2. This leaves seven routes that are identified as "unprofitable. These are Phoenix-Palm Springs, Phoenix-San Diego, Phoenix-Tucson, Tucson-El Paso, Tucson-Kansas City, Tucson-Palm Springs, and TucsonSan Diego. I have directed Roger Mallet to examine the pattern of air service that these routes received during July 1973, the mid-point of the year the ATA chose to simulate. Appendix A lists all flights in both directions together with the flight itineraries. As you can see, with only one exception (American Airlines Flight 374 from Tucson to San Diego), all flights offered during July 1973 were segments of longer flights. They are thus, subject to the cost and revenue allocation problems I mentioned in my memo of April 29. This means that any indication by the model of "unprofitability" needs to be examined very carefully. Let us focus particularly on the Tucson-Phoenix segment. During July 1973 there were 18 flights per day between Phoenix and Tucson and 17 flights per day between Tucson and Phoenix. (These totals do not include four daily nonstop flights in each direction by an unsubsidized commuter carrier which has total freedom to enter and exit the market at will.) All of those 35 daily flights were offered as segments of longer flights. In many cases, these flights clearly were "feed" flights. For example, American Airlines Flight 177, BostonNew York/JFK-Phoenix-Tucson is clearly operated as a means of providing single-plane service to both cities from the East Coast. It is no more run by American solely because of its Phoenix-Tucson segment than it is because of its Boston-New York segment (which I am certain is "unprofitable" judged by itself). or This example reflects the fact that airlines do not schedule flights on a city-pair basis. They are scheduled on a total flight basis. That is, what the scheduler tries to achieve is a situation in which the revenues attributable to a particular flight equal exceed the costs attributable to that flight. In deciding whether or not to include a particular segment (e.g., Tucson-El Paso segment on American's Flight 116 with an itinerary consisting of San Francisco-PhoenixTucson-El Paso-Chicago-Newark), the scheduler decides 3 what additional traffic he can pick up by including the segment (he doesn't particularly care whether this is local Tucson-El Paso traffic or whether it travels beyond El Paso it is the revenue he is concerned about) and what additional costs American will incur by serving the segment. This additional cost may or may not be related to what the CAB terms "direct operating expense." "Direct operating expense" is merely an accounting convention. If attributable revenues exceed attributable costs, the segment is included. If not, it is deleted. Deleting such a segment on a given flight would only require CAB approval if it represented the carrier's only service at either of the two cities under authority of a given "route segment. It is important to note that CAB approval often is not required either to add the first nonstop or delete the last nonstop in a given city-pair market. For example, nothing in American Airlines' certificate currently compels it to offer nonstop service in any of the Arizona city-pairs mentioned in the ATA simulation. Indeed, during July 1973, American eliminated, without seeking CAB approval, the Palm Springs segments from Flights 159, 221, and 622, thus ending nonstop service by American between Palm Springs, on the one hand, and Phoenix and Tucson, on the other. To see how widespread this carrier authority to add and delete service currently is, I had Mike Roach check the certificates of the carriers serving the Arizona markets. He informs me that, perhaps with the single exception of Frontier's Phoenix-Tucson service, the carriers involved are currently free to terminate service over all the indicated segments. In other words, American, under its current certificate, has no obligation to provide service in any of the six city-pairs where it currently is certificated (it does not hold a certificate allowing it to fly between Kansas City and Tucson). This brings me to my final point service reductions on segments that the ATA model predicts would be retained. As you can see, in many cases, the hypothetical reductions indeed would be massive. ChicagoPhoenix service would be cut from 18 to 8 flights per day. Again it is vital to be clear about what is and is not being simulated by the ATA model. As I indicated in my April 29 memo, the ATA model simulates the response of a monopolist. I am not surprised that a monopolist, if free to drop all flights he didn't wish to offer, and protected from entry by other carriers, would choose substantially to curtail service. He might even drop the amount of service that the ATA model indicates. However, the current air transport system is not a monopoly. It is a competitive system where carriers face controlled prices, are free to determine the amount of service they wish to provide at these controlled prices, and are even free to exit from many city-pairs without explicit permission of the CAB if they find their operations unprofitable. The fact that they currently choose to operate large numbers of flights and routes which the simulation indicates are "unprofitable" indicates not that there is crosssubsidization in the system, but that the simulation is faulty in the way it determines whether a particular route or flight if "profitable" or "unprofitable." |