Risk Spillovers between China’s Carbon and Energy Markets

Abstract

In recent years, with the intensification of global warming and the greenhouse effect, the global consensus has focused on efficient, clean, low-carbon, and green development as a means of achieving new economic growth. China, as a major carbon emitter, has been at the forefront of efforts to reduce carbon emissions. The establishment of the carbon emissions trading market, commonly known as the “carbon market”, provides an economic solution for reducing carbon emissions in both the carbon and energy markets. As China’s carbon market continues to grow rapidly, fluctuations in the energy or carbon markets caused by information shocks can easily spread between the two markets, leading to increased interconnectedness. Moreover, the spillover effect of the volatility between China’s carbon market and energy market is not constant, and the intensity and direction of this effect vary depending on different market volatility levels and periods. Therefore, it is crucial to conduct a comprehensive study on the characteristics of the volatility spillover effect between China’s carbon market and energy market and to fully understand the mechanism of energy regulation on carbon prices. This research will have significant practical implications for promoting the establishment of a well-functioning internal price transmission mechanism between China’s carbon market and energy market. This study took the risk spillover between the carbon market and energy market as the research object and systematically combed through its pricing mechanism and spillover impact. Through constructing the DY overflow index model based on a VAR model and generalized variance decomposition method, this study explored the linkage between China’s carbon and energy markets, i.e., the linkage of price fluctuations between China’s energy and carbon markets, as well as the time-varying nature of inter-market spillovers, and provides suggestions on the risk control of price fluctuations between the carbon and energy markets.

1. Introduction

Since the first industrial revolution, energy, as the driving force of industrial civilization, has played a pivotal role in the development of countries and regions around the world. In the late 1860s, the second industrial revolution led to the rapid development of industries and technology, with fossil fuels as the main source of energy triggering the greenhouse effect, and thus, affecting the ecological system of mankind and causing significant impacts on human economic and social sustainable development [1].

The carbon market is modeled after the aggregate control and trading regulation that was used to successfully reduce sulfur pollution in the 1990s. Enterprises can buy or sell a certain number of carbon credits in the carbon market. Carbon credits and carbon trading are authorized by the government, with the aim of gradually reducing the total amount of carbon emissions and mitigating their impact on climate change [2].

The United Nations Framework Convention on Climate Change (hereinafter referred to as the Convention), which was signed by the United Nations Conference on Environment and Development in 1992, stipulates that developed countries should take the lead in reducing emissions and provide financial and technical support to developing countries [3]. In 1997, the third meeting of the Convention signed the Kyoto Protocol, which established a binding and flexible carbon emission control and carbon emission quota market [4,5,6,7,8]. In 2000, the EU released the Green Paper on Greenhouse Gases, which stipulates the trading of carbon dioxide emission rights, and thus, realizes the control of greenhouse gases. In 2003, the EU issued the Emissions Trading Directive and took the lead in setting up the Carbon Emissions Trading System (EU ETS) in 2005, which requires 11,000 energy-intensive high-emissions companies to trade their carbon dioxide emission rights, covering 46% of Europe’s carbon dioxide emissions [9,10,11,12,13,14,15]. According to the 2020 Global Carbon Market Progress Report, by the end of 2020, the world had formed 21 carbon markets, covering 16% of the global share of carbon dioxide emissions [1]. After 2020, the Paris Agreement became the general platform for global emissions reduction.

Excessive carbon emissions pose a serious threat to the ecological environment. As the world’s largest energy consumer and carbon emitter, China attaches great importance to the energy and environmental problems caused by carbon emissions and actively participates in the construction of the global climate governance system. The carbon market is recognized as an effective policy tool for controlling global carbon emissions, which will have a broad and far-reaching impact on the energy supply and demand structure. Also, it is of great significance to the international carbon emissions governance to establish and continuously maintain a sound nationwide carbon market to facilitate China’s national economic growth to gradually get rid of the growth mode of high energy consumption, high pollution, and high capital consumption (Tyndall Centre for Climate Change Research).

Since 2011, the National Development and Reform Commission (NDRC) has issued the Notice on Carbon Emissions Trading Pilot Work, identifying nine provinces and municipalities, including Beijing, Tianjin, Shanghai, Chongqing, Wuhan, Guangdong, and Shenzhen, to launch carbon emissions trading pilots one after another, adopting online trading in the form of both listed agreements and block agreements. In December 2017, the National Development and Reform Commission (NDRC) issued the National Carbon Emissions Trading Market. On 22 September 2020, President Xi made a solemn promise to the world during the general debate of the 75th United Nations General Assembly, stating that China would increase its national autonomous contribution, adopt more vigorous policies and measures, strive to achieve peak carbon dioxide emissions by 2030, strive to realize the “carbon footprint” by 2060, and continue its efforts in the area of carbon emissions trading.

