Abstract: This article systematically analyzes the productivity effect of market segmentation and its realization mechanism as well as specific change path from both theoretical and empirical perspectives，which not only overcomes the shortcomings of the inverted Ushaped hypothesis in theoretical logic and empirical analysis，but also finally solves the policy paradox it brings It is found that market segmentation will directly improve local productivity on the one hand，and indirectly hinder the improvement of local productivity by increasing the number of enterprises in the region on the other hand Under the joint effect of these two mechanisms，the impact of market segmentation on local productivity shows significant time heterogeneity，and will reduce the level of longterm productivity On this basis，the above theoretical hypothesis is verified by using the panel data variable coefficient regression model and dynamic spatial lag model，and it is found that the above conclusions are still robust even considering the spatial spillover effect of regional productivity Therefore，rather than bear the longterm efficiency loss caused by market segmentation，it is better to gradually eliminate market segmentation，and realize the smooth flow of domestic circulation

Key words: market segmentation, total factor productivity, domestic circulation

摘要： 本文从理论和经验两个层面系统分析了市场分割的生产率效应、实现机制和具体变动路径，既克服了倒“U”型假说在理论逻辑和经验分析上的不足与缺陷，又解决了其所带来的政策悖论。研究发现，市场分割一方面直接促进本地生产率提升，另一方面则通过增加区域内企业数量间接阻碍本地生产率提升。在这两种机制的共同作用下，市场分割对本地生产率的影响呈现出明显的时间异质性，并且会显著降低长期稳态下的生产率水平。在此基础上，分别使用面板数据变系数时点回归模型和动态空间滞后模型验证了上述理论假说，同时发现即使在考虑地区生产率空间溢出效应的情况下，上述结论依然是稳健的。因此，与其承受市场分割带来的长期效率损失，不如以短痛去长痛，逐步消除市场分割，实现国内大循环的畅通。

关键词: 市场分割, 全要素生产率, 国内大循环