Advance purchase behaviors of air tickets

The advance purchase behaviors of air passengers are essential to develop revenue management strategies of airlines, which should be carefully studied. Based on this, this study aims to empirically investigate the advance purchase behaviors for airline tickets based on the airline transaction data of Taipei-Macau (TPEMFM) route in 2011. In order to model the advance purchase behaviors, multinomial logit models are used. To facilitate model development, the advance purchase horizon is divided into five time periods by three segmentation methods, including equal time periods, time periods with equal number of purchases and time periods according to professional judgment. Several factors contributing to advance purchase behaviors are examined, including price, flight schedule (time of day, day of week, and months of year) and fare class preferences. The estimation results show that the model with segmentation of equal time periods performs best in terms of adjusted rho-square and AIC indices. It is also found that air passengers tend to purchase tickets earlier for the flights in the morning and in hot season, suggesting the fare and seat inventory control should be varied for different flights.