*Article* **Impact on Carbon Intensity of Carbon Emission Trading—Evidence from a Pilot Program in 281 Cities in China**

**Wanlin Yu and Jinlong Luo \***

School of Economics, Shandong University of Technology, Zibo 255000, China

**\*** Correspondence: qqn123n@163.com; Tel.: +86-132-8069-0015

**Abstract:** China's carbon emissions trading scheme (ETS) is an institutional arrangement that China intends to explore as a means of energy conservation and emission reduction. It is the core of China's goal of achieving carbon peaking and carbon neutrality. This paper regards the introduction of pilot carbon emission trading policies as a quasi-natural experiment. Propensity Score Matching (PSM), Differences-in-Differences (DID), and spatial Durbin methods were used to evaluate the policy effects of pilot carbon emission trading policies on the carbon intensity of Chinese cities. We empirically tested the impact mechanism using the panel data of 281 cities at the prefecture level and above in China from 2006 to 2019. The results show that (1) the pilot policy of carbon emission trading has significantly reduced the carbon intensity of Chinese cities and shows characteristics of heterogeneity; (2) the dynamic effect test shows that the mitigation effect of the pilot carbon emission trading policy has increased gradually with time; (3) the mediation effect shows that the pilot carbon emission trading policy alleviates urban pollution in the region by improving the level of environmental governance and jointly reduces urban carbon intensity by increasing the level of green technology innovation; (4) the Durbin test suggests that pilot carbon emissions trading policy enforcement can significantly improve the carbon intensity of the area surrounding the city. In summary, the national carbon emissions trading market appears to be a successful experiment that also can contribute to China's sustainable development. Its promise in achieving the "double carbon" target provides important policy implications.

**Keywords:** carbon emission trading pilot; carbon intensity; green technology innovation; environmental governance level

### **1. Introduction**

In the context of economic globalization, climate change is a major challenge for the survival and development of mankind in the 21st century, while the economic development of countries around the world always comes at the cost of energy consumption [1]. The "Statistical Review of World Energy" released by BP shows that global energy demand grew 2.9 percent in 2018, while carbon emissions rose 2.0 percent to reach their highest point in the 21st century. Global primary energy consumption grew 2.9 percent, almost double the average growth rate of 1.5 percent over the past decade [2]. At the same time, carbon emissions from energy consumption grew by 2%, also the highest in years. The new carbon emissions amounted to 600 million tons, which is equivalent to adding a third of the emissions produced by the planet's passenger cars. Therefore, it is of great significance to implement effective means to achieve rapid carbon peaking and net zero emissions [3].

As the world's largest developing country, China has become the world's largest carbon emitter. China's carbon dioxide emissions reached 11.3 billion tons in 2021, accounting for 33 percent of the global total [4]. The Chinese government has announced its intentions to undertake increasingly forceful measures with the goal of achieving a carbon peak before 2030 and carbon neutrality by 2060 [5]. This demonstrates China's determination to achieve

**Citation:** Yu, W.; Luo, J. Impact on Carbon Intensity of Carbon Emission Trading—Evidence from a Pilot Program in 281 Cities in China. *Int. J. Environ. Res. Public Health* **2022**, *19*, 12483. https://doi.org/10.3390/ ijerph191912483

Academic Editors: Paul B. Tchounwou, Taoyuan Wei and Qin Zhu

Received: 4 August 2022 Accepted: 27 September 2022 Published: 30 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

its "dual carbon" goal of carbon emissions and carbon neutrality and to actively undertake the corresponding obligations of its international treaty obligations.

Carbon taxes and emissions trading systems are internationally recognized as effective tools to reduce carbon emissions. According to China's current situation, in order to ensure people's livelihoods, China temporarily does not tax carbon dioxide emitted by coal and natural gas used by individuals. For China, the ETS has become the main tool to reduce carbon emissions.

In December 1997, the Kyoto Protocol was adopted as the first additional agreement to the United Nations Framework Convention on Climate Change (UNFCCC). As part of that agreement, market mechanisms were recognized as a new path to reduce greenhouse gas emission—that is, the right to emit carbon dioxide became regarded as a commodity, thus forming the basis of carbon trading systems [6]. The European Emissions Trading System (EUETS), the world's largest carbon market, came into operation in 2005. The scheme imposes emission limits on member countries; the sum of national emission allowances does not exceed the emissions allowed under the Protocol. The allocation of emission allowances takes into account factors such as historical emissions, projected emissions, and emission standards of member countries [7].

