Industrial structure upgrading and technological capability in China – based on the perspective of industrial structure depth

ABSTRACT

This paper focuses on measuring the status of industrial structure upgrading in China using a scientific and reasonable approach. We introduce a novel metric called industrial structure depth (ISD), which incorporates inter-industry proportional relationships and intra-industry technological capability, surpassing traditional indicators. Through ISD, we comprehensively assessed industrial structure upgrading across thirty provinces, sub-industries, and the national level in China from 2002 to 2022. The findings reveal a general rise in ISD at the provincial and national levels, followed by a subsequent decline. The shift in the gradient distribution of ISD among provinces is notable, with rapid development observed in certain central provinces. From the perspective of industry categories, manufacturing emerges as a significant influencer of China's overall ISD fluctuations. Furthermore, we dissect the ISD into the domestic intermediate goods technology-driven aspect and the production process technology-driven aspect, which allows for a detailed analysis of the factors driving changes in China's industrial structure upgrading. Overall, this paper sheds light on the crucial role of technological capability in promoting China's industrial structure upgrading and provides insights into the dynamics of industrial structure upgrading fluctuations across different stages.

KEYWORDS:

• Industrial structure upgrading
• sustainable development
• input-output table
• technological capability
• trade war

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Data from the National Bureau of Statistics of China.

2 $\mathrm{PROD}{Y}_{\mathrm{it}}=\sum _j{\displaystyle \frac{x_{\mathrm{kij}}}{x_{\mathrm{ki}}}\mathrm{PROD}{Y}_{\mathrm{ij},t}}$. Where $\mathrm{PROD}{Y}_{\mathrm{it}}$ represent the technical complexity of industry i in year t. $\mathrm{PROD}{Y}_{\mathrm{ij},t}$ represents the technical complexity of product j that belongs to industry i in year t, $x_{\mathrm{kij}}/x_{\mathrm{ki}}$ represents the proportion of the export value of product j in industry i belonging to region k to the total export value of industry i in region k. $\mathrm{PROD}{Y}_{\mathrm{ij},t}=\sum _m{\displaystyle \frac{X_{\mathrm{mij},t}/X_{\mathrm{mi},t}}{{\sum }_{\mathrm{mi},t}(X_{\mathrm{mij},t}/X_{\mathrm{mi},t})}{Y}_{m,t}}$. Where $X_{\mathrm{mij},t}$ represents the export value of product j that belongs to industry i, in year t and region m. $X_{\mathrm{mi},t}$ represents the export value of industry i, in year t and region m and $Y_{m,t}$ represents the per capita real GDP of region m in year t.

3 Trade data and per capita GDP data for different countries are used to calculate PRODY.

4 To ensure data continuity, the missing data for the periods 2002–2008, 2009–2015, and 2016–2022 have been substituted using the provincial IOTs data from 2007, 2012, and 2017, respectively, based on the proximity principle (Sun et al., 2017). Similarly, at the national level, the missing data for the IOTs in 2003, 2004, 2006 and 2008, 2009 and 2011, 2013, 2014 and 2016, 2019, 2021 and 2022 are replaced by the input–output table data in 2002, 2005, 2007, 2010, 2012, 2015, 2018 and 2020, respectively.

5 The relevant data for Tibet, Hong Kong, Macao Special Administrative Region, and Taiwan Province are lacking, so they are deleted.

6 The classification method comes from the National Bureau of Statistics of China.

7 The data is calculated through input–output tables at the national level in China.

8 The estimation model of panel data is selected through the Hausman test, and finally, the two-way fixed effect model is selected for regression.

9 Specifically, to ensure the continuity of the data, we referred to the method of Sun et al. (2017) and substituted missing data for the periods 2002–2004, 2005–2009, 2010–2014, and 2015–2022 using the national IOTs data from 2002, 2007, 2012, and 2017, respectively.

10 The t-test on the time-series data of the two types of ISDs shows that we cannot reject the null hypothesis that the means of the two types of ISDs are the same (p-value is 0.9859).

Additional information

Notes on contributors

Heer Wang

Heer Wang received the BS degree in Economics and Science from Southwestern University of Finance and Economics, Chengdu, China, in 2019. She is currently working toward the PhD degree in Economics with the School of Economics, Zhejiang University, Hangzhou, China. Her research interests include industrial structure upgrading, agricultural and rural economy, and labour mobility.