HubBub Bubbling Over Hub Study

<B> HubBub Bubbling Over Hub Study</B>

In "Researchers Call Hub Premium A Myth," (<I>BTN</I>, Oct. 11) Darryl Jenkins and Robert Gordon discussed the airport hub premium study funded by Northwest Airlines, which solely focused on its hubs. The study's central finding was that nonstop travelers to Detroit, for example, pay about 4 percent less than passengers connecting at Detroit en route to a final destination. The Business Travel Coalition has the following concerns with the study: 1) the source of funding; 2) the statistical methodology used; and 3) the requirement for independent expert access to the authors' data.

With Northwest providing funding for the study, how can the authors expect the results to have credibility with serious thinkers in the industry or policy makers in Washington? Some press referred to the study as an "independent analysis." With this exercise, are we resetting the standards for "rigorous independent academic evaluation?"

Gellman Research Associates, a research and consulting firm specializing in aviation economics, reviewed the Jenkins-Gordon study and has its own concerns. Frank Berardino, president at GRA, gave the following observations to an electronic newsgroup:

1) Only Northwest's fares are included in the analysis, so the full competitive alternatives available to consumers are never part of the comparison. Therefore, how do we know if the fares are relatively high or low?

2) The authors exclude revenues paid to commuter affiliates for connections via the hubs. Instead, they prorate revenues--how, we're never told. And importantly, the authors convert the trip to a hub nonstop, which may affect the connect market results.

3) Data from tables in the study indicate the average connecting fare is 6 percent higher than the average nonstop fare. However, the average connecting passenger goes 66 percent farther.

On Sept. 23, Aviation Daily quoted Jenkins as stating, "This is the most advanced and detailed analysis that anyone's ever been able to do about airline pricing and it dispels the myth that there is a hub premium." This statement defies logic because the authors did not compare fares at dominated hubs with those at non-dominated hubs.

To indirectly attempt to establish that fares for nonstop and connecting passengers at a dominated hub are somehow comparable proves little. The more direct approach is to compare fares at dominated hubs with fares at competitive airports as the following BTC testimony before the U.S. House Judiciary Committee on May 19, 1998, illustrates.

BTC stated, "U.S. DOT analysis, as well as other data, show airfare levels are linked to market share distribution and the presence of low-fare competitors. For example, Northwest has a 75 percent market share at Detroit and passengers pay 35 cents per mile in the 251 to 500 mileage block. In contrast, the number-one airline at Los Angeles, Seattle and Kansas City has a market share of 30 percent, 31 percent and 20 percent, respectively; passengers pay 19 cents per mile on average. Low-fare airlines' market share at these three competitive airports is 12.1 percent, 9.8 percent and 27.1 percent, respectively. At Detroit, low-fare competitor market share is 2.9 percent."

Another of the authors' statements, that results were "startling and surprising," would be a red flag for most economists that the study's premises and/or data were suspect. "The authors never report their regression analysis results, and their whole hub premium comparison is based on them," said Berardino.

Importantly, the results fly like arrows thick and fast at the conclusion of virtually every credible analysis regarding hub premiums since deregulation, including those by the GAO and DOT, and even by academics and organizations that have defended major airlines.

To my knowledge, Jenkins and Gordon have refused to release their data. The BTC urges them to share these data publicly so that study results may be independently replicated and validated.

<I>Kevin P. Mitchell
Chairman
The Business Travel Coalition
Lafayette Hills, Pa.