Resources Policy 91 (2024) 104712

Contents lists available at ScienceDirect

Resources Policy
journal homepage: www.elsevier.com/locate/resourpol

A wavelet analysis of the relationship between carbon emissions rights and

crude oil prices in China
Huidian Long a, b, *, Yafei Cao a
a
School of Mathematics and Statistics Guangdong University of Foreign Studies, Guangzhou, 510420, PR China
b
Business School University of Shanghai for Science and Technology, Shanghai, 200093, PR China

A R T I C L E I N F O A B S T R A C T

Keywords: In this paper, we aim to explore the relationship between the carbon emissions and oil prices from 2014 to 2021
Carbon prices in China using the Morlet continuous wavelet and the maximal overlap discrete wavelet transform. The results
Oil prices indicate that there is a positive correlation in the period from 2014 to 2018 with carbon prices leading. After
Wavelet analysis
2019, the two series show a positive correlation in the medium and high-frequency domains with oil prices
leading.

1. Introduction serves as the theoretical basis for future carbon taxes. Coase (1959,
1960) was the first one to propose a course of action to eliminate the
As the world’s largest energy producer, consumer, and carbon di­ inefficiencies associated with externalities through bargaining among
oxide emitter, China has proposed the “dual carbon” goals of controlling affected parties. A prerequisite of such a bargaining process is a clear
the total output and intensity of carbon emissions to address increas­ delineation of property rights (Coase, 1959). In his seminal paper on
ingly urgent pollution issues. As a market-based mechanism, carbon social costs, Coase (1960) demonstrates that when transaction costs of
trading has obvious advantages in achieving these goals. From the bargaining, contracting, monitoring etc. are zero, mutual benefits could
establishment of carbon trading pilot projects in seven provinces and be obtained for the polluting parties to pay the polluted parties for the
cities in 2011 to the formal launch of the carbon market in 2017, China’s pollution if the latter possess the right of clean air. According to this
carbon trading market has expanded rapidly, with a cumulative trading theory, the externalities of the environment can also be internalized by
volume of nearly 300 million tons and an accumulated transaction clarifying the property rights of environmental resources. Carbon
volume exceeding 10 billion yuan. However, compared with the mature emission trading based on the Coase property right theorem is a
carbon market in the EU, China’s carbon market is still in its initial powerful means to realize the optimal allocation of environmental re­
stages, and the research on carbon trading is still very limited. There­ sources. Now, the carbon trading market has gradually evolved into a
fore, developing an in-depth understanding of the influencing factors of new type of international commodity trading market. As a commodity,
carbon prices in China is essential in establishing a national carbon the market situation determines the equilibrium price of carbon emis­
trading market. sion rights.
The theoretical basis of carbon trading is to solve the environmental Crude oil is one of the most important strategic resources and is
externalities of greenhouse gas emissions through market mechanisms. known as the “blood” of modern economic and social systems. Since
Neoclassical economics believes that externalities have originated from crude oil consumption is the main source of carbon emissions, there is an
a market failure in which private costs are not equivalent to social costs, inherent link between the crude oil and carbon trading markets (Li et al.,
thus failing to produce an efficient outcome for society as a whole. This 2018). The mechanism through which oil prices are transmitted to
school of thought urges the government to intervene in the matter, to carbon prices is relatively complex and can be divided into direct and
eliminate the effects of externalities. Particularly, Pigou (1920) recom­ indirect effects. The direct effects are that crude oil prices can promote
mends that governments penalize the polluter by imposing an amount of industrial agglomeration and accelerate technological spillover effects
tax equivalent to the cost of harm to affected parties. This method of (i.e., information exchange) among industries, thereby motivating the
correcting externalities is called the “Pigou tax” scheme, which further energy industry to be greener and more efficient (Alshehry and

* Corresponding author. School of Mathematics and Statistics Guangdong University of Foreign Studies, Guangzhou, 510420, PR China.
E-mail address: [email protected] (H. Long).

