ATL | Atlanta | 77.6% | 78.8% | -1.2 pts |
| CLE | Cleveland | 79.8% | 81.0% | -1.2 pts |
| JAX | Jacksonville | 78.9% | 80.1% | -1.2 pts |
| BNA | Nashville | 77.3% | 78.3% | -1.0 pts |
| ABQ | Albuquerque | 82.4% | 83.2% | -0.9 pts |
| SMF | Sacramento | 82.3% | 83.3% | -0.9 pts |
| CMH | Columbus | 81.6% | 82.1% | -0.6 pts |
| IAD | Washington | 80.5% | 81.0% | -0.6 pts |
| PIT | Pittsburgh | 81.5% | 82.1% | -0.6 pts |
| DEN | Denver | 75.7% | 76.0% | -0.4 pts |
| SAN | San Diego | 77.5% | 78.0% | -0.4 pts |
| SBA | Santa Barbara | 81.9% | 82.3% | -0.4 pts |
| CHS | Charleston | 80.3% | 80.6% | -0.3 pts |
| HOU | Houston | 78.5% | 78.9% | -0.3 pts |
| ILM | Wilmington | 81.4% | 81.7% | -0.3 pts |
| MSY | New Orleans | 78.5% | 78.7% | -0.2 pts |
| RDU | Raleigh–Durham | 79.2% | 79.4% | -0.2 pts |
| PDX | Portland | 83.1% | 83.2% | -0.1 pts |
| SFO | San Francisco | 78.6% | 78.7% | -0.1 pts |
| AUS | Austin | 78.0% | 78.0% | 0.0 pts |
| CLT | Charlotte | 74.8% | 74.8% | 0.0 pts |
| SAT | San Antonio | 82.3% | 82.3% | 0.0 pts |
| OAK | Oakland | 82.1% | 82.0% | +0.1 pts |
| LAX | Los Angeles | 81.5% | 81.4% | +0.2 pts |
| PHX | Phoenix | 80.7% | 80.3% | +0.4 pts |
| BUR | Burbank | 83.3% | 82.7% | +0.6 pts |
| BOS | Boston | 77.9% | 77.2% | +0.7 pts |
| ONT | Ontario | 84.2% | 83.5% | +0.7 pts |
| TUS | Tucson | 85.2% | 84.2% | +1.0 pts |
| MYR | Myrtle Beach | 81.7% | 80.5% | +1.1 pts |
| SJC | San Jose | 85.2% | 84.1% | +1.1 pts |
| SLC | Salt Lake City | 85.0% | 84.0% | +1.1 pts |
| SNA | Santa Ana | 83.4% | 82.1% | +1.3 pts |
| LGA | New York | 79.8% | 78.2% | +1.6 pts |
| DCA | Washington | 78.1% | 76.4% | +1.7 pts |
| LAS | Las Vegas | 79.8% | 77.8% | +2.0 pts |
| HNL | Honolulu | 87.5% | 85.1% | +2.3 pts |
| SEA | Seattle | 81.6% | 78.9% | +2.7 pts |
| ANC | Anchorage | 86.3% | 83.4% | +3.0 pts |
Figures describe flights departing this airport, judged on arrival at their destination (the same basis as the "On-time arrivals" figure on the airport’s own profile page). Holiday period combines all five windows at that airport. Ranked from the largest holiday-period decline to the largest improvement.
What this does and does not explain
The most likely reason cancellations mostly fall during the highest-visibility travel weeks is that airlines plan for them: spare aircraft and crews are positioned specifically because the demand and the media attention are both predictable months in advance, and because a cancellation on Thanksgiving Wednesday costs an airline more in rebooking, compensation and coverage than the same cancellation on a random Tuesday in March. Delay is a different problem, because it is driven by how much traffic the airspace and the airports can absorb at once, and that does not go away just because an airline would prefer it to. Memorial Day’s weak showing lines up with the start of severe convective weather season in much of the country; the Christmas run sits in the middle of winter storm season. Thanksgiving and Labor Day fall in comparatively quiet weather windows, which is likely part of why they perform best. But this dataset has no weather field to confirm that directly, so it is offered as the most plausible explanation rather than a measured one.
Pages for what this report covers
Airports: Fort Lauderdale (FLL) · Anchorage (ANC) · Fort Myers (RSW) · Detroit (DTW) · Milwaukee (MKE) · St. Louis (STL) · Omaha (OMA) · Chicago (ORD) · Newark (EWR) · Buffalo (BUF)
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/. This report is a departures measure throughout: one record is one scheduled departure from one of the airports FlyerIntel covers, judged on-time by its arrival at its destination: the same basis as the "On-time arrivals" figure on that airport’s own profile page, not the arrivals-at-this-airport basis used by the connection-airport report. Holiday windows are computed, not hand-picked: Thanksgiving runs the Wednesday through the Sunday around the fourth Thursday of November; Memorial Day and Labor Day weekends run the Saturday through the Monday around the last Monday of May and the first Monday of September, the dates fixed by law; the Christmas run is December 23 through January 2; Independence Day is July 1 through 5. A window is included only where it falls entirely inside the loaded period, so a holiday can be measured over a different number of occurrences than the others; over the period behind this report it is 3 to 4, depending on the holiday. "An ordinary day" is every scheduled departure outside all five windows. The per-airport table includes an airport only where its combined holiday-period flights meet the site’s standard time-bucket threshold of 120 flights, which every covered airport clears.
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 five windows are a modeling choice (a fixed number of days around each holiday), and a different width would shift the figures somewhat, though not the direction of any comparison in this report. The dataset has no weather observations and no measure of how full a flight was, so neither can be checked directly against the patterns described here. "On time" and "cancelled" describe the flight, not the passenger: a cancellation that successfully rebooks someone onto a later flight the same day is counted the same as one that does not.
Sources
- FlyerIntel analysis — FlyerIntel · source
- Reporting Carrier On-Time Performance — US Department of Transportation, Bureau of Transportation Statistics · source
Full FlyerIntel methodology · Report a problem with this analysis