Improving the discrimination power through compromise solution approach: Malaysian banks as illustration

The conventional DEA gives each DMU an extreme flexibility in selecting its own weights in order to get its optimal efficiency score, which may result in a relatively high number of efficient DMUs, and prevents DEA from being a robust approach in determining the most efficient unit. Moreover, this f...

Full description

Bibliographic Details
Main Authors: Hanafi, Hocine, Larbani, Moussa
Format: Conference or Workshop Item
Language:English
English
Published: 2014
Subjects:
Online Access:http://irep.iium.edu.my/42113/
http://irep.iium.edu.my/42113/
http://irep.iium.edu.my/42113/2/improving_discrimination_compiled.pdf
http://irep.iium.edu.my/42113/3/DEA2014-P253.pdf
Description
Summary:The conventional DEA gives each DMU an extreme flexibility in selecting its own weights in order to get its optimal efficiency score, which may result in a relatively high number of efficient DMUs, and prevents DEA from being a robust approach in determining the most efficient unit. Moreover, this flexibility hampers a common base for comparison and leads to a weak discrimination power among all DMUs as well as unrealistic weights allocation among inputs and outputs for all or part of DMUs. In this paper we show that these drawbacks can be satisfactorily overcome by compromise solution approach (Kao and Hung, 2005). We extend the Kao and Hung’s approach to the BCC model and scale efficiency computation. To compare the proposed models with the conventional DEA models, a sample of 30 Malaysian commercial banks is examined for the period 2006 to 2011, in terms of efficiency scores, ranking and returns to scale.