The carbon market is a market-based means of carbon emissions reduction, and the carbon price is an effective medium to measure the carbon market [16,17,18,19,20,21,22,23]. The carbon market price has a non-normal distribution, with sharp peaks, thick tails, and other typical financial time series characteristics, and correlation with the energy market. The risk effect spillover mechanism between the two lies in the fact that price fluctuations in the energy market can affect carbon prices through changes in industrial CO₂ emissions. When the carbon market price changes affect the cost of industrial processes, enterprises are likely to adjust their energy consumption and optimize their energy-saving and emissions-reducing technologies, which, in turn, affects the energy market [23,24,25,26,27,28,29,30].

In addition, the outbreak of the COVID-19 epidemic at the end of 2019 was a large-scale emergent event that has affected the world. The outbreak of the epidemic reduced energy demand in the international market, leading to a decline in energy prices, which, in turn, seriously affected carbon market prices.

In summary, in the context of global integration, the carbon market and energy market price risk spillover can easily cause cross-country, cross-market, and cross-industry systemic risks. Strengthening related research is conducive to optimizing resource allocation, stabilizing prices, and controlling systemic risks [15,31,32,33,34,35]. In view of this, this study focused on the risk spillover effect between the energy market and carbon market by effectively identifying, measuring, and controlling the risk spillover characteristics between the energy market and carbon market as the core issue. The significance of this research mainly lies in the following: First, in terms of theoretical significance, most of the previous studies on the carbon market focused on the carbon market of the EU countries and other developed economies, and the studies on the risk of China’s carbon market lag far behind the development of the carbon market. This study tried to provide some reference for China to standardize the carbon market trading system; establish a reasonable price stabilization mechanism; improve the identification, measurement, prevention, and control mechanism of carbon and energy market risks; and enhance the function of carbon market in regulating the energy consumption structure through this research. Second, in terms of the practical significance of this study, the systematic internal or external events and the non-systematic events may lead relevant industries to experience significant fluctuations, which, in turn, may lead to systematic fluctuations through chain reactions.

2. Literature Review

Greenhouse gas emissions reduction is a common challenge faced by the whole world, and the carbon emissions trading market restricts carbon emissions through market-oriented means, such as “total control” and “quota trading”. A large amount of previous research focused on the operation mechanism of the carbon market and the principles, impacts, risk identification, and control of carbon market risk formation.

There is a price correlation between the energy market and the carbon market. Coal, oil, and natural gas are China’s main energy demand, and their price changes have a direct effect on the carbon market. Coal combustion is the main cause of carbon emissions, and natural gas is a clean energy source. Due to the relative concentration of coal reserves in China, especially in some provinces, it became the main fuel in China in the early days. As the regulation of carbon emissions in China has been strengthened, the price of carbon has risen to stimulate companies to switch to cleaner sources of energy, creating an energy substitution effect. Carbon markets regulate global carbon emissions through economic methods, and their marginal effect is to promote the global low-carbon economic transition [7]. Under the income effect and substitution effect, enterprises are stimulated to realize the goal of reducing carbon emissions by adjusting the energy consumption structure and upgrading technological innovation [8]. The carbon emissions trading market utilizes market-oriented approaches, such as “comprehensive control” and “cap-and-trade”, to effectively limit carbon emissions. Previous research has predominantly concentrated on understanding the operational mechanisms of the carbon market, as well as the principles, impacts, and methods of identifying and controlling the risks associated with carbon market formation [36,37].

