Flights with Two Hour Ground Delays Were Twice as Likely to Cancel This Summer


I know that I said I wouldn’t bother writing about ground delays again until some new, interesting data came out, and now, we have some. The funny thing is that I didn’t even see it. It was only brought to my attention by the DOT as part of the agency’s campaign to convince the world that the ground delay rule is having no ill effects, but clearly we read the data differently. While the DOT sent this data to me as proof in its favor, what I saw was a doubling of the rate of cancellations when flights were held on the ground for more than two hours.

Here’s the chart with the data that DOT spokesperson Jill Zuckman sent me:

Ground Delays Over Two Hours

It really shows how data interpretation can vary widely. According to Jill, “the data shows that there were fewer cancellations involving flights that experienced tarmac delays of more than 2 hours during May-August 2010 when compared to the same period in 2009.” Well yeah, that’s true, but more importantly, it shows that when airplanes were delayed for two hours on the ground, they were twice as likely to cancel this year than last. Now that to me seems like proof that the ground delay rule is causing cancellations, no? I mean, without the rule, you would assume that the historical percentage would still be the rate we’d see today.

But let’s go back to the aggregate number. Jill and the DOT look at this and say that the “decline can be attributed to the fact that the average number of 2+ hour tarmac delays was much lower in the first 4 months of the post rule period than in the same 4 month period in 2009.” Yes, that’s true. But why is that the case? Airlines have probably started canceling flights further in advance. Or maybe the weather just wasn’t as bad this summer as it was last. (I already dug in and showed that earlier this summer.) There are a lot of reasons why cancellations numbers can fluctuate, but in general it’s been hard to pin down the reasons on a single event until this nugget of data was brought to my attention.

Let’s think about this. It’s no surprise that the number of 2 hour ground delays is down year over year, right? I mean, when you have harsh penalties kicking in at 3 hours, of course you’re going to make changes to prevent anything staying out there over 3 hours. We knew that, but my hypothesis has been that it will negatively impact cancellations. That seems to be the case here. It looks like once airplanes are sitting on the ground for over two hours, it means there’s probably a weather problem or some other operational event that’s preventing airplanes from getting to the gate. Last year, of those flights, 4 to 8 percent were canceled in any given month. But this year, it’s 10 to 14 percent. Why would that be? Because the 3 hour rule is forcing airplanes to come back and cancel.

Without having the airlines give us specific numbers, this is the closest I’ve seen to something showing that there without a doubt have been more cancels this year because of the rule.

Get Cranky in Your Inbox!

The airline industry moves fast. Sign up and get every Cranky post in your inbox for free.

23 comments on “Flights with Two Hour Ground Delays Were Twice as Likely to Cancel This Summer

  1. Of note is my take on your interpretation: you are simply stating that the cancellation rate has gone up for these affected flights. You are NOT stating that this is a bad thing.

    A lot of us do NOT want to be stuck in a middle seat on a plane for a long+indeterminate delay…

  2. Some of us would rather be stuck in the middle seat during a medium to long delay if it allows me to get to my destination in that day (or night). If the flight is canceled then I am waiting till later in the day (best case) or more likely sometime the next day! And remember since most domestic flights are at 80+% capacity where are all those passengers gonna fit? If you have a canceled flight you will not get to your destination for quite some time. and the airlines owes you NOTHING. Not a room overnight, not food, nothing, if it is weather related of course.
    So have fun sleeping in the airport or a hotel on your dime, but you don’t have to be in the middle seat during your delay! enjoy……

  3. Well, one thing that I would point out is that the total number of delays > 2 hours is pretty darned minuscule. Does further analysis suggest that these delays are largely confined to a handful of airports? It wouldn’t surprise me if the NY-metro + ATL are the culprits here again.

    BTW, Brett, I don’t know what airline-supplied numbers you want to see. The stuff you would want to look at (raw operational data) should be published in the BTS Airline on-time performance table. Is it not?

