【作者】 王莹；

【导师】 胡汉辉；

【作者基本信息】 东南大学 ， 产业经济学， 2019， 硕士

【副题名】基于两阶段DEA模型及贝叶斯模型平均方法

【摘要】 经济转型与升级时期的中国,经济增长越来越依赖于技术创新。高技术产业作为知识经济时代中技术创新最为活跃的产业,已经成为转型时期中国经济增长的新动力和关键点,更是国际科技和经济竞争的重要高地。提高高技术产业创新效率从而提高产业竞争力,对于提升我国创新能力、调整产业结构、转变经济发展方式都将产生重大的影响。近年来,我国投入了大量的人力与资金并出台了一系列政策措施为高技术产业的发展保驾护航。《国家中长期科学和技术发展规划纲要（2006-2020年）》明确提出“以促进全社会科技资源高效配置和综合集成为重点^[1]”。因此,判断现实的效率水平,厘清影响创新效率的因素,以便有针对性的对资源配置进行调整优化,是新形势下提高我国高技术产业创新能力的十分重要且紧迫的任务。本文试图评价我国高技术产业的创新效率现状,通过厘清有关的影响因素,寻找改进我国高技术产业创新效率的主攻方向。具体而言,本文基于技术创新与效率评价理论,通过对国内外有关高技术产业、技术创新、效率分析等相关文献的梳理与总结,首先搭建了一个高技术产业创新效率研究框架,在此基础上建立追加中间投入的两阶段DEA模型;然后,分地区对我国高技术产业的创新效率进行测度与分析;其次,在同等效率水平下,运用贝叶斯模型平均方法对影响高技术产业创新效率的主要因素进行分析,探究造成效率差异的原因并探索改进的方向;最后,针对两阶段DEA模型和贝叶斯模型平均方法的分析结果,提出相应的政策建议。本文的主要工作及结论如下:首先以国内外理论与文献为基础,从高技术产业的概念与界定入手,对创新效率评价、贝叶斯模型平均方法及其应用研究两方面的相关文献进行梳理与分析,重点介绍了本文主要运用的DEA模型与贝叶斯模型平均方法的应用研究,由此奠定本文实证分析的理论基础。然后,本文开展了模型构建与实证研究。构建高技术产业两阶段效率分析框架,基于框架构建追加中间投入的两阶段高技术产业创新效率评价DEA模型,再运用该模型分地区对我国高技术产业的创新效率进行测度与分析。分析结果表明,我国高技术产业创新效率尚不十分理想,整体平均值仅为0.501,区域间的创新效率存在显著差异性;高技术产业技术研发阶段效率与成果转化阶段效率存在不协调的现象,自2008年开始,技术研发阶段的效率均高于成果转化阶段的效率,综合创新效率主要由成果转化阶段效率所主导;技术研发阶段与成果转化阶段脱节,我国大部分地区的高技术产业都属于高技术研发效率、低成果转化效率的发展模式。其次,在创新效率测度与评价的基础上,引入模型的不确定性,在同等效率水平下,运用贝叶斯模型平均（BMA）方法对影响高技术产业创新效率的主要因素进行分析,探究效率产生差异的原因以及提升的方向,同时设置不同模型先验假设对实证结果进行稳健性检验。分析结果表明,研发经费投入、企业规模以及政府支持程度为最具解释能力的影响因素,研发经费投入与企业规模对高技术产业综合创新效率具有显著正向影响,政府支持程度则恰恰相反。最后,根据相关理论和实证分析结果,从研发投入、资源配置、企业规模等角度提出了具有一定针对性的对策和建议,并提出了进一步研究的展望。本文的研究成果是国家自然科学基金“新型城镇化约束与引导下的产业转移的模式、路径与组织研究”（71473037）的一部分。有关“采用贝叶斯模型平均方法探讨高技术产业创新效率的主要影响因素”的研究成果曾在2018年10月于荷兰阿姆斯特丹大学召开的国际会议“第五届全球创新创业会议（GEIC）”上宣读汇报,受到主持人的肯定和与会者的好评。更多还原

