Abstract: This paper uses enterprise data to empirically test the influencing factors of Chinese hightech industry productivity, and concludes that the growth of assets, human capital, corporate profits, and enterprise scale have significant effect on productivity. Government subsidies, capital labor ratio,and inventory play negative roles.Foreign direct investment, trade credit, sales cost, market concentration, and industrial chain status have little effect. Different from the normal expectation of people, R&D investment on productivity may not work or play a negative role.This paradoxical conclusion reflects the special circumstances in China.The motive of many enterprises  R&D investment is to rent for an access to government subsidies and monopoly interests.Lack of protection for intellectual property rights makes many enterprises unwilling to put a lot of money in the use of technology research and development. Bank loans and government subsidies seriously distort and impede productivity. This paper suggests that the problems arise in Chinese financing system and government funded system, which not only fails to promote enterprise efficiency, but also inhibits the development of enterprise efficiency.

Key words: high-tech industry , productivity , LP semi-parameter method

摘要： 本文使用微观层面的企业数据，对中国高技术产业生产率的影响因素进行实证检验，发现资产增长、人力资本、企业利润、企业规模对生产率具有显著的促进作用，而资本劳动比、企业存货起明显的负向作用，外国直接投资等变量几乎不起作用。研发投入、银行贷款与政府补贴的实证结论与人们的正常预期不同，研发投入对生产率要么不起作用、要么起负向作用，银行贷款与政府补贴对生产率起着扭曲与阻碍作用，中国的一些特殊情况导致了上述结论。

关键词: 高技术产业 , 生产率 , LP半参数方法