Next Article in Journal
Forecast-Aided Converter-Based Control for Optimal Microgrid Operation in Industrial Energy Management System (EMS): A Case Study in Vietnam
Previous Article in Journal
Experimental Investigation into the Energy Performance of a Biomass Recuperative Organic Rankine Cycle (ORC) for Micro-Scale Applications in Design and Off-Design Conditions
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Price Volatility Spillovers in Energy Supply Chains: Empirical Evidence from China

School of Economics and Management, Nanjing Tech University, Nanjing 211816, China
*
Author to whom correspondence should be addressed.
Energies 2025, 18(12), 3204; https://doi.org/10.3390/en18123204
Submission received: 17 May 2025 / Revised: 13 June 2025 / Accepted: 17 June 2025 / Published: 18 June 2025

Abstract

Based on the theoretical framework of Multivariate Stochastic Volatility (MSV), this paper combines the Dynamic Generalized Correlation (DGC) model with the t-distribution, establishes the DGC-t-MSV model, and employs the Markov Chain Monte Carlo (MCMC) algorithm based on the Bayesian principle for efficient estimation to investigate the price volatility spillover effects in China’s energy supply chains. The results of this study indicate the following: (1) The upstream crude oil spot price has a positive spillover effect on the midstream freight price. The downstream diesel market price, 92 gasoline market price, and 95 gasoline market price all exert positive volatility spillovers on the midstream crude oil freight price. (2) The volatility spillover effect between the upstream power coal price and the midstream coal freight price exhibits unidirectionality, and the volatility is transmitted from the power coal price to the coal freight price. (3) The upstream natural gas price and the midstream liquefied natural gas market price display asymmetric characteristics. Among them, the upstream natural gas price has a unidirectional and more pronounced positive volatility spillover effect on the midstream liquefied natural gas market price.

1. Introduction

Against the backdrop of the continuous evolution of the global energy landscape, China is in a crucial period of accelerating its energy transition. China has long relied on non-renewable conventional energy such as crude oil, coal, and natural gas. Fluctuations in their prices significantly impact the supply and demand in the industrial sector [1,2,3] but also generate complex and far-reaching chain reactions in the supply chain system and the overall economy [4,5]. For example, in 2020, under the influence of the pandemic, the sharp decline in international oil prices led to a sudden drop in China’s crude oil import costs [6]. This not only affected the profits of the domestic refining industry but also spread to downstream manufacturing through the supply chain, thus having a certain impact on the overall economic operation. This event highlights the profound impact of energy price fluctuations on China’s supply chain and economic system.
In addition, in the context of the supply chain, the energy supply chains are mainly divided into different segments, such as upstream supply extraction, midstream transportation or energy transformation and downstream energy utilization [7]. These links are interconnected and interlocked, and any disruption or interference in the supply chain will have a tremendous influence on China’s economy [8,9]. Therefore, this article investigates the spillover effects of price volatility in China’s energy supply chains, which contributes to a more in-depth and comprehensive understanding of the operating mechanism of China’s energy supply system.
At present, research on energy price volatility spillovers mainly pays attention to three aspects. First, it examines price volatility spillovers within international energy markets—specifically, crude oil, coal, and natural gas [10,11,12,13]. For instance, Dhifaoui et al. [14] demonstrate highly nonlinear spillover correlations between crude oil and natural gas supply chain prices. Przekota [15] focuses on price volatility spillovers between Poland’s crude oil and diesel fuel markets, while Zhao et al. [16] analyze energy market spillovers (crude oil, coal, natural gas) under geopolitical hazards. Second, the analysis extends to price volatility spillovers at cross-supply-chain levels [17,18]. For instance, Hasanli [19] identifies significant nonlinear spillovers between crude oil and natural gas supply chains, while Guo and Zhao [20] examine price volatility transmission across China’s oil and coal supply chains. Third, it analyzes the spillover effects of energy price fluctuations from the perspective of external shocks. For example, He and Sun [21] concluded that geopolitical conflicts have an asymmetric effect on oil price fluctuations. Pan et al. [22] determined that shocks on the oil supply side are an important factor contributing to the rise in China’s oil prices. Current research confirms that price volatility spillover effects exist among China’s three major energy supply chains (crude oil, coal, natural gas). Nevertheless, most studies adopt cross-industry perspectives that treat supply chains holistically, failing to sufficiently address intra-chain price volatility spillovers across segments. In light of this, this paper takes the upstream, midstream, and downstream sections of the energy supply chains as the starting point to comprehensively investigate the price volatility spillover effects of China’s three major energy sources: crude oil, coal, and natural gas.
Most prior studies have employed the traditional DCC-MSV and GC-MSV models when analyzing volatility spillovers in financial markets [23,24]. However, the parameter estimation of the GC-MSV model is relatively complex and has strict requirements for the quality and quantity of data. Meanwhile, it may be restricted by the model assumptions. On the other hand, the drawback of the DCC-MSV model lies in its relatively complex model construction, and the results are highly sensitive to the selection and adjustment of parameters. To overcome the aforementioned bottlenecks, numerous scholars have explored the combination of models based on the normal distribution. For example, Wang et al. [25] and Yu et al. [26] used the combined model, the DGC-t-MSV model, to conduct an analysis of volatility spillover correlation. They also demonstrated that this combined model could enhance flexibility and robustness when capturing complex dependence structures. Meanwhile, Zhang et al. [27] utilized the DGC-MSV-T portfolio model to investigate the spillover effects of crude oil futures and spot prices, as well as those of stocks.
In this paper, we choose the DGC-t-MSV model rather than the DGC-MSV-T model. Our reasons are as follows: first, the DGC-t-MSV model can better capture short-term fluctuations and instantaneous changes, while the DGC-MSV-T model is not sensitive enough to short-term fluctuations; second, the DGC-t-MSV model performs excellently when dealing with high-frequency data and can reflect market changes in a more timely manner. Inspired by these endeavors, the DGC-t-MSV model is utilized to investigate the spillovers of price volatility within the Chinese energy supply chains in this paper.
In summary, this paper makes use of the DGC-t-MSV combined model. Based on this, it applies the Bayesian principle and the Markov Chain Monte Carlo (MCMC) algorithm. Then, it conducts an in-depth investigation of the price volatility spillover effects within China’s energy supply chains. The key contributions of this research paper are as follows: (1) The majority of previous studies have predominantly concentrated on analyzing crude oil, coal, and natural gas across the entire supply chain. In contrast, this paper takes China’s three significant energy supply chain markets as its starting point to thoroughly analyze the price volatility spillover effects in the upstream, midstream, and downstream energy markets of China. (2) Based on the DGC-t-MSV combined model, this paper employs the MCMC algorithm and the Bayesian principle. With the assistance of the Gibbs sampling technique, it more precisely captures the price spillover effects in each segment of China’s energy supply chains. (3) The following novel findings are presented: the upstream crude oil spot price has a positive spillover effect on the midstream freight price; the market prices of downstream diesel, No. 92 gasoline, and No. 95 gasoline have a more pronounced positive fluctuation spillover effect on the midstream crude oil freight price; the volatility spillover effects from the upstream power coal price to the midstream coal freight price and from the upstream natural gas price to the midstream liquefied natural gas market price are unidirectional.