【Abstract】 As China is in the period of economic transformation and upgrading,economic growth is increasingly dependent on technological innovation.Known as the most active technology innovation industry in the knowledge economy era,high-tech industry has become a new driving force and key point for China’s economic growth during the transition period,and it is also of great importance for international technology and economic competition.Improving the innovation efficiency of high-tech industries and enhancing industrial competitiveness have a major impact on increasing China’s innovation capabilities,adjusting industrial structure,and transforming economic development methods.In recent years,China has invested a large amount of labour and funds and introduced a series of policy measures to support the development of high-tech industries.The Outline of the National Medium-and Long-Term Science and Technology Development Plan(2006-2020)clearly states that “it is of great significance to promote the efficient allocation and comprehensive integration of science and technology resources in the whole society”.Therefore,the important and urgent task for improving the innovation capability of China’s high-tech industry under the new situation is to judge the actual efficiency level,clarify the factors affecting the innovation efficiency,in order to adjust and optimize the allocation of resources in a targeted manner.In recent years,China has invested a large amount of labour and funds and introduced a series of policy measures to support the development of high-tech industries.The Outline of the National Medium-and Long-Term Science and Technology Development Plan(2006-2020)clearly states that “it is of great significance to promote the efficient allocation and comprehensive integration of science and technology resources in the whole society”.Therefore,the important and urgent task for improving the innovation capability of China’s high-tech industry under the new situation is to clarify the factors affecting the innovation efficiency,judge the actual efficiency level,in order to adjust and optimize the allocation of resources in a targeted manner.This thesis attempts to evaluate the current situation of China’s high-tech industry innovation efficiency,and find out the main direction of improving the innovation efficiency by clarifying relevant influencing factors.Specifically,based on the theory of technological innovation and efficiency evaluation,related literatures on high-tech industries are sourted out and summarized,technological innovation,and efficiency analysis both in China and abroad.First,a reasonable research framework is established,along with a two-stage DEA model with additional intermediate inputs;Then,the innovation efficiency of China’s high-tech industry is measured and analyzed by region;Afterwards,under the same efficiency level,the Bayesian model averaging method is used to analyze the main factors affecting the innovation efficiency of high-tech industries,explore the causes of efficiency differences and establish the direction of improvement;Finally,the corresponding policy recommendations are proposed to the analysis results of the two-stage DEA model and the Bayesian model averaging method.The main work and conclusions are as follows:First,based on domestic and foreign theories and literatures,this thesis starts with the concept and definition of high-tech industry,and analyzes the relevant literatures on innovation efficiency evaluation,Bayesian model averaging method and their applied researches.The focal point of this part is applied research of the DEA model and the Bayesian model averaging method,which sets the theoretical foundation of the empirical analysis.Secondly,model construction and empirical research is carried out.By constructing a two-stage efficiency analysis framework for high-tech industries,a two-stage high-tech industrial innovation efficiency evaluation DEA model based on the framework is build,and then use the model to measure and evaluate the innovation efficiency of China’s high-tech industries.The analysis results show that the innovation efficiency of China’s high-tech industry is not very satisfactory,the overall average is only 0.501,and the innovation efficiency between regions is significantly different;Therefore,there is an inconsistency in the efficiency of the technology R&D stage and the achievement transformation stage in the high-tech industry.Since 2008,the efficiency of the technology research and development stage is higher than the production efficiency,and the comprehensive innovation efficiency is mainly dominated by production.With the inconsistency between the efficiency of the technology R&D stage and the achievement transformation stage,the high-tech industry in most parts of China belongs to the development mode of High-tech R&D efficiency and low efficiency of transformation.Third,based on the measurement and evaluation of innovation efficiency,the uncertainty of the model is introduced At the same level of efficiency,the Bayesian Model Average(BMA)method is used to analyze the main factors affecting the innovation efficiency of high-tech industries,explore the causes of the differences in efficiency and the direction of improvement,and set the different models of prior hypotheses to conduct a robustness test in empirical results.The analysis results show that R&D capital investment,enterprise scale and government support are the most influential factors.R&D capital investment and enterprise scale have a significant positive impact on the comprehensive innovation efficiency of high-tech industries,and the degree of government support has the opposite impact.Finally,according to the relevant theoretical and empirical analysis results,this thesis put forward some specific countermeasures and suggestions,and the prospect of further research from the perspective of R&D investment,resource allocation,enterprise scale,etc.The research results of this thesis are part of the National Natural Science Foundation’s “study on the model,path & organization of industry transfer under the constrains & guidance of new-type urbanization”(71473037).The research results of “The influencing Factors Analysis of Chinese High-tech Industry Innovation Performance——Based on the Bayesian Model Averaging(BMA)Method” were presented at the 5th Global Innovation and Entrepreneurship Conference(GEIC)held at the University of Amsterdam,Netherlands in October 2018.The report was praised by the host and scholars during the meeting.更多还原