Abstract: Base on the negative binomial model and provincial dynamic panel data from Chinese 21 provinces’ agricultural sub-categories patents data in 2005-2012,This paper study industry upgrading and linkage effects on agricultural innovation capability of China from the perspective of inter-industry and intra-industry, using SYS-GMM and LSDVC, We find: At the national level, industrial structure optimization and upgrading have obvious positive industry linkage effects, non-agricultural R&D investment have positive technology spillover effects, but agricultural R&D have negative spillover effects. At the regional level, agricultural R&D, non-agricultural R&D and non-agricultural trade have negative technology spillover in Eastern and Central regions, agricultural R&D and non-agricultural R&D have positive industry linkage and technological spillover effects respectively. We should take full advantage of industrial upgrading and linkage continue to enhance the agricultural science and technological innovation capacity for China's modern agriculture and the “three rural issues”.

摘要： 本文以2005-2012年21个省份22个农业专利子类组成的省级平衡面板数据，基于能够消除超散布性和偏离泊松分布的负二项分布模型，使用系统广义矩估计(SYS-GMM)和动态纠偏最小二乘虚拟估计(LSDVC)同时从产业间和产业内分析影响我国农业科技自主创新能力的产业关联效应和技术溢出效应。研究发现：全国层面，产业结构升级和非农FDI分别具有显著正向产业关联效应和技术溢出效应，而农业FDI产生了负向技术溢出；地区层面，东部和中部地区非农FDI、农业FDI和非农进出口均存在负向技术溢出，西部地区非农FDI和农业FDI分别产生了正向产业关联效应和技术溢出。我国发展现代农业和解决“三农问题”应充分利用产业关联和技术溢出不断增强农业科技自主创新能力。

关键词: 农业创新能力, 产业关联, 技术溢出, 负二项分布