2. Research Hypotheses and Model Formulation

2.1. Research Hypotheses

The energy supply chain system exhibits the network characteristics of high interconnectivity and multi-level coupling. There are significant two-way transmission mechanisms and deep interdependent relationships among the upstream, midstream, and downstream links. The specific transmission paths are shown in Figure 1.
Currently, the factors influencing crude oil price fluctuations are complex and variable. Among them, geopolitical conflicts have a significant impact on international crude oil price fluctuations through an asymmetric mechanism [21]. This impact is further magnified under the shock of crude oil supply and is transmitted to the price fluctuations of the domestic crude oil supply chain [22]. Wen et al. [28] found, based on the time-series analysis method, that there is a lead–lag relationship between the prices in the crude oil futures market and the prices of related products in the upstream, midstream, and downstream of the crude oil supply chain. The results of the Granger causality test show that changes in crude oil futures prices play a leading role in the prices of the crude oil supply chain, which provides empirical support for the price discovery function of the crude oil futures market.
In addition, from the systematic perspective of the crude oil industrial chain, as the main driving factor, crude oil price fluctuations transfer the impact to the prices of other commodities in the industrial chain through the lead–lag relationship of prices in horizontal and vertical fields, forming a complex multi-dimensional price correlation network [29]. In the upstream segment of the crude oil supply chain, crude oil price fluctuations directly affect the costs and expectations of crude oil extraction and production. In the midstream segment, the production costs and processing profits of refineries will be significantly adjusted due to changes in crude oil prices. In the downstream consumption segment, as an important terminal of the crude oil supply chain, there is a close transmission relationship between the gasoline market and crude oil price fluctuations [7]. In particular, rising crude oil prices will gradually be transferred to the gasoline market through the refining process of refineries, leading to corresponding adjustments in gasoline retail prices. At the same time, the seasonal changes in gasoline market demand and the macro-economic environment will also have a feedback effect on crude oil price fluctuations, forming a complete price transmission mechanism in the industrial chain [30]. This price transmission mechanism reflects the systematic characteristics of the crude oil industrial chain and highlights the systematic impact of crude oil price fluctuations on the entire crude oil supply chain. Based on the above theoretical analysis, this study proposes the following research hypothesis:
H1. 
The crude oil supply chain in China is characterized by complexity and two-way features.
In the coal market, China’s coal prices have always been controlled by the government to a certain extent, and coal prices have a heterogeneous spill-over effect on China’s industrial economic development [3]. Among them, thermal coal accounts for about 60% of China’s coal consumption and dominates the coal supply chain [31]. Due to its large share, the price fluctuations of thermal coal have an impact on the prices of the downstream coal-fired power market, which are transmitted through the price transmission mechanism between the upstream and downstream of the supply chain [32]. The systematic nature of the coal supply chain ensures the spread of thermal coal price fluctuations throughout the chain, affecting the profitability and cost structure of downstream industries. Based on this, this paper proposes the following hypothesis:
H2. 
The volatility spillover of the coal supply chain in China shows a one-way characteristic.
In the natural gas market, China’s dual-carbon goals of achieving carbon peak (by 2030) and carbon neutrality (by 2060) highlight the importance of natural gas utilization and market risks [33]. As a clean fossil energy source, natural gas plays an important role in the energy structure transformation. However, its utilization is also accompanied by significant market risks. Data shows that China’s high dependence on natural gas imports [34] leads to a relatively high concentration in the upstream market, and oil companies may affect market efficiency by exercising market power. Under this market pattern, price fluctuations in the international natural gas market will be transmitted to the domestic downstream market through the import link, thereby driving up the domestic natural gas market price [35]. Based on this, this paper proposes the following hypothesis:
H3. 
There is asymmetry in the volatility spillover of the natural gas supply chain in China.
Based on the proposed hypotheses, this paper conducts a detailed analysis and verification of the data in the third chapter.

2.2. Model Formulation

The aim of this paper is to study the dynamic correlations and volatility spillover effects among the prices of different links in China’s energy (such as crude oil, coal, and natural gas) supply chain—that is, how the price fluctuations in one link affect the fluctuation risks of prices in other links. Traditional methods struggle to fully characterize such complex dynamic volatility and interdependent structures. Therefore, this paper introduces the multivariate stochastic volatility model (MSV) and its extended form, the DGC-t-MSV model. Among them, the MSV model is an advanced model suitable for financial time-series analysis, which can effectively characterize the volatility and dynamic correlation of data [10,25], thus more naturally fitting the “volatility clustering” effect and complex dynamic dependencies. Due to the high degree of nonlinearity and dimensionality of the DGC-t-MSV model, based on the Bayesian method, this paper combines the Markov Chain Monte Carlo algorithm (MCMC) to generate samples approximately, thus effectively solving the high-dimensional computation problem [36].

2.2.1. Basic MSV Model

w t = d i a g ( exp ( f t / 2 ) ) ψ t , ψ t ~ N ( 0 , ψ )
f t + 1 = μ + d i a g ( ϕ 11 , ϕ 22 ) ( f t μ ) + ξ t , ξ t ~ N ( 0 , d i a g ( σ ξ 1 2 , σ ξ 2 2 ) )
Among them, w t represents the yield rate sequences of various markets in the upstream, midstream, and downstream of the supply chain; f t describes the volatility sequence of the market in the supply chain at time t; ψ t follows a t-distribution with a mean of 0, a variance of ψ , and degrees of freedom of d, and it does not change over time; and μ is the long-term mean of the logarithmic volatility. The historical numerical pairs of ϕ 11 and ϕ 22 have an impact on future fluctuations. The closer the coefficients of ϕ 11 and ϕ 22 are to 1, the greater the impact of historical fluctuations on current fluctuations. ξ t is the error term of the state transition equation, and σ represents the standard error of the volatility interference.

2.2.2. MSV Model with the Incorporation of Granger Causality (GC-MSV)

The GC-MSV model is established on the foundation of the basic MSV model with the addition of a one-way volatility spillover term for each link between supply chains:
f t + 1 = μ + ϕ 11 ϕ 12 ϕ 21 ϕ 22 ( f t μ ) + ξ t , ξ t N ( 0 , diag ( σ ξ 1 2 , σ ξ 2 2 ) )
ϕ 12 and ϕ 21 , respectively, represent the volatility spillover effects among the upstream, midstream, and downstream markets of the supply chain. When at least one of ϕ 12 and ϕ 21 is significantly non-zero, it can be determined that there is a certain price volatility spillover relationship between them.

2.2.3. DCC-MSV Model

For the dynamic correlation of each index between supply chains, this paper adopts the Dynamic Conditional Correlation (DCC) model. In this model, not only is the volatility time-varying; the correlation coefficient is also time-varying:
w t = diag ( exp ( f t / 2 ) ) ψ t , ψ t ~ N ( 0 , R t )
R t is the correlation matrix over time and is defined by the following DCC process:
R t = diag ( Q t ) 1 Q t diag ( Q t ) 1
The dynamics of Q t evolve as follows:
Q t = ( 1 α β ) Q ¯ + α ϵ ψ t 1 ψ t 1 T + β Q t 1
α and β are non-negative parameters to be estimated, and Q is the unconditional correlation matrix of the residuals. The DCC-MSV model captures the dynamic correlation between supply chains over time.

2.2.4. DGC Multivariate Stochastic Volatility Model with Time-Varying Property

After constructing the DGC-MSV model, the t distribution is incorporated to more comprehensively capture the leptokurtic and fat-tailed characteristics of various numerical values within the supply chains. This model not only considers the time-varying nature and directionality of the relationships between supply chains but also accounts for the time-varying features of the spillover effect:
ρ t = exp ( r t ) 1 exp ( r t ) + 1
The time-varying correlation factor ρ t is modeled by the following state transfer equation:
ψ , t = 1 ρ t ρ t 1
r t + 1 = v 0 + v a c ( r t v 0 ) + σ i o t , o t ~ N ( 0 , 1 )
Here, ρ t reflects the dynamic correlation that changes over time, ranging from −1 to 1. This coefficient varies with time and is used to indicate the strength of the time-varying correlation in the upstream, midstream, and downstream markets of the supply chain during different time periods. r t is a parameter describing the dynamic correlation; ψ , t represents the time-varying correlation coefficient matrix, which is used to test the mean spillover effect among the markets at different levels (upstream, midstream, and downstream) within the supply chain; v 0 and v a c are parameters describing dynamic correlation, representing the change speed and persistence of the time-series; and σ i is the standard deviation of the noise term, representing the random shock amplitude and change degree of volatility.