    1. That is correct. These numbers are absolutely tiny in the scheme of things. While I don’t think the delays themselves are confined to those congested airports, they’re almost always related to them. I mean, a flight could be destined for NYC and then stuck elsewhere. This also becomes a problem when weather causes diversions. Problems definitely happen out in diversion airports.

      As for what data I want, it’s pretty simple. I want the airlines to say how many flights have been canceled due to the 3 hour rule that wouldn’t have been canceled otherwise. Nobody seems willing to put that out there.

      1. Brett … I’m not sure that number exists. What airline is going to say “we had 5 flights last month where we would’ve preferred to leave the passengers stranded on the tarmac for four hours rather than bring them back to the terminal”? There’s almost no way to spin that into a positive story.

  4. This is one topic that will always have people saying the ground rule is good and those saying it’s not.

    People are always quick to say they don’t want to sit on a plane on the ground for hours, but they never comment much on how they will feel about not going at all (if they are outbound from home city) to taking days to get to where they are going which could be home.

    I don’t think there will be a solution to this until humans can control the weather or just ‘beam’ from place to place like on Star Trek.

  5. this is a bad analysis. A) of course airlines are going to cancel more flights that are delayed for more than two hours, it would be dumb not to do that and risk paying huge fines. B) the total number of flights with a ground delay of more than 2 hours went DOWN by 50%. C) the total number of flights with ground delays of more than two hours that were canceled is DOWN as well (slightly). D) now you can mess around with the percentages as you did or you can simply say:
    THE NUMBER Of flight who were affected by long ground delays went DOWN significantly and the NUMBER of flights with long ground delays and cancellations went DOWN as well…….

    1. Yeah, you can make that statement, but you can’t say that it’s because of the three hour rule.

      Without the rule, ground delays were still likely to go down significantly this year thanks to a variety of reasons including weather and greater awareness. I like a lot of the parts of the rule including the requirements for action plans in every city. That would greatly contribute to a reduction in ground delays here.

      So we can’t say how much of the difference this year is due to a 3 hour rule and how much is due to external factors. What we can say, however, is that those flights that did see ground delays for over 2 hours were twice as likely to cancel this year, and that would be because of the 3 hour rule.

  6. Tough to really draw conclusions here, given that you are working with only two datapoints. As you say, the big drop in ground delays might be all driven by weather. Really the DOT should be providing an average rate of ground delays for a number of years going back, so that we can get a real sense of a historical average for comparison’s sake. Even then, you need a way to control for the impact of weather (airport-days of bad weather might be a metric?).

    Also, as Michael said, the DOT can still claim that the total # of cancellations is down slightly, even if the rate is up.

    1. I think the point opponents are making is that for every flight that is canceled, that’s a lot more people inconvenienced than a flight that sits on the tarmac for a few hours and ultimately gets to its destination. Saving 150 people from an extra hour on the tarmac means 150 people are twice as likely to be left to fend for themselves whilst they are rebooked on a later flight, on the same day if they are lucky.

    2. I agree, Ed. It’s really hard to draw conclusions because we can’t get the full picture. That’s why I just looked at this little corner of flights because it seems like we can draw a conclusion on this little piece. But until we get better data from both sides, we can’t know for sure how to judge this in its entirety.

  7. One must take this situation into perspective. If you are held captive in an aircraft for 2, 3 or more hours not only is it a major inconvenience (food, water available, etc?) you may loose your connections, hotel reservation, etc. So, short of getting home on time you may be screwed even if you take off hours late. In many cases the airlines have brought this situation on themselves with the continuous abuses to passengers. Their “fix” may not be very acceptable, so each travler better know his Rights and the obligation of a disrupted trip by the airline. Each airline may have a different response, so check it out – BEFORE you book or travel.

  8. As a former airline PR practioner, I think 10 canceled flites draw much less media attention than a single flite that sits on the taxiway for 3 hrs or more and yields countless first-person accounts of horrendous conditions on board … no water, no a/c, no information, outrageous confinement and the ultimate headline grabber, STRANDED, etc etc. So I’d say on balance the new rules are working just fine.

