Impact of omitted variable and simultaneous estimation endogeneity in choice-based revenue management systems

Abstract

The demand models used in revenue management systems applied to air transport are affected by endogeneity. This effect is particularly important in the Discrete Choice Models used in Choice Based Revenue Management (CBRM). Still, the sources and the magnitude of the problem have not been completely studied. In this study, we use simulations to analyze the effect of endogeneity in CBRM air transport applications. We replicate different conditions that generate endogeneity described in the literature: the omission of explanatory variables in the behavioral model and the simultaneous determination between fares and demand. On the one hand, results show that simultaneous determination endogeneity is not a concern if the behavioral model includes complete information of the process, without omitting variables. On the other hand, endogeneity affects the estimation and application of CBRM when variables are omitted, a situation that worsens when the omission of variables is combined with the simultaneous determination between fares and demand. Based on these results, we discuss some practical implications and derive some recommendations for the use of CBRM in air transport.