2.2.5. Model Estimation Methods

Suppose that the random parameter space K , t T is a discrete set T = 1 , 2 , 3 , 4 , . From the Markov chain construction, it follows that
p K 0 = k 0 | K 1 = k 1 , K t = k t = p K 0 = k 0 t p K i = k i | K i 1 = k i 1
In Equation (10), p X t = x t | X t 1 = x t 1 represents the transition probability of the t sample given the (t − 1) sample. As time t increases, the Markov chain gradually converges to the posterior distribution.
Here, Gibbs sampling for high-dimensional operations is selected, and the calculations are carried out through appropriate full conditional distributions. Suppose there is an n-dimensional random vector K = k 1 , k 2 , k 3 k n . The target joint distribution is π k , and when the values of other variables are known, the conditional distribution of variable K i is π K i | k i , where k i represents the values of other variables except k i . The specific steps of the Gibbs sampling process are as follows.
We use k 0 = ( k 1 0 , k 2 0 , , k n 0 ) as the initial point for Gibbs sampling. k t 1 denotes the value of t 1 iterations.
(1) We take k 1 t from the full-fledged conditional distribution π k 1 | k 2 ( t 1 ) , , k n t 1 ;
(2) We take k 2 t from the full-fledged conditional distribution π k 2 | k 1 ( t 1 ) , k 3 t 1 , k n t 1 ;
……
(n) We take k n t from the complete conditional distribution π k n | k 1 ( t ) , k 2 t , , k n 1 t .
The transfer probability of transfer from k to k can be
p k , k = π k 1 | k 2 , , k n π k 2 | k 1 , , k n π k n | k 1 , , k n 1
Moreover, let K = ( K 1 , K 2 ) follow a multivariate normal distribution; i.e., N 0 0 , 1 ρ ρ 1 , 0 < ρ 1 .

3. Analytical Research Based on Empirical Data

3.1. The Selection of Sample Specimens and the Preprocessing of Relevant Data

The time span of the sample in this study is from 1 August 2015 to 31 July 2024, and all the data are daily data. The specific price indicators selected are as follows. For the Chinese crude oil supply chain, in the upstream, the Crude Oil Spot Price (COSP) is selected; in the midstream, the coastal Crude Oil Freight Price (COFP) is selected; and in the downstream, the Diesel Market Price (DMP), 92 Gasoline Market Price (92GMP), and 95 Gasoline Market Price (95GMP) are selected. For the Chinese coal supply chain, in the upstream, the Power Coal Price (PCP) is selected; in the midstream, the coastal Coal Freight Price (CFP) is selected. For the Chinese natural gas supply chain, the upstream Natural Gas Price (NGP) and the midstream Liquefied Natural Gas Market Price (LNGMP) are selected. The sample data are sourced from the Wind Database, the National Bureau of Statistics, and the Choice Database.
Considering potential heteroskedasticity in the nine time-series, we first apply a natural logarithmic transformation to the market price data to reduce volatility. Subsequently, we compute the logarithmic returns by taking first-order differences and scaling the results by 100. This approach effectively eliminates potential trend components and extracts return information. Meanwhile, considering that some markets (such as the coal and natural gas markets) have obvious seasonal characteristics, their prices show strong seasonal fluctuations due to heating demand. The first-order difference processing method we adopt can, to a certain extent, eliminate the seasonal components in the data. The difference operation subtracts the data of adjacent time points, thus weakening the impact of periodic changes caused by seasonal patterns. This processing method not only helps to eliminate seasonal factors but also can more clearly reveal the real fluctuations and trends in the data, making the subsequent analysis results more reliable.
Based on the descriptive statistics of the price-index returns presented in Table 1, several key observations can be made. First, the skewness values of the sample index returns are all non-zero, and the kurtosis values are greater than 3. This indicates that these data deviate from the standard normal distribution. Second, the J-B statistics are extremely large, and all values are far greater than the critical value (approximately 5.99) at the 5% significance level. This provides strong evidence that the data significantly deviate from the normal distribution and exhibit characteristics of leptokurtosis and fat tails. In particular, for the COFP series, its extremely high kurtosis value and the corresponding extremely large J-B statistic both show a highly significant deviation from the normal distribution (large skewness and large excess kurtosis will directly lead to a large Jarque–Bera value). Third, the high kurtosis and fat-tailed features suggest a relatively high probability of extreme values. The high-risk and asymmetric nature of price volatility in the energy supply chain market is reflected by this, which also suggests that this market is vulnerable to a variety of external risks, such as geopolitical conflicts, policy reforms, and supply–demand imbalances.
Therefore, considering the aforementioned statistical characteristics, this paper demonstrates that opting for the DGC-t-MSV portfolio model is extremely viable and productive. Reliable model-based support for the subsequent empirical analysis can be provided by this model, which is capable of efficiently grasping the high volatility and asymmetry of the return series.
On this basis, as can be seen from Table 2, the logarithmic return of the index meets the stationarity test, but there is a ARCH effect in the heteroscedasticity test. This indicates that from an economic perspective, the DGC-t-MSV model can capture specific characteristics that the ARCH model cannot achieve. Therefore, the DGC-t-MSV model used in this paper is more suitable for the data and research objectives of this paper.
Figure 2 depicts the time–return series of nine energy supply chain price indices in China. Figure 2a–e illustrate the volatility of the time–return series of China’s energy crude oil supply chain. Notably, the crude oil spot price exhibits more pronounced volatility around the midpoint of the time-series. This is strongly associated with the COVID-19 global health crisis. The research conducted by Sharif et al. [6] offers robust empirical evidence regarding the consequence of the COVID-19 pandemic on the volatility of the energy commodity index. Conversely, the crude oil freight price demonstrates a high level of stability. This phenomenon reflects the well-established system and mature policies in China’s crude oil transportation market.
Figure 2b shows that the coastal crude oil freight price grew at an exceptionally high rate in 2022. Our analysis identifies three primary drivers for this increase: first, concentrated inventory-replenishing demand from domestic refineries following their operational resumption after adjustments to epidemic prevention policies; second, extreme weather conditions, particularly typhoons that caused port delays and closures in East and South China; and third, oil-hoarding behavior induced by high international oil prices. Additionally, year-end surges in international VLCC pricing levels to over USD 100,000 per day led to the diversion of some coastal shipping capacity. When combined with reduced turnover efficiency due to port epidemic controls, these factors further elevated the short-term coastal crude oil freight price. Data from the Shanghai Shipping Exchange indicate that the Coastal Crude Oil Freight Index (CTFI) rose from 1000 points in January to 1400 points by December.
Figure 2f,g illustrate the volatility of the time–return series of coal supply chain prices in China. As can be observed from Figure 2f,g, the volatility of the coal market price increased significantly between 2021 and 2022, indicating the instability of the coal market price. The reasons for this instability can be attributed to the following factors. On the coal supply side, environmental protection regulations have compelled some coal mine supply enterprises to undergo rectification or shut down, impeding the release of production capacity. On the coal demand side, as China’s environmental protection policies have tightened, the development of high-energy-consuming enterprises has been restricted, resulting in a decrease in coal consumption demand. Consequently, the market balance has been thrown off-kilter due to the limitations on both the supply and demand sides. Consequently, coal prices have faced pressure, leading to heightened instability and volatility in the underlying returns.
Figure 2h,i, respectively, depict the volatility of the time-series returns of the natural gas price and the liquefied natural gas market price. By analyzing Figure 2h,i, it is evident that compared with the earlier stages of the time-series, the volatility of the natural gas price and the liquefied natural gas market price has been more pronounced since 2018. The reason for this phenomenon can be attributed to the following factors. As China’s industrial structure undergoes in-depth adjustment and continuous upgrading, the demand for clean energy, such as natural gas, in emerging industries has gradually increased. Moreover, in industrial production and power generation, natural gas acts as a crucial alternative to coal. The unstable coal price makes the price advantage of natural gas more conspicuous, which exerts an important influence on the supply–demand relationship and price trend of the natural gas market.
Figure 3 shows the comparison of the time return series of crude oil, coal, and natural gas in China. From Figure 3, the correlation characteristics among the energy price series can be clearly observed. The specific analysis is as follows:
First, in China’s crude oil supply chain market, the Crude Oil Spot Price (COSP) in the upstream market exhibits synchronized volatility and correlation with the 92 Gasoline Market Price (92GMP) and 95 Gasoline Market Price (95GMP) in the downstream market. The underlying reasons for this phenomenon can be attributed to two primary factors: On the institutional level, China’s refined oil pricing mechanism is linked to international crude oil prices. According to the Petroleum Price Management Measures issued in 2016, the retail price ceiling of refined oil products follows a “10-working-day adjustment mechanism”. The domestic refined oil pricing formula is directly tied to international crude oil prices, creating a structured price transmission channel. From a cost perspective, crude oil costs account for over 70% of the total production costs of refining enterprises (as reported in Sinopec’s 2022 annual report), while refining technology costs (such as catalytic cracking efficiency) remain relatively rigid. Consequently, fluctuations in COSP can be rapidly transmitted to the downstream gasoline market (92GMP and 95GMP), leading to synchronized price movements. This synchronization reflects the combined effect of regulatory pricing mechanisms and market-driven cost transmission dynamics.
Second, in China’s coal supply chain, the power coal prices in the upstream market and the coal freight prices in the midstream market are somewhat correlated, and the price fluctuation is large and lagging. The reason for this is that coal is one of the major energy sources on which China’s economic advancement experiences a degree of dependence, and due to the remarkable changes in its provision, consumption, and expenditure, the prices across all segments of the supply chain system undergo substantial fluctuations. The implementation of stricter environmental protection policies has forced several coal mines to cease operations. This disruption has thrown off the equilibrium between supply and demand, resulting in substantial price fluctuations within the supply chain. In addition, during coal demand season, coal prices rose to drive increased demand for transportation, but the coastal coal transportation is mostly based on long-term contracts, and the contract period of the tariff is fixed. Under the combined effect of these factors, upstream power coal prices and midstream coal freight price showed a complex correlation with high volatility and obvious lag.
Third, the volatility of natural gas prices in the upstream market and the liquefied natural gas market prices in the midstream market is significant, with a lack of synchronization and high uncertainty. The reasons for this are evident at three levels: (1) The international environment level—China is more significantly affected by the fluctuations in the global natural gas marketplace. Factors such as the instability of international natural gas prices and geopolitical situations are conveyed into the Chinese market via the import route. (2) The energy structure level—China is in a period of energy transition, and natural gas, as a relatively clean energy, has seen its demand continue to climb. However, price uncertainty will be heightened due to a high level of reliance on foreign natural gas, as the expansion pace of domestic natural gas output struggles to keep up with the surge in market need. (3) The infrastructure development—Uneven natural gas infrastructure in China, with limited transportation and storage capacity in some areas, can also exacerbate price volatility in localized areas.

