Measuring the innovation efficiency of Chinese higher education institutions by network DEA

In this paper, we proposed network DEA models and used them to measure the relative efficiency of the innovative activities of 34 universities sponsored by Project 985 considering both 'knowledge production process' and 'knowledge commercialisation process'. The main results this study has obtained are as follows: firstly, the average technical efficiency was lower than the average scale efficiency. So technical inefficiency should be attracted more attention. Secondly, there were 21 universities which had increasing returns-to-scale. So the universities with increasing return-to-scale should increase innovative inputs to acquire better performance. Thirdly, there were 21 universities whose E-RAMI lower than E-RAMI, which mean that the universities' innovation efficiency will decrease with increasing R&D expenditure.