eThesis - School of Business electronic theses

The research in revenue management in airline industry is usually limited to problems that only involve a single airline. Due to the emergence and growth of airline alliances, it is becoming increasingly important to understand revenue management in alliance contexts. In this thesis, I discuss different revenue-sharing models in airline alliances and analyze and compare these models both theoretically and through numerical simulations. I also describe how economics commonly approaches this type of problems using mechanism design and auction theory and propose that there is no efficient mechanism that generally ensures incentive compatibility in the examined framework of dynamic deterministic single object auction under interdependent values.

Most of the revenue-sharing mechanisms that I discuss in this thesis do not adjust the revenue-sharing rule during time, but consider it static throughout the booking horizon. This type of mechanisms are popular in practical usage, but a significant limitation with any static revenue-sharing mechanism is that such a mechanism does not adjust revenue splits during time to reflect the actual value of the remaining seat inventory. Several models also assume complete information sharing among alliance partners or some other form of centralized decision-making. This contradicts the current practice of airline alliances where airlines individually maximize their own revenue given the revenue-sharing rule by individually controlling their capacity and booking processes.

To test the practical potential of the revenue-sharing mechanisms, I simulate the performances of the mechanisms in a small-scale numerical experiment under both static and dynamic capacity control policies. The two examined dynamic revenue-sharing mechanisms perform notably well in all the simulations, but require the airlines to truthfully share their true actual opportunity costs of capacity across booking horizon. Among static mechanisms, two fare-based revenue-sharing mechanisms almost consistently yield the best performance. Mileage-based revenue-sharing mechanism, the default policy of the International Air Transportation Association, yields significantly worse results than the other examined mechanisms in most simulations.

The observations in this thesis position the fare-based revenue-sharing mechanisms as appealing candidates to be pursued in real-life alliances. The use of dynamic revenue-sharing mechanisms could lead to additional revenue earning possibilities and these mechanisms are thereby worthy of further examination.