A Dynamic Oligopoly Game of the US Airline Industry: Estimation and Policy Experiments

This paper studies the contribution of demand, costs, and strategic factors to the adoption of hub-and-spoke networks in the US airline industry. Our results are based on the estimation of a dynamic oligopoly game of network competition that incorporates three groups of factors that may explain hub-and-spoke networks: (1) travelers may value the services associated with the scale of operation of an airline in the hub airport; (2) operating costs and entry costs in a route may decline with the airline's scale of operation in the origin and destination airports (e.g., economies of scale and scope); and (3) a hub-and-spoke network may be an effective strategy to deter the entry of other carriers. We estimate the model using data from the Airline Origin and Destination Survey with information on quantities, prices, and entry and exit decisions for every airline company in the routes between the 55 largest US cities. As methodological contributions, we propose and apply a method to reduce the dimension of the state space in dynamic games, and a procedure to deal with the problem of multiple equilibria when using a estimated model to make counterfactual experiments. We find that the most important factor to explain the adoption of hub-and-spoke networks is that the cost of entry in a route declines importantly with the scale of operation of the airline in the airports of the route. For some of the larger carriers, strategic entry deterrence is the second most important factor to explain hub-and-spoke networks