Applying an improved three-stage DEA model to evaluate the innovation resource allocation efficiency in industrial enterprises

Yongfeng Zhu,Zilong Wang,Jie Yang
DOI: https://doi.org/10.1108/k-03-2022-0385
IF: 2.352
2022-09-13
Kybernetes
Abstract:Purpose The existing three-stage network Data Envelopment Analysis (DEA) models with shared input are self-assessment model that are prone to extreme efficiency scores in pursuit of decision-making units (DMUs) efficiency maximization. This study aims to solve the sorting failure problem of the three-stage network DEA model with shared input and applies the proposed model to evaluate innovation resource allocation efficiency of Chinese industrial enterprises. Design/methodology/approach A three-stage network cross-DEA model considering shared input is proposed by incorporating the cross-efficiency model into the three-stage network DEA model. An application of the proposed model in the innovation resource allocation of industrial enterprise is implemented in 30 provinces of China during 2015–2019. Findings The efficiency of DMU would be overestimated if the decision-maker preference is overlooked. Moreover, the innovation resource allocation performance of Chinese industrial enterprises had a different spatial distribution, with high in eastern and central China and low in western China. Eastern China was good at knowledge production and technology development but not good at commercial transformation. Northeast China performed well in technology development and commercial conversion but not in knowledge production. The central China did not perform well in terms of technology development. Originality/value A three-stage network DEA model with shared input is proposed for the first time, which makes up for the problem of sorting failure of the general three-stage network model.
computer science, cybernetics
What problem does this paper attempt to address?