我国自主研发投入配置效率低是自主创新能力与创新绩效不高的重要原因之一。文章从自主研发内外部支出角度出发，将知识存量纳入分析框架，利用2009~2017年我国28个省市的数据构建面板门槛模型，研究发现：自主研发内外部支出均存在显著的知识存量门槛效应，相比高知识存量地区，低知识存量地区内外部支出对创新绩效的边际产出贡献更大。分区域分析发现，西部地区与其他地区不同，其高知识存量地区外部支出对创新绩效的边际产出贡献更大。本文的研究为创新主体合理分配研发投入尤其是政府干预提供了依据。

The low allocation efficiency of independent R&D investment in China is one of the important reason for the low independent innovation ability and innovation performance.From the perspective of internal and external expenditure on independent R&D, this paper brings the knowledge stock into the analysis framework, and constructs a threshold model using the data of 28 provinces and cities in China from 2009 to 2017. The study finds that there is a significant knowledge stock threshold effect in internal and external expenditure on R&D, and low knowledge stock areas contribute more to the marginal output of innovation performance than high knowledge stock areas. The sub regional analysis shows that in the external expenditure model with knowledge stock as the threshold, the western region is different from other regions, and its high knowledge stock regions contribute more to the marginal output of innovation performance. The research of this paper provides a basis for the innovation subject to reasonably allocate R&D investment, especially government intervention.

F062.9