https://doi.org/10.1016/j.resourpol.2024.104712
Received 3 March 2023; Received in revised form 15 October 2023; Accepted 18 January 2024
Available online 6 March 2024
0301-4207/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
H. Long and Y. Cao Resources Policy 91 (2024) 104712

Belloumi, 2015). The indirect effects are that oil prices affect carbon the short and long terms, changes in oil prices explained changes in
prices through factors such as economic development, industrial carbon emissions, while GDP did not explain their growth. A VECM
upgrading, and technological progress, etc. (Zhang and Wei, 2010). specification depends only on the time domain and the constancy of the
Hence, crude oil prices are essential for explaining the overall level of parameters, whereas the application of the wavelet method allows us to
carbon emissions (Maghyereh et al., 2019). Given the interrelationships study both short and long run frequencies and thereby recognize how
between financial markets, carbon prices can also lead to fluctuations in the relationships among the key variables have evolved in the short,
oil prices. Carbon pricing mechanisms have a direct impact on the gas medium, and long terms.
industry and have become an important factor affecting the strategic Our paper is most closely related to the works of Tansuchat and
development and investment decisions of oil companies (Peng et al., Chang (2022), Wu et al. (2020) and Agbanike et al. (2019). These
2022). The carbon and oil markets are becoming more and more inter­ studies, using a variety of econometric methods that include the
connected for the price and volatility spillover effects. non-parametric, regression and co-integration approaches, tend to sug­
Although scholars have researched the relationship between carbon gest that a relationship exists between the two markets of our concern.
emissions and crude oil prices extensively, they have not yet reached a We aim to build on the abovementioned works, especially the nonlinear
consensus on its direction and magnitude. For example, Alberola et al. causality approach of Geng et al. (2017) and, to some extent, the cor­
(2008) and Zou (2018) found that crude oil prices positively impact relation analyses of Geng et al. (2016a,b), with the variables being
carbon prices using the multiple linear regression method and VECM dis-aggregated into their various frequencies based on EEMD. We do this
model. Khalifa et al. (2015) found that crude oil prices negatively affect by relying on coherency and phase differences in a wavelet analysis,
carbon prices using the nonlinear autoregressive distributed lag model. which allows us to delineate a time-varying causal relationship across
In addition, most of these studies focus on mature carbon trading mar­ frequencies between carbon prices and crude oil prices.
kets such as those in Europe and the United States rather than that in This paper has made the following contributions to the existing
China. It should also be noted that with rapid industrialization and ur­ literature. Firstly, we explore the relationship between carbon and oil
banization in China, its demand for crude oil has risen sharply and it has prices in China by using wavelet analysis and expand the research in
overtaken the US as the world’s largest net importer of oil. related fields. When assessing the variables influencing carbon pricing,
Many scholars have studied the relationship between crude oil and most current research centers on the developed EU carbon trading
carbon prices in China. For example, Wang and Hu, 2018 found that market. By emphasizing how crude oil prices affect the carbon price in
crude oil prices mainly affected carbon prices in the high-frequency China, it provides Chinese experience for China to establish a unified
range using ensemble empirical mode decomposition (EEMD). Ji and carbon emission market. Secondly, we creatively introduce wavelet
Fangang (2021) and Wen and Chen (2022) used Daqing crude oil prices analysis to investigate the dynamic relationship between crude oil prices
to study the relationship between the crude oil and carbon markets in and carbon prices. This model can overcome the shortcomings of the
China. The results show that oil price shocks positively impact China’s traditional econometric model and rolling window and capture the time-
carbon price. However, the research on the impact of international varying relationship among economic variables. The application of
crude oil prices on China’s carbon price is still very limited. wavelet analysis provides new evidence for the correlation between oil
In this paper, we aim to explore the relationship between interna­ and carbon prices from the perspective of time-varying and time delay.
tional crude oil and carbon prices in China for the period from 2014 to Finally, our results more vividly describe the intensity and direction of
2021. We advance the existing literature by conducting a wavelet time-varying spillover effects between crude oil price and carbon price
analysis on the relationship between the variables of interest. Wavelets at different time points. It provides experience for carbon trading market
are a time-varying methodology that include both the time and fre­ participants on how to avoid the impact of risks in advance.
quency domains. This method requires the decomposition of a time se­ The remainder of this paper is structured as follows. In the following
ries into time-frequency space, which allows researchers to identify the section, we describe our data set and provide its descriptive statistics.
dominant modes of variability and the variations in those modes over Section 3 presents an overview of the methodology applied. In Section 4,
time (Torrence and Compo, 1998). Wavelets are often used in we discuss the results of the empirical model. Finally, Section 5 provides
geophysical applications, where time series data may be nonstationary a few concluding remarks and the policy implications of our findings.
or contain dominant periodic signals, which can vary in terms of both
amplitude and frequency over the longer time frames (Torrence and 2. Data
Compo, 1998). For example, to explore the relationship between natural
gas and crude oil prices in the US between 1997 and 2017, Tiwari et al. Considering that trading is most active in all markets in Guangdong
(2019) applied a wavelet method and found that significant or high Province, we choose the carbon trading data of Guangdong Province as
wavelet coherency was observed during 2000–2015 for periods of 3–4 our primary data set. We use monthly data from March 2014 to
years. In the economics literature, wavelet analysis has been used to November 2021 for all of our variables. Carbon prices (in dollars per
study time series data that contain combinations of components oper­ ton) are the monthly average price of the carbon emissions quota in
ating at different frequencies. Boubaker and Raza (2017) combined Guangdong Province (GDEA). Oil prices are the monthly average spot
wavelet decomposition with the ARMA (1,1)-GARCH (1,1) model to price (in dollars per barrel) of Daqing crude oil (DO) and Brent crude oil
analyze the relationship between oil prices and stock prices in BRICS (BO) from the Wind database. The following figure (Fig. 1) provides the
countries. Monge et al. (2017) studied the relationship between US descriptive statistics for the full sample.
crude oil production and WTI crude oil prices between 2000 and 2016 in
the time-frequency domain by applying a wavelet tool. They observed 3. Methodology
higher frequencies between 2003 and 2009, which suggested that a
short-term relationship was existed, and lower frequencies for the period We apply the following modeling methods and their combinations:
between 2009 and 2014, which indicated the presence of a long-term the continuous wavelet transform (CWT), the cross-wavelet transform,
component in the relationship between crude oil production and crude wavelet coherence, phase difference theory and the discrete wavelet
oil prices. transform.
Our paper is related to the paper-Zou (2018). Zou (2018) analyzed
the impact of the relationships between US oil prices, carbon emissions 3.1. Continuous wavelet transform
and GDP using data for the period between 1987 and 2017. Then he
applied ARIMA, VAR, and VECM to establish a synthesis integration The objective of wavelet analysis is to determine the frequency
model and verify the robustness of the model. He concluded that in both content of a variable with a view toward extracting the temporal vari­

