Measuring and Explaining the Relative Efficiency of Container Terminals by Means of Bayesian Stochastic Frontier Models

Abstract

A terminal is efficient if it produces a maximum output (container traffic in TEU) for given inputs (terminal superstructure). Traditional studies on container terminal efficiency tend to focus on partial productivity measures such as TEU per crane. These instruments do not assess the overall efficiency of terminal operations, as they only look at specific aspects of the terminal operation process. This paper presents an approach for the measurement of container terminal efficiency based on Bayesian Stochastic Frontier modelling. A Bayesian Stochastic Frontier Model is a stochastic and parametric method for analysing productive efficiency. The stochastic production function takes into account statistical noise (composed error) and models the level of inefficiency via a non-negative disturbance. As such, the level of (in)efficiency of different terminals can be compared. The model is tested using a sample of 36 European container terminals, supplemented with four Asian container ports. In general, north European container terminals attain a slightly higher degree of efficiency than southern terminals. Terminals located in hub ports on average are considered as more efficient than those in feeder ports.