Air Traffic Delays Are Awful, Everything Else is Worse

Economic growth in general and deregulation in particular have brought about lower real airline ticket prices and substantial increases in passenger traffic. Infrastructure hasn’t kept pace and air traffic control hasn’t kept pace. And so we face congested airports and airspace. This frustrates travelers, who live in a world of tradeoffs but take what they have as given and long for what they’ve given up. Richard Posner is one such business traveler.

And the usually sober, sometimes brilliant, and certainly prolific judge and scholar offers up an unusually misguided rant on why he believes “airline service is so bad” over at the Becker-Posner Blog, Becker rebuts.

Posner lays out a myriad of unsupported gripes and mixes in blatantly incorrect facts, leading him to circle around some rather silly notions about re-regulating the aviation industry. Becker does a nice job offering an alternative theory of the world. I bring you the alternate set of facts. Hopefully together we can reconstruct a bit of sense about what’s wrong with the industry and how to correct it.

Here are Posner’s grips:

Airline delay has increased in the last five years, and the statistics understate the amount of delay because airlines have increased scheduled flight times–the flight from Chicago to Washington used to be scheduled for an hour and a half; now it is scheduled for two hours. Flights are horribly crowded, food and beverage service has deteriorated in first class and virtually disappeared in coach, and the incidence of mislaid baggage has increased.

Now, I’m not unsympathetic to rants about air travel. I write this as I sit waiting for a delayed flight to Phoenix. And I’m flying USAirways so I certainly empathize with complaints about degraded service levels. But my woes this evening don’t substitute for a more robust understanding of aviation when setting policy.

Delays are a real problem, if they can be alleviated at an acceptable cost there’s presumably real economic productivity to be captured here at least if you attribute airport wait times as dead times. Although with wireless internet, handheld devices, and a bit of strategy I find that airport transit times can be among my most productive. Still, I wouldn’t choose a delay unless I’m flying first class on Thai Airways out of Bangkok and can enjoy my delay in the spa. Delays are the problem worth focusing on.

Posner’s complaint about lack of food rings a bit hollow. True, airlines no longer advertise cooked to order steaks onboard domestic flights, but that was an artifact of a bygone era in which airlines couldn’t compete on price so they were forced to compete on amenities. Air travel has simply become mass transportation in a way that it wasn’t during the regulated era. Posner complains prices are too low, and if you regulated higher prices the competitive market would – judging from history – respond with food. But that’s hardly a tradeoff worth defending.

Airports contain ever more convenient to-go meal options, and airlines have replaced complimentary food with “buy on board” in many cases. Continental remains a leader in providing complimentary meals on domestic flights, but this is no longer the rule — United doesn’t even offer a complimentary meal in coach for its Chicago – Honolulu service! But that doesn’t mean there’s no food available, it just isn’t included in the price of your ticket. Virgin America has gotten quite a bit of mileage out of its innovation offering passengers the ability to ‘place their order’ through the onboard entertainment system. Plus the market has responded in many airports with an increasingly array of takeaway meals for a wide variety of tastes.

As Posner narrows his focus to delays, he offers that ‘others’ have claimed a variety of causes:

high fuel costs that have contributed to deferred maintenance that results in cancellations, the failure of the Federal Aviation Administration to upgrade the air traffic control system so that it can handle more traffic with less spacing between aircraft, more turbulent weather perhaps due to global warming, and crowded aircraft that result in delays in boarding and hence in departure.

