Abstract: Based on the inter-provincial panel data of high-tech industries from 2009 to 2015, this paper. For the first time, this paper considers the environmental effects and random errors. Besides, it uses the SFA model to decompose and adjust the offset of input sample. Moreover, it combines the super-DEA technology to more effectively and truly reflect the innovation-driven development of high-tech industry and its spatial-temporal heterogeneity and source system. The results show that environmental factors have a significant heterogeneity effect on the performance of innovation-driven development of high-tech industry. On the whole, the adjusted innovation-driven performance of high-tech industry is still at a relatively low level from 2009 to 2015, and there is considerable room for development. The improvement of pure technical efficiency is significantly better than scale efficiency, which is the main source of impetus for the improvement of innovation-driven development performance. The performance shows a certain degree of growth in time, but the spatial difference is significant. It is urgent to lead industrial transformation of innovation-driven structure.

Key words: innovation-driven development performance, spatial-temporal heterogeneity, high-tech industry, environmental effects, random errors

摘要： 基于2009~2015年高技术产业省际面板数据，首次纳入高技术产业创新驱动的环境效应和随机误差考虑，利用SFA模型进行分解修正调整样本投入的偏移量，进一步结合超效率技术改进，更为有效真实地反映高技术产业创新驱动发展成效及其时空异质和来源体系配置特征。研究发现，环境因素对高技术产业创新驱动发展绩效产生显著异质影响。总体上，调整后2009~2015年高技术产业的创新驱动整体效率还处于相对较低水平，存在较大发展空间。纯技术效率改进显著优于规模效率，是其创新驱动发展绩效提升的主要动力来源。高技术产业创新驱动发展绩效时间上呈现一定程度的增长，而空间差异显著，亟待进行引领产业创新驱动结构转型。

关键词: 创新驱动发展绩效, 时空异质, 高技术产业, 环境效应, 随机误差