The EU emissions trading system uses "Cap-and-Trade" rules. In order to limit the total amount of greenhouse gas emissions, administrative permits for emissions are bought and sold. The three major principles are the total trade principle, decentralized governance mode, and development characteristics. Under the EUETS, EU member state governments must agree to national emission caps set by the EUETS. Within this cap, companies can sell or buy additional credits in addition to their allocated emissions, provided that overall emissions fall within a specific quota. Firms that emit excess emissions beyond their allocated or purchased allotment are penalized, while those with surplus allowances can keep the emissions for future use or sell them to other firms. The EUETS has played an exemplary role in the world's development of carbon trading markets.

China's carbon market construction started with local pilots [8,9] based on the EUETS. In 2011, the Chinese government listed seven provinces and cities, including Beijing, as pilot areas of the ETS. In 2013, these pilot carbon markets began online transactions. The aim of the program is to cost-effectively reduce greenhouse gas emissions of enterprises in the pilot provinces and cities. The goals include training talent and accumulating experience to lay the foundation for a national carbon market [10]. At present, a national carbon market has started with the power generation industry (2225 enterprises). Eight industries with high energy consumption, including power, petrochemical, chemical, building materials, steel, non-ferrous, paper-making, and civil aviation, will be included in the national carbon market. It is expected to gradually include another seven industries over the 14th Five-Year Plan period.

The carbon emission trading scheme (ETS), regarded as a vital market-driven carbon mitigation instrument, could trigger technology innovation and accelerate a green economic transition [11]. In 2015, China's CO<sup>2</sup> emission from fossil fuel consumption was about 9 billion tons. During the 14th Five-Year Plan period, overall carbon intensity is expected to decrease by 18% and energy consumption per unit of GDP will be reduced by 13.5 percent. Now, the ETS has introduced a system innovation. How to reduce the carbon intensity of cities? What are the pathways that affect carbon intensity? This study will evaluate the ETS policy from the perspective of regional carbon emissions. A thorough review of the pilot policy's impact on carbon emissions, and its relationship to China's overall development, will provide valuable experience for China's efforts to deepen the reform and transformation of its pattern of economic development.

The rest of this study will be divided into the following parts. Part 2 is a literature review. Part 3 is a theoretical hypothesis. Part 4 is the data and empirical framework. Part 5 is the regression analysis. Part 6 further analyzes the mediating effect and spillover effect. Part 7 concludes and makes policy recommendations.

#### **2. Literature Review**

Because carbon emissions cause negative externalities [12], the arguments of Pigou [13] suggest government intervention through means such as taxation. Coase [14] held the opposite opinion, believing that the government should regulate property rights and allow the market to respond to externalities. In both cases, the instruments of the market are used to address externalities. Dales [15] proposed commercializing pollution on the basis of Coase, arguing that the pollution caused by companies is the property of the government, and that businesses should be able to buy and sell freely in the market. This was the embryonic form of the modern emissions trading system.

As mentioned above, although China's ETS has borrowed some practices from the EUETS, it is different. First, the EUETS consists of a "three-pillar" system of "carbon trading", "carbon tax", and "carbon border tax". This is slightly different from a carbon emission quota, which is the basis of carbon emission trading in China. Second, the EU emissions trading scheme adopted a cap-and-trade principle. That is, on the premise that the total amount of emissions does not exceed the allowable upper limit, each emission source can adjust its emissions through exchange of permits. The upper limit will be reduced year by year. By contrast, carbon trading in China is divided into a primary market and a secondary market. The primary market is mainly for "quota creation", which is managed by national authorities and entrusted to agencies to create and distribute carbon emission rights quotas. The participants in the secondary market are mainly enterprises and financial institutions. Third, the trading rules published by the Shanghai Ring Exchange have price fluctuation limits within a daily limit. The EU carbon price, on the other hand, has no price limit. Carbon prices in the European Union have risen rapidly in recent years, more than doubling from pre-pandemic levels. Fourth, the industry coverage of the EU carbon trading system, which started with the power industry and energy-intensive industries, gradually expanded to the transportation sector and the production of specific products such as steel and cement. At present, China's carbon emission trading market is focused on the electric power industry [16,17].

Existing research on emissions trading can be broadly divided into two categories. The first category focuses on assessing the efficiency of the ETS design, including the effectiveness of a carbon price in reducing emissions [18,19], the controllability of transaction costs [20,21], and the rationality of quota allocation [22,23]. The second category focuses on how the ETS affects macroeconomic variables. This study is in the second category.