3.2. Model Convergence Analysis and Testing

Figure 4 shows the iterative fluctuation of time-series data of the Crude Oil Spot Price (COSP), Coastal Crude Oil Freight Price (COFP), Diesel Market Price (DMP), 92 Gasoline Market Price (92GMP), 95 Gasoline Market Price (95GMP), Power Coal Price (PCP), Coastal Coal Freight Price (CFP), Natural Gas Price (NGP), and Liquefied Natural Gas Market Price (LNGMP). According to the index iterative fluctuation chart (see Figure 4), the time-series data maintains relatively stable fluctuations up and down in the long run, and there is no obvious upward or downward trend. The autocorrelation chart converges to zero rapidly (see Figure 5). It can also be confirmed by the fact that the MC errors in the subsequent volatility spillover table are all less than the standard deviation, indicating that the MCMC iterative process has reached stability.
In addition, this study uses the Gibbs sampling method to perform 80,000 iterations on the nine parameters of the DGC-t-MSV model. Among them, the first 10,000 iterations are used as “burn-in period” samples and discarded to ensure the stability of parameter estimation. To evaluate the convergence of the model, the Gelman–Rubin statistic graph is used.
As shown in Figure 6, in the DGC-t-MSV model constructed based on the price return rate series of various indices of China’s energy supply chain, as the number of iterations increases, the statistical results of the Gelman–Rubin diagnostic graph gradually approach 1. This result indicates that the MCMC algorithm based on Gibbs sampling shows good convergence. It also further demonstrates that the DGC-t-MSV model has good adaptability when used to study the price volatility spillover effect of China’s energy supply chain.

3.3. Mean Spillover Effect Analysis of the Model

In this section, to more accurately measure the average spillover effect, the dynamic spillover effect among different variables is gauged using the price dynamic correlation coefficient, as presented in Table 3.
Table 3 presents a statistical description of the dynamic correlation coefficients among prices within the supply frameworks of China’s three major energy sources: crude oil, coal, and natural gas. As shown in Table 3, there are differences in the dynamic correlations among energy prices across China’s energy supply chains. The specific analyses of these differences are as follows.
First, during the operation of the crude oil supply chain, the average value of the dynamic correlation coefficient between the upstream COSP and the midstream COFP is −0.0302. The range reaches 0.9106, reflecting substantial variation in their relationship. The reason is that geopolitical conflicts have disrupted the crude oil supply, driving up the international and Chinese crude oil spot prices, while the transportation market has reduced regional operations due to related risks. This has led to an oversupply of coastal transportation capacity, thus depressing the freight rates and forming a negative correlation. Moreover, affected by the rise in international oil prices, the Chinese crude oil spot price has also increased. However, since the global pandemic, China’s economic growth has slowed down in some stages, and the import volume and transportation demand have shown a fluctuating downward trend, resulting in a decline rather than an increase in freight rates, and thus forming a phased reverse change with the upstream crude oil spot price. This indicates that the spillover effect within the Chinese crude oil supply chain is deeply influenced by factors such as geopolitical conflicts and changes in the international crude oil market.
Second, in the coal supply chain market, the mean in the dynamic state of the correlation parameter between the upstream PCP and the midstream CFP is 0.0130. However, the range between the two is as high as 0.9209, suggesting significant uncertainty in price fluctuations between upstream power coal prices and midstream coastal coal transportation prices. A confluence of elements, including the balance between market supply and demand, global coal price levels, and the overarching macro-economic landscape, leads to this situation. It is in the natural gas provision network that the calculated average of the dynamic correlation parameter between the price of upstream NGP and the market LNGMP stands at 0.0026, which suggests that the correlation is rather weak. However, the range reaches 1.0324, the highest among all datasets, reflecting large price fluctuations and a lack of a stable linkage mechanism. The cause behind this phenomenon is that multiple intricate factors have an impact on China’s natural gas supply chain market, such as international gas supply, domestic pipeline transportation, storage and peak-shaving capacity, and seasonal consumption differences. These factors make it difficult to stabilize the correlation between upstream natural gas prices and midstream liquefied natural gas market prices.
Figure 7 depicts the dynamic correlations among the prices of the three energy supply chains in China over time. Table 4 presents the results of the statistical significance tests (p-values) for the correlations shown in Figure 7. This provides a solid statistical basis for the subsequent in-depth analysis of their change characteristics and underlying causes. The strength criteria for Spearman’s rank correlation coefficient are similar to those for Pearson’s correlation coefficient. Generally, the following ranges are used to interpret its strength: 0.00–0.19, extremely weak correlation; 0.20–0.39, weak correlation; 0.40–0.59, moderate correlation; 0.60–0.79, strong correlation; 0.80–1.00, very strong correlation. The following is an in-depth analysis of the change characteristics and underlying causes of the dynamic correlations of prices at each link around these three supply chains of crude oil, coal, and natural gas.
First, in the initial stage of the time-series of China’s crude oil supply chain, there is a dynamic correlation among the upstream COSP, the midstream COFP, the downstream DMP, the 92GMP, and the 95GMP. In a market environment characterized by relatively balanced crude oil supply and demand, stable policies, and the absence of major trade frictions, upstream price changes are rapidly transmitted to downstream sectors, which is consistent with the principles of market economics. However, in the middle stage of time-series analysis, the price correlation among various links in the crude oil supply chain gradually decreases. This phenomenon is closely associated with domestic and foreign policy adjustments, the China–US trade war, the Russia–Ukraine conflict, and the global COVID-19 pandemic [37,38]. Specifically, policy adjustments have modified supply–demand relationships and price formation mechanisms. Trade wars and military conflicts have increased price uncertainty, whereas the global pandemic significantly reduced energy demand and disrupted price transmission mechanisms. Collectively, these effects illustrate how external shocks destabilize China’s crude oil market. By the end of the time-series analysis, the correlation among variables gradually recovers and stabilizes. This recovery reflects the adaptation of market participants to the new policy environment and economic conditions after medium-term fluctuations.
Second, as shown in Figure 7e, the correlation within the coal supply chain exhibits slight fluctuations in the initial period. China possesses abundant coal reserves and maintains a stable supply, with its coal market being relatively self-contained and largely insulated from international geopolitical volatility. In the middle stage of this time-series, the correlation suddenly rose to about 0.06 and then quickly turned into a negative correlation. This change can be attributed to China’s recent domestic policy reforms. In recent years, China has vigorously implemented policies on new energy and clean energy. For example, the Action Plan for Reaching Peak Carbon Emissions before 2030 has been introduced to guide the transformation of the energy consumption structure, resulting in a decline in the market demand for coal [39,40]. Meanwhile, Chinese coal production enterprises have adjusted their production plans in line with the “dual carbon” goals issued domestically, leading to the gradual withdrawal of some production capacities from the market. The reversal of the supply–demand relationship has caused the above-mentioned changes in the correlations among various links of the coal supply chain.
Third, as illustrated in Figure 7f, the dynamic correlation between upstream NGP and midstream LNGMP evolved as follows: initially negative, it gradually stabilized at approximately 0.02, followed by a surge to a peak of around 0.05. This trend reflects China’s evolving natural gas policies and market dynamics. For example, the “Administrative Measures for the Pricing of Natural Gas Pipeline Transportation (Trial)” promulgated in 2016 marked the beginning of the market-oriented reform of natural gas. It allowed the prices of natural gas sources and sales to be determined by the market. Specifically, China’s 2015 market-oriented pricing reforms for natural gas triggered heightened volatility in upstream prices. However, midstream price adjustments lagged behind upstream fluctuations due to rigid regulatory controls and delayed market reactions, leading to a persistently low correlation coefficient (below 0.1) during this phase. By 2021, China had deepened its natural gas price liberalization reforms: government interventions were substantially scaled back, while market mechanisms played an increasingly dominant regulatory role. Furthermore, China’s natural gas import prices became more tightly coupled with international benchmarks as global energy market integration intensified. These developments collectively reshaped the correlation patterns within China’s natural gas supply chain.