2
H. Long and Y. Cao Resources Policy 91 (2024) 104712

where δt is the uniform step size and * denotes a conjugate complex

number. By transforming the wavelet scale s and localizing along the
time index n, a map showing the fluctuation characteristics of the time
series at a certain scale and its variation with time—that is, the wavelet
power spectrum—can be obtained. From the expression above, the
wavelet power spectrum that measures the variability in both time and
⃒ ⃒2
frequency is defined as ⃒Wx (s)⃒ .This power spectrum expresses the
m
degree of fluctuation of the time series in a given wavelet scale and time
domain (Lafrenière and Sharp, 2003). By time-averaging the wavelet
power spectrum in a certain period, we can obtain the global wavelet
spectrum. Since the full wavelet spectrum can show a measure of the
background spectrum, the peaks of the local wavelet spectrum can be
Fig. 1. Carbon emissions and oil prices. verified. Because of this characteristic, the periodic fluctuation charac­
teristics of the time series and its intensity can be clearly distinguished
from the full wavelet spectrum. Since the wavelet transform assumes
ations in this frequency content (David Labat, 2005). A wavelet is a
that the data are cyclic, when we deal with a finite time series, the
function with a mean of zero localized in both time and frequency. It
wavelet edge effects occur in the power spectrum—that is, errors occur
grows quickly and decays within a limited period (Fan and Gençay,
∫ ∫ at the beginning and end of the power spectrum. To cope with this
2010), thereby satisfying the conditions that φ(η)dη = 0 and
challenge, the cone of influence (COI), which represents the wavelet
|φ(η)|2 dη = 1.We can characterize a wavelet by its localization in time spectrum region and its corresponding edge effect, is introduced. Here,
(Δt) and frequency (Δω or the bandwidth). Thus, for a CWT of series we take the COI as the area in which the wavelet power drops to e− 2 of
x(t): the value at the edge.
∫ +∞ (t − τ )
1
Wx (τ, s) = √̅̅̅̅̅ x(t)ψ ∗ dt (1) 3.2. Wavelet coherence
|s| − ∞ s

