Ability to Recover Full Costs through Price Discrimination in Deregulated Scheduled Air Transport Markets

Abstract

The regulatory reforms of domestic airline markets and regional markets such as the Europe Economic Area and the increased number of Open Skies agreements have led to a move away from the administrative structure of fare setting to one dominated by market forces. Within this latter framework the initial market power and price leadership exercised by former flag carriers and charter airlines is gradually being eroded by the emergence of low‐cost carriers among other factors. A major problem encountered in supplying pre‐committed scheduled services in a competitive market is that of full cost recovery. It has been argued, mainly from studies using aggregate data, that the ability of major carriers to recover fixed costs has deteriorated in deregulated markets as barometric price leadership is replacing that of dominant firm price leadership. This paper uses disaggregate data to examine the pattern of fares set by airlines as they sell seats up to the time of departure of a service. In particular, it analyses the impact of the entrance of low‐cost carriers into a small country—Portugal—focusing on the changes that are occurring in Lisbon and Porto airports. As such it looks at the way the price of a particular product—an airline service—varies in different market environments and whether price discrimination (yield‐management) is any longer a viable approach to full cost recovery.

Notes

1. When market entry and fare regulation were more prevalent, many countries clearly defined markets, e.g. in the 1960s the British Overseas Air Corporation had the monopoly of UK scheduled routes outside of Europe, British European Airways the monopoly in Europe, and a variety of carriers competed for charter business.

2. For details of the power and extent of price discrimination in airline markets in the initial aftermath of US deregulation, see Borenstein and Rose (1994).

3. The issue here is the fixed cost of providing a particular service rather than the more traditional one used in industrial organization theory that involves only the cost of fixed plant. One can use Doganis (2002) back‐of‐the‐envelope calculations to give guidance to their relative importance. Doganis argues that variable direct operating costs embody fuel costs, variable flight and cabin crew costs, airport and en‐route charges, and passenger service costs, such as handling fees paid to concessionaires, and depend on the type of aircraft used and the route being flown. These would be avoided if a flight were cancelled, along with some portion of averaged out engineering and maintenance costs, that in aggregate amount to something like 30–45% of total operating costs. Fixed direct operating costs cover depreciation and aircraft leasing costs and insurance as well as fixed flight and cabin crew costs. These come to 25–30% of total costs and are not avoided in the short‐term if a flight is cancelled. Finally, there are indirect operating costs. Some of these are route specific, which are escapable (about 5–10%), and others are not (25–35%). These costs include station and ground costs at the destination airport, and sales and advertising costs.

4. As Tirole (1988) says ‘What is worse than a monopoly? A chain of monopolies’.

5. Dresner (2004) provides an examination of this type of price leadership in the US market.

6. Dodgson et al. (1991) provide a comprehensive discussion of the various ways of defining predation and the difficulties of isolating predatory behaviour in practice.

7. Alamdari and Fagan (2005) look at the deviations and their implications for European carriers that deviate from the principles first adopted by Southwest.

8. This is a medium‐sized airport located centrally in England that has been a traditional facility for charter services to Europe and as a base for scheduled services offered by bmi. It has a significant catchment area and is on the main North–South motorway with good road access to the East of England.

9. Unlike most European carriers, Air Berlin offers both a frequent mileage programme on interconnecting services on a hub‐and‐spoke basis.

10. The fall in the fare for the 19 May flight after 26 May was, as ‘Noted’, due to the return flight being fully booked. The same pattern of fares is seen in the Air Berlin service between Porto and Frankfurt that is provided through Palma.

11. For example, taking an F‐test can be conducted for the null hypothesis: Then X is the said to Granger‐cause Y if some of the α’s are statistically different from zero. If it can be ascertained that α₀ is significant, there exists a contemporaneous causality between the variables. Furthermore, if when regressing X on its past values and current and past values of Y some or all of the current or past values of Y are significant, there exists feedback between the variables. Unidirectional causality exists when it can be shown that one variable Granger causes the other, but not the other way around. For implementation, it is necessary to estimate the restricted equation and compare sum‐of‐squared residuals of the restricted and unrestricted equations. When the test statistic is greater than the specified critical value, the null hypothesis that X does not Granger‐cause Y is rejected, i.e. X does Granger‐cause Y.

12. Only services where there are full data sets are considered; situations where data were spliced because a flight was full before take‐off are excluded from the considerations. Given the very large number of calculations conducted, and to make the presentation of material tractable, the set of parameters (e.g. F‐statistics) is not included here, but is available from the authors on request.