The deferred maintenance piece is silly, there’s no evidence at all that maintenance necessary for aircraft operations has been systematically delayed and also no evidence that an increase in delays is in any way attributable to such practices. There have been many carriers deferring updates to their aircraft interiors (so you get fewer cleanings and pen marks on seats) but this is not due to fuel costs, but rather the overall economic environment which led to a rash of bankruptcies throughout the decade (United’s 2002 bankruptcy certainly predated current fuel price levels, as did USAirways’ two bankruptcies). And deferring optional maintenance prior to departure actually contributes to overall on-time performance…

Posner explains marginal cost pricing and then concludes that it is the cause of airline delay:

Prices tend to be pushed down to a level at which the airlines find it difficult to finance the purchase of new planes. As the existing planes age, equipment failures become more frequent, contributing to delays and cancellations. Airlines prefer delays to cancellations, because they get to keep the fares, and they resist raising prices to reduce congestion because that will make it more difficult to fill the planes, and an empty seat is, as explained, very costly in revenue forgone. Furthermore, airline service is quite uniform across airlines, which makes travelers more sensitive to airline prices than, say, to hotel prices, since hotels compete in many other dimensions besides price.

There are several mistakes here.

First, the idea that higher fares would mean newer planes, and that new planes would mean fewer delays (tell that to Singapore, the launch customer for the A380). Northwest, which operates the oldest fleet of the major US airlines, doesn’t have appreciably more delays than other carriers with younger fleets. Most delays aren’t mechanical in any case.

And to say that airlines don’t compete in areas other than price is silly. Obviously they compete on schedule (frequency of flights, total travel time to destination). Continental still offers meals. United offers “economy plus” with extra legroom. JetBlue pioneered inflight TV, and Airtran inflight XM Radio. United lets you listen to air traffic control. Virgin America has mood lighting and a minibar, not to mention a relatively affordable first-class cabin with pretty good food and decent seats for a domestic flights (and now access to the lovely Virgin Clubhouses for an added charge). Airlines have varying reliability, and this has little to do with maintenance. I’ve long done my best to avoid Delta’s express carrier Atlantic Southeast, other airlines outcompete them on the ability to deliver me to my destination on-time.

Deregulation has brought 11-figure annual savings, with ticket prices perhaps 40% lower in real terms than before deregulation. The real culprit to Posner is too many people flying, and he seems to attribute zero value to such incremental flying as a discretionary trip to visit family.

The cost of delay is a deadweight loss, whereas a higher price would be merely a wealth transfer to the sellers and would finance an increase in supply.

Empirically we know that higher prices reduce the total number of trips. Posner may hypothesize a world in which every seat on every plane is taken, with more people waiting to board who can’t, and so a higher price leaves us with the same number of trips. But that’s not the world we live in.

And thus he argues for higher prices (to promote efficiency from lower delays, ignoring the economic losses of the passengers shut out of frequent air travel. Of course he somehow seems to think that he – or a government regulator – can objectively determine what the ‘right’ price (which isn’t the profit-maximizing price, but the one which offers the optimum load factor and airport congestion).

Another aspect of lumpiness that should be noted is the difficulty of adjusting prices to different passenger time costs. Business travelers have higher time costs than leisure travelers, but there are not enough business travelers to fill a plane of efficient size, and even if there were, no one airline could significantly reduce the problem of delay, just as no one driver can affect traffic congestion by reducing the number of his trips.

But airlines do discriminate based on time costs. They offer both non-stop and connecting flights, with lower yield inventory at times only available on the connections (and varying these sorts of decisions nearly real-time based on demand). Airlines offer priority routes. They choose when to substitute an aircraft, pulling a plane off a leisure destination and substituting it for a mechanically delayed aircraft flying a route with a heavier concentration of high-paying time-sensitive business travelers. Airlines even have a say with Air Traffic Control over which of their delayed flights to give priority to, and which to cancel.

Meanwhile, the solutions that Posner offers don’t point towards price discriminating across different time preferences. Rather, his solutions use government policy to institutionalize his own time preferences into airfares.

Here Posner shows he’s calculating only one side of the equation:

I am not aware that the delay costs of airline service, and the costs of the other disamenities (the very crowded airplanes and slow boarding and deplaning in coach) in the current market, have been quantified, but assuming that they are, as I suspect, very substantial, the question arises what if anything should be done to alleviate the problem.

So if the costs are high, we should consider action. But the benefits are high, too, in the form of greater access to air travel. Skybus may have ceased operations, but if airlines are more like public buses than ever before the utility of all the riders on the bus need to be counted.