where s and τ are the scale and location parameters, ψ ((t − τ) /s) is the Since our intention is to measure the extent of synchronization be­
“mother” wavelet function, which is possibly complex-valued, and * tween two given time series, it is informative to use the coherence be­
denotes the complex conjugate. From the formula, we can see that the tween them. Wavelet coherence (WTC) analysis can simultaneously
result of the wavelet transform is a binary function of τ and s, which is analyze the correlation between two signals in the time-frequency
consistent with the fact that the wavelet transform is a time-frequency domain. Given two time series x(t) and y(t) with wavelet transforms
domain transform. x(t) is the original time domain signal and ψ (t) is Wx (τ, s) and Wy (τ,s), the cross-wavelet spectrum is defined as Wxy (τ,s) =
the wavelet basis. There are many kinds of wavelet bases and different Wx (τ, s)Wy∗ (τ, s).Following Torrence and Webster (1999), we define the
wavelet bases have different characteristics, so choosing the appropriate wavelet squared coherency as:
wavelet base for the wavelet transform is an important step. These ⃒ ( −1 )⃒
⃒S s Wxy (τ, s) ⃒2
wavelet bases are of finite length in the time domain, which is why they R2 (τ, s) = ( ) ( ⃒ ⃒2 ) (4)
are called “wavelets.” τ and s are the translation and scaling of the S s− 1 |Wx (τ, s)|2 ⋅S s− 1 ⃒Wy (τ, s)⃒
wavelet basis, and the negative half of the absolute value of s is used for
energy normalization. Intuitively, this formula calculates the orthogonal where S( ⋅) denotes smoothing in both time and scale and s− 1 is the
values of the original signal and the different scale transformation and energy density conversion factor. The R2 (τ, s) is between 0 and 1, with a
translation transformation versions of the wavelet basis. When high (low) value indicating a strong (weak) relationship. Hence, one can
analyzing time series, smooth and continuous wavelet amplitudes are distinguish the regions in the time-frequency space by analyzing the plot
desirable, so using a nonorthogonal wavelet function is appropriate in of the wavelet squared coherency, where the link is stronger and can
our research context. In addition, to obtain the amplitude and phase of identify both time- and frequency-varying features. Additionally, we can
the time series, it is necessary to choose the complex-valued wavelet also compute the wavelet phase, which captures the lead-lag relation­
because it has an unquantifiable component and can express the phase ship between the variables in the time-frequency space. The wavelet
well (Torrence and Compo, 1998). Morlet wavelets are not only non­ phase difference is defined as:
orthogonal, but also an exponential complex-valued wavelet tuned by ( ( ))
Gaussian. We employ the Morlet wavelet function given by: − 1 I Wxy (τ, s)
φ(τ, s) = tan ( ) (5)
t2
R Wxy (τ, s)
(2)
1
ψ ω0 (t) = π− 4 eiω0 t e− 2

