Renewable and Sustainable Energy Reviews 79 (2017) 1–8

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

Research on the efficiency of carbon trading market in China MARK

⁎
Xin-gang Zhao, Lei Wu , Ang Li
School of Economics and Management, North China Electric Power Univ., Beijing 102206, China

A R T I C L E I N F O A BS T RAC T

Keywords: In 2014, China proposed medium and long-term low carbon development goals in China-U.S. Joint Statement
Market efficiency on Climate Change that the emission of carbon dioxide would reach its peak and the proportion of non-fossil
Economic performance energy accounted for the primary energy consumption would increase to 20% in 2030. In order to achieve these
Efficient markets hypothesis goals, the unified carbon emission trading system should be put into effect by 2017, the implementation of the
Carbon trading market
unified carbon emission trading system depends on the effectiveness of the current carbon trading market in
Carbon trading system
China. On the basis of the effective market theory and fair game model, the unit root test and the run test are
developed to analyze the carbon emission market of four representative cities in China. The results show that (1)
the carbon trading market in China has only achieved weak efficiency, while the semi strong efficiency and the
strong efficiency have not been reached; (2) with the expansion of the market scale, the increase of trading
volume, the carbon trading market would converge from the state of inefficiency to weak form efficiency
gradually, and the carbon trading market in China shows signs of restoring market efficiency.

1. Introduction would reach its peak and the proportion of non-fossil energy accounted
for the primary energy consumption would increase to 20% in 2030. To
To cope with climate change, the international community for- make this possible, the unified carbon emission trading system should
mulated and adopted “the United Nations framework convention on be put into effect by 2017 [5].
climate change” and “the Kyoto protocol” under the framework. In this Therefore, it is critical to complete the construction of current
context, as a major carbon emitter, China is facing an unprecedented carbon market and to accumulate experience in order to guarantee the
pressure to reduce its carbon dioxide emission [1]. In order to ease the completion of the national carbon emission market, and fulfilling the
pressure of China, emissions trading scheme(ETS)is thought to be an goal of China's carbon reduction and energy structure optimization,
effective measures. It is also considered an important mechanism to thus, the study of the efficiency of China's carbon trading market has
address climate change, and an effective measure to reduce global become an urgent priority. However, Carbon trading system in China is
greenhouse gas emissions [2,3]. still in the experimental phase, and there are lots of problem existed in
In June 2012, the national development and reform commission the market, such as low market participation, the defective of mechan-
issued “interim measures for voluntary greenhouse gas emissions isms in pricing, supervision, rewards and punishment. Liu Liwei et al.
trading management”, to pursue voluntary emissions trading program (2014) considered that the prominent problem of China's carbon
in China, which is regarded as, lays a foundation for the further trading market is market segmentation, low participation and the
implementation of the mandatory carbon trading mechanism. defective of pricing mechanism, which seriously affect the expansion of
Subsequently, the carbon emission rights exchanges in Beijing, market trading volume, price formation and the development of a
Shanghai, Tianjin, Guangdong and other provinces were set up in unified market [6]. Zhou Wenbo et al. (2011) analyzed China's carbon
succession, and the compulsory carbon emission rights trading pilot trading market from the perspective of macroeconomic and micro-
work were started up [4]. In June 2013, the carbon emission exchange economic respectively, they believe that the main problems are the
in Shenzhen took the lead, and then Shanghai and Beijing also started unreasonable carbon emission trading price, the defective of govern-
at the end of November in 2013, which means China's carbon trading ment supervision and the initial allocation of emission allowance, etc
market has achieved a key step from theory to practice and come to the [7]. Huang Ping et al. (2014) compared the development status of
substantial operation phase. In the year of 2014, China proposed domestic and international carbon trading market and point out that
medium and long-term low carbon development goals in China-U.S. the main problems of China's emerging carbon trading markets are the
Joint Statement on Climate Change that the emission of carbon dioxide lack of incentive mechanism, market fragmentation, the defective of

⁎
Corresponding author.
E-mail address: [email protected] (L. Wu).

http://dx.doi.org/10.1016/j.rser.2017.05.034
Received 20 July 2016; Received in revised form 16 February 2017; Accepted 10 May 2017
1364-0321/ © 2017 Elsevier Ltd. All rights reserved.
X.-g. Zhao et al. Renewable and Sustainable Energy Reviews 79 (2017) 1–8