3.4. Analysis of Volatility Spillovers in the Model

This paper captures and analyzes the volatility spillover correlation and timing variability among time-series with the help of the DGC-t-MSV model, and this part introduces the volatility persistence parameter to quantify the degree of influence of past volatility on current and future volatility. Table 5 shows the simulated data estimates of price volatility spillovers among China’s crude oil, coal, and natural gas energy supply chains. A specific analysis of the relevant content is available.
The Chinese energy market is highly concentrated, and government regulation and policy changes have a significant impact on market fluctuations. The DGC-t-MSV model can dynamically capture the spillover effects of these exogenous shocks on price fluctuations. In addition, the price fluctuations in the Chinese energy market usually exhibit non-linear and non-normal characteristics. The DGC-t-MSV model uses the t-distribution to describe the residuals, which can better fit these data characteristics, thereby improving the applicability and accuracy of the model. The specific analysis is as follows.
First, there is a two-way volatility spillover effect between the upstream Crude Oil Spot Price (COSP) and the midstream coastal Crude Oil Freight Price (COFP), but the transmission process shows distinct asymmetry. Quantitative analysis shows that in terms of transmission intensity, the average spillover value from COSP to COFP (0.0251) is higher than that of COFP to COSP (0.0114), revealing that upstream price fluctuations have a stronger radiation effect on midstream transportation costs. In terms of volatility characteristics, the standard deviation of the COFP to COSP direction (0.0115) is 2.74 times that of the COSP to COFP direction (0.0042) (Monte Carlo error: 0.0081 > 0.0030), highlighting the high-frequency oscillation attribute when the midstream freight price impacts the upstream price. In terms of directional stability, the 97.5% quantile of COFP to COSP (−0.0006) is close to the zero-value threshold, while the corresponding quantile of COSP to COFP (−0.0211) falls deeply into the negative value range (95% confidence interval: −0.0223 to −0.0006), further confirming the obvious uncertainty of the former effect. The reason is that the supply–demand relationship in the upstream of China’s crude oil supply chain is more rigid than that in the midstream transportation market. The upstream COSP is affected by global supply–demand, geopolitics, and macro-regulation [41]. The changes in these factors are relatively slow and predictable. However, the midstream coastal COFP is affected by micro-factors such as ship supply–demand, port congestion, and fuel prices, resulting in short term and violent oscillation characteristics of its price feedback to the upstream. Therefore, the upstream COSP has a more significant spillover effect on the midstream coastal COFP.
Second, in China’s downstream crude oil market, the Diesel Market Price (DMP), the 92 Gasoline Market Price (92GMP), and the 95 Gasoline Market Price (95GMP) all have positive volatility spillover effects on the midstream coastal crude oil freight price (COFP). However, the feedback from COFP to the downstream market shows a structural asymmetric suppression. The specific analysis is as follows: (1) Statistical fragility of the suppression effect: the 95% confidence intervals of COFP to 92GMP ([−0.0203, 0.0020]) and COFP to 95GMP ([−0.0202, 0.0061]) both cross zero, indicating that although the increase in freight prices briefly suppresses gasoline price fluctuations through the cost transmission channel (such as delaying the refineries’ price adjustment decisions), it does not form a continuous spillover path. (2) The reasons for the weak negative correlation between the two (for COFP to 92GMP) are as follows: regional market segmentation—the differentiation of inventories between coastal and inland refineries leads to an increase in transmission noise; policy smoothing effect—the national refined oil price adjustment mechanism (Chapter 4 “Improving the Energy Reserve System” of the “14th Five-Year Plan for a Modern Energy System”) weakens the intensity of price fluctuation transmission through the strategic reserve’s purchase and sale adjustment. (3) Structural amplification effect of downstream volatility: the standard deviation of the downstream prices’ response to COFP fluctuations increases significantly (DMP: 0.0187 → 92GMP: 0.0170 → 95GMP: 0.0262), which is consistent with the high-order volatility transmission model. These results verify hypothesis 1.
Third, in China’s coal market, the volatility spillover between upstream PCP and midstream CFP exhibits asymmetric directions (positive and negative). A 1% rise in PCP drives a 17.64% surge in CFP (median coefficient = 0.1764). Since the 97.5% confidence interval is [0.0148, 0.3380], this indicates that there is a transmission effect between the upstream PCP and the midstream CFP. Notably, this pathway carries the highest volatility risk in the entire dataset (standard deviation = 0.1701), indicating that freight prices are acutely sensitive to coal price fluctuations. Conversely, the reverse transmission (from CFP to PCP) is practically non-existent. This result verifies hypothesis 2. This phenomenon can be explained by two factors: first, the Chinese government actively stabilizes power coal prices to ensure the reliability of critical supplies (e.g., electricity), which isolates the upstream market from midstream fluctuations; second, China’s coal market is dominated by Long-Term Agreement (LTA) contracts. Concentrated supply and pre-negotiated LTA prices (fixed during the contract period) further diminish the prominence of midstream CFP impacts on upstream PCP [42].
Fourth, in China’s natural gas supply chain, the average volatility spillover effect from upstream NGP to midstream LNGMP is 0.1465. A unidirectional and pronounced positive volatility spillover effect exists from upstream natural gas prices to midstream liquefied natural gas market prices. Specifically, a 1% increase in NGP drives a 14.65% rise in LNGMP (median coefficient = 0.1465; 97.5% confidence interval [0.0070, 0.2860]). Conversely, the negative spillover effect from midstream LNG market prices to upstream natural gas prices remains marginal. This result verifies hypothesis 3. Three factors explain this phenomenon: First, the price formation mechanism differs between the two markets: upstream NGP is policy-driven and cost-based, resulting in price rigidity, while midstream LNGMP is market-responsive and lacks reverse influence on upstream pricing. Second, China’s cost-plus pricing model directly transmits upstream NGP increases to midstream costs, amplifying the positive spillover from NGP to LNGMP. Third, natural gas import and supply plans are governed by long-term contracts and stability mandates. These pre-negotiated terms reinforce the dominance of upstream price dynamics over midstream LNGMP.