From the perspective of energy conservation and emission reduction, earlier studies mostly used scenario simulation to evaluate carbon emission trading. In terms of energy saving, most scholars have used data simulation analysis. It has been found that ETS can effectively reduce the consumption of non-renewable energy [24,25]. In terms of emission reduction, Zhang et al. [26] simulated ETS implementation in China and found that interregional commodity exchanges can alleviate carbon emissions, based on China's provincial panel data [27]. The simulations were analyzed in the case of both unconstrained and constrained countries to assess the potential effectiveness of ETS in China. The study found that ETS had the potential to reduce carbon intensity by 20.06% without having a negative effect on GDP.

The development of the ETS systems in Europe and China provides the opportunity to turn the simulation into reality. Most studies have found that ETS has reduced carbon in pilot areas in China. Computable General Equilibrium (CGE) and Difference in Difference (DID) models have been the main empirical evaluation methods used in recent years. Liu et al. [1], using a regional CGE model, found that the Hubei province pilot ETS reduced carbon emissions by about 1% in 2014. In an empirical study, Yucai et al. [28] used DID to model the effect of the pilot ETS on energy conservation and emissions reduction; the results showed that regulated industry energy consumption in the ETS pilot areas decreased by 22.8% and carbon emissions by 15.5%.

Some scholars also have studied the possible economic losses caused by the implementation of ETS. Most scholars have found that EUETS has had no adverse effect on

corporate profits and social welfare [29]. For China's carbon trading market, however, Wang and Pan [30] found that the implementation of ETS has led to a 0.28% decline in GDP. This is because China's economic development has been dependent on natural resources. Hubler et al. [31] found that the economic losses of ETS in China may be around 1%.

In conclusion, the existing papers mainly study the impact of ETS on energy saving, emission reduction, and economic loss. However, there are few studies on comprehensive macro indicators, such as urban carbon emission intensity. Urban carbon emission intensity is defined as the ratio of CO<sup>2</sup> emissions to GDP in a city within a year. This indicator has been widely used to evaluate China's "double carbon" target [32].

In China, most studies on ETS use a CGE model or a DID model. There are a number of limitations with these studies. CGE modeling is subject to defects such as difficulty in meeting the assumptions on which it is premised, strong subjectivity of parameter setting, and difficulty in determining whether its feedback mechanism measures real effects. The DID model requires homogeneity of the sample, whereas in reality, there is heterogeneity in relevant characteristics between the treated and control localities. In addition, most of the relevant studies start from the provincial level, while implementing carbon emission trading policies depends more on whether urban units can strictly implement the orders of their superiors. Further, earlier studies have ignored the influence of spatial factors on carbon intensity, although spatial factors have an important impact on carbon intensity and neglecting spatial factors may lead to bias in simulation results.

Against this background, this study makes the following contributions. First, the introduction of pilot carbon-emission trading policies is regarded as a quasi-natural experiment. This allows the use of a PSM-DID model estimation method to assess the impact of ETS on urban carbon intensity. The quasi-natural experiment not only meets the requirements of a DID model, but also ensures optimal matching because of the large samples. This gives more credibility to the research conclusions. Second, this study focuses on carbon intensity at the city level. Considering that cities are an important part of local government institutions in China, this makes the policy effect more plausible. Third, this study uses spatial Durbin to test the spillover effect of ETS on surrounding areas, thus going beyond the previous focus on the local area, which has ignored the surrounding area. This provides a more complete picture of the impact of emissions trading policies.

#### **3. Theoretical Background**

The carbon emission trading system is mainly an exercise of the "Porter hypothesis," which holds that appropriate environmental regulation can encourage enterprises to carry out more innovative activities [2]. These innovations will increase the productivity of firms, thereby offsetting the costs of environmental protection and reducing total carbon emissions at the societal level. Theoretically, the system is dominated by the government, which uses market mechanisms to promote energy conservation and emission reduction [33]. First, the ETS sets a relatively strict carbon allowance for each company. Within this limit, companies can carry out free carbon emissions. The excess needs to be purchased from the carbon emissions retained by other companies. In essence, carbon permits have become a commodity [34]. Because firms aim at profit maximization, companies make good use of free credits while trying to avoid exceeding that limit; otherwise, high production costs will be incurred. Second, ETS can promote corporate emission reduction because firms will sell unused emissions credits if the carbon price is higher than the firm's marginal cost of emission reduction [35]. Therefore, a market-based trading system can be effective in mitigating carbon emissions.