fairness and a series of systemic problems [8]. In addition to that, due Tang et al. (2013) discussed the relationship between the EUA futures
to the immaturity of the market, China's carbon trading market reveal a market and spot market based on the unit root test and the cointegra-
series of systemic problems, such as inaccuracy of quota allocation, tion test. They conclude that the EUA futures market is efficient within
imperfect trading mechanism, lagging legislation, etc [9,10]. China 1 month. Furthermore, the impact of the price would continue for 3
should build a mature carbon trading market as soon as possible. months, examined by a vector error correction model (VECM) [30].
To solve these problems, scholars in China studied on the efficiency However, some scholars believe that the European carbon trading
of China's carbon trading market, which mainly divided into two market is inefficient. For example, Daskalakis and Markellos (2008)
stages: CERs OTC and the pilot carbon trade. In the research on early used the empirically test to examine the weak form efficiency in the
CERs etc., Wang Can et al. (2005) used the global carbon emissions European carbon market and have found that the European carbon
trading partial equilibrium model to analyze the market potential of market is inefficient [31].
China's CDM [11]. Hong Juan et al. (2010) described the long-run Carbon finance is a whole new area for Chinese enterprises and
equilibrium relationship of CERs spot and CERs futures and the even management authorities, there are many deficiencies in carbon
correlation between CERs and EUAs through the ADF test, cointegra- emissions measurement, carbon asset pricing, carbon trading system
tion analysis, granger causality test and impulse response function construction and market supervision. The study of the efficiency of
[12]. In the research on carbon trading market, Wang Qian (2014) used China's carbon trading market is conducive for China to speed up the
a series of variance ratio tests and the ADF tests to investigate whether familiarity of the rules of carbon trading market, in-depth under-
the China's carbon trading market is efficient based on the data from standing of the operation and effect, is also conducive to the establish-
the pilot trading market and the results shows that part of the pilot ment of a unified national carbon market. The establishment of the
trading market achieved the weak efficiency [13]. Tang Ling et al. unified national carbon market can help reduce carbon emissions,
(2015) developed a multi-agent model to investigate the impact of develop low-carbon economy, help to reduce people's dependence on
different CET design, with a view of finding the most suitable CET fossil fuels, thus promoting the development and utilization of renew-
design for China [14]. Zhao Xingang et al.(2016)analyzed the market able energy [32]. At the same time, there are intersections between the
efficiency of China's carbon trading market from price, trading volume, carbon emission rights and the renewable energy certificates, and the
market liquidity,and information transparency, they found that the policies of the two are also affected [33]. The opinions of the carbon
market efficiency is not satisfactory [15]. trading market in this paper also have reference for the policy of
The researches on the efficiency of the European carbon trading renewable energy. In short, this study is conducive to the promotion of
market mainly focus on the volatility and the predictability of the carbon renewable energy development and utilization, and its policy recom-
price. Chesney and Taschini (2008) point out that discounting the price mendations has a guiding significance for the development of renew-
of carbon options is the martingale process about the correlation filtering able energy policy.
[16]. Richard Starkey (2012) made a critical survey on personal carbon This paper investigates the efficiency of present China's carbon
trading to discuss the efficiency and effectiveness of carbon market [17]. trading market and to provide suggestions to the unified national
Daskalakis et al. (2009) point out that the EU ETS spot prices show a carbon trading market. The carbon price of four typical cities during
jump and non-stationary behavior [18]. Paolella and Taschini (2008) 2013–2015 are selected as sample data in this paper and the empirical
believe that carbon prices cannot be predicted by model [19]. Feng at al analysis is employed based on unit root test and runs test. This paper is
(2011) used the random walk model to investigate the volatility of carbon organized as follows:Section 2 introduces the related policy and system
price, they found that the carbon prices cannot fully reflect all historical framework of China's carbon trading market, Section 3 analyzes the
information and show a short-term memory [20]. While Another part of efficiency of China's carbon trading market based on three hypotheses
the scholars mainly discussed liquidity, volume volatility, and price of EMH, the empirical analysis is presented in Section 4, followed by
clustering of European carbon trading market [21–24]. Takashi conclusions and policy implications in Section 5.
Kanamura (2016) examines the role of carbon swap trading and energy
prices in volatilities and price correlations between the EU and Kyoto 2. The situation and the framework of the carbon trading
Protocol emissions trading schemes. Yves Rannou, Pascal Barneto(2016) market in China
analysis the relationship between European carbon market efficiency and
their volatility/liquidity. Anca Claudia Balietti (2016) believe that the 2.1. The development course
trading activity-volatility link is relevant for evaluating the efficiency of
the EU ETS. China's carbon trading development process can be divided into
Though the literature on the price dynamics of CO2 allowances as two phases. In the first phase, China was given priority to the
part of the EU ETS is steadily increasing, the issue of carbon trading in development of CDM and gradually became the largest country in
relation to the EMH has not been addressed so far. Some scholars the world in CDM project. In the second phase, China started to
believe that the European carbon market is efficient. Seifert et al. develop carbon trading system. In 2008, China's national development
(2008) presented a stochastic equilibrium model to test the efficiency of and reform commission (NDRC) put forward to the work of building
European carbon market by using empirical test and summarized the carbon trading pilot for the first time [34]. In 2009, a target to control
martingale characteristic of European carbon market, thus the conclu- greenhouse gas emission was announced officially by Chinese govern-
sion that the European carbon market is informational efficient is ment that, by year 2020, the carbon dioxide emission per unit GDP will
drawn [25]. Amelie Charles et al. (2011) confirmed that the EU ETS is drop by 40–45% compared with 2005 [35]. In October 2010, China's
efficient from the perspective of the execution cost [26]. Daskalakis G State Council proposed to establish and to improve the system of
(2013) also confirmed the weak form efficiency of the EU ETS is based emissions trading [36]. In November 2011, the national development
on the simple technical analysis rules and naive forecasts [27]. While and reform commission approved the plan of carrying out the pilot
Alberto Montagnoli et al. (2010) used the variance ratio tests to verify work of carbon trading in Beijing, Tianjin, Shanghai, Chongqing,
the efficiency of the EU ETS, and they conclude that the EU ETS is Hubei, Guangdong, Shenzhen and other municipalities. On June 18,
periodic efficient. The first and the third phases are not of weak 2013, the first carbon trading pilot of China was launched in Shenzhen
efficiency while the second phase achieved the weak efficiency [28]. [37,38]. In August 2013, the state council confirmed the work of
Feng Xiaoying (2014) also studied the efficiency of the EU ETS. The building carbon trading pilot as a way to push the development of
results came out that the market efficiency of EU ETS has different marketing mechanism [39]. By the end of the July 2014, all these seven
phase features and the trading market efficiency is related to not only major carbon trading pilot have started operating and the unified
the market mechanism, but also the stability of policies [29]. Bao-jun carbon emission trading system will be put into effect in 2017 [5].

