The purpose of this note is to provide background to the study of the airline industry by briefly discussing four important economic aspects of the industry: (1) the nature and measurement of airline costs; (2) economies of scope and hub-and-spoke networks; (3) the relationship between yields and market characteristics; and (4) the S-curve effect. The Appendix to this note contains a glossary of key terms used throughout the discussion.
Airline costs fall into three broad categories:
flight sensitive costs which vary with the number of flights the airline offers. These include the costs associated with crews, aircraft servicing, and fuel. Once the airline sets its schedule, these costs are fixed. traffic-sensitive costs which vary with the number of passengers. These include the costs associated with items such as ticketing agents and food. Airlines plan their expenditures on these items in anticipation of the level of traffic, but in the short run, these costs are also fixed. fixed overhead costs which include general and administrative expenses, costs associated with marketing and advertising, and interest expenses.
The largest category of costs is flight-sensitive. An important point about an airline’s cost structure, and a key to understanding the nature of competition in the industry, is that once an airline has set its schedule, nearly all of its costs are fixed and thus cannot be avoided. Because it is better to generate cash flow to cover some fixed costs, as opposed to none at all, an airline will be willing to fly passengers at prices far below its average total cost. This implies that the incidence of price wars during periods of low demand is likely to be greater in this industry than in most.
There are two alternative measures of an airline’s average (or, equivalently, unit) costs:
cost per available seat mile (ASM)
cost per revenue passenger mile (RPM)
Cost per ASM is an airline’s operating costs divided by the total number of seat-miles it flies. (An available seat mile is one seat flown one mile.) It is essentially the cost per unit of capacity. Cost per RPM is the airline’s operating costs divided by the number of revenue-passenger miles it flies. (A revenue passenger mile is one passenger flown one mile.) It is essentially the cost per unit of actual output. These two measures are related by the formula:
Cost per RPM = cost per ASM ( load factor
where load factor is the fraction of seats an airline fills on its flights. In the end, it is cost per RPM that an airline must worry about, for it must cover its cost per RPM to make a profit.
Airlines differ greatly in both their costs per ASM and costs per RPM. For example, in 1992 Southwest had a cost per ASM of 7.00 cents, while USAir had a cost per ASM of 10.90 cents. Similarly, Delta had a cost per RPM of 15.33 cents while American had a cost per RPM of 13.81.
Differences across airlines in cost per ASM reflect differences in:
1) average length of flights (cost per ASM declines with distance).
2) fleet composition (cost per ASM is smaller with bigger planes).
3) input prices, especially wage rates.
4) input productivity, especially labor.
5) overall operating efficiency.
Differences across airlines in cost per RPM reflect differences in cost per
ASM plus differences in load factor. Two airlines might have very similar costs per ASM, but quite different costs per RPM because of differences in load factor. For example, in 1992 USAir and United’s cost per ASM differed by less than 2 cents (USAir 10.90, United 9.30), but their costs per RPM differed by nearly 5 cents (USAir 18.54, United 13.80) because of USAir’s lower overall load factor (USAir .59, United .67)
Economies of Scope and Hub-and-Spoke Networks
Economies of scope play an important role in shaping the structure of the U.S. airline industry. The source of economies of scope in the airline industry is the hub-and-spoke network. In hub-and-spoke network, an airline flies passengers from a set of “spoke” cities through a central “hub,” where passengers then change planes and fly from the hub to their outbound destinations. Thus, a passenger traveling from, say, Omaha to Louisville on American Airlines would board an American flight from Omaha to Chicago, change planes, and then fly from Chicago to Louisville.
In general, economies of scope occur when a multiproduct firm can produce given quantities of products at a lower total cost than the total cost of producing these same quantities in separate firms. If “quantity” can be aggregated into a common measure, this definition is equivalent to saying that a firm producing many products will have a lower average cost than a firm producing just a few products. In the airline industry, it makes economic sense to think about individual origin-destination pairs (e.g., St. Louis to New Orleans, St. Louis to Houston, etc.) as distinct products. Viewed in this way, economies of scope would exist if an airline’s cost per RPM is lower the more origin-destination pairs its serves.
To understand how hub-and-spoke networks give rise to economies of scope, it is first necessary to explain economies of density. Economies of density are essentially economies of scale along a given route, i.e., reductions in average cost as traffic volume on the route increases. Economies of density occur because of two factors: (1) spreading flight sensitive fixed costs and (2) economies of aircraft size. As an airline’s traffic volume increases, it can fill a larger fraction of seats on a given type of aircraft and thus increase its load factor. The airline’s total costs increase only slightly as it carries more passengers because traffic-sensitive costs are small in relation to flight-sensitive fixed costs.
