A two-stage DEA model considering shared inputs and free intermediate measures

• Allocation of inputs and intermediate measures in efficiency evaluation is modeled.
• Overall efficiencies evaluated by conventional and proposed model are decomposed.
• Traditional model over-evaluates the efficiency of process with complex structure.
• Conventional two-stage model sub-evaluates the efficiency of the same process.
• Proposed model revise efficiency score by considering intermediate measures structure.

The internal structure of decision making unit (DMU) and the structure of system inputs become two key elements that guide the recent methodological developments of data envelopment analysis (DEA). But the allocation problem of intermediate measures does not arouse attention in this movement. This negligence will lead to biased efficiency measurement of overall system as well as individual stages. The paper proposes a two-stage DEA model considering simultaneously the structure of inputs and intermediate measures in efficiency evaluation and decomposition. The methodology extends previous work in which the intermediate measures play only a linking role in DMU efficiency forming. The study follows an empirical analysis which shows that the proposed model is able to revise the efficiency scores over-evaluated by traditional model and sub-evaluated by conventional two-stage model. The application results support the arguments that assignment of intermediate measure has strategic implications for decision makers and should therefore be fully examined from the expert and intelligent systems perspective.