2
X.-g. Zhao et al. Renewable and Sustainable Energy Reviews 79 (2017) 1–8

2.2. Carbon trading trial system isms. Fig. 2 displays the framework of carbon trading market and its
main participation subjects are consisted of government, emission
Building the unified carbon emission trading system is a long-term reduction enterprise or individual, financial institutions, intermediaries
process, which can be divided into four stages. The first stage is and carbon investors.
preparation, which means to complete the construction of carbon Government plays the key role in goal setting, legislative support,
emission quota cap-and-trade system at the local level by the end of encouraging and guiding the transaction, standardizing and super-
2015. The second stage is the improvement for operation during the vising the market and participators behavior. Reduction enterprise or
period from 2016 to 2020. The target of this stage is to accomplish the individuals are divided into the demander and supplier. The entities
construction of carbon emission quota trading system at the national with insufficient allowances act as main demand-side to buy allowan-
level, which requires to convert from the intensity target to the total ces. The other covered entities act as main supply-side to sell
amount control target. The third stage is the expanding stage, which allowances in the market. Suppliers can also participate in the carbon
covers the period from 2020 to 2030. The main objectives are to build trading market in the form of carbon financial derivatives and carbon
the carbon trading system under the absolute amount at the national sinks. Financial institutions can participate in the carbon market
level and the institutional, regulatory and market operation will be directly by providing financing service; they can also supply other
mature. The fourth stage is the deepening stage with the target of financial supports and innovative researches to join the market
building the trading system under the aim of absolute quantified indirectly. The duties and responsibilities of intermediary institutions
emissions at national level. Besides, this system has the capacity of in the carbon market are providing consulting services, providing the
connecting with other countries and dominating the international carbon consulting certification for carbon suppliers, providing consulting for
market. Till now, China's carbon trading system is gradually established carbon sequestration investors, and offering verification testing for
and improved, which realizes the development from the local trial corporations and individuals whose carbon emissions is excessive.
market to national unified market and then to the global market.
At present, China's carbon trading market is just in the experi-
3. The analysis of China's carbon trading market efficiency
mental phase. Since 2011, some provinces and cities attach great
importance to the work of carbon trading pilot and carry out the basic
According to the modern efficient market theory, the effectiveness
works in succession, for example, formulating local laws and regula-
of the carbon market refers to all the information associated with the
tions, determining the target and scope of total amount control,
carbon emissions that can be reflected in terms of price. Fama (1970)
establishing the system of greenhouse gas measurement, reporting
divided the market efficiency into three levels according to the
and verification (MRV), etc [15]. It can be formed in Fig. 1 as a
information reflected by prices.
complete basic framework of carbon trading pilot system.
1) The weak form efficiency.
In general, carbon trading pilot system is mainly divided into six
In the case of weak form efficiency, the market price has fully
modules, local government has define the total target and scope, quota
reflects the price information of all past history. This implies that
allocation and management, MRV system, implementation system,
investors cannot earn abnormal profit by exploiting past information.
trading system and capacity building respectively by the rules and
2) The semi strong form efficiency.
regulations. It can be concluded that the trial carbon trading system is
In the case of the semi strong form efficiency, market prices can
to prepare for the construction of unified carbon emission trading system.
fully reflect the public information, which implies that investors cannot
earn abnormal profit by exploiting public information.
2.3. Carbon trading basic framework 3) The strong form efficiency.
In this scenario, market prices can fully reflect all information about
China's carbon trading market is a free market whose target is to the company's operation, such as history information, public informa-
achieve greenhouse gas (GHG) emissions reduced by market mechan- tion and internal undisclosed information.

Fig. 1. The basic framework of the Pilot carbon trading system.

Source: authors.

3
X.-g. Zhao et al. Renewable and Sustainable Energy Reviews 79 (2017) 1–8

Fig. 2. The basic organizational framework of carbon trade [15].

Source: authors.

