A frontier-based hierarchical clustering for airport efficiency benchmarking

Abstract

Purpose: When large samples are used to estimate airport efficiency, clustering is a necessary step before carrying out any benchmarking analysis. However, the existing literature has paid little attention to developing a robust methodology for airport classification, instead relying on ad hoc techniques. In order to address this issue, this paper aims to develop a new airport clustering procedure. Design/methodology/approach: A frontier-based hierarchical clustering procedure is developed. An application to cost-efficiency benchmarking is presented using the cost function parameters available in the literature. A cross-section of worldwide airports is clustered according to the relevant outputs and input prices, with cost elasticities and factor shares serving as optimal variable weights. Findings: The authors found 17 distinct airport clusters without any ad hoc input. Factors like the use of larger aircraft or the dominance of low-cost carriers are shown to improve cost performance in the airport industry. Practical implications: The proposed method allows for a more precise identification of the efficiency benchmarks, which are characterized by a set of cophenetic distances to their "peers". Furthermore, the resulting classification can also be used to benchmark other indicators linked to airport costs, such as aeronautical charges or service quality. Originality/value: This paper contributed to airport clustering by providing the first discussion and application of optimal variable weighting. In regard to efficiency benchmarking, the paper aims to overcome the limitations of previous papers by defining a method that is not dependent on performance, but on technology, and that can be easily adapted to large airport datasets.