- Email: [email protected]
Mitigation potential of carbon dioxide emissions in the Chinese textile industry
Applied Energy 113 (2014) 781–787
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Mitigation potential of carbon dioxide emissions in the Chinese textile industry Boqiang Lin a,b,⇑, Mohamed Moubarak c a
New Huadu Business School, Minjiang University, Fuzhou 350108, China Collaborative Innovation Center for Energy Economics and Energy Policy, China Institute for Studies in Energy Policy, Xiamen University, Xiamen, Fujian, 361005, PR China c School of Energy Research, Xiamen University, Xiamen 361005, China b
h i g h l i g h t s We employed Johansen cointegration technique and scenarios analysis. The carbon intensity in the Chinese textile industry can be reduced by 60%. The CO2 emissions reduction potential is estimated to be 44.8 million tons by 2025. Improving the technology and labor productivity will drive down the CO2 emissions. Energy substitution is a promising method to cut down CO2 emissions.
a r t i c l e
i n f o
Article history: Received 31 March 2013 Received in revised form 6 June 2013 Accepted 4 August 2013 Available online 4 September 2013 Keywords: Carbon intensity CO2 emissions Textile industry China
a b s t r a c t We estimated the reduction potential of carbon dioxide emissions in the Chinese textile industry by forecasting the carbon intensity (CO2 emissions/industrial value added) in different scenarios. The Johansen co-integration technique was employed in order to establish the long term equilibrium equation. Three scenarios (Business As Usual (BAU), medium and optimum) were designed to estimate the future trend of carbon intensity in the Chinese textile industry. The results showed that energy price, energy substitution, labor productivity and technology have significant impact on the carbon intensity. Estimated to 1.49 t CO2/10,000 yuan in 2010, we found that for the BAU scenario, the carbon intensity will decrease to 0.5 and 0.29 t CO2/10,000 yuan by 2020 and 2025 respectively. For the medium scenario, carbon intensity will decline to 0.12 t CO2/10,000 yuan. Yet by the optimum scenario, the intensity is expected to considerably decrease to 0.05 t CO2/10,000 yuan by 2025. Using the BAU forecast as baseline, the quantity of reduction potential in carbon dioxide emissions is estimated to be 44.8 million tons CO2 by 2025. Considering this huge potential, we provided policy suggestions to reduce the level of CO2 emissions in the Chinese textile industry. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction The carbon dioxide emissions in China has been increasing geometrically over the past thirty years. According to the 2010 estimates, the quantity of CO2 emissions in China was equivalent to 25.1% of the global emissions [1]. In order to mitigate it, the Chinese government made a commitment to reduce considerably, the levels of emissions for the coming years and decided to give much priority to economic growth balanced with less pollution. According to the 12th Five-year plan presented in 2011, China is expecting by 2015 to reduce the energy consumption per GDP by 16%, to increase the share of non-fossil fuel in the total energy ⇑ Corresponding author at: New Huadu Business School, Minjiang University, Fuzhou 350108, China. Tel.: +86 0592 2186076; fax: +86 0592 2186075. E-mail address: [email protected] (B. Lin). 0306-2619/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apenergy.2013.08.010
consumption by 11.4% and to cut down the carbon dioxide emissions per unit of GDP by 17%, compared to 2010. The realization of this plan will substantially boost China’s contribution to the mitigation of carbon dioxide emissions at global level. In this paper, we analyzed the potential to reduce emissions in the Chinese textile industry. China became the world’s largest producer of textile materials. Consequently due to this industrial expansion, the total energy consumption and CO2 emissions grew annually by 4% and 2% respectively from 1985 to 2010. Energy consumption in the textile industry was estimated to be 73.45 mtce (million tons coal equivalent) in 2010; which was equivalent to 2.41% of the total energy consumed in China. In the meantime, carbon dioxide emissions represented 3.7% of the total emissions from the manufacturing sector [2]. The energy consumption was mainly driven by coal consumption due to its predominance as energy source employed in the production processes. Despite its progressive
782
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
decline in proportion to the total energy consumed, the dependency on coal is not expected to drastically change in the short term (Fig. 1). Coal consumption increased from 18.68 mtce to 29.26 mtce during the study period. Moreover, coal has a high coefficient factor in term of CO2 emissions. Thus, a combination of coal and other fossil fuels, namely coke and oil products led to high level of CO2 emissions in the textile industry. As illustrated in Fig. 1, the CO2 emissions were mainly due to the coal consumption; which represented an important proportion of the total energy consumed. Through a series of questions orienting our research, we estimated the carbon intensity trends and the carbon dioxide reduction potential in several scenarios: (i) what variables have influence on carbon intensity in the Chinese textile industry? (ii) What are their degree of influence? and (iii) What will be the carbon intensity and is the reduction potential of CO2 emissions? Answers to these questions will have significance for this paper; it will provide policy methods to reduce the carbon intensity, contribute to provide industrial strategy for the Chinese textile industry by cutting down CO2 emissions while improving the industrial competitiveness. 2. Literature review There have been several reports on carbon dioxide emissions reduction. These reports have been conducted using different approaches, leading to mixed conclusions. One popular and widely used method is the decomposition analysis. It has some convenience in studying the CO2 emissions change. This method is based on analyzing the contribution of selected factors to the change in emissions. Sun et al. [3] studied the change of CO2 emissions in China’s iron and steel industry and concluded that energy consumption was the most important factor leading to a decrease in CO2 emissions. Later, Xu et al. [4] analyzed the case of China’s cement industry and Wang and Liang [5] focused on 12 key economic sectors. Similarly, Hatzigeorgiou et al. [6] analyzed the energy-related CO2 emissions in Greece from 1990 to 2002. From these studies, we denote that decomposition approach is convenient to study change in CO2 emissions, however it is not appropriate for forecasting the future emissions level and reduction potential. The data envelopment analysis (DEA) model has been also used in several studies. Guo et al. [7] focused on finding out the emissions performance of Chinese provinces and their respective reduction potential. Later, Choi et al. [8] attempted to determine the potential reduction and efficiency of CO2 emissions in China. According to their results, the average marginal abatement cost of CO2 emissions is about $7.2. However, the limitation of this study is due to the fact that there is no specific function form while using the nonparametric method.