The efficient market theory is based on three hypotheses, which can there are individual investors and investment institutions exist in the
be summarized as rational participants, no opportunity cost and no market. They have exact required limits of carbon emissions. Besides,
information cost. These hypotheses will be used in this section to profit is the only purpose for them to participate in the trade, which
analyze the efficiency of the China's carbon trading market, and make a means, their investment behaviors are irrational,so that the irrational
preliminary judgment. speculation will result in the market stability reduction as well as the
carbon price fluctuations.
3.1. The analysis of rational participants
3.2. The analysis of information cost
China's carbon trading market is currently in the experimental
phase, as well as, seven experimental units are independent individuals China's carbon trading market is just in the experimental phase that
and each experimental unit has its own unique situation. Each there are no unified regulations on the disclosure of market informa-
experimental unit has difference in the sense of leadership and division tion. Therefore, all experimental units disclosure the information
of responsibilities, which result in great difference in the release time selectively based on their own circumstances, making it a huge gap
and content of the same category of laws between different pilot on the transparency of information in every trading market. For
projects. For example, in the handling of excess emissions, Shenzhen instance, Guangdong and Hubei announced the information about
made provision the first place that the reduction enterprise will be fined quota volume, proportion of the quota, free quotas and paid quotas,
three times according to the market average price of illegal carbon while other pilots are reticent about this. The low transparency may
emissions; in November 2013, enterprises in Shanghai had fined cause the masses of mistrust of the market, which makes it possible
30000–150000 RMB, if they excess the emission restrictions; Beijing that some enterprises want to enter the market. However, they might
made it clearly that the units of over-quota emissions would be fined 3– be evicted because of the misunderstanding on the market. Besides,
5 times according to the market average price on December 27, 2013. some investors cannot make decisions because of the information
These differences will make it hard for participants to form reasonable asymmetry, which is also the consequence of low transparency of the
expectations, which lead to irrational investment behavior. trading market.
All the carbon trading experimental units are in the presentation Information asymmetry is also embodied between emission reduc-
stage, whether trading system or trading pattern is mature and most tion enterprises and policy makers. In the transfer process of carbon
enterprises and investors’ understanding of carbon emissions trade are trading policy, due to the non-uniformity of interests at all levels,
still in the training phase, which result in the irrational investment. At emission reduction enterprises usually take their own loss and profit as
the same time, in addition to the enterprises with high reductions, a counter measure parameter in the implementation process of carbon

4
X.-g. Zhao et al. Renewable and Sustainable Energy Reviews 79 (2017) 1–8

trading policy. The flexible interpretation of policy leads to information of each market are different. Meanwhile, the market mechanism is not
asymmetry. For example, in the stage of allocation of quotas, decision yet stable. On the one hand, the maturity of carbon trading market
makers determine the distribution of weight according to the expected cannot support market participants to obtain all the public informa-
emission reduction ratio of the enterprises or input and output. tion, such as social actual emissions reductions, quotas redundancy,
However, the governance cost of the enterprises and productivity corporation emissions information,and etc.; on the other hand, the lack
belong to private information. Owing to the problem of “adverse of carbon accounting systems and the information disclosure mechan-
selection”, the behavior of concealed report and false report of ism have hindered the market participants to obtain all public
enterprises will emerge. For enterprises, it will result in low productiv- information. Therefore, the current China's carbon trading market
ity, high emission production, high governance cost and low emission has not reached the semi strong form efficiency. However, market
reduction capacity while getting more quotas. Therefore, the allocation conditions will be gradually satisfied with the development of market,
of resources and the price in market will be distorted and the emission rational market participants will increase, the problem of information
effect will be reduced. The information asymmetry phenomenon will be asymmetry will be eased, transaction costs will be reduced sharply, and
more common in the auction system. Enterprises cannot determine the the weak efficient market is a progressive realization process.
existing quota price based on the total amount of emission in the future Therefore, it will be analyzed that the weak form efficiency of China's
market. Thus, it is difficult to make an accurate judgment on the initial carbon trading market in the next section.
price of the quota. At the same time, the process of carbon trade often
involves multiple government departments, such as fiscal, financial, 4. Empirical analysis
environmental protection, and etc. and each department cannot get rid
of the concept of individual interests. Besides, they cannot achieve 4.1. Methodology
efficient communications in terms of functions, resources, information,
interests, and etc. There are some problems of "selective filter" in the This paper, having built the efficient market model of China's
process of information transfer between the various departments, carbon trading market based on the "fair game" model, will test the
which will also result in information costs. weak form efficiency of carbon trading market and investigate whether
the returns in the CO2 market would subordinate a martingale
3.3. The analysis of transaction cost difference sequence. In the empirical test, the scholars generally test
whether the error term of first-order regression model of time series is
In carbon trading market, the transaction cost of enterprises mainly independent identically distributed or irrelevant by random walk
consists of two parts, which are presented in Fig. 3. The first part is the model. However, according to the requirement of the martingale
entrance part, including the application of procedures, the registration assumption, the random walk process and the martingale process are
of service system and other processes happening at the time of distinguished. The requirement of random walk process is strict,
enterprises entering the market, is the entrance cost; the second part however, the random walk process is the sufficient but not the
is the running cos, consisting of containing predicting risk, hunting for necessary conditions for the martingale process, which means if the
transaction objects, executing transactions and other occurring beha- random walk process can be satisfied, the martingale process can also
viors at the time of completion of the quote transaction, is the running be met. However, if the martingale process can be satisfied, the random
cost. walk hypothesis may not be met. Therefore, it is obviously wrong to
In China's carbon trading market, some have entrance costs while judge whether the market achieves weak form efficient on the basis of
most have operating costs. The presence of these additional costs leads such a criterion that “if the random walk is rejected, the martingale
to the decline of the market liquidity and enthusiasm. The existence of hypothesis ought to be rejected”. It also means that the simply using of
transaction costs make price cannot reflect the information reasonably, random walk process is incorrect for identifying the weak form
which will affect the efficiency of the market. In conclusion, there are efficiency of the market. Due to the unit root test's particularity, it
irrational participants, information asymmetry and transaction cost may be more reasonable to test martingale assumption compared with
existing in China's carbon trading market. The presence of these factors the random walk test. Therefore, the model (see Dai Xiaofeng et al.,
hinders the transmission of information and makes price-related 2005) [40] could be estimated as followed:
information cannot fully obtain by the participants, namely, historical yt = dyt−1 + _t (1)
information, public information and internal undisclosed information.
Therefore, China's carbon trading market has not reached the strong in which, d = 1, {_t} is a stable process, and E (_t ) = 0 ,
form efficiency. cov(t , t − s ) = _s < ∞, (s = 0, 1, 2, …, n ), so the random process
China's carbon trading market was established in 2013, and is in {yt , t = 1, 2, ⋯} is a unit root process.
the experimental phase. The enterprises category and system settings The stochastic disturbance in unit root process only needs to obey
the general stable process, defined by：
∞
_t = xt + x _ t −1 + x _ t −2 + ⋯ + = h(L )xt = ∑ hjxt −j
j =1 (2)