Qiao et al. [38] analyzed the two-way spillover effect between China’s carbon market and the energy market by combining the rolling window technique with the DY spillover index for four representative pilot carbon markets, namely, Beijing, Guangdong, Hubei, and Shanghai, and the results show that the spillover effect between China’s carbon market and the energy market had significant temporal variability and asymmetry; in addition, the spillover effect of different carbon pilot markets on the energy market had regional heterogeneity. In addition, the spillover effects of different carbon pilot markets on the energy market are regionally heterogeneous, with the volatility spillover effects of the Beijing and Shanghai carbon markets mainly coming from the crude oil futures market, the Guangdong carbon market mainly coming from the new energy market, and the Hubei carbon market mainly coming from the crude oil and electricity markets. Mehmet Balcılar et al. [1] used a Markov dynamic model and vector autoregressive model to verify that risk spillover between energy and carbon markets has significant volatility and a time-varying nature and that futures markets are relevant in controlling the risk of spot markets, which has significant implications for national policymaking and investment. Using the Diebold and Yilmaz model, which is the method of constructing a spillover index via variance decomposition of forecast errors, Wang and Guo [2] showed that there is an asymmetric spillover effect in the sequence of returns and volatility between carbon and energy markets. Roman V. Klyuev considered that when analyzing the forecasting methods at all relevant levels, there is a tendency toward the active use of neural networks. This trend has arisen because neural networks allow for taking into account the non-linear nature of power consumption data and finding non-trivial dependencies in them. All of this allows for an improvement in the accuracy of operational, short-term, medium-term, and long-term forecasting of electricity consumption [39]. Specifically, among the three major energy markets, WTI crude oil, Brent crude oil, and natural gas, the WTI crude oil market had the strongest spillover effect on the carbon market, and the natural gas market had a more prominent spillover effect on the carbon market; furthermore, he used the rolling window technique to detect the time-varying characteristics of the spillover effect, and the results show that major policies and events could lead to large changes in the spillover index. According to the existing literature, due to the late establishment of China’s carbon trading market, the domestic and foreign literature mostly focuses on the more mature EU carbon market, and mostly qualitatively analyzes the correlation between the carbon market and the energy market through the establishment of a model; however, most of the research on China’s carbon market has only explored the factors influencing the price of carbon, the carbon financial pricing mechanism, and the construction of a nationwide carbon market and a carbon financial market, while the internal and external factors of the carbon market and the energy market are not discussed, and the carbon market and the energy market are not discussed in detail [40,41,42,43,44,45,46]. However, most studies on China’s carbon market have only explored issues such as the carbon price influencing factors, carbon financial pricing mechanism, and construction of a national carbon market and carbon financial market, and have not examined the internal and external systemic risk spillover effects of the carbon market and energy market. Therefore, it is of strong theoretical and practical value to systematically investigate the linkages and time-varying nature of China’s carbon and energy markets and to propose risk spillover identification, measurement methods, and risk control recommendations.

As a result, the marginal contribution of this study is mainly reflected in two aspects. First, in terms of the research object, this study took the risk spillover effect between China’s carbon market and energy market as the research object and systematically explored the linkage relationship and dynamic characteristics between China’s carbon market and energy market, as well as the time-varying nature of price fluctuations. Second, in terms of research methodology, this study evaluated the gross, directional, and net spillovers of the risk between the two markets through the spillover index model proposed by Diebold and Yilmaz [4], which combines qualitative and quantitative methods to judge the correlation and time-varying nature of the risk between the two markets, and enriched the assessment methods of the existing inter-market risk spillover research.

3. The Model and Data

3.1. Model

In recent years, research on the spillover effects between the carbon and energy markets’ volatility has received extensive attention from academics and the industry, and relevant quantitative analysis methods and assessment models have been widely used. This study empirically analyzed the risk spillover effect between China’s carbon market and energy market based on the spillover index of Tang [47]. First, we constructed a stable vector autoregressive (VAR) model with N variables and P orders [48,49]:

$$x_t={{\displaystyle \sum }}_{i=1}^p{\varnothing }_i{x}_{t-i}+{\epsilon }_t$$

(1)

where x_t is an N-dimensional variable that satisfies smoothness and obeys the vector autoregressive equation shown in Equation (1); ε_t is an N-dimensional vector that is independently and identically distributed and obeys a mean of 0 and variance of ∑. The above vector autoregressive (VAR) form can then be transformed into a vector moving average (VMA) form:

$$x_t={{\displaystyle \sum }}_{i=1}^{\infty }{A}_i{\epsilon }_{t-1}$$

(2)

When i < 0, A_i is an N-dimensional unit matrix, and A_i satisfies the above recursive equation when i is other values. On this basis, the standard error of prediction under a given prediction horizon h is decomposed, and the degree of contribution of each variable to the standard error of prediction can be calculated.

In order to avoid the Cholesky decomposition in the model that leads to the orthogonalization of the residual terms affected by the order of the variables, Diebold and Yilmaz [4] adopted the generalized error decomposition method; in this framework, the variance contribution of the prediction error is rearranged and estimated, which is calculated as follows:

$${\theta }_{ij}\left(H\right)={\delta }_{ij}^{-1}\frac{{{\displaystyle \sum }}_{h=0}^{\phi -1}{\left(e_i^{’}{A}_h\sum e_j\right)}^2}{{{\displaystyle \sum }}_{h=0}^{\phi -1}{e}_i^{’}{A}_h\sum A_h^{’}{e}_j}$$

(3)

where ∑ is the variance matrix of the prediction error vector ε_t and δ_ij is the standard deviation of the error term of the jth equation. And e_i is an N-dimensional vector with the ith element being 1 and all other elements being 0. Since under the generalized error decomposition method, the contribution of all variables to the equation does not necessarily add up to equal 1, in order to make the different θ_ij(H) comparable, it is necessary to standardize θ_ij(H) in the following way:

$${\tilde{\theta }}_{ij}\left(H\right)=\frac{{\theta }_{ij}\left(H\right)}{{{\displaystyle \sum }}_{j=1}^N{\theta }_{ij}\left(H\right)}$$

