Innovation policy in China: nationally promulgated but locally implemented

ABSTRACT

In China, policies are often announced at the national level but implemented locally. Innovation policies are no exceptions. This article studies China’s 50% R&D tax deduction policy, a key innovation policy promulgated by the central government in 2006. We find that the degree to which the policy was implemented during the period of 2006–2009 varied across the provinces and industries, which in turn had significant impacts on local firms’ innovation output. The findings of the heterogeneity in local implementation of national innovation policies are of important implications for both policy makers and innovation scholars.

KEYWORDS:

• Innovation policy
• local implementation
• R&D
• tax deduction

JEL CLASSIFICATION:

• O3
• L5

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ R&D expenditure in high/new technologies is denoted as ‘qualified R&D expenditure’ in the 50% R&D tax deduction policy. The high/new technologies are specified by the ‘Guideline on key high technology fields for prioritized industrialization (2007)’ (http://www.casted.org.cn/web/index.php?Page=1&NewsID=3773) and the ‘State supported high/new technology fields’

(http://www.mofcom.gov.cn/aarticle/bh/200805/20080505534363.html).

² See http://www.chinatax.gov.cn/n810341/n810765/n812171/n812675/c1190645/content.html.

³ Innovation policies are well justified because firm R&D is of public good nature (it is difficult for a firm to appropriate the full benefit of its R&D). For example, in 2011, 26 OECD countries provided tax incentives for firm R&D, up from 18 in 2004 and 12 in 1995. Non-OECD countries, including Brazil, India, Russia, Singapore and South Africa also provide similar fiscal incentive policies.

⁴ See Hall and Van Reenen (2000), Ientile and Mairesse (2009), Mohnen and Lokshin (2009), Mairesse and Mohnen (2010). Most studies focus on developed economies, except for Yang, Huang, and Hou (2012) on Taiwanese manufacturing firms and Özçelik and Taymaz (2008) on Turkish manufacturing firms.

⁵ These six 2-digit high/new tech industries include chemical, communication, drug and medical, instruments, specialized equipment and transportation. We identify these industries based on the high tech industry directory issued by the China’s National Bureau of Statistics in 2002

(http://www.bjstats.gov.cn/zdybz/tjbz/hyfldm/gmjjhyflxgwj/200612/t20061207_78074.htm).

⁶ Since not all the six high/new tech industries exist in each of the 31 provinces in China, there are in total 144 ratio observations (at the province by high/new tech industry level) in the data. Of these observations, 12 have missing values from the survey data and another 12 have values larger than 0.125 (which could be erroneous). We dropped these 24 observations from the analyses.

⁷ A positive correlation between R&D expenditure and R&D outputs has been well documented (see Hall, Griliches, and Hausman 1986).

⁸ We aggregate firm registration types in the data into four ownership types: state or collective ownership, private ownership, Hong Kong/Taiwan/Macao ownership, and foreign ownership.

⁹ Given that the two levels of clusters, province and industry, are not nested within each other, the adjusted variance-covariance matrix of the estimated coefficient equals to the sum of the variance-covariance matrices clustered at province and industry levels respectively, minus the variance-covariance matrix clustered at province-by-industry level.

¹⁰ In the data, there are also two firms that changed their provincial location between 2005 and 2009 and 125 firms that changed their industries. These firms might be outliers and thus are excluded in our analyses.

Additional information

Funding

This work was supported by The National Social Science Fund of China [grant number 17CJY007]