Effects of Carbon Trading Pilot on Carbon Emission Reduction: Evidence from China’s 283 Prefecture-Level Cities

Abstract

Empirical evidence demonstrates that market-driven carbon trading scheme (ETS) is a crucial instrument for China to control environmental pollution. Based on the panel data of China’s 283 prefecture-level cities from 2006 to 2017, this research investigated the transmission mechanism, direct and indirect effects of ETS on carbon emission intensity (CEI) using difference-in-differences (DID) model, propensity-score-matched difference-in-differences (PSM-DID) model at national, regional, and local levels (cities with different industrial characteristics). The results demonstrated the mediating effects of total energy consumption, energy consumption structure, and industrial structure upgrading in the incentive role of ETS on CEI reduction. Moreover, ETS directly and effectively reduced CEI at the national level, while the spatial heterogenous effects were identified at regional and local levels, which emphasises the necessity and importance of unified carbon trading market establishment and classified governance.

Keywords:

• Carbon trading scheme;
• quasi-natural experiment;
• mediating effect;
• spatial spillover effect;
• spatial heterogeneous effect

Disclosure statement

The authors declare that they have no known competing personal relationships or financial interests that could have appeared to influence the work reported in this paper.

Data availability statement

Publicly available datasets were applied in this research. The data can be collected from the official website: China Statistical Yearbook (http://www.stats.gov.cn/english/Statisticaldata/AnnualData/).

Notes

1 https://www.bp.com/en/global/corporate/energy-economics/statistical-review-of-world-energy/downloads.html.

2 Rapid urbanisation industrialisation are also the main drivers of China’s water-energy demand problems (Chen & Warren, 2011).

3 See http://www.gov.cn/zwgk/2010-10/18/content_1724848.htm/.

4 See http://www.gov.cn/zwgk/2012-01/13/content_2043645.htm/.

5 See http://www.gov.cn/zwgk/2012-01/13/content_2043645.htm/.

http://www.ndrc.gov.cn/zcfb/zcfbtz/201201/t20120113_456506.html/.

http://qhs.ndrc.gov.cn/zcfg/201206/t20120621_487133.html/.

6 See http://www.ndrc.gov.cn/zcfb/zcfbtz/201201/t20120113_456506.html/.

7 See http://www.gov.cn/xinwen/2016-10/23/content_5123277.htm.

8 Where ${\mathcal{J}}_i$ is the proportion of industry $\mathrm{i}$’s (i = 1, 2, 3) output value to GDP.

9 Hydroelasticity is currently the most important renewable energy in China (Li, 2012).

10 The interprovincial price index of province where the city belongs was used to adjust the monetary indicators to the 2006 constant price (Yang et al., 2021). Regional FDI amount is converted into an RMB denominated amount in terms of the actual exchange rate.

11 We treat column (4) in table 2 as the benchmark regression model because carbon emission intensity is core dependent variable in this study.

12 See http://www.gov.cn/2011lh/content_1825838.htm.

13 To testify the reliability, we run the data by adding an additional variable as control variable: ${\mathrm{target}}_{12\text{th Five}-\text{Year Plan}}\times \mathrm{lnpgdp}$ (${\mathrm{target}}_{12\text{th Five}-\text{Year Plan}}$ is the carbon emission reduction commitments stipulated by the Chinese government in the National 12th Five-year Plan), as shown in Appendix C. Fortunately, the results are consistent with the initial parallel trend test results, suggesting that our conjecture of the “anticipation policy effect” is reasonable. In fact, the relevant policy effects on carbon reductions have become evident since 2011 (Hu et al., 2020; Zhang et al., 2020).

14 We use Bootstrap method instead of the Sobel test. The reasons lie in threefold (Preacher & Hayes, 2004): First, the Sobel test requires that the specific mediating and total mediating effects in the model obey normal distribution. Second, the Sobel test requires a large sample. Third, the calculation process of the Sobel test is very sophisticated and needs manual computation. Thus, to derive a robust result and avoid the shortcomings of the Sobel test, the better method to solve the structural equation modelling is the Bootstrap method, because the confidence interval of coefficient product is more accurate than obtained by Sobel model (Cheung, 2007; Lau & Cheung, 2012). Meanwhile, we also report the Sobel test results, as shown in Appendix D.

15 To overcome the shortcomings of the traditional adjacent spatial weight matrix, known as the 0-1 matrix, this paper adopts the temporal-spatial weight matrix, which is the combination between the temporal weight matrix and economic-geographical spatial weight matrix (the product of geographical distance weight matrix and the diagonal matrix of the proportion of per capital GDP (Yao et al., 2021)) (The detailed information is shown in Appendix E).

16 In the benchmark regression, we added the interaction term ${\mathrm{region}}_{\mathrm{n}}\times \mathrm{year},$ where China was divided into four economic zones. Unfortunately, the northeastern region must be included into the eastern region when performing Eq. (5), because no pilot regions belong to the northeastern region during the research period.

17 See http://www.gov.cn/gongbao/content/2013/content_2441018.htm.