The impact of passenger mix on reported "hub premiums" in the U.S. airline industry

Southern Economic Journal, Oct, 2005 by Darin Lee, Maria Jose Luengo-Prado

Although Table 1 demonstrates that average prices to and from hubs tend to be higher than those at other large U.S. airports, studies such as Morrison and Winston (1995) have shown that a simple comparison of average fares across airports can be misleading because it fails to control for a number of factors known to impact fares, such as the proportion of passengers traveling nonstop (presumably a higher-quality product), the proportion of travelers choosing to purchase more flexible--and more expensive--unrestricted tickets, average distance, frequent flyer awards, and market density. Table 2 illustrates how some of these factors varied across the networks of the six largest hub-and-spoke airlines in 2000. For example, Table 2 shows that a considerably higher proportion of passengers flying to and from hubs fly nonstop compared to nonhub passengers. Likewise, on average, hub passengers tend to fly shorter distances than their nonhub counterparts. Table 2 is also consistent with the assertion made by airlines that a greater proportion of passengers traveling to and from hubs purchase premium tickets compared to passengers in markets neither originating nor terminating at an airline's hub (i.e., the system remainder).

One question that comes to mind is whether or not hub cities generate proportionally more business travelers or whether the higher proportion of premium tickets is simply the result of airlines using their sophisticated yield-management systems to make fewer discount seats available on flights to and from their hubs, thereby "forcing" significant numbers of passengers to purchase less restrictive--and significantly more expensive--premium tickets than they would otherwise want to purchase. For the purposes of our analysis, we believe that it is reasonable to treat the fare class a passenger travels in (i.e., restricted coach or premium) as exogenous. The primary reasoning behind our assumption is as follows. First, it is well understood that hub-and-spoke carriers have successfully created techniques to differentiate between business and leisure passengers (such as 14- or 21-day advance purchase requirements, ticket refundability, or Saturday-night stay requirements). Therefore, carriers have (especially in 2000, the year of our data) driven a large wedge between the average restricted coach fare and average premium fare in most markets. Indeed, in our sample of markets, the average premium fare is nearly three times as expensive as the average restricted coach fare. (12) The large gap that exists between restricted coach and premium fares implies--in our view--that few leisure passengers would be willing to purchase premium tickets even if no restricted coach fares were available. (13) Rather, passengers faced with such a large difference in fares would likely choose to fly on another carrier or take a connecting rather than a nonstop flight. Although we believe it is fairly unlikely that the "average" restricted coach passenger would be willing to pay the "average" premium fare rather than fly on another carrier or take a connecting flight, what about passengers at the top end of the restricted coach passenger distribution? Is it possible that they might be forced to purchase premium tickets as a result of inventory management? Again, we believe the data support our assumption of fare class exogeneity. This is because there is a fairly significant, discrete jump between the highest restricted coach fares and the lowest premium fares in most markets. In particular, we compute the ratio of the 20th percentile premium fare to the 80th percentile restricted coach fare for each of the markets in our dataset. The mean value of this ratio for the carriers in our dataset was: American (1.26), Continental (2.34), Delta (1.84), Northwest (1.97), United (1.57), and US Airways (1.24). This implies that, on average, passengers paying the lowest premium fare in a market are still paying 70% more than the passengers paying the highest restricted coach fare. (14)