Establishing a carbon emission trading system can force enterprises to innovate, and technological progress is one of the three major factors affecting the environment [36]. Green technology innovation depends on increasing investment in such innovation. The emissions trading system encourages companies to actively develop and apply green technologies [2]. Companies that invest more resources in reducing carbon emissions can sell surplus carbon emission credits to high-carbon emission enterprises and obtain high

profits [37]. Firms will tend to accelerate the process of green technology development in order to achieve higher profits. This is the incentive effect. Conversely, for high carbon emission enterprises, it is necessary to buy carbon emission credits from sellers, which will increase production costs, compress profit margins, and reduce these firms' competitiveness. Under this pressure, enterprises have to carry out technological innovation [2]. This is the punishment effect.

The incentive effect and punishment effect of market-based environmental regulation such as ETS give the government more tools for environmental governance. Because carbon dioxide does not harm health or production in the short run, and because it is costly to enforce non-market forms of governance, it has been difficult for the focus of environmental governance to shift quickly in the direction of reducing carbon emissions. By encouraging innovation and providing opportunities for profit, ETS has effectively improved the environmental governance level while ensuring that normal activities and production can continue.

Improved environmental governance can promote a change of regional energy structure. In particular, ETS has the potential to reduce coal consumption [28]. This paper applies the new economic geography to evaluate such changes. Firms will always look for the optimal location in order to maximize profits [38]. Theoretically, a carbon emission trading system should have policy spillover effects [39], including alleviating regional carbon emissions. It can also encourage high-tech enterprises to continue to innovate through its incentive mechanism. However, it will also cause a large number of polluting enterprises to incur high production costs due to its punishment mechanism. This is because polluting enterprises in the region face increased production costs due to the need to buy carbon emission rights, which reduces their profits. If there is no ETS policy in the surrounding areas, polluting enterprises are expected to migrate to the surrounding areas. Conversely, high-tech firms from surrounding areas are expected to migrate to the ETS area in order to increase their profits by selling carbon credits. The transfer behavior of the two types of enterprises can reduce carbon emissions in the ETS region while increasing emissions in the area around the ETS [40]. The theoretical background of this study is shown in Figure 1. Accordingly, the following hypothesis is proposed:

**Hypothesis 1 (H1).** *The carbon emissions trading system has reduced the carbon intensity of the pilot cities in China*.

**Hypothesis 2 (H2).** *Green technology innovation is one mechanism through which the carbon emissions trading system reduces regional carbon intensity*.

**Hypothesis 3 (H3).** *Improving urban environmental governance is another mechanism through which the carbon emission trading system alleviates regional carbon intensity*.

**Hypothesis 4 (H4).** *The carbon emission trading system has increased the carbon intensity of the areas surrounding the pilot cities*. *Int. J. Environ. Res. Public Health* **2022**, *19*, x FOR PEER REVIEW 6 of 20

**Figure 1.** Theoretical background of the study. emission data, and is not based on carbon trading schemes, it should be considered post **Figure 1.** Theoretical background of the study.

**4. Data and Methodology**  *4.1. Data Sources* 

2021).

*4.2. Variable Selection* 

duced during energy consumption.

௧ = ∑ ௩௧

Most of the data in this study come from data already publicly available in China, including the National Bureau of Statistics (https://data.stats.gov.cn/, accessed on 15 August 2021), the China City Statistical Yearbook, the China Energy Statistical Yearbook, and the Statistical Yearbooks of 281 cities. The data were drawn from reports on social development, including statistics on the main energy consumption of local industries above city size, total industrial output value, urbanization rate, etc. Patent data were derived from the State Intellectual Property Office (https://www.cnipa.gov.cn/, accessed on 20 August

Urban carbon intensity is based on the total amount of carbon emissions to be measured. In this study, a material balance algorithm was used to calculate the total carbon emissions. Carbon emissions are estimated using the chemistry of carbon dioxide pro-

௩௧ is the annual actual consumption of the V type of energy in the city in year t. According to the 26 fossil fuels listed in the China Energy Statistical Yearbook, they are combined into nine final energy consumption types: coal, coke, crude oil, gasoline, kerosene, diesel, fuel oil, natural gas, and electricity. Because electricity is not a direct energy source, the concept of secondary energy reflects the fact that electricity is produced by consuming other energy; therefore, this study will not measure electricity separately. ௩, ௩, ௩ are the energy calorific value conversion coefficient, carbon emission coefficient, and carbon oxidation factor, respectively. The data come from the average low calorific value of the China Energy Statistical Yearbook and IPCC (2006). As (44/12) is known to be the ratio of carbon dioxide to carbon molecular weight, the carbon dioxide emissions of 281 cities in China from 2006 to 2019 can be calculated. Since this study uses historical CO2