Fig. 1. Percentage of growth of coal consumption, total energy consumption and CO2 emissions.
Hasanbeigi et al. [9] applied the bottom-up CO2 abatement cost curves model which is a variant of the energy conservation supply curves methodology to determine the potentials and costs of CO2 mitigation through abatement technologies from 2010 to 2025 in the Thai cement industry. They found that the abatement potential represented 15% of the total cement industry emissions in 2005. Similarly, Worrell et al. [10] analyzed the cost-effective energy efficiency and the carbon dioxide emissions mitigation potential in the US iron and steel industry between 1958 and 1994. They found a total cost-effective reduction potential of 3.8 GJ/t with a payback period of three years or less. Despite the wide use of this method, it has some limitations because it is focused only on the existing technology. Using the scenario analysis method, Ke et al. [11] studied the energy saving and CO2 emissions reduction potential in China’s cement industry. They estimated that the carbon reduction potential represents 3.2–4.4 gigatonnes during 2011–2030 and that the energy efficiency is the key policy measure to drive down carbon emissions intensity in the cement’s industry. Ari and Aydinalp Koksal [12] analyzed the reduction potential of carbon dioxide emission from the Turkish electricity sector. Based on four scenarios, they argued that a significant decrease in the amount of CO2 emissions from electricity generation can be achieved if the share of the fossil-fueled power plants is lowered. Later, Özer et al. [13] used the Long-range Energy Alternatives Planning system (LEAP) model to investigate mitigation potential of emissions in the electricity sector. They found that cumulative CO2 emission reduction between the BAU and Mitigation Scenarios from 2006 to 2030 is 903 million tons. He et al. [14] attempted to forecast the future CO2 emissions from fossil energy combustion in China from 2010 to 2020; similar to Li et al. [15] who focused their research on shanghai city (China). Henriques et al. [16] analyzed the carbon dioxide (CO2) emissions reduction potential from energy consumption in Brazilian industrial sector. Their conclusions showed that by improving energy efficiency, recycling and energy substitution, it will be possible to reduce carbon dioxide emissions by 43% in the industrial sector during 2010–2030 period. However, an important variable such as price was not considered. Far to be exhaustive, several methods and variables have been used to determine the carbon emissions change and reduction potential. However to the best of our knowledge, studies conducted on reduction potential of CO2 emissions in the Chinese industries are limited. Therefore, we extended the research on Chinese textile industry by conducting an in-depth study on mitigation potential of CO2 emissions. 3. Methodology and data sources 3.1. Co-integration approach The Johansen co-integration test has a property to determine the long term relationship among variables. It is employed to determine a long run relationship among carbon intensity, technology, energy price, labor productivity and energy substitution. Prior to the regression analysis, we verified that the variables are stationary in order to avoid spurious results. So, to determine whether these variables have the same integration order, the unit root tests were applied. The most common tests used are Augmented Dickey–Fuller (ADF) tests [17], Phillips–Perron (PP) tests [18] and Kwiatkowski–Phillips–Schmidt–Shin (KPSS) tests [19]. Eq. (1) is the Augmented Dickey–Fuller (ADF) test representation to test the unit root hypothesis.