where {xt } is independent identically distributed, and E (xt ) = 0 ,

D(xt ) = e2 , L is the lag factor. h(L ) is the lag of infinite order polynomial,
∞
whose related coefficients meet the condition that ∑ j =0 j hj < ∞,
therefore, _t is the white noise. Under the condition of the existence
of time series variance, if E (ut ) = E[E (ut It−1)] = 0 , E (ut , ut − j ) =
E[E (ut ut −1 It −1)] = 0 , E (ut 2 ) = e2 , the martingale process is a white
noise. Of course, the time series, which meet the white noise, must
meet the martingale process under linear condition. In conclusion, the
unit root test can meet the requirements of martingale assumption and
it could be used to verify the weak form efficiency of market is
Fig. 3. Transaction cost structure [29]. reasonable.
Source: authors. In this paper, the empirical test is divided into two processes:

5
X.-g. Zhao et al. Renewable and Sustainable Energy Reviews 79 (2017) 1–8

(1) Using ADF test to investigate the weak form efficiency of China's sample data for Beijing shows a less irregular CO2 price pattern. The
carbon trading markets. The model can be estimated as followed: CO2 prices of Beijing are relatively stable, varying around 50 RMB and
p the price only fluctuates once in June 2014, which rises to 80 RMB, and
Δxt = (p − 1)xt −1 + ∑ θiΔxt −i + xt then, drop back to 50 RMB soon and subsequently remain stable. The
i =1 (3) CO2 prices of shanghai fluctuated between10 to50, On the whole, the
carbon trading prices reveal a downward trend. The CO2 prices of
where p is the number of lags, In order to ensure xt is the white
Tianjin were stable, beginning with a price of about 30 RMB in
noise, the size of the best lag item( p ) can be determined by AIC.
December 2013 and then varied between 20 and 50 by the end of
(2) Using run test to test the process that if China's carbon trading
September. In general, the CO2 prices of Shenzhen changed dramati-
market have achieved the weak form efficiency. The biggest
cally, and the prices of the other three cities were relatively stable.
advantage of run test is the ability to overcome the defect that
Table 1 is the descriptive statistics. According to it, the CO2 prices
the serial correlation test is easily affected by extreme value and
of Beijing and Shenzhen are higher, Shenzhen carbon trading market
troubled by whether limited variance exists. To use the random
has the highest average price, which reached of 54.59RMB, and the
walk assumption, when the sample size is larger, the total runs (Q)
maximum price is 122.97 RMB. The average prices of Tianjin and
obey the normal distribution. The test statistics are defined as
Shanghai are relatively lower, which are 25.38 RMB and 32.21 RMB.
follows:
The return distribution of these four carbon market are not subject to
Q − E (Q ) strict normal distribution, the skewness of Beijing, Tianjin, and
Z=
e (4) Shenzhen are positive which implies the returns are flatter to the right
N + 2n1n2 2n1n2(2n1n2 − N ) compared to the normal distribution. The negative skewness implies
where E (Q) = , e= , and N is the total number of
N N 2(N − 1) that the returns of Shanghai are flatter to the left compared to the
days that the carbon price changes, n1 is the number of days that carbon
normal distribution. The kurtosis indicates that the return distributions
price rise, n 2 is the number of days that carbon price fall. If test
of Beijing and Tianjin have sharp peaks and fat tails compared to a
statistics Z is larger than the critical value, then the carbon trading
normal distribution, while the return distributions of the other two
market doesn't achieve the weak form efficiency. This paper, using runs
cities have thin tails.
test, is mainly to test how China's carbon trading market achieved the
weak form efficiency by annual methods. Moreover, this paper is
4.3. Results
aiming to confirm the growth of China's carbon trading market, to
mature the regulation and to improve the laws and regulation. Run test
In this section, we tested for the EMH hypothesis for the carbon
can provide a clear framework for the evaluation of the development of
trading market in China using ADF test and Runs Test. Under the null
China's carbon trading market.