The first Posner solution is allowing price collusion on the part of the airlines, but he rejects that because prices might go up too much (the baby bear ‘just right’ price increase is the amount necessary to keep other people off planes so Posner can fly with elbow room, but not so much that airlines earn too much money from passengers with inelastic demand).

His next solution is re-regulation.

Another possibility would be to return to the regulatory system administered by the Civil Aeronautics Board before the deregulation of the airline industry in 1978. The CAB did not regulate rates, but it controlled entry into city pairs and used that control to limit entry to the point that flights were frequent and uncrowded.

Of course here Posner just doesn’t know what he’s talking about. The Civil Aeronautics Board absolutely did regulate prices as well as routes. Meanwhile, there are certainly more flights today than there were in 1978. The analysis in this section flows entirely from economic theorizing about a market in which entry is limited but prices aren’t regulated, which doesn’t describe reality at all.

In 1976, CAB Chairman John Robson pushed to “experiment” with price competition, and Continental was allowed to lower its intra-Texas fares to match and even undercut Southwest’s. Incidentally, Southwest introduced a two-tiered pricing structure… $13 each way fares matching Continental and $26 business fares which included a free fifth of alcohol. I vaguely recall Southwest becoming the largest liquor distributor in the state of Texas in 1977. (Folks whose employers were paying naturally selected the higher priced fare with the personal benefit to them. Heh. Southwest was also creative in pushing bookings to their airline by rewarding the secretaries who booked travel for their bosses with the equivalent of frequent flyer credits.)

But the idea that even when the CAB didn’t permit airlines to compete on the basis of price (and for many routes, didn’t allow them to compete at all) the notion that airlines didn’t compete in other areas is silly. Rather than driving down price, they drove up costs. There was even a time when the CAB got into the business of regulating the thickness of onboard sandwiches to try to clamp down on the myriad ways in which airlines tried to compete against each other in a world where they couldn’t adjust their prices.

When the CAB recommended deregulation, it did so unanimously. Ralph Nader favored it, and Ted Kennedy sponsored it. His committee chief counsel was Stephen Breyer. It was a consumer issue, not a free markets issue. And we got democracy in the skies. In a world of tradeoffs, not everyone likes it.

Penultimately, Posner postulates encouraging mergers as a solution. But that will lead to reductions in capacity, so while fares may rise they won’t solve his crowded planes a bit. And he acknowledges new entrants will keep the skies just as crowded.

After all that, his final punchline – which he doesn’t much analyze – is congestion pricing. Hardly a panacea, one the whole a reasonable part of an overall toolkit. It would help congestion by encouraging airlines to operate fewer flights with larger aircraft at peaks times across to and from the busiest airports. (Fewer flights on larger planes means less accommodation of the time preference of business travelers Posner wants…)

But even increased average fares (which, depending on how airline scheduling adapts, isn’t a foregone conclusion) wouldn’t generate the investment in new aircraft Posner wants to see or bring back meals to coach. The revenue would go to airport authorities. And economic expansion might well bring us back to congested airports in short order. It’s a modest solution hardly befitting the apocalyptic rant in the post.

So what are the real problems?

The existing array of airports serving US cities couldn’t possibly be built as-is today. With far greater environmental scrutiny and NIMBY opposition, it’s incredibly difficult to expand airport infrastructure. Similar to the US experience, the London-Heathrow terminal 5 project spent more time in its public comment phase than the entire much-larger Beijing terminal 3 took to go from proposal to completion. I’m not saying I prefer the Chinese model, but the difference illustrates how cumbersome infrastructure issues are in the modern Western political context.

Bob Poole does much good work on what could be accomplished through reform of the air traffic control system. He also includes congestion pricing among his solutions along with certain capital improvement projects and a variety of specific reform measures that hold a good deal more promise than re-regulating the airline industry. I prefer re-conceptualizing our whole approach to air traffic control, making flying planes more like driving cars aided by the kinds of technology for navigation and automated collision avoidance which couldn’t even be dreamt of decades ago when our current approach was put into place. My proposal won’t happen, of course, because the air traffic controllers stand to lose and they have both a keen interest in the outcome, financial resources through current dues, and the narrative of safety to bludgeon home their point.