Dzt ¼ b0 þ a0 t þ a1 zt1 þ
m X bi Dzt1 þ et i¼1
ð1Þ
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
With, zt representing the variable at time t; Dzt1 the zt1 zt2; et as the disturbance with a mean 0 and a variance r2; t the linear time trend and m is the lag order. The null hypothesis H0 is: a1 = 0 in Eq. (1), if a1 is significantly less than zero, the null hypothesis of a unit root is rejected. However, the ADF tests may lack efficiency when applied to a small sample size. The PP test uses the same models as the ADF test, but is remarkably insensitive to the heteroscedasticity and the autocorrelation of the residuals. Both ADF and PP tests require assumptions that tested sequence may contain a constant term and trend variables, the application has a certain degree of inconvenience. To strengthen it, KPSS test is also applied; which is more appropriate for small samples when it chooses a lower lag truncation parameter [20]. Therefore, we used all three test methods and compared their results. We found that the variables are stationary and consequently proceeded to test for cointegration. There are two most popular methods used to establish co-integration relationship: Engle and Granger [21] and Johansen and Juselius method [22,23]. The difference between these approaches is that the Engle–Granger procedure is mostly applied to a single equation co-integration test, while the second approach can detect the existence of co-integration among variables and also determine the number of co-integration vectors. As being the most convenient for this paper, we applied the Johansen–Juselius method trace test and maximum eigenvalue test to determine the number of cointegrating vectors in our model. 3.1.1. Stability Once the co-integration relationship is determined and the parameters are estimated, it is imperative to test for robustness of the model. After the error correction model has been built, the Pesaran and Pesaran [24] test is applied to the cumulative sum of recursive residuals (CUSUM) and the cumulative sum of square (CUSUMSQ) to examine the parameter stability. The model is qualified stable if the plots of CUSUM and CUSUMSQ statistics are confined within the 5% critical bounds of parameter stability. 3.1.2. Scenarios and variables We explored three carbon dioxide reduction scenarios. The feasibility of these scenarios is justified by the fact that for each variable, the future trend is based on the historical data from 1985 to 2010. The scenarios are: 1. The BAU (Business As Usual) scenario, which is based on the historical trend observed for each variable. According to this scenario, the future values of the variables will be similar to their respective average growth rate observed during the period 1985–2010. For The carbon dioxide reduction potential, we considered the BAU scenario as a base line for policy actions. 2. The optimum scenario represents the most committed scenario to reduce the carbon intensity. For this scenario, the future growth rate of each variable is considered to be the highest historical value observed since 2001. There are important growths observed during the period prior to 2001. However, it seems not realistic to expect that it will be possible to reach this anterior performance due to important transformations that occurred in several aspects of the Chinese economy, such as structural change since 2001 and the important transformation that occurred in the Chinese textile industry since the country joined the world trade organization (2001). For instance, exportation of textile and clothing product considerably increased during that period. The promotion of industrial upgrading and setting up structure regulation optimization have become priorities. Furthermore, reduction of carbon dioxide emissions and promotion of energy saving became major concerns.
783
3. The medium scenario is considered to be the average policy commitment. For this scenario, the future values of the variables will be based on the average growth of BAU scenario and the optimum scenario. The growth rate of variables is more important than in BAU scenario; meaning that the policy strength of this scenario is weaker than the optimum scenario but stronger than the BAU scenario. In this paper, four factors are integrated to detect their respective weights on the carbon intensity: Energy Price; it is expected to have an opposite trend to the carbon intensity [25]. This means that a growth of price leads to a reduction in energy consumption and carbon intensity. Because of lack of availability of data, raw material and fuel price index is considered [26]. Energy substitution; it is the proportion of non-fossil fuel energy consumed in the Chinese textile industry. An increase in share of energy consumption from non-fossil fuels is expected to reduce the carbon intensity and the CO2 emissions. This includes energy such as nuclear power, hydropower and renewable energy. This variable has an important impact on the carbon intensity [27,28]. Labor productivity; efficiency in labor productivity contributes to reduce the carbon intensity in textile industry. We assume that labor productivity is negatively correlated to the carbon intensity. It is represented by the industrial value added per capita. Increase in aggregate carbon intensity is driven primarily by an increase in activity [29]. Technology; it is considered as a key element resulting in carbon intensity reduction [30,31]. The Research and Development brings new techniques and methods in the production system which leads to reduced energy consumption and improvements of efficiency. In this paper, the technology is represented by the expenditure for science and technology as percentage of total revenue from sales. Furthermore, there are other variables which may have also impact on the trend of carbon intensity. The CO2 emissions trading scheme in China is an interesting factor which will have impact on the trend of carbon intensity. However, it is still at early development stage in China. It is confronted with the difficulty of the availability and accuracy of CO2 emissions data from participating manufacturers. There is also controversy about defining emission limits under diversified energy structures. Therefore we ignored this variable in our analysis. Adjustment for high added value products in the textile industry is also an important factor to drive down carbon intensity. The optimization of the product structure such as the design, brand value and apparel products may help to reduce the energy consumption, therefore to decrease the carbon emissions level. However, due to the lack of available data and the complexity of the Chinese textile industry, we do not consider this variable as key factor in this study.
3.2. Data sources We used data covering the period from 1985 to 2010 in this study. The data were mainly collected from China Statistical Yearbook [32], China Statistical Yearbook on Science and Technology [33], CEIC China Database [34], China Energy Statistical Yearbook [2] and were calculated based on 2005 constant prices. For stability of the data, we used the natural logarithm of the variables carbon intensity (LnCI), energy price (LnEP), energy substitution (LnES), labor productivity (LnLP) and technology (LnTD). The coefficients of the CO2 emissions of the various energy sources (coal, coke, crude oil, diesel, fuel oil, gasoline, kerosene and natural gas) were chosen based on the estimations made by the Intergovernmental Panel on Climate Change [35]. CO2 emissions were estimated by multiplying
784
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
Table 1 Unit root test. Series
LnCI LnEP LnES LnLP LnTD 4LnCI 4LnEP 4LnES 4LnLP 4LnTD
Table 2 Lag selection criteria.