where ℜ and ℑ are the real and imaginary components, respectively. The
where t is dimensionless time and ω0 is dimensionless frequency. For result provides information about the lead-lag relationship between the
optimal balance, we set ω0 = 6, as suggested by Torrence and Compo two series. A phase difference of zero indicates that the time series move
(1998), such that the wavelet scale s is equal to the Fourier period (λ; λ = together at the specified frequency. The range of values φ(τ, s) is [ − π, π],
1.03s) (Torrence and Webster, 1999). The scale and periodic terms of and the phase difference ranges represent different relationships be­
Morlet wavelets can be substitute for each other. Morlet wavelets have a tween sequence x and sequence y. When the phase difference is zero, the
good balance between time and frequency localization (Grinsted et al., time series move together at the specified frequency. If φ(τ,s) ∈ [0, π /2],
2004). In addition, Morlet wavelets also contain more vibration infor­ then the series move in-phase, with time series y leading time series x.
mation, and their wavelet power can contain positive and negative On the other hand, if φ(τ, s) ∈ [ − π /2，0], then time series x is leading.
peaks within a broad peak (Torrence and Compo, 1998). The discretized When φ(τ,s) ∈ [ − π， − π /2] ∪ [π /2，π], there is an anti-phase relation.
version of Eq. (1) for time series {xn : = 1, 2, ⋯N } is given by: If φ(τ, s) ∈ [ − π， − π /2], then time series y is leading; time series x is
√̅̅̅̅ N − 1 ( leading if φ(τ, s) ∈ [π /2，π].
δt ∑ δt)
Wmx (s) = xm ψ ∗ (m − n) m = 1, 2, ⋯N − 1 (3)
s n=0 s

3
H. Long and Y. Cao Resources Policy 91 (2024) 104712

4. Results and discussion The cross-wavelet power spectrum reflects the leading and lagging
relationships according to the arrow direction and the phase difference
In this paper, our objective is to examine the relationship between represents the specific leading and lagging time between two time series
carbon emissions rights and crude oil prices. To do so, we use monthly in a certain period. The results of wavelet coherence, partial WTC and
data from March 2014 to November 2021. Fig. 2 has six panels in which associated phase differences are presented in Fig. 3. The colors used in
we present the results of a continuous wavelet power spectrum. The left this work range from dark blue (no coherency) to red (strongest wavelet
side is the average wavelet power plot, with the ordinate axis repre­ coherency). In Fig. 3, Panel a.1 presents the results of wavelet coherency
senting periods (in months), the abscissa axis representing the average between carbon emissions and Daqing crude oil prices, while Panel b.1
wavelet power, and the red dots representing the 5% significance level. shows the results of partial wavelet coherency between carbon emis­
The continuous wavelet power spectrum is shown on the right side, with sions and BO prices. Panel a.2 presents the 8-16-month fluctuation cycle
red indicating stronger power, blue indicating weaker power, and black phase differences associated with the wavelet coherency exhibited by
ridges indicating the strongest power. It is evident from Fig. 1 that there Panel a.1, and Panel a.3 presents the 16-32-month fluctuation cycle
are some common features in the wavelet power of the correlation be­ phase differences. Similarly, Panel b.2 presents results related to the
tween the carbon emissions and crude oil prices. Specifically, the com­ phase differences associated with the wavelet coherency exhibited by
mon features in the high (or significant) wavelet power of the two time Panel b.1, and Panel b.3 presents the 16-32-month fluctuation cycle
series are evident in 1-4-year scales occurring in 2014–2021. In addi­ phase differences. The wavelet coherency between carbon emissions and
tion, it can be seen that the fluctuations in domestic and international oil prices in Panels a.1 and b.1 show that there is evidence of significant
crude oil prices are roughly the same and have a common 4-year cycle in wavelet coherency during 2013–2021 for the 8-16-month fluctuation
the low-frequency range. However, the observed similarity between the cycle. We also notice that there is an increase in the wavelet power of the
patterns in the two series in these periods may merely be a coincidence coherency between 2013 and 2015, that is, the size of the red region
rather than the result of a cause-and-effect relationship. To analyze the increases and high wavelet coherency power is observed during the 8-
relationship between them, we use the cross-wavelet transform and 16-month period. If we analyze the results according to the phase dif­
phase difference associated with each model. ferences, we find from Panels a.2 and b.2 that during 2013–2018, the 8-

Fig. 2. Continuous wavelet transform.