(4)

where N is the total number of dimensions of the variable. In turn, the normalization obtained by the calculation satisfies the following equation:

$${{\displaystyle \sum }}_{j=1}^{N~}\theta \left(H\right)=1$$

(5)

$${{\displaystyle \sum }}_{i,j=1}^N\tilde{\theta }\left(H\right)=N$$

(6)

$\tilde{\theta }ij\left(H\right)$ is defined as the variance contribution, whose value is the contribution of variable j to the total prediction error variance of variable i and takes a value between 0 and 1. When the value is higher, it reflects that variable j exerts a more significant influence on the total prediction error variance of variable i. Further, by summing up $\tilde{\theta }ij\left(H\right)$, the core indicator total spillover index can be constructed.

The total spillover index measures the strength of the mutual spillover effect between industries in explaining the total forecast error variance, reflecting the degree of the market contagion effect. When the total spillover index is relatively high, it indicates that the mutual spillover effects among variables are large and the variables are relatively closely linked to each other. The total spillover index S(H) is obtained by summing up $\tilde{\theta }ij\left(H\right)$, except for i = j, and dividing by the total dimension N:

$$S\left(H\right)=\frac{{{\displaystyle \sum }}_{i,j,i\ne j}^N{\tilde{\theta }}_{ij}\left(H\right)}{N}\times 100$$

(7)

The total spillover index S(H) is an average indicator, reflecting the average level of linkages between variables. And by rearranging the above formula, indicators reflecting the spillover effect of a single variable on other variables and indicators reflecting the spillover effect of a single variable on other variables can be obtained [50,51,52,53,54,55,56].

3.2. Data

Risk spillover between the carbon market and energy market is mainly manifested as risk spillover of return and volatility. So far, China has initially formed an operating national carbon market framework consisting of trading platforms such as the Shanghai Environment and Energy Exchange, Shenzhen Carbon Emission Exchange, Beijing Carbon Emission Rights Electronic Trading Platform, Guangzhou Carbon Emission Exchange, Hubei Carbon Emission Rights Trading Center, Tianjin Carbon Emission Rights Exchange, Chongqing Public Resources Trading Network, and Fujian Carbon Emission Trading Network. Combined with China’s energy consumption structure and the availability of price indices, this study chose Beijing, Shanghai, Guangdong, Shenzhen, Tianjin, Hubei, and Chongqing, which are representative carbon markets, as the research object, and selected the price indices of the four energy sources with a large proportion of consumption, namely, coal, crude oil, natural gas, and liquefied natural gas, from the databases of Wind and Choice. The sample intervals for both the carbon market and energy market indicators started on 1 January 2018 and ended on 31 December 2022. To maintain the consistency of the time series and avoid unnecessary data deletion, we calculated the return rate of each market on each trading day: $R_{t,i}=\frac{\left(P_i-P_{i-1}\right)}{P_{i-1}}$ (t indicates the week and i indicates the day). And then, we obtained the realized volatility statistic for week t: ${\sigma }_t^2={\displaystyle \sum }_{i=1}^{W_t}{R}_{ti}^2$, where Wt indicates the trading days in week t.

Table 1. Descriptive statistics and series stationary test.

Carbon Market	Variable	Mean Value	Standard Deviation	Minimum Value	Maximum Value	Skewness	Kurtosis	JB Statistic	ADF Trace Statistic
Beijing	Return	0.0025	0.0295	−0.1298	0.1834	0.8697	11.9360	7241.4953 ***	−22.189 ***
Shanghai		0.0025	0.0343	−0.1159	0.3643	4.1816	43.9120	435,637.9111 ***	−22.019 ***
Guangdong		0.0184	0.1003	−0.1609	1.2222	6.8721	68.9707	20,391.8219 ***	−18.844 ***
Shenzhen		0.0416	0.1768	−0.2281	2.0433	6.4619	59.4070		−17.465 ***
Tianjin		0.0009	0.0174	−0.0977	0.2143	4.8494	70.4897		−16.381 ***
Hubei		0.0008	0.0154	−0.0855	0.0832	−0.3251	11.8631	436,340.7191 ***	−21.896 ***
Chongqing		0.0027	0.0466	−0.1667	0.1992	0.2011	6.4314		−15.367 ***
Beijing	Volatility	0.0180	0.0318	0.0000	0.3026	3.3376	22.2667		−12.550 ***
Shanghai		0.0247	0.1821	0.0000	3.3175	17.0259	306.2195		−18.747 ***
Guangdong		0.3182	2.4024	0.0000	42.7835	15.9799	279.5159		−17.824 ***
Shenzhen		0.9918	7.6705	0.0000	134.1203	15.3668	261.0747		−16.251 ***
Tianjin		0.0087	0.0655	0.0000	1.1480	15.4196	263.2136		−18.332 ***
Hubei		0.0073	0.0146	0.0000	0.1739	5.6689	53.5639		−15.164 ***
Chongqing		0.0284	0.0453	0.0000	0.2914	2.1499	8.0544		−9.320 ***

Footnote: *** indicates significance at the levels of 1%, respectively.

provides the descriptive statistics of the rate of return and volatility of carbon emissions trading prices in each market.

