Stupid Government Tricks: Airline On-Time Percentages

DOT requires airlines provide shoppers with broad data about the performance of a particular flight, by number: what percentage of the time it arrived on-time, was delayed by 30+ minutes, or flat-out cancelled. The airline is required to provide the most recent month of data to you that it has reported to the U.S. DOT — meaning the data is several weeks stale. Most do it such that if you’re shopping online, you just click the flight number to see the data. The catch is, you shouldn’t click. You will waste moments of your life you can never get back, reading data that is, almost certainly, completely useless to you, said uselessness only clarified very vaguely in fine print*.

Three of the biggest reasons the DOT’s rule fails you are below:

Weather Variance and Disasters

Weather is possibly the biggest single factor in flight delays, in that it both delays flights directly and causes further down-route issues, like outbound aircraft arriving late. In an aviation network as complex as the U.S. or continental Europe, small weather changes (like a big storm in the northeastern U.S. that happens one year and not another) dramatically change year-to-year on-time rates. Even if your weather doesn’t change from December to February, weather in the rest of the country/world does, and that impacts delay percentages. You can have a perfect day in San Francisco, and if your plane is stuck in a blizzard in the northeast, you’re not getting out on time.

The only weather-related factor that an individual airline can even change is routing: where it puts its hubs, for example. Every potential flight path has some vulnerability to weather or geology; earthquakes, landslides, snow, heavy rain, floods, extreme heat, severe thunderstorms and even tropical weather like hurricanes, impact different parts of the United States, at different times of the year. If you’re booking a March or May flight, you don’t really care about delays in December from winter weather in the northeast. Similarly, if you’re booking a December flight, you don’t care about the airline with a North Carolina, Florida or Texas hub that got walloped by a hurricane and had major delays. Year-ago data would help, here, but of course, that’s not what the DOT requires, in part because flight paths can change considerably in that time.

Airport Issues

Taking a flight that goes through San Francisco? Good luck getting in or out on time, no matter what your airline or flight number is, and trust me: you’ll need it. With only a 77.9% on-time departure rate in Travel and Leisure’s 2014 survey, SFO is among the worst airports in the country for delays. That is, in part, due to SFO’s unpredictable, legendary fog. SFO’s morning-hour fog delays make it the worst airport in America for weather delays. Another factor is SFO’s runways are too short to safely land large jets in the low visibility conditions often seen there. Runway work designed to fix that problem caused, you guessed it, more delays. Such management issues are out of airlines’ control, and will hurt any airline that flies through a particular airport. If you look at flights into/out of SFO, you’ll be lucky to find an on-time arrival rate north of 70% for a mid-morning flight.

Sample Size

With only a month of data, and only for one flight number, a fluke problem will decimate the flight’s on-time percentage when the data is reported. Even for a hypothetical flight that ran every day in a month, a single delay moves the on-time rate by three points. A weather event a few days in duration, like a strong severe weather system, a winter storm or a hurricane, could easily knock a full ten points off the percentage, then vanish in the following month’s data.


The U.S. government requires carriers to produce statistics for your use as a consumer. The carrier pays employees to publish these statistics with your fare dollars, and the government pays employees to police them with your tax dollars. Yet these statistics provide data with very little correlation to whether your flight will arrive or leave on time, or whether a particular airline is generally a good one to fly between your arrival/departure points. Weather, a huge factor in delays, varies tremendously month-to-month and even day-to-day, and its effect is unpredictable. Airport management issues like runway work are also beyond the airline’s control. The handful of factors that truly are within airline control make such a microscopic dent in the results that the numbers are virtually useless to you as a consumer.

For some significant percentage of consumers, the data will be worse than useless, as they might choose an airline based on a factor that is out of date and suffer delays they shouldn’t have. For example, choosing a carrier based on winter-time statistics for a spring or summer flight, when the carrier performed best in the winter and has a poorer record at other times of year.

In any case, even useless data is worse than nothing, because you’re spending money toward its production every time you fly; that’s money you could be saving if not for the costs of producing the data.


About Matt Murphy

I'm a software engineer with a certain company you've probably heard of. In my free time, I'm an advocate for individual liberty and free markets.
This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s