Overbooking, yield management and demand forecast: an empirical study of the arrivals in airline reservation systems

The practice of overbooking in air transportation is a result of a broader market strategy, known as Yield Management (Revenue Management), which aims at optimizing the total revenues of airlines. Given this complex environment for decision making, one has that the investment in improving demand forecast capabilities, and more specifically, in understanding the process of arrival of requests in Reservation Systems, is potentially competitive advantage enhancing and ultimately improves market performance. The present paper aims at developing an empirical model of the arrivals of requests in airlines' (and, more generally, service firms') Reservation Systems. By making use of a database consisting of actual information collected from an airline operating the Rio de Janeiro - São Paulo shuttle route, here we make use of parametric estimation of a model of reservation arrivals. The time series modeling implemented here can be regarded as within a theoretical framework for demand forecast. Main results are indicative of an autoregressive process of arrivals in which lagged variables along with differences have high predictive power of current arrival levels. This paper is therefore able to fill a gap found in the literature regarding forecast methods applied to corporate Reservation Systems.