April 21, 2008
Airline Service
The Decline of Airline Service-Becker
The increase in flight delays is just one aspect of a general decline in airline service. A few prominent examples of this decline are the elimination of meals in economy on domestic flights, more lost baggage, and dirtier seats and toilets. Posner gives several reasons why service has declined, with concentration on flight delays. I will add one reason to those he mentions that has been very important but is usually overlooked; namely, that the decline in service is an integral part of the substantial fall in the cost of airline travel after the airline industry was deregulated.
The crucial point is that the lower income and leisure flyers, and families with children that now make up the vast majority of customers on domestic routes, and a substantial fraction on international flights, prefer lower prices to better quality service and higher prices. On this interpretation it is no surprise that Southwest Airlines achieved remarkable success by offering low prices with no meals, and only non-reserved economy seats. The revealed preference of the majority of flyers has been that they prefer lower prices to meals and many other amenities that were once standard on flights in the United States. American Airlines and other more traditional airlines have been forced to react to the competition from the no frills airlines by, for example, eliminating meals, and sometimes eliminating first class seats and even reserved seats, so that they too can offer lower prices.
The same considerations explain the greater crowding on flights since airlines can offer much cheaper seats with average loads of 80 percent rather than say 50 percent. The reason is that the cost of handling the additional 30 percent of passengers is very low compared to the high cost of owning and operating planes, and compared even to the reduced revenue they receive from these passengers. Most persons who now fly prefer lower prices with more crowded planes and greater delays in boarding to the high prices and low occupancy rates that prevailed prior to deregulation.
The same trade off between price and service applies to the flight delays discussed by Posner. These delays are stochastic in that they vary from day to day according to different probability distributions of delay times. When delays persist, airlines are forced by the regulatory authorities to increase their scheduled flight times to recognize that typical travel times are longer. Airlines can cut down their delays by having greater amounts of deicing equipment during winter in the event of unexpected icing of planes, greater reserves of crews in case some crew members call in sick or get stranded, a larger number of backup aircraft in case some aircraft develop mechanical problems, and so on for other determinants of delays.
I am arguing that many passengers prefer the combination of low fares and greater delays on average to higher fares and fewer delays because it is cheaper to operate planes when inventories of people and equipment are smaller. Consider as an analogy a clothing store that has a large inventory of merchandise. It charges customers a lot for the suits and dresses it sells partly to help finance the convenience to customers of having many types of clothing to choose from. More generally, sellers often offer customers goods that have at least two dimensions: price and size of inventories. Prices generally are higher when inventories are greater so that delays in buying particular merchandise due to unexpected surges in demand are shorter because of the greater inventory of goods.
Of course, not all passengers are identical, and some of them prefer shorter delays and higher prices to the present situation. A number of airlines have been started in order to cater to these customers by providing exclusively first class or business class seats. These airlines have generally not been successful, probably because there are not enough of these customers to support commercial airlines that only offer first class service. Instead more affluent flyers have opted to buy shares in or outright ownership of private planes where service is far better and delays are much shorter. Private planes even have shorter delays from air traffic congestion because they frequently use less crowded airports.
Air traffic delays at major airports can be reduced through varying pricing of take offs and landings with time of day and the density of traffic, as discussed by Posner. Implementation of such sophisticated pricing is always difficult with publicly owned resources like airports and roads. So the privatization of airports could make possible greater flexibility in pricing the use of airports to take account of congestion.
The high fixed cost and low variable cost of airlines do not imply that airlines must be unprofitable in a competitive environment. Southwest and Jet Blue are two examples of airlines that have done quite well by offering low prices and limited service. Airlines made considerable profits in the latter part of the 1990s, but then the industry was hit first by 9/11, and then by the sharp run up in the cost of their fuel. Fuel costs have become more important to airlines than labor costs.
The hotel industry offers insights into the airline industry. This industry is more competitive than the airline industry since entry into the hotel industry is even easier than into the airline industry, and hotels also have high fixed costs of operations and relatively low variable cost of servicing individuals who occupy their rooms. Hotel prices adjust in the short run to the level of demand, but there are times when it is hard to get rooms, and people with reservations are denied rooms because of overbooking. Yet hotels have on average been quite profitable, which has stimulated various booms in the building of new hotels. To be sure, airlines have been exposed to major shocks, such as deregulation, terrorism attacks, and expensive fuel, but there is no intrinsic structural reason why the airline industry should be less profitable than the hotel industry.
Why Is Airline Service So Bad? Posner
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.
Delay is the main problem, and the one that I shall focus on. Many culprits have been named--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. But all these seem to me to miss the point. Persistent delay is usually the result of a failure to use price to equate demand and supply. When demand increases in advance of an increase in supply, failure to raise price results in buyers' incurring cost in the form of delay rather than in the form of a higher price. 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.
Some delay in the provision of services is unavoidable because of fluctuations in demand; it usually is wasteful to increase supply to the point at which every spike in demand can be accommodated without rationing (i.e., queuing, delay). But the persistent delays that airline passengers have been encountering for many years now cannot be explained by demand uncertainty. The delays impose enormous costs, particularly but not only on business travelers. The value of Americans' time is high.
So why are airline prices so low? The answer may lie in the lumpiness of airline service. (This was pointed out many years ago by the Chicago economist Lester Telser, and was repeated last week by Holman Jenkins in the Wall Street Journal.) The fixed costs of modern passenger aircraft are very high, but the marginal costs--the costs of carrying one more passenger if the plane is not full--are very low. At any price above marginal cost, the airline is better off selling a ticket than flying with the seat empty. Competition between airlines will therefore exert strong downward pressure on price. 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.
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.
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.
One possibility would be to allow the airlines to agree on minimum prices: in other words, to exempt the airlines from section 1 of the Sherman Act, which forbids competitors to agree on prices. The problem is that the airlines would fix a profit-maximizing minimum price, and it probably would exceed the price necessary to reduce congestion to the optimal level. Moreover, any increase in the price level would attract inefficient entry.
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. If a flight was canceled or delayed, it was usually easy to get a seat on another flight leaving soon. But with entry tightly limited, prices were above the competitive level; planes were not just uncrowded, they flew nearly empty. Prices have fallen sharply since deregulation. Competition has also led the airlines to adopt a variety of cost-saving measures. Pilots' wages are now much lower. Before deregulation, the powerful pilots' union (powerful because of the enormous costs of a work stoppage to a company that cannot produce for inventory and thus make up some of the revenue that it loses from a strike) was able to extract some of the airlines' regulation-enabled cartel profits, in the form of supracompetitive wages for pilots.
Another option would be to encourage, or at least place no antitrust or other obstacles in the way of, mergers between airlines. If there were only two airlines on every route, tacit collusion between them would probably keep prices high but not so high as if there were a single airline or an explicit price-fixing agreement. But any increase in prices would attract entry, pushing prices back down. Moreover, mergers often result in higher rather than lower costs.
A better alternative than any I have discussed thus far would be a heavy tax on airline transportation, with the tax rate varying according to the contribution of a particular route, time, or type of plane to congestion (for example, in general large planes would be taxed less heavily per passenger than small ones, because for a given number of passengers there are fewer big planes to clog the airways and runways than there would be small ones). To the extent effective, the tax would eliminate the deadweight cost of congestion.