ADF
PP
KPSS
Lag Log L
No trend
With trend
No trend
With trend
No trend
With trend
0.411 1.995 0.436 1.054 1.142 3.330b 3.693b 5.113a 2.489 4.584a
1.138 2.213 2.328 1.617 1.794 3.262c 3.959b 4.966a 2.446 4.530a
0.449 1.954 0.401 0.349 1.148 3.292b 2.674c 5.120a 2.445 4.592a
1.513 1.335 2.328 1.623 1.866 3.218 2.671 4.694a 2.446 4.537a
0.735a 0.619a 0.733a 0.735a 0.674a 0.121c 0.230c 0.114c 0.144c 0.097c
0.120b 0.165a 0.118c 0.115c 0.127b 0.114c 0.094c 0.144b 0.143b 0.074c
0 1 2
consumption of individual fuels by their CO2 emissions coefficients. Similar to Lin et al. [36], we used the following relation to determine the carbon dioxide reduction potential:
ð2Þ
CRPt is the carbon dioxide reduction potential in period t; CIBAUt is the carbon intensity from the business as usual scenario in period t. CISCENARIOt is the carbon intensity of period t for optimum and medium scenarios, respectively. VAt is the value added in textile industry for period t. 4. Empirical analysis 4.1. Unit root tests The Augmented Dickey–Fuller (ADF), Phillips–Perron (PP) and Kwiatkowski–Phillips–Schmidt–Shin (KPSS) tests were applied in levels and in first differences (Table 1).
ð0:07775Þ
AIC
SC
HQ
43.34087 NA 2.82e08 3.195072 2.949644 3.129960 156.6918 170.0265a 1.89e11a 10.55765a 9.085087a 10.16698a 177.3132 22.33975 3.72e11 10.19276 7.493056 9.476530
Table 3 Johansen co-integration test. Hypothesized No. of CE(s) b
None At most At most At most At most
1b 2b 3 4
Eigen value
Trace Statistic
0.05 Critical value
p-Valuesa
0.973500 0.746096 0.585954 0.429147 0.016030
148.5794 65.07531 33.54692 13.26603 0.371681
69.81889 47.85613 29.79707 15.49471 3.841466
0.0000 0.0006 0.0177 0.1054 0.5421
Max-Eigen
0.05
Statistic
Critical value
p-Valuesa
33.87687 27.58434 21.13162 14.26460 3.841466
0.0000 0.0147 0.0654 0.0814 0.5421
Hypothesized No. of CE(s) b
None At most At most At most At most
1b 2 3 4
Eigen value 0.973500 0.746096 0.585954 0.429147 0.016030
ð0:04059Þ
According to the results, the null hypothesis of a unit root could not be rejected in levels for all the variables tested. Exception is observed with the KPSS test, revealing that the null hypothesis could be rejected for the variables (Table 1). However, despite this inconsistency, it was safer and reasonable to accept the null hypothesis for the level series [37] and then to reject the null hypothesis at the first differences. 4.2. Co-integration analysis An unrestricted vector autoregressive [38] was built to determine the cointegration relation among variables (LnCI, LnES, LnEP, LnTD and LnLP). Using the Akaike information criterion (AIC), Schwarz information criterion (SC), sequential modified LR test statistic (LR), Final prediction error (FPE), and Hannan–Quinn information criterion (HQ), a lag of 1 was selected (Table 2).
83.50406 31.52839 20.28089 12.89435 0.371681
a MacKinnon–Haug–Michelis probability values, MacKinnon–Haug Michelis (1999). b Rejection of the hypothesis at the 0.05 level.
Following this, we proceeded to the determination of the number of co-integration relations and estimation of the co-integration equation. The results of the rank tests for the logarithm of the variables are given in Table 3. The trace statistic and the maximum eigenvalue statistic revealed a unique cointegration equation among the logarithm of the variables. Thus, the co-integration equation can be formulated as:
LnCI ¼ 2:523464 0:224566 LnEP 0:248886 LnES 0:458779 LnLP 0:439590 LnTD ð0:03365Þ
FPE
LR: sequential modified LR test statistic (each test at 5% level). FPE: Final prediction error. AIC: Akaike information criterion. SC: Schwarz information criterion. HQ: Hannan–Quinn information criterion. a Lag order selected by the criterion.
Note: Eviews7 has been used for the tests. D Indicates series in first difference. The critical value for ADF with constant (and trend) at the 5% significance is 2.986 (3.603) and at the 1% significance is 3.724 (4.374). The critical value for KPSS with constant (and trend) at the 5% significance is 0.463 (0.146) and at the 1% significance is 0.739 (0.216). KPSS, Kwiatkowski et al. [19], refers to testing the null hypothesis of stationarity against the alternative of unit root. PP refers to Phillips and Perron [18] unit root test. a Significance at 1%. b Significance at 5%. c Significance at 10%.
CRPt ¼ ðCIBAUt CISCENARIOt Þ VAt
LR
ð0:03315Þ
ð3Þ
The numbers in parentheses are the resulting standard errors. Some comments can be formulated from Eq. (3). Eq. (3) indicates that there exists a long-term equilibrium relationship among variables carbon intensity, energy price, labor productivity, energy substitution and technology. (a) Eq. (3) confirms that as expected the variables energy price, labor productivity, energy substitution and technology are negatively correlated to the carbon intensity. In the long run, these variables contribute to considerably reduce the carbon intensity level. According coefficients values, a 1% increase in energy price, energy substitution, labor productivity and technology will result to a decline of carbon intensity by 0.22%, 0.24%, 0.45%, and 0.43% respectively.
