Upper and lower bounds for non-radial efficiencies in data envelopment analysis: a vector approach

This paper presents a new approach to evaluate non-radial DEA efficiencies. This approach uses vector formulation to non-radial efficiencies. We propose the construction of an efficiency vector that leads to an index, which is an upper bound for the non-radial efficiency module. We also perform the evaluation of a lower limit for the same index. The developments here proposed are for DEA BCC model, as for the CCR model the efficiencies are the same independently from the projection region. The proposed score is applied in a numeric example.