Abstract
This paper aims to evaluate environmental efficiency of transportation systems by integrating a multiple-output stochastic frontier analysis (SFA) model with a structural equation model (SEM) to flexibly incorporate cause-effect relationships among various factors influencing traffic emissions. The SEM is firstly specified to illustrate such complex interrelations and the calculated latent variables are taken as the inputs to SFA model. A SFA model, named the distance function approach which can accommodate multiple inputs and multiple outputs, is adopted to calculate efficiency scores in transport sectors at different cities. To examine the model performance, an empirical study is carried out considering three types of emissions, i.e., CO, VHC and NOx, based on the Millennium Cities Database. The effectiveness of the proposed model is confirmed. Efficiencies in transport sectors are compared and it is found that cities in developed countries do not consistently show higher efficiency than other developing cities.