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
785
Fig. 2. (a) Plot of cumulative sum of recursive residuals (CUSUM) and (b) the CUSUM of square (CUSUMSQ).
(a) The labor productivity and technology are the most influencing factors; meaning that these variables had predominant impact on the decrease of carbon intensity during the past 25 years in the Chinese textile industry. This implies that improving the technology and efficiency in labor productivity are the most important aspects driving down the CO2 intensity. 4.3. Test of stability We verified the stability of our model by employing the method proposed by Pesaran and Pesaran [24]. For a certainty in the estimations made, we applied the test of stability. The relevance of this test is to determine whether or not there are over changes in the estimated parameters; because in case of no consistency, the validity of conclusions may be biased. Fig. 2(a and b) illustrate plots of CUSUM and CUSUMSQ statistics of the stability tests. The curves are within the 5% critical bounds indicating that model and the parameters of Eq. (3) satisfied the stability test, and then valid. 4.4. Carbon intensity projections The error correction model applied reflects a short-term relationship among the variables, and the cointegration model indicated the existence of a long-term relationship. Therefore, we conducted analysis in order to predict the reduction potential of carbon dioxide emissions of the Chinese textile industry up to 2025. Similar approach to Lin et al. [39], we used the co-integration Eq. (3) to estimate the amount of carbon dioxide emissions that can be reduced from the textile industry. The first step for the estimation was to determine a measure of the prediction by substituting the historical data of the variables (CI, ES, EP, TD and LP) into the co-integration Eq. (3) in order to obtain fitted trend of carbon intensity for the period 1985–2010. Fig. 3 illustrates the average relative error for the fitted trend compared to the actual trend. Following this, we proceeded to the estimation for future values of carbon intensity up to 2025. For the BAU scenario, the carbon intensity was estimated based on the average annual growth rate of the variables during the study period 1985–2010; which are TD = 7%, ES = 5%, EP = 5% and LP = 13%. Therefore, the forecasts of carbon intensity are 0.86 t CO2/10,000 yuan by 2015, 0.5 t CO2/10,000 yuan by 2020 and 0.29 t CO2/10,000 yuan by 2025. This finding implies that, based on the current trends of CO2 emissions, energy consumption and value added in the BAU case, the carbon intensity in the Chinese textile industry will decline to 0.29 t CO2/10,000 yuan by 2025; equivalent to 1/5 of the level observed in 2010 (1.49 t CO2/
10,000 yuan). It is therefore possible to determine the carbon intensity trend under different scenarios and the related reduction potential of CO2 emissions. 4.5. Future carbon reduction potential from scenarios The reduction potential of CO2 emissions is based on three scenarios; the business as usual, medium and optimum scenarios. The BAU scenario as described above is policy driven. The optimum and the medium scenarios are formulated to speed up the CO2 emissions reduction. The optimum scenario is the most active one in term of carbon dioxide mitigation policy. According to that scenario, the future growth rate of each variable influencing the carbon intensity will be similar to the highest historical value observed since 2001; which were ES = 11% in 2002, EP = 7% in 2010, LP = 15% in 2009 and TD = 29% in 2007. For the medium scenario, the growth rates of variables are the average of values between the BAU and active scenarios (Table 4). Consequently, we determined the future trends of the carbon intensity by substituting the values of the variables into the cointegration Eq. (3), (Table 5). Table 5 shows that the carbon intensity in the China’s textile industry will considerably decline during the period 2010–2025. The decrease rate differs and is more important when evolving from the BAU to the medium and to the optimum scenarios. In the BAU scenario case, the carbon intensity in the Chinese the textile industry will drop from 1.49 t CO2/10,000 yuan (the level in 2010) to 0.29 t CO2/10,000 yuan by 2025. For the other two scenarios, the carbon intensity will decrease to 0.12 t CO2/10,000 yuan with the medium scenario and 0.06 t CO2/10,000 yuan for the optimum scenario by 2025. Accordingly, Eq. (2) is used to determine the reduction potential of carbon dioxide emissions by scenarios. The previsions from the
Fig. 3. Curves for actual, fitted and forecasted trends of carbon intensity.
786
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
Table 4 CO2 emissions reduction scenarios. Variables
BAU (%)
Medium scenario (%)
Optimum scenario (%)
EP ES LP TD
5 5 13 7
6 8 14 18
7 11 15 29
Table 5 Carbon intensity forecasts by scenarios (in t CO2/10,000 yuan). Scenarios
2015
2020
2025
BAU Medium scenario Optimum scenario
0.86 0.65 0.5
0.5 0.28 0.17
0.29 0.12 0.06
Fig. 4. Reduction potential of carbon dioxide emissions in the Chinese textile industry.
