The role of productivity in China's growth: new estimates

Abstract

The impressive growth of the Chinese economy in the past three decades has attracted a lot of attention and hence the research interest of many scholars and policy makers. One of the focuses in the literature is the role of productivity in China's economic growth. This paper aims to revisit the debate about the role of productivity in China's growth, to provide an updated estimate of productivity growth and hence to make a contribution to the understanding of China's economic growth in recent years. Its objective is to propose and apply a growth accounting technique to assess economic performance in China, in particular the role of technological progress in China's recent growth. The findings about the latter may have important policy implications for the sustainability of China's economic growth in the future.

Keywords:

• total factor productivity
• growth accounting stochastic frontier approach

JEL codes:

• O11
• C23

Acknowledgements

The author would like to thank Jinghai Zheng, Zheng Yuxin and participants of China Productivity Workshop (Tsinghua University, 2007) for helpful comments and suggestions, comments from two anonymous referees are also appreciable.

Notes

Notes

1. See, for example, the World Bank (1993), Krugman (1994), Young (1994) and Kim and Lau (1994).

2. Examples include Ito and Krueger (1995), Young (1995), Fu et al. (1999) and Wu (2002), to cite a few.

3. See, for instance, Sachs and Woo (1997), the World Bank (1997a), Maddison (1998), Wang and Meng (2001) and Chow (2002).

4. Examples include Borensztein and Ostry (1996), Fleisher and Chen (1997), Woo (1998), Chow and Lin (2002), Wang and Yao (2003), Young (2003), Wu (2003) and Zheng and Hu (2006).

5. The Japanese and German figures are estimated by Dougherty and Jorgenson (1996).

6. For more detailed surveys of economic growth in China, see the World Bank (1997a), Bramall (2000), Chow (2002), Garnaut and Song (2003) and Wu (2004).

7. For surveys of the measurement of economic openness, see Harrison (1996).

8. China officially became a WTO member in November 2001.

9. In agriculture, the state sector is very small. However, private ownership of land has ceased to exist in China since 1949.

10. Reviews of stochastic frontier models are presented by Lovell (1996), Greene (1997), Coelli et al. (1998) and Kumbhakar and Lovell (2000), to cite a few.

11. See, for example, Färe et al. (1994), Lovell (1996) and Coelli et al. (1998).

12. There are of course many other environmental factors affecting economic performance. For example, Sala-i-Martin (1997) identified more than 60 variables that may be related to economic growth. Due to data constraints (requirement of data for 29 regions and over 13 years), only three factors are considered in this study.

13. Examples of the one-step literature include Kumbhakar et al. (1991), Huang and Liu (1994) and Battese and Coelli (1995).

14. In total, there are 31 regions in mainland China. In this study, Tibet is excluded due to missing data. In 1997 Chongqing became a municipality supervised directly by the central government and is included in Sichuan province in this study.

15. See Appendix A for the classification of Chinese regions.

16. This rate of depreciation is chosen for consistency with the estimation of domestic capital stock (see Appendix B). For most Chinese regions, the initial value of foreign capital stock in 1978 was assumed to be zero.

17. It is argued that this indicator of economic reform or marketisation may be biased due to its ignorance of many other factors. The author is currently investigating other measures of marketisation for the regions, such as the one developed by the National Centre for Economic Research, Beijing (Wang 2004).

18. There are many different ways of measuring economic openness. For a brief survey, see Harrison (1996) and Edwards (1998).

19. It is noted that the consistency and efficiency of the estimates from some of the models is an issue of controversies (Battese and Broca 1997). All models in this study are estimated using FRONTIER 4.1 developed by Coelli (1996).