hypothesis the series follows a random walk and have the unit root. In
this paper, the ADF test was conducted by Eviews 6.0 and the Runs
4.2. Data Test was conducted by SPSS 20.0.
The results of ADF test are shown in Table 2, from it, we can see the
The price data in this paper comes from China's carbon emissions test statistics of Beijing, Shanghai, Shenzhen, Tianjin are greater than
trading website. The reason why we choose the carbon price of Beijing, the critical value under any significant level. The test statistic cannot
Tianjin, Shanghai, and Shenzhen as sample data is that the data of reject the null hypothesis for all cities. To test the efficiency of all cities,
these cities are more plentiful and the trading volumes of these four we use three different significant levels: 1%, 5% and 10%. The test
cities are larger. The sample covers the period from 18 June 2013 to 23 statistic for Beijing is −2.135, the test statistic for Tianjin is −2.445, the
September 2015. This yields 6276, 412, 628 and 813 observations test statistic for Shanghai is −1.222, and the test statistic for Shenzhen
respectively. The total number of observations is 2479. is −2.161. Table 2 indicates that the test statistic cannot reject the null
Fig. 4 shows the price for both these carbon trading markets hypothesis. Thus, we can summarizes the results that the data is not
respectively from 2013 to 2015. In which, Shenzhen's carbon trading smooth, the price change of carbon trading market is in line with the
market has the highest carbon price, Beijing ranks the second and the martingale process and the carbon trading market in China has
prices of Tianjin are the lowest. Fig. 4 also shows the price fluctuations achieved weak form efficiency. Market investors cannot predict the
for all these markets. The volatility of Shenzhen market is the biggest, price according to the historical information. They cannot get excess
starting off with a price of about 30 RMB in June 2013. There is a fast profits either.
increase approaching 120 RMB by the mid of October 2013, then a Because Shenzhen's carbon exchange is established earlier, it has
sharp downward appears adjustment to a price close to 80 RMB. Since relatively completed sample data in 2013. The data samples of Beijing,
then, the prices reveal a downward trend. In general, the CO2 prices of Tianjin and Shanghai in 2013 are small which have no reference
Shenzhen are not stable, varying between 30 RMB and 120 RMB. The significance. Table 3 shows that the associated probability parameter of
Shenzhen's carbon exchange is less than the critical value in 2013, so
the test statistic rejects the null hypothesis, which means that Shenzhen
carbon trading market did not reached the weak form efficiency in
2013. Table 3 also shows that the associated probability parameter of
Shenzhen's carbon exchange is larger than the critical value in 2014
and 2015. The test statistic cannot reject the null hypothesis. Thus,
Shenzhen's carbon trading market has reached the weak form efficiency
in 2014 and 2015. We believe that our results are closely linked with
the development of the Shenzhen carbon trading market. Shenzhen
carbon trading market just started in 2013, and it has imperfect market
mechanism. There are less experience and historical information to
refer to for both the policy makers and market participants. Because
carbon trading market has the problems of irrational participants,
information asymmetry and transaction costs, the price cannot cor-
Fig. 4. The graph of price trend. rectly reflect the information, thus, this phase of Shenzhen carbon
Source: http://www.tanpaifang.com/. trading market did not reached the weak form efficiency. While in the

6
X.-g. Zhao et al. Renewable and Sustainable Energy Reviews 79 (2017) 1–8

Table 1
Descriptive statistics.
Source: authors.

Means Median Standard deviation Skewness Kurtosis Max Min

Beijing carbon price 51.93 52 6.49 0.93 7.27 77.0 35.0

Tianjin carbon price 25.28 25.03 6.48 0.65 4.35 50.11 11.2
Shanghai carbon price 32.21 33.03 9.59 −0.41 2.73 48.0 9.50
Shenzhen carbon price 54.59 47.5 18.94 0.47 2.44 122.97 21.47

Table 2
The results of ADF test.
Source: authors.