4. Conclusions

Based on the framework of the DGC-t-MSV combined model, this paper applies the Markov Chain Monte Carlo (MCMC) algorithm and Bayesian principles to estimate parameters using Gibbs sampling. Building on this foundation, nine representative key price indices from the three major energy markets (crude oil, coal, and natural gas) were selected as observation sequences to represent China’s energy supply chain. The price volatility spillover effects across China’s energy supply chains were empirically investigated. The main conclusions are as follows:
(1) In the upstream, midstream, and downstream markets of China’s crude oil supply chain market, there are price volatility spillover effects. Specifically, the upstream Crude Oil Spot Price (COSP) has a positive spillover effect on the midstream Crude Oil Freight Price (COFP) (mean = 0.0251; 95% CI [−0.0291, −0.0211]), while the reverse transmission is not significant (95% CI of COFP to COSP is [−0.0223, −0.0006]). In addition, research has found that the downstream diesel market price has a positive volatility spillover effect on the midstream coastal crude oil freight price. According to the “Administrative Measures for the Collection of Risk Reserve for Oil Price Regulation” of China in 2023, when the diesel wholesale price deviates from the control range, priority should be given to providing subsidies to midstream transportation enterprises (such as state-owned tanker companies), for example, through targeted tax and fee reduction policies, to mitigate the impact of downstream price fluctuations on the midstream market. This policy recommendation helps to enhance the resilience of the supply chain and provides an effective regulatory tool for the government during periods of price fluctuations.
(2) The volatility spillover of China’s coal supply chain presents a one-way characteristic: the upstream PCP dominates the transmission to the midstream CFP. Specifically, the volatility of the upstream power coal price has a positive spillover effect on the midstream freight price (mean = 0.1764; 95% confidence interval [0.0148, 0.3380]), which causes the two to change in the same direction. The reverse transmission of the midstream freight price to the upstream coal price volatility is not significant (mean = −0.0012; 95% confidence interval [−0.0024, −0.0001]). This conclusion indicates that the midstream market is highly sensitive to upstream fluctuations but lacks the ability to adjust. The price risks in the supply chain are mainly transmitted through a “top-down” path. In addition, this conclusion also validates the “supply-guarantee warning line” of Shaanxi Coal Industry Group in 2022. It is recommended that the government mandatorily implement the “three-coverage” policy for long-term coal supply agreements (i.e., covering major users, key regions, and critical periods) to reduce the high sensitivity of the midstream market. This policy recommendation helps to improve the stability of the supply chain and provides a reference for the government’s regulation during periods of coal price fluctuations.
(3) There is an asymmetry in the volatility spillover of China’s natural gas supply chain: the upstream Natural Gas Price (NGP) has a strongly positive spillover effect on the midstream Liquefied Natural Gas Market Price (LNGMP) (mean = 0.1465; 95% confidence interval [0.0070, 0.2860]), whereas the midstream market price shows a weak but negative spillover effect on the upstream price (mean = −0.0235; 95% confidence interval [−0.0457, −0.0012]). The comparison of asymmetric intensities shows that the intensity of positive spillovers is approximately 6.2 times that of negative spillovers, and its 95% confidence interval is wider, further confirming the dominant influence of the upstream sector on the midstream sector. Although the Shanghai Petroleum and Natural Gas Exchange launched the “Imported LNG Window Period Bidding Transaction” mechanism in 2024, the market-oriented reform of natural gas still needs to be continuously promoted to hedge against the risks of upstream price fluctuations. It is recommended that the government further improve the natural gas price formation mechanism and promote market-oriented reforms. For example, by establishing a more flexible price adjustment mechanism or introducing more market participants, the risk-resistance ability of the supply chain can be enhanced.
The research findings on the price volatility spillover effect of China’s energy supply chain in this paper not only clarify the transmission path of price fluctuations among various links in China’s energy supply chain but also deeply reveal the mutual influence mechanism between different links. This research discovery has important practical significance for policymakers and market investors.
For policymakers, this research provides a scientific basis for decision-making, which helps them formulate precise policy measures to alleviate the diffusion effect of price shocks and maintain the stable operation of the energy market. At the same time, based on the research results, policymakers can dynamically adjust market intervention measures, optimize the regulatory framework, and enhance the resilience of the supply chain (for example, in 2022, the National Development and Reform Commission piloted the policy of “10% flexible production capacity release of coal mines” in Inner Mongolia and Shanxi; in 2023, this mechanism successfully stabilized price fluctuations on three occasions). Through these measures, policymakers can effectively prevent the emergence of systemic risks. For market investors, this research provides a theoretical reference for understanding the price transmission efficiency of the supply chain, helping them better identify investment opportunities and risks and make more scientific and effective investment layout decisions. By grasping the transmission law of price fluctuations, investors can enhance the risk-resistance ability of their investment portfolios, optimize resource allocation, improve investment efficiency, and ultimately maintain a competitive edge in market fluctuations (for example, ENN Energy conducted the first domestic long-term hedging of LNG on the Shanghai Crude Oil Futures Exchange to enhance its price buffer capacity).
However, due to the complexity and dispersion of data in the middle and downstream of China’s energy industry chain, including the prices of different coal types, the prices of gas gate stations in different cities, and the prices of various types of coal warehouses, this has brought challenges to the related research. In the subsequent research, an in-depth exploration will be carried out to understand how multi-category data and relevant factors affect the transmission patterns of price fluctuations in the downstream segments of coal- and natural gas-related energy industries.

Author Contributions

J.W., conceptualization and methodology; Y.S., writing—original draft preparation and analysis; L.W., calculation and supervision. L.W., Y.S. and J.W. contributed equally to this work and are co-first authors. All authors have read and agreed to the published version of the manuscript.

Funding

This research is supported by National Natural Science Foundation of China (No. 71971111) and Jiangsu Province’s Youth Science and Technology Talent Support Project (No. JSTJ-2024-438).

Data Availability Statement

The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare that they have no competing interests.