4
H. Long and Y. Cao Resources Policy 91 (2024) 104712

Fig. 3. Wavelet coherency and phase differences.

16-month fluctuation cycle phase differences are between [0, π /2], thus
Table 1
indicating that both series move cyclically and the carbon emissions
Cross-wavelet results for each variable.
price leads oil price peaks throughout period, except in 2017 when
phase differences are between [− π/2,0], which indicates that natural Variable Frequency
gas production is leading during this period. We also find that during High frequency Medium frequency (16–32 months)
2019–2021, the 8-16-month fluctuation cycle phase differences are be­ (8–16 months)
tween [ − π /2,0], thus indicating that both series move cyclically and the 2014–2016 2014–2017 2019–2021
oil price leads the carbon emissions price. Last, the 16-32-month phase
(GDEA, In phase In phase In phase
differences indicate that the carbon emissions price leads the oil price DO) GDEA leads DO by GDEA leads DO by GDEA lags DO by
before 2019 and that the oil price leads after 2019. The details for the about 1.5 months about 1.5 months about 1.5 months
cross-wavelet and phase difference results between the carbon emissions (GDEA, In phase In phase In phase
BO) GDEA leads DO by GDEA leads DO by GDEA lags DO by
and crude oil prices are listed in Table 1.
about 1.5 months about 1.5 months about 2.25 months
In general, the prices of carbon emissions and oil are interdependent
in the medium- and high-frequency states, and this interdependence is a
dynamic process. The correlation between the two series is significant in positive correlation in the medium- and high-frequency domains. Dur­
the period from 2014 to 2017, during which the arrow in Fig. 3 mainly ing this period, the arrow in Fig. 3 mainly points to the lower left, thus
points to the upper right, thus indicating that there is a positive corre­ indicating that the change in the oil price during this period led to the
lation and a lead-lag relationship between them. In this period, the price change in the price of carbon emissions. The likely reasons for this
of carbon emissions is leading. One possible reason for this finding is as finding are that China’s carbon market matured after 2019 and that
follows. The oil price shows a sagging tendency during this period, while fluctuations in the oil price only have a short-term impact on energy
the domestic carbon market is only recently establishing itself and the consumption. When the oil price is too high, it will lead to an increase in
carbon market is generally oversupplied. People are generally pessi­ coal consumption, thus increasing carbon emissions and leading to a rise
mistic about the carbon price, so the decline in the carbon price may in carbon prices.
have led to that in the oil price. After 2019, the two series also show a