As shown in Table 1, in terms of return, Shenzhen carbon market had the highest return performance, followed by Guangdong carbon market and Hubei carbon market, and the carbon market return varied greatly, with the average value of return in Shenzhen carbon market being more than 50 times of that in the Hubei carbon market. In terms of volatility, the Shenzhen carbon market had the highest volatility, followed by Guangdong, Chongqing, Shanghai, Beijing, and Tianjin, and the return and volatility levels of each carbon market basically satisfied the characteristics of high return–high risk. From the skewness and kurtosis of each carbon market, it can be seen that the return and volatility series of carbon markets were right-skewed distributions with sharp peaks and thick tails.

4. Empirical Results and Analysis

4.1. Return Spillovers

We examined the static spillover effect of risk across the Chinese carbon markets and World carbon market in Figure 1 and Figure 2, as well as the return dynamic spillover index in the Chinese carbon market and the world carbon market total volatility spillover, which included the directional spillover and net spillover of market yields. In terms of the directionality of yield spillovers, the carbon markets in Shanghai and Guangdong were asymmetric.

In

Table 2. Carbon market return spillover index table.

Carbon Market	Beijing	Shanghai	Guangdong	Shenzhen	Tianjin	Hubei	Chongqing	Internal Spillover
Beijing	98.44	0.41	0.2	0.74	0.02	0.09	0.09	1.56
Shanghai	0.05	97.1	1.1	1.46	0.03	0.08	0.17	2.9
Guangdong	0.15	1.75	95.99	0.14	0.15	1.58	0.24	4.01
Shenzhen	0.99	1.34	0.2	96.72	0.18	0.44	0.12	3.28
Tianjin	0.56	0.2	0.2	0.16	98.32	0.17	0.39	1.68
Hubei	0.2	0.68	5.8	0.26	0.2	92.01	0.85	7.99
Chongqing	0.04	1.4	0.28	1.04	0.26	0.22	96.75	3.25
External spillover	1.99	5.79	7.79	3.8	0.85	2.59	1.86	24.67
Net spillover	99.43	98.9	98.78	97.51	99.17	94.6	98.61

, specifically, by calculating the spillover indices of the yields of each carbon market, we found that the external spillover indices of Shanghai and Guangdong were 5.79% and 7.79%, respectively, while the receiving spillover indices of each market were 2.9% and 4.01%, respectively. It can be seen that all the carbon markets had a positive net spillover and were close to 100 percent, indicating that they had strong carbon market pricing ability and risk spillover effectiveness. Combining the positive and negative spillovers, the total internal and external spillover of the carbon markets in Beijing, Shanghai, Guangdong Shenzhen, Tianjin, Hubei, and Chongqing was 24.67%, indicating a strong correlation between risk spillovers of regional carbon markets.

4.2. Volatility Spillovers

Table 3. Carbon market volatility spillover index table.

Carbon Market	Beijing	Shanghai	Guangdong	Shenzhen	Tianjin	Hubei	Chongqing	Internal Spillover
Beijing	95.03	0.47	0.80	0.20	1.20	2.05	0.25	5.00
Shanghai	3.24	18.58	0.08	0.05	71.63	5.88	0.54	81.40
Guangdong	6.24	0.33	92.53	0.18	0.24	0.26	0.22	7.50
Shenzhen	3.00	0.05	0.19	96.49	0.18	0.05	0.04	3.50
Tianjin	2.31	0.95	0.10	0.14	89.63	6.22	0.66	10.40
Hubei	0.58	0.23	0.22	0.12	0.31	97.12	1.42	2.90
Chongqing	0.14	0.92	2.46	0.42	5.26	2.14	88.66	11.30
External spillover	15.50	2.90	3.90	1.10	78.80	16.60	3.10	17.40
Net spillover	10.50	−78.50	−3.60	−2.40	68.50	13.70	−8.20

demonstrates the spillover effects of volatility across carbon markets, with internal spillovers (excluding the outlier of Shanghai) of 5%, 7.5%, 3.5%, 10.4%, 2.9%, and 11.3% in Beijing, Guangdong, Shenzhen, Tianjin, Hubei, and Chongqing, respectively, and external spillovers (excluding the outlier of Tianjin) of 15.5%, 2.9%, 3.9%, 1.1%, 16.6%, and 3.1%, respectively. Overall, without considering the two anomalies of Tianjin and Shanghai, the Beijing and Hubei markets had the strongest net spillover effects on other markets, at 10.5% and 13.7%, respectively, which suggests that these two markets were of high systemic importance to the operation of China’s carbon market, i.e., price fluctuations and risk events generated by the Beijing and Hubei carbon markets were more likely to contribute to the overall market volatility and systemic risk. In Figure 3 and Table 3, in terms of the spillover intensity between the two markets, the Beijing carbon market had the largest spillover effect on the Guangdong carbon market (6.24%), followed by the Hubei carbon market’s spillover effect on the Tianjin and Shanghai carbon markets (6.22% and 5.88%).

