Abstract: In the implementation of the innovation-driven development strategy, technological distance is the premise that affects the rational choice of innovation models. In this paper, we use the DEA method to calculate the high-tech industry’s Malmquist index to represent technical distance. Then we choose the nonlinear panel threshold regression model and empirical analysis which could inspect technical distance nonlinear threshold character and study the internal correlation and influence mechanism of independent R&D and collaborative innovation on innovation performance improvement. As the threshold effect of technical distance, the following findings are established: (1) The impact of independent R&D on innovation performance has a triple-threshold relation, the trend is positive and sustaining rise before the third threshold and change to fall after that; (2) The impact of collaborative innovation on innovation performance has a double threshold effect, and the influence trend shows a transition from negative to positive as the technical distance decreases. When the technical distance is high, independent R&D should be the main approach to improve the performance of innovation. As the technology distance shrinks, it will be turned into an innovation model base on independent R&D and supplemented by collaborative innovation to improve the innovation performance in high-tech industries.

摘要： 在创新驱动发展战略的实施中，技术距离是影响创新模式合理选择的前提。本文选取非线性的面板门槛回归模型，采用DEA方法计算Malmquist指数代表高技术产业技术距离并将其非线性门槛性质引入该模型，研究自主研发和协同创新对创新绩效提升的内在关联和影响机理。实证结果表明，自主研发对创新绩效的影响存在以技术距离为门槛的三重门槛效应，影响趋势随着技术距离的降低（门槛值的升高）出现先升高后降低的正向效果；协同创新对创新绩效的影响存在以技术距离为门槛的双重门槛效应，且影响趋势随着技术距离的降低呈现出由负向到正向的转变。在技术距离较大时，应以自主研发作为创新绩效的提高的主要方式，随着技术距离的缩小转变为以自主研发为主、协同创新为辅的创新模式来提高高技术产业创新绩效。