ADF test statistic (C,T,K) 1% significant level 5% significant level 10% significant level

Beijing carbon price −2.135 (C,0,3) −3.444 −2.866 −2.569

Tianjin carbon price −2.445 (C,0,4) −3.446 −2.868 −2.57
Shanghai carbon price −1.222 (C,0,1) −3.44 −2.866 −2.569
Shenzhen carbon price −2.161 (C,0,6) −3.438 −2.865 −2.569

Note: (C, T, K) is the type of test, in which C indicates the intercept term, T indicates the trend term, K represents the lag order.

the EMH.
Table 3 For the sample periods 18 June 2013 to 23 September 2015, the
The results of run test. EMH is tested through ADF test. The results show that China's carbon
Source: authors.
trading market have achieved the weak form efficiency, while the semi
Beijing Tianjin Shanghai Shenzhen strong form efficiency and the strong form efficiency have not been
reached. The results of Runs tests show that the China's carbon trading
2013 −2.501 market was inefficient when market is immature and in an emerging
(0.012)
state but efficient when market is relatively mature and reached a
2014 1.082 −3.467 −3.009 −0.991 certain scale. This suggests that the weak form efficiency of China's
(0.279) (0.001) (0.003) (0.322) carbon trading market is implemented gradually, and the carbon
trading market in China shows signs of restoring market efficiency.
2015 −0.385 −0.509 −0.624 1.518 Several important policy implications emerge from our research.
(0.700) (0.611) (0.533) (0.129)
First, Paying attention to improve enterprise's recognition of carbon
trading and ability of carbon asset management, strengthening the
year of 2014 and 2015, carbon trading market has reached a certain communication with companies to improve their level of understanding
scale. With the trade volume increasing gradually, the market transac- the carbon trade policy. Second, improving the carbon trading informa-
tions became active and market liquidity was greatly strengthened. tion disclosure system, promulgating carbon accounting standards, the
Correspondingly, with market mechanism becomes mature, so the openness, transparency and standardization of carbon trading market
phase of Shenzhen's carbon trading market has reached the weak form information make the information of each enterprise comparable and
efficiency, as well as, the market prices tend to be more reasonable. credible, which is convenient for investors to make investment decision-
The development of Tianjin's carbon trading market and Shanghai's making and social supervision. Third, constructing the interim carbon
carbon trading market are similar to Shenzhen. The market has just trading system, developing a differentiated market mechanism according
started in the year of 2014, but has not reached the weak form to the main objectives and tasks of the carbon market in each period.-
efficiency. While in the year of 2015, with the development of market, fourth, building a multi-level carbon finance market system, to promote
both Tianjin and Shanghai reached the weak form efficiency. Overall, the stability of carbon finance prices and enhance the enthusiasm of
the efficiency of each trial market was gradually improved with the enterprises. Finally, in order to improve the liquidity of carbon emissions
time. The results of runs test for each trial market reflects the rights,the local government should regard the carbon emission quota as a
development process of carbon trading market. resource for the production, rather than the assessment indicators of
In general, the results of ADF test show that the test statistic cannot carbon emissions management.
reject the null hypothesis for all cities, and the carbon trading market in
China has achieved weak efficiency. The results of run test show that
the carbon trading market gradually altered from inefficiency status to Acknowledgment
weak efficiency with the expansion of the market scale, the increase of
trading volume, the enhance of the market liquidity, and the constant This paper is supported by the Beijing Municipal Social Science
improvement of the mechanism. Foundation (Grant No. 16JDYJB031), the National Natural Science
Foundation of China (Grant No. 71273088) and the Energy Service
Company Energy Design and Research Institute of State Grid
5. Conclusions and policy implications
Corporation (Grant No. SGJNSJ00SCTS1600079).
This paper tests the efficiency of the carbon trading market in
China. The market efficiency is generally important. It is relevant for
carbon risk management and investment strategies. However, in the Appendix A. Supporting information
relatively young China's carbon market, the number of trading volume
is very limited for the period in our sample. It is this infrequent trading Supplementary data associated with this article can be found in the
that has negative implications for the statistical inferences for testing online version at doi:10.1016/j.rser.2017.05.034.

