Abstract: Industrial digitization & digital industrialization are regarded as the main battlefield of technological and economic competition by various countries and their enterprises. However, it is still poorly understood by the academic community how these modern digital technology will affect the innovation division of labor between upstream and downstream enterprises in the industrial chain. This paper tries to reveal the mechanism in the dynamic process of digitization by referring to the theory of "social environment-heterogeneous labor matching structure", the mirror hypothesis and system integration hypothesis of organizational economics. The paper divided the digital development level of automobile emission control system into three stages: "component collaboration", "human-machine dialogue" and "automatic control", and has examined the relative probability of supplier and manufacturer choosing product architecture innovation or product component innovation at different stages of digital development using a binary selection model with USPTO invention patent application data from 1970 to 2018. The results show that the proportion of component suppliers choosing architecture innovation increases significantly when the industrial chain shifting from "component collaboration" to "human-machine dialogue", and then decreases further shifts to "automatic control", but no evidence has been found that it has decreased to the level of "component collaboration" stage. Therefore, digitalization will encourage more suppliers to choose product framework innovation, and thus intensify the competition degree of innovation divisionof labor within the industrial chain. Combined with mechanism analysis, dominant cognitive labor and its dominance over productive labor, transactional labor and R&D labor will become the key source of competitive advantage of enterprises in the digital supply chain era, while the shortening of industrial chain and the blurring of organizational boundaries may be important by-products of industrial chain digitalization. The research conclusion can provide important enlightenment for enterprises to participate in supply chain innovation division competition, and organizational form and management mode adaptation reform in the digital era.

摘要： 产业数字化和数字产业化被各国及其企业视为技术经济竞争的主战场。然而，数字技术现在及将来如何影响产业链上下游企业间的创新分工，学术界仍然知之甚少。借鉴“社会环境—异质性劳动匹配结构”理论、组织经济学的镜像假说和系统集成假说等理论框架，在数字化的动态过程中尝试揭示这一机制。将汽车尾气排放控制系统的数字化发展水平区分为“组件协同”、“人机对话”和“自动控制”三个阶段，使用1970—2018年USPTO发明专利申请数据，利用二值选择模型检验供应商和制造商在不同数字化发展阶段选择产品构架创新或产品组件创新的相对概率。结果表明，当产业链从“组件协同”阶段转向“人机对话”阶段时，选择构架创新的组件供应商数量占比显著提高，在进一步向“自动控制”阶段转变时有所下降，但尚未发现降低至“组件协同”阶段的证据。因此，数字化将激励更多供应商选择产品构架创新，进而加剧产业链内部创新分工的竞争程度。结合机制分析，拥有占优的认知性劳动及其对生产性劳动、交易性劳动和研发性劳动的支配能力，将成为供应链数字化时代企业竞争优势的关键来源，而产业链短化和企业组织边界模糊化可能是产业链数字化的重要副产品。研究结论可为数字化时代企业参与供应链创新分工竞争、组织形式及管理模式适应性变革提供重要启示。