As a result, the airline’s cost per RPM falls as flight-sensitive fixed costs are spread over a larger traffic volume. As traffic volume on the route gets even larger, it becomes worthwhile to substitute larger aircraft (e.g., 300 seat Boeing 767s) for smaller aircraft (e.g., 150 seat Boeing 737s). A key aspect of this substitution is that the 300 seat aircraft flown a given distance at a given load factor is less than twice as costly as the 150 seat aircraft flown the same distance at the same load factor. The reason is that doubling the number of seats and passengers on a plane does not require doubling the number of pilots or flight attendants or the amount of fuel.
Economies of scope emerge from the interplay of economies of density and the properties of a hub-and-spoke network. To see how, consider an origin-destination pair – say, Indianapolis to Chicago – with a modest amount of traffic. An airline serving only this route would use small planes, and even then, would probably operate with a low load factor. But now consider an airline serving a hub-and-spoke network, with the hub at Chicago. If this airline offered flights between Indianapolis and Chicago, it would not only draw passengers who want to travel from Indianapolis to Chicago, but it would also draw passengers from traveling from Indianapolis to all other points accessible from Chicago in the network (e.g., Los Angeles or San Francisco). An airline that includes the Indianapolis-Chicago route as part of a larger hub-and-spoke network can operate larger aircraft at higher load factors than an airline serving only Indianapolis-Chicago.
As a result, it can benefit from economies of density to achieve a lower cost per RPM along the Indianapolis-Chicago route. In addition, the traffic between Indianapolis and the other spoke cities that will fly through Chicago will increase load factors and lower costs per RPM on all of the spoke routes in the network. The overall effect: an airline that serves Indianapolis-Chicago as part of a hub-and-spoke network will have lower costs per RPM than an airline that only serves Indianapolis-Chicago. This is precisely what is meant by economies of scope.
Relation Between Airline Yields and Market Characteristics
An airline’s yield is the amount of revenue it collects per revenue passenger mile. It is essentially a measure of the average airline fares, adjusting for differences in distances between different origins and destinations. Airline yields are strongly affected by the characteristics of the particular origin-destination market being served. In particular, there are two important relationships: Shorter distance markets (e.g., New York-Pittsburgh) tend to have higher yields than longer distance markets (e.g., New York-Denver). Controlling for differences in the number of competitors, flights between smaller markets tend to have higher yields than flights between larger markets.
The reasons for relationship 1) are summarized in Figure 1.
higher cost per RPMlower load factor
higher cost per RPM
Cost per ASM generally falls as distance increases. This is because, say, doubling trip mileage does not require doubling key inputs such as fuel or labor. Thus, shorter flights have higher cost per ASM than longer flights, and airlines must achieve higher yields to cover these higher costs. In addition, shorter distance flights generally have lower load factors than longer distance flights, which implies a higher cost per RPM for shorter distance flights, again requiring higher yields. Why are load factors lower for shorter flights?
The reasons has to do with the greater substitution possibilities that consumers have in short-distance markets (e.g., car of train travel are more viable options). In short –distance markets, we would therefore expect that some fraction of time-sensitive travelers (e.g., vacationers) would travel on these alternative modes, so short distance flights would have a higher proportion of time-sensitive travelers (e.g., business persons) than longer distance flights. Competitive pressures thus force airlines to offer more frequent flight schedules in short-distance markets, which leads to lower load factors.
The reason for relationship 2) has to do with the economies of density discussed earlier. Smaller markets will have lower traffic volumes, and airlines will generally operate smaller aircraft at lower load factors, increasing costs per RPM and yields.
The S-Curve Effect
The S-curve effect refers to a phenomenon whereby a dominant carrier’s market share (share of RPM) in a particular origin-destination market tends to be greater than the carrier’s share of capacity (share of ASM). Thus, for example, if United offers 70% of the seats flown between Denver and San Francisco, and Continental flies the remaining 30%, then the S-curve effect says that United’s share of the actual traffic in this market will be greater than 70% and Continental’s will be less than 30%. This translates into an S-shaped relationship between “share of capacity” and “market share,” as shown in Figure 2.
The S-curve effects stems from two sources. First, an airline with a greater share of capacity in a market is likely to have greater visibility in that market, so passengers are likely to contact it first. Second, an airline with a greater capacity share is likely to have more frequent – and thus more convenient – departures. This, too, works to boost its share of the actual traffic.
The S-curve phenomenon makes capacity an important competitive weapon in the rivalry among airlines. An airline with the financial resources to purchase aircraft and airport gates to achieve a dominant capacity share on key routes is likely to win the fight for market share. This suggests that, in general, it will be very difficult for a small carrier to challenge a dominant carrier at a hub airport, unless the small carrier can achieve significant cost advantages unrelated to scale. The history of competition in the post-deregulation airline industry seems to bear this out.