௩ୀଵ × ௩ × ௩ × ௩ × 44/12 (1)

Through screening and matching, this paper selected panel data of 281 cities in China from 2006 to 2019 as the research object. A total of 37 cities at the prefecture level and

#### **4. Data and Methodology**

#### *4.1. Data Sources*

Through screening and matching, this paper selected panel data of 281 cities in China from 2006 to 2019 as the research object. A total of 37 cities at the prefecture level and above were designated as pilot carbon emission trading cities. These 37 cities constitute the experimental group, and the remaining cities were analyzed as the control group. Most of the data in this study come from data already publicly available in China, including the National Bureau of Statistics (https://data.stats.gov.cn/, accessed on 15 August 2021), the China City Statistical Yearbook, the China Energy Statistical Yearbook, and the Statistical Yearbooks of 281 cities. The data were drawn from reports on social development, including statistics on the main energy consumption of local industries above city size, total industrial output value, urbanization rate, etc. Patent data were derived from the State Intellectual Property Office (https://www.cnipa.gov.cn/, accessed on 20 August 2021).

#### *4.2. Variable Selection*

Urban carbon intensity is based on the total amount of carbon emissions to be measured. In this study, a material balance algorithm was used to calculate the total carbon emissions. Carbon emissions are estimated using the chemistry of carbon dioxide produced during energy consumption.

$$\mathcal{C}arbon\_{it} = \sum\_{\upsilon=1}^{n} \mathcal{Q}\_{\upsilon t} \times \mathcal{W}\_{\upsilon} \times M\_{\upsilon} \times R\_{\upsilon} \times 44/12 \tag{1}$$

*Qvt* is the annual actual consumption of the V type of energy in the city in year t. According to the 26 fossil fuels listed in the China Energy Statistical Yearbook, they are combined into nine final energy consumption types: coal, coke, crude oil, gasoline, kerosene, diesel, fuel oil, natural gas, and electricity. Because electricity is not a direct energy source, the concept of secondary energy reflects the fact that electricity is produced by consuming other energy; therefore, this study will not measure electricity separately. *Wv*, *Mv*, *R<sup>v</sup>* are the energy calorific value conversion coefficient, carbon emission coefficient, and carbon oxidation factor, respectively. The data come from the average low calorific value of the China Energy Statistical Yearbook and IPCC (2006). As (44/12) is known to be the ratio of carbon dioxide to carbon molecular weight, the carbon dioxide emissions of 281 cities in China from 2006 to 2019 can be calculated. Since this study uses historical CO<sup>2</sup> emission data, and is not based on carbon trading schemes, it should be considered post hoc analysis, and therefore calculation errors caused by different ways of allocating emission reduction targets can be avoided.

By referring to relevant literature and considering the actual situation [34], the following control variables were selected to conduct propensity matching scores: regional economic development level (PGDP), industrial structure (IND), urban population (PP), degree of openness to the outside world (OPEN), efficiency of financial development (FS), scale of financial development (FD), and government environmental intervention (WODK). The specific calculation methods are shown in Table 1.

**Table 1.** Description of variables.


#### *4.3. Model Setting*

The reference measurement model of this paper is set as follows:

$$\text{Carbon}\_{\text{il}} = \mathfrak{a}\_0 + \mathfrak{a}\_1 \text{treated}\_{\text{il}} \* \text{time}\_{\text{il}} + \sum\_{i=1}^{N} \mathfrak{beta}\_j \text{control}\_{\text{il}} + \mathfrak{\mu}\_i + \gamma\_t + \varepsilon\_{\text{il}} \tag{2}$$

where *i* represents the individual city and *t* represents the year. *Carbonit* is the carbon emission of city *i* in year *t*. The year dummy variable *timeit* takes a value of 0 before the introduction of the carbon emission trading pilot policy (the policy impact point is set as 2013) and 1 after the establishment, and *treatedit* is a group dummy variable. The ETS pilot cities are assigned a value of 1, non-ETS pilot cities are assigned a value of 0, and *treatedit* ∗ *timeit* is the interaction term of the two and takes the value of 0 or 1. Here, 1 represents the pilot cities after 2013, and 0 represents non-pilot cities and the pilot cities before 2013. The coefficient *α*<sup>1</sup> before the interaction term of *treatedit* ∗ *timeit* is an important explanatory variable that represents the policy effect of emissions trading on urban carbon intensity. This paper will introduce various control variables affecting urban carbon intensity into the multi-stage DID regression. The bidirectional fixed effect of city and year will be introduced for further analysis.
