<B> Researchers Call Hub Premium A Myth</B>
By Robert J. Gordon and Darryl Jenkins
There is little in aviation that has caused more consternation--and media attention--than the pricing policies of the major network airlines in their hubs. Despite the acknowledged benefits of airline deregulation over the past 20 years, enormous criticism has been leveled at the network carriers for fares they charge to residents of hub cities.
At issue is the perceived difference between the fares an airline charges residents at its hub cities, compared with the fares it charges on the rest of its system. This perceived fare differential has come to be known as a "hub premium."
The essence of the hub premium argument is this: Since a network carrier controls the major share of traffic in and out of a hub, it exploits its so-called "monopoly power" by charging hub-city residents higher fares than other passengers for travel to/from the hub.
By contrast, virtually everyone agrees that consumers who choose one-stop flights enjoy the full benefits of competition. If a passenger is traveling, say, from Newark to Los Angeles, and is willing to include a stop in the itinerary, that person has a choice of flying perhaps seven or eight different airlines. And because of the rich array of choices, passengers can be confident of receiving a competitive fare.
When the airline industry collectively lost $13 billion in the early '90s, the hub premium theory received scant attention. But when the industry returned to profitability several years ago, the theory was readily accepted as dogma--too readily. The problem was, no one had done a valid study using extensive proprietary data for a multi-year test period.
As industry researchers, we believed the hub premium theory was overblown. We concluded that if we could compare a major carrier's nonstop hub fares to its one-stop fares through its hubs, we could test the hub theory.
Because the premium is claimed to be particularly large in Minneapolis/St. Paul, where Northwest Airlines has 62 percent of the origin and destination traffic, Northwest was a likely candidate for the study. Accordingly, we approached Northwest and asked whether it would be willing to cooperate in an in-depth study by providing us with uncensored access to a complete data set of its prices for all three of its hubs--Minneapolis/St. Paul, Detroit and Memphis--for the years 1996 through 1998, and by defraying the costs of our work.
Northwest agreed. Of critical importance to us, Northwest agreed to give us not only complete access to its proprietary pricing data, but also complete academic freedom to analyze the data and formulate our conclusions without any direction or interference on its part.
After completing the initial portion of what we anticipate will be a longer, more detailed examination of the hub pricing policies of other major network airlines across the industry, we concluded:<ul><li>Using the complete database for fares and fare classifications for the Northwest Airlines domestic system for the years 1996-98, we are unable to find evidence of a hub premium. The hub premium theory is a myth. Instead, we find there is a "hub discount" of about 4 percent when length of trip and major fare categories are held constant.<li>Our finding of a modest hub discount is absolute. It makes no allowance for the lower quality of connecting service compared with nonstop travel.<li>Some specific fare categories do exhibit the presence of a slight hub premium, but this finding is canceled out by other categories that exhibit a slight hub discount. There is no Northwest hub pricing policy that makes its fares consistently higher than fares for connecting traffic of similar mileage.<li>Other researchers who have identified and measured a hub premium have actually identified and measured little more than unique buying patterns that exist in different cities; that is, the different mix of full-fare and excursion travel from one city versus another.</ul>
In addition, we find wholly invalid prior DOT and GAO studies that found the existence of a hub premium, because they rely upon "average fares." Average fares cannot by used for this purpose. They produce results that are severely skewed by the purchasing patterns of relatively few passengers.
Certainly as a result of our research, this method should be discontinued immediately as it misleads policy makers and the public as to what the real facts are and leads to incorrect and even harmful policy choices.
<I>Robert J. Gordon is an economist and the Stanley G. Harris Professor in the Social Sciences at Northwestern University. Darryl Jenkins is executive director of the Aviation Institute at George Washington University.