China’s catch-up to the US economy: decomposing TFP through investment-specific technology and human capital

Abstract

We implement a neoclassical growth model that incorporates investment-specific technology (IST) modifying capital investment in the law of motion of capital and bifurcates productivity into human capital and total factor productivity (TFP) in the production function. We focus on the role of changes in the quality-adjusted price of investment goods on China’s growth by comparing the effects of IST and human capital on the decomposition of US and Chinese productivity. The results show that both human capital and IST play an important role in the decomposition of US TFP. For China, human capital accounts for an increasingly higher portion of Chinese TFP for the period 1952–2009; however, IST contributes to the explanation of TFP only after the 1979 reforms. The analysis is extended by considering the impact of IST in the consumer’s investment decision and by projecting both countries’ GDP while modelling unbalanced Chinese growth using catch-up. Our model predicts that the Chinese economy will surpass the US economy in 2024.

Keywords:

• China
• investment-specific technology
• human capital
• convergence

JEL Classification:

• O32
• O47
• O53

View correction statement:

Corrigendum

Acknowledgements

We gratefully acknowledge the helpful comments of two reviewers.

Notes

¹ As discussed in detail below, investment-specific technology (IST) is the inverse of the quality-adjusted price of investment or capital goods.

² This is commonly known as ‘absorptive capacity’; see Blalock and Gertler (2008) and Alfaro et al. (2010).

³ For additional studies analysing Chinese growth, see Urata (1987), Garnaut and Ma (1993), Li (1994) and Chen and Feng (2000).

⁴ There are two classes of empirical studies analysing human capital and growth. The first class measures the portion of output growth that is explained by human capital (Jorgenson and Fraumeni, 1992; Mankiw et al., 1992; Hall and Jones, 1999). The second class implements regression analysis to identify human capital (among other variables) as a potential source of growth (Barro, 1991; Murphy et al., 1991; Levine and Renelt, 1992; Bils and Klenow, 2000; Bassanini and Scarpetta, 2002). These studies generally find that education serves as a strong proxy for measuring human capital.

⁵ Note that this function is linearly homogenous, satisfies the standard assumptions of positive, but diminishing marginal products, and the Inada conditions.

⁶ Heston et al. (2011) provide a detailed explanation of data methodology in Appendix 6.1. Also, Jones (1994), Hsieh and Klenow (2007) and Alfaro and Ahmed (2010), discuss Penn World Tables (PWT) data construction. According to Heston et al. (2011), China earned two out of four for data quality.

⁷ Greenwood et al. (1997) calculate the US depreciation rate for structures (0.056) and equipment (0.124). Their motion equation for structures does not include IST whereas the equipment motion equation does. We are not able to use these numbers because, while PWT data allows us to account for IST in both the United States and China, we are not able to distinguish structures from equipment.

⁸ Our calculation of the growth rate of the Chinese capital stock is 9.07% for the period 1979–1995, which is similar to those found in the literature. Over the same period, Hu and Khan (1997), The World Bank (1997), Maddison (1998) and Zhang (2008) calculated the Chinese capital stock to grow at 7.70%, 7.90%, 8.86% and 10.49%, respectively.

⁹ Key properties of the model that allow us to define a balanced growth path are constant intertemporal elasticity of substitution inherent in the preference structure, production follows constant returns to scale and satisfies positive and diminishing marginal product and the Inada conditions, and capital earns its marginal value product.

¹⁰ Catch-up growth has been used to model economic convergence in regional and international settings; it captures accelerated growth of a below average economy and the dampened growth of an above average economy through a spillover effect. Abramovitz (1986) links the inverse correlation between level and growth of a country to its proximity to the technology frontier. Counties farther from the frontier are able to easily implement already developed technology whereas countries on the frontier devote large resources advancing the frontier. Tamura (1991) develops an endogenous growth model to analyse the income convergence of developed economies and interstate income convergence in the United States. Lucas (1993) finds that human capital convergence, modelled through a catch-up process, played a fundamental role in the rapid growth of East Asian economies during the 1970s and 1980s. More recently, Lucas (2009) uses a mechanical catch-up model in the human capital sector to describe GDP trends for all open economies and shows that growth rates for open economies tend to converge.