Abstract: The government talents policy is an important form of promoting regional innovation activities in addition to its fiscal and taxation arrangements. Based on theoretical analysis and the provincial panel data of China from 2005 to 2017, this paper empirically analyzes the impact of government talents policy on regional innovation efficiency and innovation output by using the stepwise regression method. The results show that the government talents policy can directly improve the efficiency and output of regional innovation, and it can also indirectly improve the efficiency and output of regional innovation by changing the stock and structure of regional human capital. Moreover, it is better for the government to improve regional innovation efficiency and innovation output by changing the regional human capital stock than by changing the regional human capital structure. However, there is a Williamson spatial clustering effect on the influence of government talents policy on regional innovation, and the influence of government talents policy on regional innovation will be greatly reduced after it reaches a certain degree. Therefore, local governments should fully understand the situation of regional social and economic development and human capital and choose the appropriate direction of talents policy and the appropriate degree of effort.

Key words: government talents policy, stock of human capital, structure of human capital, regional innovation, Williamson spatial clustering

摘要： 政府人才政策是除了其财税安排以外促进区域创新的重要手段。基于理论分析，利用2005~2017年中国省级面板数据，通过逐步回归法实证分析了政府人才政策对区域创新效率和创新产出的影响。研究发现，人才政策能够直接提高区域创新效率和创新产出，同时也能通过改变区域人力资本存量和人力资本结构间接提高区域创新效率和创新产出。而且，人才政策通过改变区域人力资本存量的方式来提高区域创新效率和创新产出的效果要优于通过改变区域人力资本结构的方式。然而，人才政策对区域创新的影响存在威廉姆森空间聚集效应，人才政策力度达到一定程度后对区域创新的影响会减弱。因此，地方政府要充分了解本区域社会经济发展状况和人力资本情况，从而确定政策方向和政策力度。

关键词: 政府人才政策, 人力资本存量, 人力资本结构, 区域创新, 威廉姆森假说