Demand models for U.S. domestic air passenger markets

Abstract

The airline industry in recent years has suffered from the adverse effects of top level planning decisions based upon inaccurate demand forecasts. The air carriers have recognized the immediate need to develop their forecasting abilities and have applied considerable talent to this area. However, their forecasting methodologies still are far below the level of sophistication of their other planning tools. The purpose of this thesis is to develop a set of demand models which are sufficiently sensitive to measure the effects upon demand of policy decisions with respect to such variables as fare and technological and quality of service factors. A brief overview of transportation demand theory and a survey of recently published research in air passenger demand modeling are presented. Following these is a discussion of the economic nature of domestic air transportation passenger service indicating the demand and service attributes and how they interact in equilibrium. Based upon this background information a multi-equation econometric model is developed. The model is calibrated over subsets of a base of historical data from 180 markets over a six year time frame. The subsets are cross classifications of markets with respect to length of haul and market size. Recently developed techniques in model sensitivity analysis are applied to ensure statistical robustness, and principal components regression is employed to combat the problem of multicollinearity. Numerical examples of applications of the model are provided. The results indicate that the model performs very well in the analysis of long and medium haul markets. It is particularly effective in the higher density markets. The model is not equipped to account for the impacts upon air transportation passenger demand of competing modes, and therefore does not perform well in the analysis of short haul (less than 400 miles) markets.