‘‘12th five-year plan for textile industry’’, the average annual growth rate of value added is predicted to be 8% from 2011 to 2015. Considering its historical trend, the value added of the Chinese textile industry 1.2 trillion yuan by 2020 and 1.8 trillion yuan by 2025. To estimate the CO2 emissions reduction potential, two cases are built: Case A is the change in carbon intensity from the basis scenario (BAU scenario) to the optimum scenario. Case B is the change in the carbon intensity from the BAU scenario to the medium Scenario. Using the expected future values of carbon intensity from scenarios and the forecasted industrial value added, the potential reduction in carbon dioxide emissions are estimated for the period of 2011–2025, as illustrated in Fig. 4. If the CO2 emissions in the Chinese textile industry change from the BAU scenario to medium scenario (Case B), the quantity of carbon dioxide reduction will be approximately 16.4 million tons CO2 by 2015, 25.8 million tons CO2 by 2020, and will increase to 30.5 million tons CO2 by 2025. If more ambitious measures are implemented, meaning that the CO2 emissions change from the BAU scenario to the optimum scenario (Case A), the reduction potential will be more important. It is estimated to be about 28.1 million tons CO2 by 2015, 39.7 million tons CO2 by 2020 and will be 43 million tons CO2 by 2025.
determined by considering the historical values of the variables, energy price, energy substitution, labor productivity and technology. Moreover, the future values of the carbon intensity have been estimated based on three scenarios. The results showed that the carbon intensity will decrease from 1.4 t CO2/10,000 yuan in 2010 to 0.29 t CO2/10,000 yuan for the BAU scenario, 0.12 t CO2/ 10,000 yuan for the medium scenario and 0.06 t CO2/10,000 yuan for the optimum scenario by 2025. Furthermore, the results showed that the carbon dioxide emissions will decrease by 16.4, 25.8 and 30.5 million tons CO2 by 2015, 2020 and 2025 respectively. If more ambitious measures are implemented, the reduction potential is estimated to be 43 million tons CO2 by 2025; equivalent to a 60% reduction in CO2 compared to the 2010 emissions level. The coefficients of elasticities from the cointegration equation revealed that all the variables statistically have significant impact on the evolution of the carbon intensity trend. The significance of these coefficients provides orientations of policies to be implemented to reach the targets of carbon intensity reduction and carbon dioxide emissions mitigation. We suggest that the Chinese government continues to invest a lot into research and development to come out with environmentally-friendly cutting-edge technologies in the production and distribution of power to sustain the robust economic growth. It is also necessary for the Chinese government to actively promote the technology transfer by facilitating international cooperation among and industries. Wang et al. [40] suggested that the technology upgrading can simultaneously improve economic growth, accelerate the energy efficiency and also reduce the carbon intensity per unit of output. Improving technology will also improve the productivity, which is also a determinant for the carbon intensity. Similar to our findings, the technology and labor productivity are the key elements to focus on in order to mitigate the carbon dioxide emissions in the Chinese textile industry. We also found that energy substitution is an important factor in the efforts to cut down emissions. Coal has been and will continue to be for the coming years the main source of energy in Chinese industries including the textile. Naturally, a progressive substitution of coal by other less- polluting energy sources such as renewable energy, natural gas and nuclear power will considerably reduce the CO2 emissions. The price of fossil fuel has been controlled in the past by the Chinese government. A progressive and complete release of control on energy price will have interesting impact on the long term. Increase in energy cost will lead to more efficiency and energy conservative measures at the firms’ level; which will in return contribute to decrease the emissions. The government should establish a proportional tax policy on fossil fuel consumption, subsidies and tax relieve in order to promote the interest for less polluting alternative energy sources and influence on the energy demand. As results, there will be an important decrease in carbon intensity in the short-run and long run, which will significantly cut down the CO2 emissions. Acknowledgment The paper is supported by Newhuadu Business School Research Fund, the China Sustainable Energy Program (G-1305-18257), National Social Science Foundation of China (Grant No.12 & ZD059), and Ministry of Education (Grant No. 10GBJ013). References
5. Conclusions and policy suggestions We estimated the reduction potential of carbon dioxide emissions in the Chinese textile industry by employing a co-integration approach. In effect, the carbon intensity (CO2/value added) was
[1] EIA, U.S. Energy Information Administration. [accessed 15.03.13]. [2] China energy statistical yearbook, N.B.o.S., People’s Republic of China, Department of Energy Statistics. China Statistical Press; 2012. [3] Sun W-Q et al. Change in carbon dioxide (CO2) emissions from energy use in China’s iron and steel industry. J Iron Steel Res Int 2011;18(6):31–6.
