Capacity efficiency measurement using a three-stage DEA approach: evidence from domestic airports in Taiwan

<title>Abstract</title> This paper aims to modify the Stage 1 data envelopment analysis (DEA) of the three-stage DEA approach to allow for the fact that undesirable outputs like aircraft noise may not be freely disposable, and the efficiency measure of airport capacity use is modified to allow for an asymmetric treatment of desirable and undesirable outputs. A directional distance function is used to construct an output orientated DEA model with a constant returns to scale assumption. Stage 3 DEA is a repetition of Stage 1 DEA with observed outputs adjusting to account for the effects of variation in the operating environment, variation in statistical noise, and variation in technical-change obtained from stochastic frontier analysis (SFA) regression analysis in Stage 2. Results from the empirical analysis for the case of Taiwan's domestic airports strongly demonstrate that airport capacity efficiency performance turns out to be very sensitive to whether or not environmental impacts, stochastic noises, technical-change factor, and undesirable outputs are included.