So in the end we’re left with piecemeal tinkering, and a recognition that in a politically constrained world we have tradeoffs — and the current world is better than the one of thirty years ago.
Becker explains

I am arguing that many passengers prefer the combination of low fares and greater delays on average to higher fares and fewer delays

In fact, entire airlines such as Airtran have built their business model of “people waiting on planes” rather than “planes waiting on people” through longer connecting times in Atlanta to spread traffic out across the day and utilizing their labor more effectively as a result (this is in contrast to the standard hub-based model involves staffing for high periods of activity and paying those same staff to wait around during low periods for the hub’s next “bank” of flights).

Personally I might accept the current bargain more happily if airlines could improve the service provided by the same number of flight attendants at the same pay. Union contracts, labor laws, and culture all play a factor here. The highest average fares in the world won’t cause United or American to offer the same levels of service as Asian carriers such as Singapore, Cathay Pacific, ANA, and Asiana.

About Gary Leff

Gary Leff is one of the foremost experts in the field of miles, points, and frequent business travel - a topic he has covered since 2002. Co-founder of frequent flyer community, emcee of the Freddie Awards, and named one of the "World's Top Travel Experts" by Conde' Nast Traveler (2010-Present) Gary has been a guest on most major news media, profiled in several top print publications, and published broadly on the topic of consumer loyalty. More About Gary »

More articles by Gary Leff »


  1. My jaw dropped too when I saw Posner suggest that the CAB did not price-regulate.

    Meanwhile, did you see Robert Crandall’s NYT op-ed last week that made many of the same fallacious arguments as Posner?

  2. Responding to KipEsquire: Yes, that Crandall NYT article was bad but note that it’s not the Bob Crandall at Brookings. It’s the Bob Crandall who used to run American Airlines. Crandall’s piece recently on the WSJ op/ed page is first-rate.

  3. “Posner complains prices are too low, and if you regulated higher prices the competitive market would – judging from history – respond with food. But that’s hardly a tradeoff worth defending.”

    To say the least. According to this GAO report (, the median airline price has dropped 40% in constant dollars since 1980. Even on a relatively inexpensive flight (say, the $125 that Expedia is quoting for DCA to LGA), you’re still looking at a $50 minimum for that steak dinner.

  4. What do you think about the observation (by James Fallows and others) that widespread delay is exacerbated by the airlines’ increasing reliance on frequent-scheduling with small planes, rather than less frequent scheduling with larger planes. Obviously, the frequent nonstop flights do give a consumer advantage. (I love being able to fly directly city to city.) But they seem to misprice/misallocate costs, since 30-seat jets use the same amount of atc resources (and take up pretty much the same landing slot unit) as 200-seat jets, but benefit only a tiny fraction of the number of people.

  5. dcuser, it isn’t a particularly deep insight to say fewer flights would mean less air traffic congestion. Fewer cars would mean less freeway traffic as well.

    Congestion pricing would limit the number of flights at the most heavily taxed times, at least if the tax is high enough. And this may well make sense.

    But I think it’s a bit of a cop-put to treat the air traffic control system itself as given, not subject to improvement or alteration.

    Continued economic growth will likely mean that any reduction in flights by changing the incentives to fly small planes vs large will be short-lived. We might get fewer flights on larger planes initially, but we’ll eventually have the same or more flights all on large planes simply through continued passenger growth.

    So we need to think about pushing the boundaries of the constraints like how we do air traffic control.

  6. One reason for delays to keep in mind is that airlines use lots of smaller planes to fill up schedules at major airports, both to have a flight scheduled at any hour when a traveler wants to travel, and also to take up slots and gates that could be used by competitors.

    In contrast, in Japan, they use 747’s on crowded one hour flights.

    We could certainly change our ticket tax system to incentivize the use of larger planes, i.e., tax by the plane, not the passenger.

Leave a Reply

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