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787 [4] Xu J-H et al. Energy consumption and CO2 emissions in China’s cement industry: a perspective from LMDI decomposition analysis. Energy Policy 2012;50:821–32. [5] Wang Y, Liang S. Carbon dioxide mitigation target of China in 2020 and key economic sectors. Energy Policy 2013;58:90–6. [6] Hatzigeorgiou E, Polatidis H, Haralambopoulos D. CO2 emissions in Greece for 1990–2002: a decomposition analysis and comparison of results using the Arithmetic Mean Divisia Index and Logarithmic Mean Divisia Index techniques. Energy 2008;33(3):492–9. [7] Guo X-D et al. Evaluation of potential reductions in carbon emissions in Chinese provinces based on environmental DEA. Energy Policy 2011;39(5):2352–60. [8] Choi Y, Zhang N, Zhou P. Efficiency and abatement costs of energy-related CO2 emissions in China: a slacks-based efficiency measure. Appl Energy 2012;98:198–208. [9] Hasanbeigi A, Menke C, Price L. The CO2 abatement cost curve for the Thailand cement industry. J Cleaner Prod 2010;18(15):1509–18. [10] Worrell E, Price L, Martin N. Energy efficiency and carbon dioxide emissions reduction opportunities in the US iron and steel sector. Energy 2001;26(5):513–36. [11] Ke J et al. Potential energy savings and CO2 emissions reduction of China’s cement industry. Energy Policy 2012;45:739–51. [12] Ari I, Aydinalp Koksal M. Carbon dioxide emission from the Turkish electricity sector and its mitigation options. Energy Policy 2011;39(10):6120–35. _ S. The scenario analysis on CO2 emission mitigation [13] Özer B, Görgün E, Incecik potential in the Turkish electricity sector: 2006–2030. Energy 2013;49:395–403. [14] He J, Deng J, Su M. CO2 emission from China’s energy sector and strategy for its control. Energy 2010;35(11):4494–8. [15] Li L et al. Energy demand and carbon emissions under different development scenarios for Shanghai, China. Energy Policy 2010;38(9):4797–807. [16] Henriques Jr MF, Dantas F, Schaeffer R. Potential for reduction of CO2 emissions and a low-carbon scenario for the Brazilian industrial sector. Energy Policy 2010;38(4):1946–61. [17] Dickey DA, Fuller WA. Distribution of the estimators for autoregressive time series with a unit root. J Am Stat Assoc 1979;74(366):427–31. [18] Phillips PCB, Perron P. Testing for a unit root in time series regression. Biometrica 1988;75(2):335–46. [19] Kwiatkowski D et al. Testing the null hypothesis of stationarity against the alternative of a unit root: how sure are we that economic time series have a unit root? J Econom 1992;54(1–3):159–78. [20] Sabuhoro JB, Larue B. The market efficiency hypothesis: the case of coffee and cocoa futures. Agr Econ 1997;16(3):171–84. [21] Engle RF, Granger CWJ. Cointegration and error correction: representation, estimation and testing. Econometrica 1987;55(2):251–76.
787
[22] Johansen S. Likelihood-based inference in cointegrated vector autoregressive models. Oxford: Oxford University Press; 1995. [23] Johansen S, Juselius K. Maximum likelihood estimation and inferences on cointegration with applications to the demand for money. Oxford Bull Econ Stat 1990;52(2):169–210. [24] Pesaran M, Pesaran B. Working with microfit 4.0: interactive econometric analysis. Oxford University Press; 1997. [25] Enevoldsen MK, Ryelund AV, Andersen MS. Decoupling of industrial energy consumption and CO2-emissions in energy-intensive industries in Scandinavia. Energy Econ 2007;29(4):665–92. [26] Lin BQ. Electricity demand in the People’s Republic of China: investment requirement and environmental impact. Asian Development Bank; 2003. Working Paper (No. 37). [27] Greening LA, Davis WB, Schipper L. Decomposition of aggregate carbon intensity for the manufacturing sector: comparison of declining trends from 10 OECD countries for the period 1971–1991. Energy Econ 1998;20(1):43–65. [28] Nag B, Parikh J. Indicators of carbon emission intensity from commercial energy use in India. Energy Econ 2000;22(4):441–61. [29] Greening LA, Ting M, Davis WB. Decomposition of aggregate carbon intensity for freight: trends from 10 OECD countries for the period 1971–1993. Energy Econ 1999;21(4):331–61. [30] Graus W, Worrell E. Methods for calculating CO2 intensity of power generation and consumption: a global perspective. Energy Policy 2011;39(2):613–27. [31] Cai W, Wang C, Chen J. Revisiting CO2 mitigation potential and costs in China’s electricity sector. Energy Policy 2010;38(8):4209–13. [32] China Statistical Yearbook, N.B.o.S., People’s Republic of China, China Statistics Press, Beijing; 1990–2011. [33] China statistical yearbook on science and technology, N.B.o.S., People’s Republic of China, China Statistics Press, Beijing; 1990–2011. [34] CEIC China Database; 2013.
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Mitigation potential of carbon dioxide emissions in the Chinese textile industry Boqiang Lin a,b,⇑, Mohamed Moubarak c a
New Huadu Business School, Minjiang University, Fuzhou 350108, China Collaborative Innovation Center for Energy Economics and Energy Policy, China Institute for Studies in Energy Policy, Xiamen University, Xiamen, Fujian, 361005, PR China c School of Energy Research, Xiamen University, Xiamen 361005, China b
h i g h l i g h t s We employed Johansen cointegration technique and scenarios analysis. The carbon intensity in the Chinese textile industry can be reduced by 60%. The CO2 emissions reduction potential is estimated to be 44.8 million tons by 2025. Improving the technology and labor productivity will drive down the CO2 emissions. Energy substitution is a promising method to cut down CO2 emissions.
