ICT, human capital and productivity in Chinese cities

Abstract

This study uses a rich city-level dataset to analyse the relationship between information and communication technology (ICT) and productivity performance in China during 2003–2016. It is shown that ICT positively contributes to Chinese cities’ productivity in conjunction with other growth determinants, such as human capital, foreign direct investment (FDI), infrastructure development, financial market development, and research and development investment. An identifiable amplified effect is detected when ICT exceeds certain threshold in Chinese cities. This threshold level is reached in over a half of Chinese cities particularly cities in coastal regions. Finally, ICT is found to substitute human capital in China’s context. Since the average education level in Chinese cities is low, the finding is in line with the argument that ICT only improves productivity of high-skilled workers but worsens that of the low-skilled ones.

Keywords:

• ICT
• human capital
• productivity
• China

JEL classification::

• O47
• O33
• R10

Acknowledgements

We thank two anonymous referees for very helpful comments.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 China’s National Bureau of Statistics broadly defines S&T as ‘… activities closely related to the creation, development, dissemination, and application of scientific and technical knowledge’ and it is used to indicate intangible capital in extant studies ADDIN CSL_CITATION {"citationItems":[{"id":"ITEM-1","itemData":{"author":[{"dropping-particle":"","family":"Fleisher","given":"Belton M.","non-dropping-particle":"","parse-names":false,"suffix":""},{"dropping-particle":"","family":"McGuire","given":"William H.","non-dropping-particle":"","parse-names":false,"suffix":""},{"dropping-particle":"","family":"Smith","given":"Adam Nicholas","non-dropping-particle":"","parse-names":false,"suffix":""},{"dropping-particle":"","family":"Zhou","given":"Mi","non-dropping-particle":"","parse-names":false,"suffix":""}],"collection-title":"IZA Discussion Paper","id":"ITEM-1","issued":{"date-parts":[["2013"]]},"number":"7798","title":"Intangible knowledge capital and innovation in China","type":"report"},"uris":["http://www.mendeley.com/documents/?uuid=96d16a35-ec0d-4b62-b7c4-5f5084d9f33f"]}],"mendeley":{"formattedCitation":"(Fleisher et al., 2013)","plainTextFormattedCitation":"(Fleisher et al., 2013)","previouslyFormattedCitation":"(Fleisher et al., 2013)"},"properties":{"noteIndex":0},"schema":"https://github.com/citation-style-language/schema/raw/master/csl-citation.json"}(Fleisher et al., 2013).

1 Before 1994, telecommunications sectors were fully owned and operated by the government (Chen et al. 2005). The Ministry of Posts and Telecommunications (MPT) acted as a monopoly administration in China, and China Telecom was the dominant market player (Ward and Zheng 2016). After the then President Jiang Zemin delivered a keynote speech at the 2000 World Computer Congress, the Chinese government has put informatisation on the top agenda of economic development and declared as a national strategy in the Fifth Plenary Session of the 15th Central Committee in 2000. After that, MPT was split into a Postal ministry and the Ministry of Information Industry (MII), which administered telecommunications operation according to the principle of fairness. New players like China Netcome, China Jitong and China Railway, were established. Provinces started to enjoy a degree of de facto ‘federalism’ in telecommunications operation: Provincial telecom bureaus are located in the provinces and directly accountable to MII (Chen et al. 2005); China Telecom was split along regional lines beginning in 2002 (Ward and Zheng 2016). Driven by many short- and long-term development policies, ICT sectors in China, especially computer software and information communications, went through the ‘golden ten-years’ development.

2 China Internet Network Information Centre (CNNIC) www.cnnic.net.cn

3 ICT is the most extensive definition that integrates information technologies (IT) and communication technologies (CT). While IT stresses the role of computer hardware, software, and the relevant, CT covers products and technologies that transmit, store, retrieve, manipulate information. This paper does not cover IT relevant technologies and mainly focuses on fixed lines, mobiles, and broadband, as three most populous indicators of CT.

4 The assumption of constant returns to scale is necessary since technology evolution in an information economy would change not only the inputs (labour and capital) quantity, but more frequently, the input quality, the elasticity of substitution, and the gross output price. For each case, a production function homogeneous of degree m (m > 1) will provide a better statistical fit but bias the contributions of technological progress (Ferguson 1965).

5 The four headings are: Creation, transmission, and absorption of knowledge (the focus is knowledge creation and transfer. Research and development (R&D), trade, and FDI are reviewed as productivity determinants); Factor supply and efficient allocation (the focus is capital intensity and allocation. Human capital, physical infrastructure, financial system are reviewed as productivity determinants); Institutions, integration and invariants (Institution, integration, geography are reviewed as productivity determinants ); and Competition, social dimension and environment (Competition, social dimension and environment are reviewed as productivity determinants). In sum, the latter two is called “deep” determinants that provide mainly the long-term view on productivity progress.