4.3. Predicted Output Values for the Short Term and Long Term

According to the data presented in Table 1, it is evident that neither the independent variable nor the dependent variable could satisfy the normal distribution assumption. Thus, we can consider the time series regression model to predict the long-term and short-term output values of the return and volatility of the price index in the carbon market. This model captures the correlation and trend over time, making it suitable for forecasting time series data.

Table 4. Predictive output values of carbon market return.

Predictive Value	Posterior Stage	Posterior 2 Stage	Posterior 3 Stage	Posterior 4 Stage	Posterior 5 Stage	Posterior 6 Stage	Posterior 7 Stage	Posterior 8 Stage	Posterior 9 Stage	Posterior 10 Stage	Posterior 11 Stage	Posterior 12 Stage
Mean value	0.060	0.060	0.060	0.060	0.060	0.060	0.060	0.060	0.060	0.060	0.060	0.060
Standard deviation	0.010	0.010	0.010	0.010	0.010	0.010	0.010	0.010	0.010	0.010	0.010	0.010
Minimum value	−0.141	−0.141	−0.141	−0.141	−0.141	−0.141	−0.141	−0.141	−0.141	−0.141	−0.141	−0.141
Maximum value	0.616	0.616	0.616	0.616	0.616	0.616	0.616	0.616	0.616	0.616	0.616	0.616
Skewness	1.396	2.010	2.326	2.488	2.572	2.614	2.636	2.648	2.654	2.657	2.658	2.659
Kurtosis	23.799	30.431	32.963	33.931	34.300	34.441	34.495	34.515	34.523	34.526	34.527	34.528

and

Table 5. Predictive output values of carbon market volatility.

Predictive Value	Posterior Stage	Posterior 2 Stage	Posterior 3 Stage	Posterior 4 Stage	Posterior 5 Stage	Posterior 6 Stage	Posterior 7 Stage	Posterior 8 Stage	Posterior 9 Stage	Posterior 10 Stage	Posterior 11 Stage	Posterior 12 Stage
Mean value	1.487	1.487	1.487	1.487	1.487	1.487	1.487	1.487	1.487	1.487	1.487	1.487
Standard deviation	0.200	0.200	0.200	0.200	0.200	0.200	0.200	0.200	0.200	0.200	0.200	0.200
Minimum value	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000	−0.000
Maximum value	26.020	26.020	26.020	26.020	26.020	26.020	26.020	26.020	26.020	26.020	26.020	26.020
Skewness	10.707	10.707	10.707	10.707	10.707	10.707	10.707	10.707	10.707	10.707	10.707	10.707
Kurtosis	105.825	137.216	147.295	150.531	151.570	151.903	152.010	152.045	152.056	152.059	152.060	152.061

showcase the utilization of the ARIMA (autoregressive integrated moving average) model to predict the long-term and short-term output values of the carbon market price index return and volatility. In this model, the independent variables include the mean value, standard deviation, minimum value, maximum value, skewness, and kurtosis of the regional carbon market return and volatility. These variables served as predictors, while the output values of the time series, encompassing volatility and other relevant variables, acted as the dependent variables.

Additionally, the relationship between the carbon market and energy market was further examined through the ARIMA model. This study revealed that while correlations existed, there were significant variations in the sensitivity, asymmetry, and regional differences between different regional pilot carbon markets and energy markets. Furthermore, the magnitude of influence from different energy markets on the same pilot carbon market also differed.

4.4. Risk from Energy Markets to the Carbon Market (Coal, Oil, and Gas)

In summary, as China’s carbon market is at an early stage, there is still a certain gap between carbon prices and turnover in each region, which is partly due to the fact that there are large differences in the level of economic and social development, natural resource reserves and energy consumption structure between regions [5]. At present, China has not formed a truly unified carbon market, the carbon trading mechanism is still unsound, and the risk warning and control capacity is still relatively limited. In terms of price fluctuations in the regional carbon market, Guangdong and Shenzhen carbon markets are characterized by large price fluctuations, while the price fluctuations in the Hubei carbon market are relatively small. This is related to the extent to which different markets are affected by international carbon market prices and regional energy demand (Guangdong and Shenzhen are important windows for China’s foreign economy).

