FEASIBILITY, SUSTAINABILITY, SUBADDITIVITY, NONCONVEXITY, AND SPECIAL PLEADING.

Wall Street Journal commentator Holman Jenkins sees doom for the air carriers.

Smart businesses ask themselves how they can make themselves more popular with customers by "unbundling" the goods or services they provide. Burger joints charge extra for cheese, but not usually for lettuce and tomato. Yet fairness is surely served by letting haters of lettuce and tomato escape the burden of subsidizing those who love them?

Provided the self-selection constraint is sufficiently cheap to implement. Consider McDonald's. Their "Speedee Service System" depended on consumer willingness to purchase a standard product, and Burger King, early in the 1970s, teased them with "Hold the pickle, hold the lettuce, special orders don't upset us." I don't recall if the cheeseburger was part of the original McDonald's lineup: in the early 1960s it was 18 cents against the hamburger's 15 cents, and the cook staff had to learn the mix of orders that would not leave people waiting more than 30 seconds for their hamburger or with cheeseburgers that would have to be thrown away after going unsold for 15 minutes. Bundling can be a way of exploiting cost complementarities (this appears to be the case for packet services carrying telephone, television, and computer signals, and such cost complementarities may be present in fast food). It can also be a way of segmenting markets according to ability to pay, something the airlines have been doing for years, and something that Mr Jenkins fails to understand.

Carriers are slicing away at services to create options for the most price-sensitive fliers to avoid services they don't want to pay for: checked luggage, meals, pillows and blankets.

To some extent, they've always done this: how else describe three-class cabins? Here, however, the problem is not one of bundling, it is one of pricing each component of the service in order to encourage proper self-selection by passengers -- a dodge that I think Jules Dupuit discovered was the railroads making third class miserable in order that none of the intended first class passengers would ever ride at the much lower rate -- while ensuring a total revenue high enough to make the operation of the service worthwhile.

To keep fares low, airlines also skimp on backup planes and empty seats -- even if this makes it harder to get home if your flight is cancelled.

D'oh! How much more are the time-sensitive travelers willing to pay to be sure a spare is available? There's an upper bound on what an airline can charge, as there are hungry charter operators out there. (The best way to make a small fortune in aviation is to start with a large fortune.)

Even so, it's far from enough as fuel costs price the flying public out of the air. Herb Kelleher was speaking for Southwest when he observed a basic economic reality: "Our only competition is the car or the television set. People who fly us are people who weren't going to fly in any case." He might have been speaking for all airlines. Their fleets and networks today are designed to fly thousands of people who don't have to fly. If ticket prices fully reflected current fuel prices, by common estimate the industry would have to be 20% smaller.

A supply curve shifts to the left. There is excess demand at the old price, but (barring annoying discontinuities) there will be a new equilibrium, involving a higher price and a smaller output. Some of Southwest's passengers might well be the marginal consumers, buying at the old price but substituting to something else at the new higher price. Contrary to Mr Jenkins's lead, without the supposed subsidy (it's too late in the evening for me to spell out all the conditions for subsidies, which involve an excursion into something called stand-alone costs) from the price-sensitive coach passengers, there's no plane for the inframarginal passengers (there being no such thing as people who "have to fly" although there can be people who are willing to pay more than the cut-rate price to get on the plane.) The inframarginal passengers, who in the absence of the competition for the marginal passengers, would harvest a smaller consumer surplus, are not happy with their inframarginal status being recognized by the air carriers.

No wonder there is panic at the Business Travel Coalition, representing corporate travelers who need the discretionary grannies and backpackers to sustain a system of frequent connections even to smaller cities. "Brand name legacy carriers that we and American communities from coast-to-coast have depended upon for decades to provide us with affordable, frequent air service are running out of cash, and therefore, toward a date with bankruptcy, and even liquidation," the group warned last week.

Boo hoo. Their website has a typically vapid mission statement.

Founded in 1994, the mission of Business Travel Coalition is to bring transparency to industry and government policies and practices so that customers can influence issues of strategic importance to their organizations.

I see nothing on the main site about Passenger Rail, although I smell a quest for corporate welfare, in the form of more road and airport construction. Perhaps there are some nonconvexities in the provision of air service. Fewer passengers on fewer (smaller?) planes at lower frequencies are the logical outcome. Mr Jenkins sees an intriguing replay of Penn Central, with one amusing twist.

Because even in the unlikely development that regulators would welcome capacity-shrinking mergers, the immediate costs and disruption would likely bankrupt the combined carriers before any benefits materialize. Continental and US Airways, looking at United as a potential life raft, looked closely and now see a ticket straight to Chapter 11.

Let's also dispense with any idea that Southwest, which continues to be profitable thanks to prescient fuel-price hedging, offers an answer. Southwest, though not much of a "low-fare" carrier anymore, sticks to its business model of mostly flying point to point.

Of the 50,000 city pairs served by the U.S. airline industry, only 3% can sustain such direct service. If you want to fly from most places in America to most places in America, you need your own plane -- or a legacy carrier's hub and spoke system.

That is, if there are any benefits to mergers. A combination of suboptimal-sized weak companies tends to create a suboptimal portfolio of capacity in a weak company: Penn Central, LTV Steel, Stroh's Beer. Then there's that hub-and-spoke reference. The rigamarole of getting cleared to land, landing, taxiing to the gate, unloading passengers, loading passengers (there's a reason air carriers have a "boarding process"), taxiing to the runway, getting cleared to take off, and heading to the next airport renders a multi-stop flight a losing proposition (in the time it takes to carry out all those steps, a Wolverine is nearly to Kalamazoo, even without benefit of the 110 mph track, and the Wolverine will unload and load through more than one door.) On the other hand, many of the speed advantages of direct flights from regional airport to hub are dissipated in the delays of changing planes at the hub.