Relationship Between Environmental Regulations and Global Value Chains in Chinese Manufacturing

Abstract

While active participation in GVCs can stimulate economic growth, it has also been noted to exacerbate environmental pollution. To address environmental concerns, most countries have strengthened environmental regulations, which in turn may affect the positioning of countries within GVCs. This study empirically analyzes the impact of environmental regulations imposed by the Chinese government on the positioning of the manufacturing sector within GVCs of Chinese manufacturing industries from 2006 to 2018.

The findings of the research are as follows: Firstly, the strengthening of environmental regulations has an inverted U-shaped impact on the overall GVC position of Chinese manufacturing sector, demonstrating the combined effects of cost effect and innovation compensation effect. Specifically, when environmental regulations are lower than a certain level, strengthening regulations shifts the GVC position upstream. However, when regulations are more stringent than the level, further strengthening them leads to a downstream shift in GVC position. Secondly, it seems that China's current level of environmental regulation is below the apex of the inverted U-shaped curve, and therefore, it is anticipated that further strengthening of environmental regulations in China will result in a further upstream shift in the GVC position. Thirdly, these analytical results are particularly applicable to technology-intensive industries.

KEYWORDS:

• The manufacturing industry
• GVC position
• environmental regulation
• technology-intensive industries

JEL Classifications:

• F18(Trade and Environment)
• F64(Environment)

Disclosure statement

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

Notes

1 The authors have a keen interest in the Korean economy, which is why comparisons have been made between Korea and the United States. Furthermore, China and the United States are the largest economies globally in terms of economic size and trade volume. Additionally, these three countries are at significantly different stages of economic development, making them intriguing subjects for comparison. Therefore, they have been presented in this study.

2 In previous studies, there is a tendency to include control variables such as FDI, labor force, and trade dependency in the analysis. While we also considered these variables in our study, high correlations among them often lead to multicollinearity issues. To address this, we selected the two variables with the least multicollinearity and the highest frequency of use as control variables. Additionally, at the outset of this study, we attempted to analyze using industry-specific dummy variables. However, including interaction terms of dummy variables in models up to the quadratic level made the models overly complex and led to a decrease in degrees of freedom. Consequently, we opted to analyze the data by separating it into industry-specific categories.

3 UIBE provides estimates of production length using GVC-related data from the OECD.

4 Correlation coefficients between EPI and GVC_pos are as follows.

Table

5 In this study, while performing Pooled OLS, White test was conducted for heteroskedasticity testing, and Wooldridge test was conducted for autocorrelation testing, revealing the presence of both heteroskedasticity and autocorrelation. For instance, in the models presented in Table 3, the coeeficient of AR(1) model of the error term yielded identical values of 0.935 for both model 1 and model 2, indicating autocorrelation. Due to space constraints, detailed information is omitted.

6 Model 1 estimates a linear function, while Model 2 estimates a quadratic function. In this study, the main focus is on estimating Model 2, while Model 1 serves merely as a comparative reference. Therefore, we do not extensively focus on interpreting Model 1 in our analysis. Regarding Model 2, the positive coefficient value of the EPI linear term indicates that the maximum point of the position index is determined by positive values of EPI. This aligns with reality and is a plausible result.

7 For example, if the right side of equation (3) is transformed into a perfect square trinomial form, it would be approximately -4.034(EPI-0.6605)${}^{\wedge }$2 + 1.760.

8 Using a representative 60% of data, the average value of China's environmental regulation index over the 13-year study period is 0.54.

Additional information

Funding

This work was supported by Chungnam National University; National Research Foundation of Korea: [NRF-2021S1A5B8096365].