In Figure 4 and Figure 5, the analysis of the directional spillover index and net spillover index diagrams between the seven regional carbon markets and energy markets in China revealed several key findings. First, the spillover effects between the carbon markets and energy markets exhibited significant time variability. The size and direction of the spillover, both in terms of directional spillover and net spillover, changed over time. Moreover, the time-varying directional spillover index demonstrated that the spillover effect between China’s carbon markets and energy markets was bidirectional and asymmetric. While the carbon market in different regions of China may not be significantly affected by volatility spillover from the energy market during certain energy price shocks, overall, the spillover index from the energy market to China’s carbon market was notably higher compared with other periods of energy market shocks, such as crude oil, coal, and natural gas. In addition, the investigation period also revealed variations in the time-varying spillover effects of carbon markets in different regions on energy markets.

5. Conclusions and Strategy Suggestions

5.1. Conclusions

This study employed the DY spillover index model to examine the spillover effect between China’s carbon market and the energy market, aiming to investigate the impact of energy market price fluctuations on China’s carbon market. By analyzing various inter-market spillover indices, the empirical findings yielded the following conclusions. First, the results based on the full sample revealed a two-way volatility spillover between the markets, yet the characteristics and net spillover relationship differed across different carbon markets and energy markets. In terms of the spillover characteristics, the Shanghai carbon market exhibited the closest linkage with the coal market, while the Beijing and Guangdong carbon markets were more susceptible to fluctuations in the crude oil market. Regarding the net spillover relationship, the Shanghai carbon market acted as a net importer of energy market volatility spillover, whereas the Beijing, Guangdong, and Hubei carbon markets functioned as net importers of energy market spillover. Furthermore, the findings based on the rolling window analysis revealed that the volatility spillover between China’s carbon market and the energy market exhibited significant time-varying characteristics. Both the inter-market total spillover and directional spillover exhibited strong time-varying patterns in terms of the spillover size and direction. This suggests that the information transmission mechanism between China’s carbon market and the energy market was highly uncertain.

Moreover, a closer examination of the various types of spillover indices between the markets indicates that the overall level of total spillover between the domestic carbon market and the energy market shows a significant upward trend during periods of energy market shocks. This implies that the spillover effect of energy market price fluctuations on China’s carbon markets was notably higher during these periods. However, it is important to note that the degree of influence varied across different pilot carbon markets in China, highlighting certain regional differences in their susceptibility to energy market fluctuations.

Therefore, effectively identifying and controlling the risk spillover effects between the carbon market and the energy market is of great significance to the market participants. For policymakers, it is conducive to avoiding or suppressing large price fluctuations in the carbon and energy markets, and thus, controlling systemic risks within and between industries. For investors, this is beneficial for stabilizing the return on investment portfolios and has a certain guiding significance for making optimal investment decisions under specific conditions. For high-energy-consuming enterprises, this is conducive to the judgment of energy and carbon price trends by business operators and stabilizes the development of enterprises.

5.2. Strategy Suggestions

In order to address the issue of greenhouse gas emissions, it is crucial for the central government to assume the role of a policy guide and prioritize the carbon market as a key component. This entails stabilizing carbon market prices and effectively coordinating efforts to address environmental challenges while promoting the country’s economic and social development. The government should fulfill its responsibilities in terms of policy guidance, administrative services, and market regulation to mitigate and control the formation and spillover effects of the risks associated with the carbon market.

In China, the National Climate Change Leading Group takes the lead in addressing climate change issues, while the National Development and Reform Commission (NDRC) is responsible for the overall planning and management of the carbon market. The NDRC has developed a comprehensive program for the establishment of the Chinese carbon market, taking into account the national context and drawing from internationally recognized operational mechanisms and management experiences, such as those observed in the EU ETS.

Based on the findings and conclusions of this study, the following policy recommendations are proposed to facilitate the stable operation of China’s carbon trading system:

(1)

Establishing a unified national carbon market is imperative. Currently, China’s carbon emission trading market is still in its nascent stage and lacks a comprehensive framework and robust mechanisms. There is a need to enhance risk identification, early warning systems, and control mechanisms for effective management.

(2)

Harnessing the power of policy guidance and market mechanisms at the national level is crucial. The carbon market should be at the forefront, with a focus on stabilizing market prices and aligning environmental concerns with the country’s economic and social development goals. The government should assume the roles of policy guidance, administrative services, and market supervision to mitigate and control the risks associated with the carbon market.