5
H. Long and Y. Cao Resources Policy 91 (2024) 104712

5. Conclusions References

As a commodity with financial attributes, carbon assets are affected Agbanike, T.F., Nwani, C., Uwazie, U.I., Anochiwa, L.I., Onoja, T.-G.C., Ogbonnaya, I.O.,
2019. Oil price, energy consumption and carbon dioxide (CO2) emission: insight into
by many complex factors. Accordingly, the relationship between carbon sustainability challenges in Venezuela. Lat. Am. Econ. Rev. 28 https://doi.org/
pricing and the energy market is expected to be dynamic. Revealing this 10.1186/s40503-019-0070-8.
dynamic connection has three benefits for market decision-makers and Alberola, E., Chevallier, J., Chèze, B., 2008. The EU emissions trading scheme:
Disentangling the effects of industrial production and CO2 emissions on carbon
participants. First, China’s carbon trading system is not only an impor­ prices. SSRN Electronic J. https://doi.org/10.2139/ssrn.1133139.
tant tool through which policymakers can achieve the emissions Alshehry, A.S., Belloumi, M, 2015. Energy consumption, carbon dioxide emissions and
reduction goals, but also helps market policymakers in formulating economic growth: The case of Saudi Arabia. Renew. Sustain. Energy Rev. 41,
237–247. https://doi.org/10.1016/j.rser.2014.08.004.
policies by clarifying how emissions quotas are priced. Second, it is of Boubaker, H., Raza, S.A., 2017. A wavelet analysis of mean and volatility spillovers
great significance for enterprises to adjust their emissions and use of between oil and BRICS stock markets. Energy Econ. 64, 105–117. https://doi.org/10
fossil fuels. Third, estimating the response of the carbon emissions price .1016/j.eneco.2017.01.026.
Coase, R.H., 1959. The federal communications commission. J. Law Econ. 2 (Oct), 1–40.
to several influencing factors is necessary for strengthening risk man­
Coase, R.H., 1960. The problem of social costs. J. Law Econ. 3 (1), 1–44.
agement practices. In this paper, we analyze the relationship between Fan, Y., Gençay, R., 2010. Unit root tests with wavelets. Econom. Theor. 26, 1305–1331.
the prices of carbon emissions and crude oil between 2013 and 2021 by https://doi.org/10.1017/S0266466609990594.
applying a wavelet methodology, which allows us to study the dynamic Geng, J.-B., Ji, Q., Fan, Y., 2016a. The impact of the North American shale gas revolution
on regional natural gas markets: evidence from the regime switching model. Energy
correlations of the two series in a time-frequency space. The results Pol. 96, 167–178. https://doi.org/10.1016/j.enpol.2016.05.047.
indicate that there is a high wavelet coherency between the variables of Geng, J.-B., Ji, Q., Fan, Y., 2016b. How regional natural gas markets have reacted to oil
interest during 2013–2021 on a frequency of 8–32 months. Between price shocks before and since the shale gas revolution: a multi-scale perspective.
J. Nat. Gas Sci. Eng. 36, 734–746. https://doi.org/10.1016/j.jngse.2016.11.020.
2013 and 2015, the wavelet power of the coherency further increased on Geng, J.-B., Ji, Q., Fan, Y., 2017. The relationship between regional natural gas markets
a frequency of 8–16 months. The results of the phase differences show and crude oil markets from a multi-scale nonlinear Granger causality perspective.
that except for a single year (2017), carbon emissions and oil prices Energy Econ. 67, 98–110. https://doi.org/10.1016/j.eneco.2017.08.006.
Grinsted, A., Moore, J.C., Jevrejeva, S., 2004. Application of the cross wavelet transform
moved cyclically between 2006 and 2015, with carbon prices leading oil and wavelet coherence to geophysical time series. Nonlinear Process. Geophys. 11
prices before 2019 and oil prices leading carbon prices after 2019. This (5/6), 561–566. https://doi.org/10.5194/npg-11-561-2004.
indicates that when carbon emissions and oil prices are in phase, they Ji, Ziteng, Fangang, 2021. Research on Central Bank policies under the background of
carbon peaking and carbon neutrality. Jianghan Tribune 3, 69–74.
move together and in the same direction. Khalifa, A., Caporin, M., Hammoudeh, S., 2015. Spillovers between energy and FX
Since the price of carbon emissions is related to the price of oil, the markets: The importance of asymmetry, uncertainty and business cycle. Energy Pol.
following suggestions are proposed. To begin, the government should 87, 72–82. https://doi.org/10.1016/j.enpol.2015.08.039.
Labat, D., 2005. Recent advances in wavelet analyses: Part 1. A review of concepts.
