A double frontier DEA cross efficiency method aggregated by evidential reasoning approach for measuring road safety performance

Abstract

Road traffic fatalities have long been a main concern of road safety experts in developing countries. Such countries annually lose a large number of their citizens due to road crashes. Globally, the Data Envelopment Analysis (DEA) has been successfully applied in road safety assessment as a powerful decision making method. It is noteworthy that all of the current studies optimistically measured the efficiency score of each Decision Making Unit (DMU) based on its position relative to only the efficient frontier. However, the efficiency value can be pessimistically determined on the basis of the anti-efficient frontier. A double-frontier model will lead to more realistic results. Meanwhile, many studies have recently implemented Cross-Efficiency Method (CEM) as a DEA extension for performance evaluation because of the unique ranking results without requiring unrealistic weight restrictions.In this regard, the present study aims to assess Iranian road safety performance by proposing a novel double-frontier CEM taking into account both optimistic and pessimistic points of view, simultaneously. Therefore, two cross-efficiency matrices are constructed by applying the standard and the inverted input-oriented CCR (Charnes, Cooper, Rhodes) models, respectively. Evidential Reasoning Approach (ERA) is then employed as a nonlinear method of combination to aggregate the transformed cross-efficiency scores to pieces of evidence. Finally, the findings confirm that the presented procedure results in a more realistic and comprehensive assessment, which is relatively in agreement with the evaluation results obtained considering both the efficient and anti-efficient frontiers.