References

  1. Zhang, C.; Mou, X.; Ye, S. How do dynamic jumps in global crude oil prices impact China’s industrial sector? Energy 2022, 249, 123605. [Google Scholar] [CrossRef]
  2. Failler, P.; Liu, Y. The complexity of energy price volatility. Energies 2023, 16, 2354. [Google Scholar] [CrossRef]
  3. Lin, B.; Lan, T. The transmission of coal price shock to Chinese industry: Sub-sectors and regions heterogeneity. Energy 2025, 316, 134471. [Google Scholar] [CrossRef]
  4. Nguyen, D.B.; Nong, D.; Simshauser, P.; Pham, H. Economic and supply chain impacts from energy price shocks in Southeast Asia. Econ. Anal. Policy 2024, 84, 929–940. [Google Scholar] [CrossRef]
  5. Cheng, Y. Forecasting the delayed impact of energy price volatility on China’s general prices based on a temporal input-output approach. Energy Strateg. Rev. 2024, 52, 101340. [Google Scholar] [CrossRef]
  6. Sharif, T.; Ghouli, J.; Bouteska, A.; Abedin, M.Z. The impact of COVID-19 uncertainties on energy market volatility: Evidence from the US markets. Econ. Anal. Policy 2024, 84, 25–41. [Google Scholar] [CrossRef]
  7. Wang, S.; Li, B.; Zhao, X.; Hu, Q.; Liu, D. Assessing fossil energy supply security in China using ecological network analysis from a supply chain perspective. Energy 2024, 288, 129772. [Google Scholar] [CrossRef]
  8. Bashmakov, I. Three laws of energy transitions and economic growth. Energy Effic. 2025, 18, 7. [Google Scholar] [CrossRef]
  9. Sassi, F. The (Un) Intended consequences of power: The global implications of EU LNG strategy to reach independence from Russian gas. Energy Policy 2025, 198, 114494. [Google Scholar] [CrossRef]
  10. He, Q.; Zhang, X.; Xia, P.; Zhao, C.; Li, S. A comparison research on dynamic characteristics of Chinese and American energy prices. J. Glob. Inf. Manag. 2023, 31, 1–16. [Google Scholar] [CrossRef]
  11. Wang, X.; Wang, J.; Wang, W.; Zhang, S. International and Chinese energy markets: Dynamic spillover effects. Energy 2023, 282, 128724. [Google Scholar] [CrossRef]
  12. Goodell, J.W.; Gurdgiev, C.; Paltrinieri, A.; Piserà, S. Global energy supply risk: Evidence from the reactions of European natural gas futures to Nord Stream announcements. Energy Econ. 2023, 125, 106838. [Google Scholar] [CrossRef]
  13. Chen, T.; Zheng, X.; Wang, L. Systemic risk among Chinese oil and petrochemical firms based on dynamic tail risk spillover networks. N. Am. J. Econ. Financ. 2025, 77, 102404. [Google Scholar] [CrossRef]
  14. Dhifaoui, Z.; Ben Jabeur, S.; Khalfaoui, R.; Ali Nasi, M. Time-varying partial-directed coherence approach to forecast global energy prices with stochastic volatility model. J. Forecast. 2023, 42, 2292–2306. [Google Scholar] [CrossRef]
  15. Przekota, A. Directions of Price Transmission on the Diesel Oil Market in Poland. Energies 2025, 18, 139. [Google Scholar] [CrossRef]
  16. Zhao, Y.; Chen, L.; Zhang, Y. Spillover effects of geopolitical risks on global energy markets: Evidence from CoVaR and CAViaR-EGARCH model. Energy Explor. Exploit. 2024, 42, 772–788. [Google Scholar] [CrossRef]
  17. Min, H. Examining the impact of energy price volatility on commodity prices from energy supply chains perspectives. Energies 2022, 15, 7957. [Google Scholar] [CrossRef]
  18. Chen, L.; Pan, L.; Zhang, K. The Dynamic Cointegration Relationship between International Crude Oil, Natural Gas, and Coal Price. Energies 2024, 17, 3126. [Google Scholar] [CrossRef]
  19. Hasanli, M. Re-examining crude oil and natural gas price relationship: Evidence from time-varying regime-switching models. Energy Econ. 2024, 133, 107510. [Google Scholar] [CrossRef]
  20. Guo, Y.; Zhao, H. Volatility spillovers between oil and coal prices and its implications for energy portfolio management in China. Int. Rev. Econ. Financ. 2024, 89, 446–457. [Google Scholar] [CrossRef]
  21. He, Z.; Sun, H. The time-varying and asymmetric impacts of oil price shocks on geopolitical risk. Int. Rev. Econ. Financ. 2024, 91, 942–957. [Google Scholar] [CrossRef]
  22. Pan, C.; Huang, Y.; Lee, C.C. The dynamic effects of oil supply shock on China: Evidence from the TVP-Proxy-VAR approach. Socio-Econ. Plan. Sci. 2024, 95, 102026. [Google Scholar] [CrossRef]
  23. Ma, M.; Zhang, J. A Bayesian analysis based on multivariate stochastic volatility model: Evidence from green stocks. J. Comb. Optim. 2023, 45, 19. [Google Scholar] [CrossRef] [PubMed]
  24. Zhang, J.; Zhuang, Y. Cross-Market Infection Research on Stock Herding Behavior Based on DGC-MSV Models and Bayesian Network. Complexity 2021, 2021, 6645151. [Google Scholar] [CrossRef]
  25. Wang, J.; Zeng, R.; Wang, L. Volatility Spillover Between the Carbon Market and Traditional Energy Market Using the DGC-t-MSV Model. Mathematics 2024, 12, 3789. [Google Scholar] [CrossRef]
  26. Zhang, J.; Zhuang, Y.M.; Liu, J.B. An Empirical Analysis of Oil and Stock Markets’ Volatility Based on the DGC-MSV-t Model. J. Math. 2021, 2021, 6270525. [Google Scholar] [CrossRef]
  27. Yu, L.; Yaacob, H.M.; Buqi, D. Multiscale Spillover transmission in chinas investment preferences using dynamic stochastic volatility framework. Inf. Sci. Lett. 2023, 12, 3283–3298. [Google Scholar]
  28. Wen, D.; Wang, H.; Wang, Y.; Xiao, J. Crude oil futures and the short-term price predictability of petroleum products. Energy 2024, 307, 132750. [Google Scholar] [CrossRef]
  29. Guo, S.; Li, H.; An, H.; Ma, N.; Sun, Q.; Feng, S.; Sun, G.; Liu, Y. Detecting the horizontal/vertical price relationship patterns in the global oil industry chain through network analysis. Energy 2024, 296, 131054. [Google Scholar] [CrossRef]
  30. Qi, Y.; Bai, J.; Liu, S. Spillover dynamics among commodities along the Chinese oil industrial chain: From the perspective of multidimensional networks. Int. Rev. Econ. Financ. 2024, 96, 103612. [Google Scholar] [CrossRef]
  31. Gao, J.; Zhang, L. Environmental regulation, market power and low-carbon development of China’s coal power industry chain—Based on both strategy and return perspectives. Energy Strateg. Rev. 2025, 58, 101651. [Google Scholar] [CrossRef]
  32. Wang, C.; Xu, G.; Sun, C.; Xu, J.; Xu, K.; Jiang, L.; Wang, Y.; Su, S.; Hu, S.; Xiang, J. Modeling and forecasting of coal price based on influencing factors and time series. J. Clean. Prod. 2024, 467, 143030. [Google Scholar] [CrossRef]
  33. Wang, T.; Qu, W.; Zhang, D.; Ji, Q.; Wu, F. Time-varying determinants of China’s liquefied natural gas import price: A dynamic model averaging approach. Energy 2022, 259, 125013. [Google Scholar] [CrossRef]
  34. Rioux, B.; Galkin, P.; Murphy, F.; Feijoo, F.; Pierru, A.; Malov, A.; Li, Y.; Wu, K. The economic impact of price controls on China’s natural gas supply chain. Energy Econ. 2019, 80, 394–410. [Google Scholar] [CrossRef]
  35. Zhang, Y.; Yu, X.; Zhang, P.; Zhang, H. Price impacts of natural gas chokepoints on China: An energy security analysis of the Russia-Ukraine conflict. Util. Policy 2025, 96, 101985. [Google Scholar] [CrossRef]