  9. .
    So, if I’m understanding this analysis properly, the new rule resulted in a 100% reduction in tarmac strandings at the cost of an increase in cancellations of two-hour+ delayed flights from 8% to 14%.

    That looks like an overall WIN.

    1. We can’t really draw that conclusion. I don’t think anyone would say that all of the reductions in ground delays were due to this rule. And we really don’t know how much of it was related and how much was due to external factors. But I feel confident tying the increase in cancellations back to the rule, because I can’t think of another reason why that would be happening.

      1. “But I feel confident tying the increase in cancellations back to the rule, because I can’t think of another reason why that would be happening.”

        This is probably true, the rule is likely a big influence in the reason for cancellation. However there is the suggestion (not so much today but in the past) that the cancellations are inconveniencing more people – and that is not valid.

        For example, let’s assume 100 people per plane, to keep it simple. That means 19600 people vs 21800 people a year ago. That’s better than a year ago. [I know we’re working with only a couple of datapoints of course].

        I think the overlooked factor is the number of flights delayed 2+ hours. Down from 3401 to 1570. Very good. This is the number to investigate. Why did the number of 2+ hour delays fall in half? It could be:
        a) better weather
        b) fewer total flights
        c) better scheduling of flights throughout the day and less overscheduling (maybe not slots, but airport gate / ramp capacity)
        d) more delayed flights held at the gate instead of dispatched and held on the ramp/tarmac etc.
        e) proactive cancellations and consolidation of flights (eg. Delta Shuttle)
        f) increased schedule padding
        g) defensive cancellation due to the rule

        I’m sure there’s more. Would be great to see if it was possible to know how each of these contributed to the decrease from 3400 to 1700 flight delays of 2+ hours.

        (a)-(b) are outside airlines’ control. (c) I don’t know it it’s real or not, but if its true that’s good for customers. (d)-(e) I would argue are positive side-effects of the 3-hour rule. (f)-(g) I would say are negative side-effects. And, it would be great to know, if (f)-(g) are really happening, or if its the other factors which I think are good outcomes of the 3-hour rule.

        Always fun to comment on this topic!

        1. I think we actually could get some of these numbers but certainly not all. I mean, we know how many flights were run at each airport and how flights are scheduled throughout the day. But this is a good model showing how many different types of reasons there are for these changes and why without more detail from airlines, we can’t really know the impact for sure of each one.

  10. the analysis that Brett is speaking too is that if your flight is delayed 2 hours (which means you have already been impacted) your odds on getting canceled all together goes up much more now than it is before the rule.

    I realize my opinion is biased by they fact that I would rather have the rules in place about food and water but not the penalty aspect. If I am onboard a plane for 3hrs and have food and water but get to my destination it is fine with me. Now if I am onboard a plane for 2 hrs then head back to the gate and get canceled. Then I am searching for another flight and (most likely) a hotel for night (at my own expense). In which case and I most impacted and inconvenienced?

  11. “The analysis that Brett is speaking too is that if your flight is delayed 2 hours (which means you have already been impacted) your odds on getting canceled all together goes up much more now than it is before the rule.”

    As well as it should–because the odds that if your flight is delayed 2 hours you will be stuck in it for 4, 5, 6, 7, 8, 9 hours have gone from something like 5% (Bret you can run the numbers) to pretty much zero.

    It’s a great WIN.

    1. Whoa, you’re waaaaay too high on that one. In August 2009, a tiny .13 percent of flights were delayed for more than 2 hours on the ground. In August 2010, it was a tinier .04 percent. No matter what, we’re still dealing with an incredibly small percentage. And once again, we can’t directly tie that cut back to the three hour rule itself because of all the other potential factors.

  12. It doesn’t make any difference if the cancellations go to TEN TIME the previous numbers. The airline’s jobs are to take people from one point to another. Sooner or later they will figure out how to do it without canceling. If they cancel, they don’t make money. And money is why they are in business.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Cranky Flier