6 If there are more than one threshold, the panel threshold model is displayed as: $\mathrm{\text{ln}}\left(y_{\mathit{\text{it}}}\right)={\alpha }_i+{\gamma }_t+{\beta }_1\mathrm{\text{ln}}\left(k_{\mathit{\text{it}}}\right)+{\beta }_2^{\prime }\mathrm{\text{ln}}\left({\mathit{\text{ICT}}}_{\mathit{\text{it}}}\right)·I\left(q_{\mathit{\text{it}}}\le {\theta }_1\right)+{\beta }_2^{″}\mathrm{\text{ln}}\left({\mathit{\text{ICT}}}_{\mathit{\text{it}}}\right)·I\left({\theta }_1<q_{\mathit{\text{it}}}\le {\theta }_2\right)+{\beta }_2^{″\prime }\mathrm{\text{ln}}\left({\mathit{\text{ICT}}}_{\mathit{\text{it}}}\right)·I\left(q_{\mathit{\text{it}}}>{\theta }_2\right)+{\beta }_3\mathrm{\text{ln}}\left(H_{\mathit{\text{it}}}\right)+\eta \mathrm{\text{ln}}\left({\mathit{X}}_{\mathit{\text{it}}}\right)+{ϵ}_{\mathit{\text{it}}}$

7 Techniques include structural equations models (Gruber et al. 2011), instrumental-variable approach (Czernich et al. 2011), Granger causality test (Chakraborty and Nandi 2003), Levinsohn–Petrin estimators (Levinsohn and Petrin 2003), stochastic-frontier production functions (Baquero Forero 2013) and dynamic GMM (Dimelis and Papaioannou 2011).

8 With high persistency, lagged values of endogenous variables contain less information and become weak instruments in a difference GMM model. Model efficiency would be improved by adding level equations in a system GMM (Blundell and Bond 1998).

9 According to ITU, fixed lines are still one of the important ICT indicators in the world, so we incorporate fixed lines into considerations in our baseline regression. However, some may argue that fixed lines are outdated to some extent. Thus, we also exclude fixed lines and use mobiles and internet only for robustness check in Section 5.

10 By now, there are in total 293 PAA cities in China, in which 257 cities are existing throughout the whole period of 2003–2016 without missing variables (CCSY, various years).

11 Intuitively, principal components of a collection of points can be understood as a sequence of unit vectors being orthogonal to each other in a real coordinate space. Each data point is projected onto principal components to obtain lower-dimensional data while preserving as much of the data’s variation as possible. Mathematically, principal components are often computed by eigen-decomposition of the data covariance matrix. Here, we decompose the covariance matrix of ICT proxies by eigen-decomposition and use uncorrelated and normalised eigenvectors as the proxy for ICT development (ICT penetration rate).

12 A covariance matrix instead of a correlation matrix is used for PCA analysis because the variables are expressed in the same units.

13 The base year is 2010 for constant GRP values and values of capital stock. The depreciation rate of physical capital stock and the price deflator is from Li and Wu (2018).

14 Coastal regions include Beijing, Tianjin, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, Liaoning and Hainan, central regions include Anhui, Heilongjiang, Henan, Hubei, Hunan, Jiangxi, Jilin, and Shanxi, and western regions include Chongqing, Gansu, Guangxi, Guizhou, Inner-Mongolia, Ningxia, Qinghai, Shaanxi, Sichuan, Xinjiang, Yunnan, and Tibet (Shen et al. 2019).

15 It can be partly explained by a self-selected inward economic transformation. Eastern cities are losing their attractiveness for foreign investment because of their increasingly expensive labour force, and western cities are lack of advantages in terms of geographical and political factors. As a consequence, central cities might in turn take place to attract more foreign investment relative to its GDP volume.

16 The results of robustness check are presented in Table A2 in the appendix.

17 The meta-study of Stiroh (2010) found a very wide range of estimates of the elasticity of output with respect to ICT capital. The estimates range from an upper end of over 25% to minus 6%. Niebel (2018) analysed ICT elasticities of 59 developed, developing, and emerging countries over 1995–2010. The elasticities of ICT range between 5% and 10% and show no significant differences between developing and developed nations.

18 For dynamic specifications, we use ‘xtendothresdpd’ in STATA17 as the command and the same instruments for endogenous variables as that in Table 5.

19 The results considering alternative lags and the ICT indicator excluding fixed lines are shown in Table A2 in the appendix.

20 Examples include occupations such as cashier, bank teller, and alike, which are being replaced by e-payment systems and ATM.

Additional information

Funding

This work was supported by Science and Technology Commission of Shanghai Municipality (grant numbers 22692192700).

Notes on contributors

Qing Li

Dr Qing LI is Assistant Professor, Department of Economics and Finance, SILC Business School, Shanghai University. She is an early career researcher with her doctoral degree in economics from the University of Western Australia. Her work has appeared in China Economic Review, Telecommunications Policy, Applied Economics and so on.

Yanrui Wu

Professor Yanrui WU is with Department of Economics, Business School, University of Western Australia. He specialises in economic development, innovation economics, energy and environmental economics. He has published widely in these fields.