7
X.-g. Zhao et al. Renewable and Sustainable Energy Reviews 79 (2017) 1–8

References correlations and volatilities between carbon markets. Energy Econ

2016;54:204–12.
[22] Rannou Y, Barneto P. Futures trading with information asymmetry and OTC
[1] Olivier JGJ, Peters JAHW. No growth in total global CO2 emissions in 2009. predominance: Another look at the volume/volatility relations in the European
Cement 2010. carbon markets. Energy Econ 2014;53:159–74.
[2] Zhenliang Liao, Xiaolong Zhu, Jiaorong Shi. Case study on initial allocation of [23] Balietti AC. Trader types and volatility of emission allowance prices. Evidence from
Shanghai carbon emission trading based on Shapley value. J Clean Prod 2014:1–7. EU ETS phase I. Energy Policy 2015;98:607–20.
[3] Yue-Jun Zhang, Yi-Ming Wei. An overview of current research on EUETS: evidence [24] Rannou Y. Liquidity, information, strategic trading in an electronic order book: new
from its operating mechanism and economic effect. Appl Energy 2010;87:1804–14. insights from the European carbon markets. Res Int Bus Financ, 39 2014; 2014. p.
[4] Decisions on Speeding up the Cultivating and Developing Strategic Emerging 779–808.
Industry.State Council of the People's Republic of China; 2010. [25] Seifert J, Uhrig－Homburg M, Wagner M. Dynamic behavior of spot prices. J
[5] China’s carbon emissions trading network.〈http://www.tanpaifang.com/〉. Environ Econ Manag 2008;56(2):180–94.
[6] Liwei Liu, Zewen Zhang. The problems and countermeasures of China's carbon [26] Charles A, Darné O, Fouilloux J. Market efficiency in the european carbon markets.
emissions trading. Contemp Econ Manag 2012;34(7):84–8. Energy Policy 2012;60(5):785–92.
[7] Wenbo Zhou, Chen Yan. The status quo, problems and countermeasures of china's [27] Daskalakis G. On the efficiency of the European carbon market: new evidence from
carbon trading market [J]. J JiangXi Univ Financ Econ 2011;3:12–7. Phase II. Energy Policy 2013;54(2):369–75.
[8] Ping Huang, Jin Zhou. The problems and countermeasures of china's carbon [28] Montagnoli A, De Vries FP. Carbon trading thickness and market efficiency. Energy
trading market from the practice of international market. Lanzhou Acad J Econ 2010;32(6):1331–6.
2014;10:140–7. [29] Xiaoying Feng. Study on market efficiency of international carbon trading market.
[9] Liu L, Chen C, Zhao Y, et al. China's carbon-emissions trading: overview, challenges Jinan University; 2014.
and future. Renew Sustain Energy Rev 2015;49:254–66. [30] Daskalakis G, Markellos , Raphael N. Are the European carbon markets efficient?.
[10] Huang Ping, Zhou Jin. The existing problems and solutions of China's carbon Social Science Electronic Publishing; 2007. p. 17.
markets from international market practice. LanZhouXueKan 2014;10:140–7. [31] Tang BJ, Cheng S, Chao G. The efficiency analysis of the European CO 2 futures
[11] Can Wang, Ji-ning Chen, Ji Zou. Potential analysis of clean development market. Appl Energy 2012;112(4):1544–7.
mechanism put into effect in China. China Environ Sci 2005;25(3):310–4. [32] Wang Liang. The dynamic relationship of consumption of renewable energy and
[12] Juan Hong, Jing Chen. Empirical research of prices relationship in international carbon emission in China. Tian Jin University; 2014.
carbon trading market. China Price 2010;1:7–11. [33] Zeng M, Zhu XL, Xue S, et al. Interactive influence between carbon emission
[13] Wang Qian, Wang Shuo. Study on market efficiency of china's carbon Trading reduction policies and renewable energy promotion Polices. East China Electr
market. Social Sci J 2014;4:30–6. Power 2012.
[14] Ling Tang, Jiaqian Wu, Lean Yu, Qin Bao. Carbon emissions trading scheme [34] Notice of Doing the Pilot Work in Low Carbon Provinces and Low Carbon Cities.
Exploration in China:a multi-agent-based model. EnergyPolicy 2015;81:152–69. National Development and Reform Commission; 2010.
[15] Zhao X, Jiang G, Nie D, et al. How to improve the market efficiency of carbon [35] The Twelfth Five-Year Plan for National Economic and Social Development of the
trading: a perspective of China. Renew Sustain Energy Rev 2016;59:1229–45. People’s Republic of China. 2011. (http://www.china.com.cn/).
[16] Chesney M, Taschini L. The endogenous price dynamics of the emission allowan- [36] Decisions on Speeding up the Cultivating and Developing Strategic Emerging
ces:an application to CO2 option pricing. Appl Math Financ 2008;19(5):447–75. Industry. State Council of the People’s Republic of China; 2010.
[17] Starkey R. Personal carbon trading: a critical survey Part 2: efficiency and [37] The 12th Five-Year Plan for National Economic and Social Development. State
effectiveness. Ecol Econ 2012;73(15):19–28. Council of the People’s Republic of China; 2011.
[18] Daskalakis G, Psychoyios D, Markellos R. Modelling CO2 emission allowance prices [38] Notice on Carrying out Pilot Emissions Trading. National Development and Reform
and derivatives: evidence from the European trading scheme. J Bank Financ Commission; 2011.
2009;33:1230–41. [39] Opinions on Speeding up the Development of Energy Conservation and
[19] Taschini L, Paolella MS. An econometric analysis of emission trading allowances. J Environmental Protection Industries. State Council of the People’s Republic of
Bank Financ 2008;32(10):2022–32. China; 2013.
[20] Zhen-Hua Feng, Le-Le Zou, Yi-Ming Wei. Carbon price volatility: evidence from [40] Xiao-feng Dau, Jun Yang, Qing-hai Zhang. Testing Chinese stock market's weak-
EUETS. Appl Energy 2011;88:590–8. form Efficiency: a root unit. Syst Eng 2005;23(11):23–8.
[21] Kanamura Takashi. Role of carbon swap trading and energy prices in price