Measuring Competition in the U.S. Airline Industry Using the Rosse-Panzar Test and Cross-Sectional Regression Analyses

We employ the Rosse-Panzar test to assess market performance in selected airport-pairs originating from Atlanta. The Rosse-Panzar test stands in the tradition of the New Empirical Industrial Organization. It is based on the comparative statics of a reduced form revenue equation. Therefore, it is less powerful than structural models, but it offers the advantage of less stringent data requirements and reduces the risk of model misspecifications. The test statistic allows us in most airport-pairs to reject both conducts consistent with the Bertrand outcome, which is equivalent to perfect competition, and the collusive outcome, which is equivalent to joint profit-maximization. Rather, the test statistic suggests that behavior is consistent with a range of intermediate outcomes between the two extremes, including, but not limited to the Cournot oligopoly. In the second part of the paper, a cross-section pricing regression complements the Rosse-Panzar test. It shows that the presence of low-cost competition in an airport-pair reduces the average fare significantly.