  36. Gunawan, D.; Kohn, R.; Nott, D. Variational Bayes approximation of factor stochastic volatility models. Int. J. Forecast. 2021, 37, 1355–1375. [Google Scholar] [CrossRef]
  37. Cheng, J.; Mohammed, K.S.; Misra, P.; Tedeschi, M.; Ma, X. Role of green technologies, climate uncertainties and energy prices on the supply chain: Policy-based analysis through the lens of sustainable development. Technol. Forecast. Soc. Change 2023, 194, 122705. [Google Scholar] [CrossRef]
  38. Chen, S.; Kuang, H.; Meng, B. The dependence structures between geopolitical risks and energy prices: New evidence from regional heterogeneity and quantile-on-quantile perspective. Energy 2024, 310, 133325. [Google Scholar] [CrossRef]
  39. Ren, S.; Jiao, X.; Zheng, D.; Zhang, Y.; Xie, H.; Zhang, R. Impact of Carbon Neutrality Goals on China’s Coal Industry: Mechanisms and Evidence. Energies 2025, 18, 1672. [Google Scholar] [CrossRef]
  40. Wang, Z.; Zhang, H.; Li, H.; Zhang, B. Towards sustainable future: Assessing the impact of coal phase-down on the sustainable development goals in China. J. Environ. Manag. 2025, 378, 124713. [Google Scholar] [CrossRef]
  41. Apergis, N.; Fahmy, H. Geopolitical risk and energy price crash risk. Energy Econ. 2024, 140, 107975. [Google Scholar] [CrossRef]
  42. Wang, T.; Wu, F.; Dickinson, D.; Zhao, W. Energy price bubbles and extreme price movements: Evidence from China’s coal market. Energy Econ. 2024, 129, 107253. [Google Scholar] [CrossRef]
Figure 1. Schematic diagram of price transmission in China’s crude oil, coal, and natural gas supply chains.
Figure 1. Schematic diagram of price transmission in China’s crude oil, coal, and natural gas supply chains.
Energies 18 03204 g001
Figure 2. Time–return series of nine energy supply chains price indices in China.
Figure 2. Time–return series of nine energy supply chains price indices in China.
Energies 18 03204 g002
Figure 3. Comparison of time–return price index series of China’s three major energy supply chains: crude oil, coal, and natural gas.
Figure 3. Comparison of time–return price index series of China’s three major energy supply chains: crude oil, coal, and natural gas.
Energies 18 03204 g003
Figure 4. Iterative volatility of China’s three major energy supply.
Figure 4. Iterative volatility of China’s three major energy supply.
Energies 18 03204 g004
Figure 5. Autocorrelation diagram.
Figure 5. Autocorrelation diagram.
Energies 18 03204 g005
Figure 6. Gelman–Rubin diagnostic chart.
Figure 6. Gelman–Rubin diagnostic chart.
Energies 18 03204 g006
Figure 7. Correlation between the dynamics of China’s crude oil, coal and natural gas energy supply chains price indices.
Figure 7. Correlation between the dynamics of China’s crude oil, coal and natural gas energy supply chains price indices.
Energies 18 03204 g007
Table 1. Descriptive statistics of the returns of nine energy supply chains price indices in China.
Table 1. Descriptive statistics of the returns of nine energy supply chains price indices in China.
NodeMeanMedianMaxiMiniSDSkewnessKurtosisJ-B
COSP−0.01070.003136.1749−25.52722.41391.121232.9372134,852.0622
LNGMP0.00230.001533.4422−28.38652.09030.214095.59311,282,835.8961
PCP−0.01380.01499.3143−17.93050.9741−1.180152.1331362,037.3924
CFP−0.00340.000830.0067−18.16661.77261.056451.1017346,866.6450
NGP−0.00320.002648.5508−31.36583.18491.083649.8140319,462.6619
COFP−0.00020.001934.5263−34.52630.83610.02511620.4936391,461,713.7542
DMP−0.00350.00037.7218−6.97750.71850.005137.0257173,227.8824
92GMP−0.00220.00479.7041−10.62760.7208−0.764961.4800512,054.8234
Table 2. Preliminary data test of index returns.
Table 2. Preliminary data test of index returns.
NodeADF StatisticADF p-ValueARCH LM StatARCH LM p-Value
COSP−59.57768710.001499.25143640.01
COFP−45.06796740.001234.73718420.01
DMP−59.06775770.001235.86823990.01
92GMP−43.37643850.00145.762402510.01
95GMP−25.85638540.00186.989332710.01
PCP−38.60646920.00119.792857310.01
CFP−64.07852150.001127.78625030.01
NGP−53.83544010.00198.474636330.01
LNGMP−19.257375410.00163.776346910.01
Table 3. Dynamic statistical depiction of correlation among price metrics in China’s crude oil, coal, and natural gas energy provision chains.
Table 3. Dynamic statistical depiction of correlation among price metrics in China’s crude oil, coal, and natural gas energy provision chains.
NodeMeanSEStdMaxMinRange95%CI
ρ C O S P C O F P −0.03020.13170.18390.4956−0.41510.9106[−0.0822, −0.0252]
ρ C O S P D M P −0.03530.19450.13060.3733−0.41960.7930[−0.1875, −0.0198]
ρ C O S P 92 G M P −0.01640.28640.12930.3533−0.40170.7551[−0.1695, −0.0382]
ρ C O S P 95 G M P −0.02180.14720.13130.3616−0.39280.7545[−0.0747, −0.0329]
ρ P C P C F P 0.01300.00220.12810.5045−0.41640.9209[0.0087, 0.0174]
ρ N G P L N G M P 0.00260.00230.12840.5664−0.46611.0324[0.0019, 0.0071]
Table 4. p-value table of time series for China’s crude oil, coal, and natural gas energy supply chains.
Table 4. p-value table of time series for China’s crude oil, coal, and natural gas energy supply chains.
NodeCOSP and COFPCOSP and DMPCOSP and GMP92COSP and GMP95PCP and CFPNGP and LNGMP
p-value0.04080.02890.02110.04560.03790.0166
Table 5. Volatility spillover among price indices in Chinese crude oil, coal, and natural gas energy supply chains.
Table 5. Volatility spillover among price indices in Chinese crude oil, coal, and natural gas energy supply chains.
VertexMeanSDMC ErrorSE2.50%5.00%Median97.50%
COSP to COFP0.02510.00420.00300.0145−0.0291−0.0289−0.0251−0.0211
COFP to COSP0.01140.01150.00810.1852−0.0223−0.0218−0.0114−0.0006
COFP to DMP−0.00380.01470.01210.0964−0.0200−0.0192−0.00380.0124
DMP to COFP0.01870.01890.01320.01670.00090.00180.01870.0365
COFP to GMP92−0.00910.00630.00830.0095−0.0203−0.0197−0.00910.0020
GMP92 to COFP0.01700.01710.01210.01630.00080.00170.01700.0332
COFP to GMP95−0.00710.01380.00980.0082−0.0202−0.0195−0.00710.0061
GMP95 to COFP0.02620.02630.01860.16510.00120.00260.02620.0512
PCP to CFP0.17640.17010.12030.09430.01480.02330.17640.3380
CFP to PCP−0.00120.00120.00090.0037−0.0024−0.0024−0.0012−0.0001
NGP to LNGMP0.14650.14680.10380.15260.00700.01440.14650.2860
LNGMP to NGP−0.02350.02340.01660.0051−0.0457−0.0446−0.0235−0.0012
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Wang, L.; Sun, Y.; Wang, J. Price Volatility Spillovers in Energy Supply Chains: Empirical Evidence from China. Energies 2025, 18, 3204. https://doi.org/10.3390/en18123204

AMA Style

Wang L, Sun Y, Wang J. Price Volatility Spillovers in Energy Supply Chains: Empirical Evidence from China. Energies. 2025; 18(12):3204. https://doi.org/10.3390/en18123204

Chicago/Turabian Style

Wang, Lei, Yu Sun, and Jining Wang. 2025. "Price Volatility Spillovers in Energy Supply Chains: Empirical Evidence from China" Energies 18, no. 12: 3204. https://doi.org/10.3390/en18123204

APA Style

Wang, L., Sun, Y., & Wang, J. (2025). Price Volatility Spillovers in Energy Supply Chains: Empirical Evidence from China. Energies, 18(12), 3204. https://doi.org/10.3390/en18123204

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop