Navigating Efficiency: Insights from One-Stage and Two-Stage DEA Modeling in the Airline Industry

Data envelopment analysis (DEA) has widely been used for airline efficiencies. Among the studies, two common approaches are considered. One approach assumes an airline as one system of inputs and outputs. The other approach, which is more preferable in the literature, assumes an airline as a two-stage network (2SN) system of inputs, intermediates, and outputs, where intermediates are outputs of the first stage and inputs of the second stage. There are also two different views when an airline is introduced as a 2SN system. The pre-view of a 2SN system relies on unobserved units whereas the post-view relies on the observed units only. In this study, we discuss the performance of eight US airlines for two decades (2001–2021) using these three approaches. We suggest the post-view of the 2SN system to represent US airlines; nevertheless, we investigate whether assuming an airline as one system or as a 2SN system with both pre- and post-views results in meaningful differences. We use a windows analysis approach to measure the overall average performance of each airline over time. Furthermore, we examine the impact of varying window lengths on both one-stage and two-stage DEA to determine if the length of the windows affects the overall efficiency score. Statistical findings highlight remarkable consistency among the three mentioned approaches. When examining the effects of windows lengths on the performance of airlines, one-stage DEA consistently maintains its efficiency scores across variations. Two-stage DEA is fairly consistent, with minor fluctuations in efficiency scores across various windows lengths.