Optimal Pricing of Flights and Passengers at Congested Airports : The Efficiency of Atomistic Charges

This paper investigates optimal airport pricing when airlines provide imperfect substitutes products, and make decisions on capacity, scheduling and pricing. We show that the first-best toll per flight may be higher than the simple market-shares formula that were recently derived for Cournot models, and approaches the atomistic toll (which ignores the airlines' internalization of self-imposed congestion) as products become closer substitutes. This increases the relevance of congestion pricing and does not require leadership behavior. We also find that an airport requires two pricing instruments to achieve the first-best outcome: per-passenger subsidies to counteract airlines' market power, and per-flight tolls to correct congestion externalities. We numerically analyze second-best policies of having only one tax instrument, as well as the performance of atomistic pricing, and find that the latter may offer a more attractive alternative than what is suggested by simpler Cournot models