Operating efficiency in Chinese universities : an extended two-stage network DEA approach

Operating efficiency of universities is widely concerned by the education community. As a non-parametric method for efficiently handling multiple inputs and outputs, data envelopment analysis (DEA) is often used for measuring the operating efficiency. However, shared input resources are often ignored in the existing DEA studies. In order to remedy the shortcoming with a focus on teaching and research processes of universities, this paper adopts an extended two-stage network DEA approach to measure the operating efficiency of 52 universities in China using a data set in 2014. The main findings show that: (1) Among the operating efficiency of 52 universities, about one third and two thirds of universities are efficient and inefficient, respectively. It may reflect some problems such as inefficient use of resources or unsatisfactory outcomes for these inefficient universities. By giving first priority to universities’ teaching or research process, we provide alternative ways for teaching-oriented or research-oriented universities to benchmark and improve their performance. (2) For the heterogeneity efficiency analysis of different universities, the operating efficiency of "non-985" universities are significantly higher than that of "985" universities, while there is only a small difference on the operating efficiency between comprehensive universities and science & engineering universities. Although the efficiency of the central and western universities is slightly better than that of the eastern universities in terms of the average efficiency, there is no significant efficiency difference among the eastern, central, and western regions statistically. Hence, to improve the operating efficiency of Chinese universities, the Chinese government should improve the financial allocation mechanism and introduce successful budget performance management. For the Chinese universities, they should formulate teaching and scientific research plans according to their own research needs and development goals.