strengthen the energy reserve system and stabilize energy prices. At
J. Hydrol. 314, 275–288. https://doi.org/10.1016/j.jhydrol.2005.04.003.
present, China’s carbon market is in its initial stages and thus it should Lafrenière, M., Sharp, M., 2003. Wavelet analysis of inter-annual variability in the runoff
learn from the experience of developed countries to improve the do­ regimes of glacial and nival stream catchments, Bow Lake, Alberta. Hydrol. Process.
mestic energy reserve system and control market volatility, which may 17 (6), 1093–1118. https://doi.org/10.1002/hyp.1187.
Li, W., Sun, W., Li, G., Jin, B., Wu, W., Cui, P., Zhao, G., 2018. Transmission mechanism
be caused by oil market waves. The energy sector should adjust the between energy prices and carbon emission using geographically weighted
availability of energy storage in response to oil price shocks to stabilize regression. Energy Pol. 115, 434–442. https://doi.org/10.1016/j.
the price of carbon. Second, the government should optimize the oil enpol.2018.01.005.
Maghyereh, A.I., Awartani, B., Abdoh, H., 2019. The co-movement between oil and clean
supply structure, limit the consumption of fossil energy, especially that energy stocks: a wavelet-based analysis of horizon associations. Energy 169,
with high carbon content, and purchase natural gas, liquefied petroleum 895–913. https://doi.org/10.1016/j.energy.2018.12.039.
gas, and other relatively high-quality clean energy sources from the Monge, M., Gil-Alana, L.A., de Gracia, F.P., 2017. U.S. shale oil production and WTI
prices behavior. Energy 141, 12–19.
international market in a structured manner. Finally, enterprises should Peng, Yun, Liang, Xiong, Li, Jia, Yi, Jiexin, Li, Hongwei, 2022. Status and trend of global
further reduce their carbon dioxide emissions from fossil energy con­ carbon pricing mechanism and its impacts on petroleum industry. China Pet. Explor.
sumption by investing in technological innovation. 27 (6), 41–53.
Pigou, A.C., 1920. The Economics of Welfare. Macmillan and Co, London.
Tansuchat, R., Chang, C.-L., 2022. Latent carbon emission pricing model for Thailand: a
Authorship contributions nonlinear autoregressive distributed lag model. Energy Rep. 8, 768–775. https://doi.
org/10.1016/j.egyr.2022.05.187.
Tiwari, A.K., Mukherjee, Z., Gupta, R., Balcilar, M., 2019. A wavelet analysis of the
The first author: Conceptualization, Methodology, Software,
relationship between oil and natural gas prices. Resour. Pol. 60, 118–124. https://
Writing-Original Draft. doi.org/10.1016/j.resourpol.2018.11.020.
The corresponding author: Conceptualization, Methodology, Soft­ Torrence, C., Compo, G.P., 1998. A practical guide to wavelet analysis. Bull. Am.
ware, Writing-Original Draft. Meteorol. Soc. 79, 61–78. https://doi.org/10.1175/1520-04771998079<0061:
APGTWA>2.0.CO;2.
The second author: Formal analysis, Supervision, Writing-Review & Torrence, C., Webster, P.J., 1999. Interdecadal changes in the enso–monsoon system.
Editing. J. Clim. 12 (8), 2679–2690. https://doi.org/10.1175/1520-0442(1999)012<2679:
ICITEM>2.0.CO;2.
Wang, Zhonghua, Hu, Biao, 2018. An empirical analysis of the factors affecting the
Declaration of competing interest carbon price in China. J. Ind. Technol. Econ. 37 (02), 128–136.
Wen, Xin, Chen, Jiajing, 2022. Technological innovation and the energy transition: A
The authors declare that they have no known competing financial literature review. Pioneer. Sci. Technol. Monothly 12, 142–148.
Wu, Q., Wang, M., Tian, L., 2020. The market-linkage of the volatility spillover between
interests or personal relationships that could have appeared to influence traditional energy price and carbon price on the realization of carbon value of
the work reported in this paper. emission reduction behavior. J. Clean. Prod. 245, 118682 https://doi.org/10.1016/
j.jclepro.2019.118682.
Zhang, Y.-J., Wei, Y.-M., 2010. An overview of current research on EU ETS: evidence
Acknowledgement from its operating mechanism and economic effect. Appl. Energy 87, 1804–1814.
https://doi.org/10.1016/j.apenergy.2009.12.019.
We would like to thank three anonymous referees for many helpful Zou, X., 2018. An analysis of the effect of carbon emission, GDP and international crude
oil prices based on synthesis integration model. Int. J. Energy Sect. Manag. 12,
comments. However, any remaining errors are solely ours.
641–655. https://doi.org/10.1108/IJESM-10-2017-0013.