(3)

Strengthening market operations, supervision, and law enforcement is essential. Currently, China’s carbon market requires enhancements in operational mechanisms, technical specifications, legislation, and law enforcement practices. It is vital to improve the binding nature and enforceability of the legal framework governing the carbon market, as well as establish a robust foundation for enforcement actions and well-trained personnel. This will ensure compliance and foster a fair and efficient national carbon market, thereby mitigating risks and spillover effects (Ministry of Ecology and Environment: Report on the First Compliance Cycle of the National Carbon Emission Trading Market, 3 January 2023).

(4)

Promote the vibrancy of financial instruments. The financial sector’s diverse range of products and trading activities can enhance the carbon market’s price discovery mechanism, stabilize carbon prices, and improve carbon asset management capabilities, thereby strengthening the ability to mitigate and address carbon market risks. This can be achieved through various financial derivatives, such as carbon financial asset portfolios, carbon futures, carbon options, and carbon asset securitization. In China, the limited number of participants in the carbon financing market and the relative scarcity of financial products and liquidity have hindered the financing capacity and flexibility of enterprises’ carbon assets. To address this, carbon market support tools, such as carbon indices and carbon insurance, play a crucial role in improving investors’ understanding of carbon market prices and providing investment guarantees. These tools also serve as important means to enhance the creditworthiness of the carbon market. Furthermore, enterprises can be incentivized to strengthen technological innovation in energy conservation and emissions reduction, as well as a transition from high-carbon energy utilization to low-carbon energy sources. This can be achieved through the income effect and substitution effect, encouraging enterprises to adopt low-carbon new energy sources instead of high-carbon energy, thereby reducing greenhouse gas emissions. Additionally, supporting enterprise development through the purchase of carbon credits ensures that energy emissions are regulated, contributing to the control of carbon emissions and promoting environmental protection alongside economic and social development.

(5)

Enhance the capabilities for identifying, assessing, and managing risks while continuously improving early warning systems. Carbon market risks stem from both demand-side and supply-side factors. It is crucial to expedite the development of a comprehensive top-level design for the carbon market, establish a fair and efficient national unified carbon market that accommodates regional diversification, and ensure the fairness and efficiency of the carbon quota allocation system. This should be accompanied by the establishment of robust carbon trading and offset mechanisms, rationalization of quota storage and liquidity mechanisms, and enhancement of the carbon market’s capacity for adjustment. Additionally, a multi-level supervision mechanism should be implemented effectively, including the enforcement of penalty mechanisms and the strengthening of information disclosure mechanisms. To mitigate information asymmetry and reduce uncertainties in carbon market prices and transactions, it is essential to improve the enterprise credit evaluation system and implement comprehensive risk management throughout the entire process. By conducting qualitative and quantitative analyses to identify systemic risks, a robust carbon market risk early warning system can be established. This system will enable effective prediction and formulation of risk control programs, thereby preventing and addressing inter-industry risk spillovers.

(6)

Exploit the self-regulatory function of the carbon market to incentivize participation and enhance the liquidity of carbon allowances. Encourage enterprises to enhance energy efficiency through technological innovation and foster the growth of non-fossil-fuel enterprises and industries. By improving energy efficiency and accelerating the shift in energy demand structure, effective carbon asset management can be achieved. In the short term, the rise in carbon prices may increase production costs and reduce the competitiveness of enterprises. However, in the long run, enterprises should proactively enhance their carbon asset management capabilities and expedite the transition on the energy demand side, driven by the income effect and substitution effect. As an emerging financial market, the carbon market plays a crucial role in regulating the consumption of fossil fuels within environmentally sustainable limits, promoting the development of low-carbon energy sources, and facilitating the green transformation of the development model. Investors in the carbon market can optimize asset allocation, stabilize portfolio returns, and minimize risk signals transmitted to the market. This can be achieved by leveraging information from risk early warning systems and risk assessment systems, enabling effective risk mitigation and control (The People’s Bank of China, et al.: Knowledge of the Opinions on the Construction of a Green Financial System. 21 August 2016).

(7)

Promote the transformation and upgrading of the energy supply and demand structure. China should enhance its cooperation and integration with the international carbon market. This can be achieved by optimizing the energy structure, reducing reliance on coal and fossil fuels, and promoting industrial optimization and upgrading. It is crucial to prevent the blind development of industries with high energy consumption and high emissions. China can learn from the carbon market trading rules of the European Union, as well as domestic and international risk early warning systems. By identifying and understanding the factors, causes, and transmission paths of risks, a comprehensive risk early warning system can be established using big data information. This will improve the ability to identify and warn investors of risks in the carbon market, preventing risk spillover from the carbon market to the energy market. Furthermore, strengthening China’s participation in the international carbon market will help to build the image of China as a responsible global power. It will also contribute to the establishment of effective mechanisms to address global warming. By actively addressing climate change and taking responsibility for reducing greenhouse gas emissions, China can enhance its reputation as a responsible major country and contribute to global efforts to combat global warming.