关键词: 高校创新效率, 核密度估计, 内外部影响因素

Abstract: Innovation is the first driving force for development and the strategic support in building a modern economic system. As one of the three basic research subjects, universities are not only the source of subversive and cutting-edge innovations, but also the key to transformation of technological innovation into growth mode innovation. And universities have extensive and far-reaching impact on industrial development, economic progress and production mode change. Therefore, university innovation is considered to be the engine of social progress and plays a key role in transforming the economic development model and overcoming the lack of growth. Under the rigid constraints of limited innovation resources, how to further improve the innovation efficiency of universities is an unavoidable problem in the process of building national innovation system. Firstly, taking R&D funds and R&D personnel as innovation input variables, and the number of scientific papers and patent applications as innovation output variables, this paper calculates university innovation efficiency of 31 provincial administrative units in China during the Twelfth Five-Year Plan period. During the sample period, and finds that the innovation efficiency of universities in China is not DEA effective as a whole, with an efficiency of 69.64%.The technical efficiency of the university innovation fluctuated greatly and increased first and then decreased, reaching the highest value in 2013.The technical efficiency of the university innovation fluctuated greatly and increased first and then decreased, reaching the highest value in 2013. The scale efficiency of innovation in universities is better than the technical efficiency. The highest value is 93.08%, the lowest value is 88.33%, and the fluctuation is not prominent. At the same time, the scale efficiency of Jiangsu, Henan, Ningxia and Xinjiang is close to 1. The pure technical efficiency of university innovation is low, the average value is only 76.64%, and the fluctuation is prominent. The overall increase is followed by the scale efficiency, indicating that there are more technical inefficiencies in the research activities of Chinese universities during the sample period, resulting in low efficiency of pure technology. Hunan, Jilin, Gansu, Guizhou, Fujian, and Inner Mongolia contributed the most to the improvement of pure technology efficiency. In addition, compared to 2013, up to 20 provinces' pure technical efficiency of innovation in universities declines in 2015.Therefore, the optimization of the structure of scientific research personnel and the rational allocation of funds for innovation activities need to be strengthened. Among them, Zhejiang and Chongqing have the largest declines. Therefore, the decline of pure technical efficiency is the direct cause for increase and decreases of the efficiency of innovative technologies in Chinese universities form some eastern and central regions. From the perspective of the eight comprehensive economic zones, the regional differences in innovation efficiency of universities are more significant, showing the highest distribution in the west and the lowest in the east. The universities in the Northwest Comprehensive Economic Zone have the highest innovation efficiency, which is 1.5 times that of the Northeast Comprehensive Economic Zone. Based on the factor input quantity and configuration structure considerations, the innovation efficiency of universities in Shanghai, Guangdong and other regions fluctuated greatly during the sample period. The innovation efficiency of universities in Heilongjiang, Beijing and Anhui was deteriorating, and the innovation efficiency of universities in Tianjin and Shanxi was significant. At the same time, the paper makes a detailed examination of the dynamic evolution trend of university innovation efficiency and decomposition index by density estimation. It is found that during the Twelfth Five-Year Plan period, the nuclear density distribution curve of university innovation technical efficiency in China has shifted to the right and presented a significant "two peak" trend, indicating that the innovation efficiency of some provinces is gathering at low values, and some of them are moving toward high values, but the overall trend of university innovation efficiency is optimized, and the provincial differences are getting smaller. The pure technical efficiency of Chinese university innovation and was “two peak” before 2013, indicating that the pure technical efficiency is bipolar. After 2013, the left-peak height of the innovative pure technical efficiency distribution curve of universities increased significantly and the width changed from flat to steep, indicating that the pure technical efficiency of university innovation has generally deteriorated during 2014-2015. In 2011-2013, scale efficiency nuclear density distribution curve of Chinese university innovation increased highly, and the width changed from flat to steep. This indicates that the overall scale efficiency of university innovation has not been significantly improved in this period, but the difference in scale efficiency between universities has decreased. The peak height of the nuclear density distribution curve drops again, and the width changes from steep to flat, indicating that the efficiency of scale deteriorates and the provincial differences of university innovation become larger.Finally, the paper uses R&D redundancy and R&D staff redundancy as the dependent variables, and uses the SFA model to examine the impact of regional innovation environment on university innovation efficiency from the perspective of the eight comprehensive economic zones. We find that economic environment, social and cultural environment, institutional environment are the main external causes affecting the innovation efficiency of universities. The imbalance of the number of elements and the configuration structure caused by the heterogeneity of innovation environment is the main external factor of the innovation elements’ redundancy in universities. In addition, the mismatch between the factors of the eastern university system, internal governance structure, performance evaluation method and achievement transformation mechanism are important internal factors for the redundancy of R&D personnel in the east. The repeated investment in R&D expenditures in the western universities and the unsound budget management has aggravated the redundancy of R&D funds in the west after the loss of talents. And then the paper proposes countermeasures to improve and optimize the innovation efficiency of universities.

Key words: innovation efficiency of universities, nuclear density estimation, internal and external influencing factors