Sean from the Budget Travel blog passes along this post on the economic cost of flight delays.

I was actually surprised by the quality of the analysis in the post about the FAA-commissioned academic study. The study itself concludes

Passengers lose about $16 billion a year because of “schedule buffer, delayed flights, flight cancellations, and missed connections.

But there are real differences between the various causes of delays, and while each is costly it’s not really useful to talk about ‘schedule buffer’ along with delays and of course missed connections are mostly a function of delays and cancellations, and are minimized by schedule buffer.

Sean points to air traffic control as the biggest driver of the problem, citing the Department of Transportation’s Inspector General, and that gets it right.

Of course plenty of operational factors drive flight delays and cancellations like crew scheduling and weather. And while technology makes it much easier to fly in bad weather than ever before, air traffic control issues here magnify the weather problem. Airport infrastructure does as well (as in San Francisco where runways are too close together to takeoff and land simultaneously in windy conditions off the bay). And local planning issues make it nearly impossible to alter those conditions. And of course the DOT’s 3-hour tarmac delay rule increases cancellations, so this study should underscore the real cost of the rule.

But blaming airlines for ‘schedule padding’ and saying it’s economically costly is just silly. Airlines don’t ‘pad schedules’ or pretend that flights take longer than they do just because they want to look good in their on-time stats. Longer scheduled flights mean a diminished ability to use each aircraft profitably. There’s greater labor expense. It’s just that airlines know that unanticipated delays are even more costly than realistic scheduling.

Just because an airline schedules flights for more time than the same trip took years ago, doesn’t mean the airlines are padding. Instead it points to a congested air traffic system that can’t push tin as quickly as it did years ago. Airlines would rather schedule flights for less time if they could meet those schedules. The problem here is predominantly air traffic, not airline scheduling.

The Budget Travel blog piece writes up the potential solution of capping flights at busy airports, even more than they are currently. But that’s both politically untenable (it pushes airlines away from smaller regional aircraft to less popular destinations but which tend to have strong political support) and because it’s a cop out. It lets the air traffic control system off the hook, it takes the ability to push aircraft through the system as a constraint rather than a problem to be solved.

And of course limiting flights has huge economic costs and undermines the point of the project which is economic gains. If you can’t take flights that are otherwise economically worthwhile to the airlines and you have to connect instead or sit around for more hours waiting for flights, those are exactly the kinds of economic costs that the study purports to criticize.

That’s perhaps a simple example of the economic fallacy, ‘the seen and the unseen’.

Another Bastiat-identified problem with the analysis is the broken window fallacy, O’Neill suggests that delays have economic benefits because stranded travelers stay in hotels, and airlines burning fuel on the tarmac benefit energy companies:

It should be said that not all flight delays are bad for the economy. If your flight is canceled and you have to pay, let’s say, $80 for a night’s stay at an airport hotel that you weren’t planning on, the hotel is profiting from your flight delay. If an airline has to spend more on fuel, that is bad for the airline but good for the energy company.

If these were valid arguments then we ought to strive for more cancellations! And we ought to break more windows so that everyone has to spend money with glass manufacturers. We could create our own shovel ready projects!

Ironically O’Neill rejects a GPS system that costs $10 billion to implement but would increase system capacity as ‘too expensive’ even while recognizing that the costs of delays are greater than that amount every single year.

But agreed, no reason to spend $10 billion if it isn’t necessary to reduce delays and the need for schedule padding. I’ve long argued for plane-to-plane communications, greater decentralization, and ultimately making planes more like cars.

One way to do this would be to increase available runways or open a new nearby airport (Peotone has its advocates). Capital construction could potentially build our way out of the problem.

It seems to me, though, that there are some good arguments for air traffic control privatization and even decentralization. Neither are sufficiently a part of policy dialogue.

Privatization has worked well in Canada and Britain.

.. Allowing planes and pilots to operate in the skies much like cars, with technology and communications that allow them to direct themselves while coordinating with each other (as a replacement for the current command and control model) offers some of the best hope for increasing the total capacity of the skies for air travel.

Anything that fails to increase capacity for aviation as an engine of economic growth can hardly be said to be acting in the name of economic growth. Now, there are other arguments which run contrary to growth, plenty of people don’t want to see more air travel especially if they live near an airport. But couching arguments for less flying (such as slot restrictions) in terms of economic improvements is ultimately disingenuous.