a r t i c l e
i n f o
Article history: Received 31 March 2013 Received in revised form 6 June 2013 Accepted 4 August 2013 Available online 4 September 2013 Keywords: Carbon intensity CO2 emissions Textile industry China
a b s t r a c t We estimated the reduction potential of carbon dioxide emissions in the Chinese textile industry by forecasting the carbon intensity (CO2 emissions/industrial value added) in different scenarios. The Johansen co-integration technique was employed in order to establish the long term equilibrium equation. Three scenarios (Business As Usual (BAU), medium and optimum) were designed to estimate the future trend of carbon intensity in the Chinese textile industry. The results showed that energy price, energy substitution, labor productivity and technology have significant impact on the carbon intensity. Estimated to 1.49 t CO2/10,000 yuan in 2010, we found that for the BAU scenario, the carbon intensity will decrease to 0.5 and 0.29 t CO2/10,000 yuan by 2020 and 2025 respectively. For the medium scenario, carbon intensity will decline to 0.12 t CO2/10,000 yuan. Yet by the optimum scenario, the intensity is expected to considerably decrease to 0.05 t CO2/10,000 yuan by 2025. Using the BAU forecast as baseline, the quantity of reduction potential in carbon dioxide emissions is estimated to be 44.8 million tons CO2 by 2025. Considering this huge potential, we provided policy suggestions to reduce the level of CO2 emissions in the Chinese textile industry. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction The carbon dioxide emissions in China has been increasing geometrically over the past thirty years. According to the 2010 estimates, the quantity of CO2 emissions in China was equivalent to 25.1% of the global emissions [1]. In order to mitigate it, the Chinese government made a commitment to reduce considerably, the levels of emissions for the coming years and decided to give much priority to economic growth balanced with less pollution. According to the 12th Five-year plan presented in 2011, China is expecting by 2015 to reduce the energy consumption per GDP by 16%, to increase the share of non-fossil fuel in the total energy ⇑ Corresponding author at: New Huadu Business School, Minjiang University, Fuzhou 350108, China. Tel.: +86 0592 2186076; fax: +86 0592 2186075. E-mail address: [email protected] (B. Lin). 0306-2619/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apenergy.2013.08.010
consumption by 11.4% and to cut down the carbon dioxide emissions per unit of GDP by 17%, compared to 2010. The realization of this plan will substantially boost China’s contribution to the mitigation of carbon dioxide emissions at global level. In this paper, we analyzed the potential to reduce emissions in the Chinese textile industry. China became the world’s largest producer of textile materials. Consequently due to this industrial expansion, the total energy consumption and CO2 emissions grew annually by 4% and 2% respectively from 1985 to 2010. Energy consumption in the textile industry was estimated to be 73.45 mtce (million tons coal equivalent) in 2010; which was equivalent to 2.41% of the total energy consumed in China. In the meantime, carbon dioxide emissions represented 3.7% of the total emissions from the manufacturing sector [2]. The energy consumption was mainly driven by coal consumption due to its predominance as energy source employed in the production processes. Despite its progressive
782
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
decline in proportion to the total energy consumed, the dependency on coal is not expected to drastically change in the short term (Fig. 1). Coal consumption increased from 18.68 mtce to 29.26 mtce during the study period. Moreover, coal has a high coefficient factor in term of CO2 emissions. Thus, a combination of coal and other fossil fuels, namely coke and oil products led to high level of CO2 emissions in the textile industry. As illustrated in Fig. 1, the CO2 emissions were mainly due to the coal consumption; which represented an important proportion of the total energy consumed. Through a series of questions orienting our research, we estimated the carbon intensity trends and the carbon dioxide reduction potential in several scenarios: (i) what variables have influence on carbon intensity in the Chinese textile industry? (ii) What are their degree of influence? and (iii) What will be the carbon intensity and is the reduction potential of CO2 emissions? Answers to these questions will have significance for this paper; it will provide policy methods to reduce the carbon intensity, contribute to provide industrial strategy for the Chinese textile industry by cutting down CO2 emissions while improving the industrial competitiveness. 2. Literature review There have been several reports on carbon dioxide emissions reduction. These reports have been conducted using different approaches, leading to mixed conclusions. One popular and widely used method is the decomposition analysis. It has some convenience in studying the CO2 emissions change. This method is based on analyzing the contribution of selected factors to the change in emissions. Sun et al. [3] studied the change of CO2 emissions in China’s iron and steel industry and concluded that energy consumption was the most important factor leading to a decrease in CO2 emissions. Later, Xu et al. [4] analyzed the case of China’s cement industry and Wang and Liang [5] focused on 12 key economic sectors. Similarly, Hatzigeorgiou et al. [6] analyzed the energy-related CO2 emissions in Greece from 1990 to 2002. From these studies, we denote that decomposition approach is convenient to study change in CO2 emissions, however it is not appropriate for forecasting the future emissions level and reduction potential. The data envelopment analysis (DEA) model has been also used in several studies. Guo et al. [7] focused on finding out the emissions performance of Chinese provinces and their respective reduction potential. Later, Choi et al. [8] attempted to determine the potential reduction and efficiency of CO2 emissions in China. According to their results, the average marginal abatement cost of CO2 emissions is about $7.2. However, the limitation of this study is due to the fact that there is no specific function form while using the nonparametric method.
Fig. 1. Percentage of growth of coal consumption, total energy consumption and CO2 emissions.
Hasanbeigi et al. [9] applied the bottom-up CO2 abatement cost curves model which is a variant of the energy conservation supply curves methodology to determine the potentials and costs of CO2 mitigation through abatement technologies from 2010 to 2025 in the Thai cement industry. They found that the abatement potential represented 15% of the total cement industry emissions in 2005. Similarly, Worrell et al. [10] analyzed the cost-effective energy efficiency and the carbon dioxide emissions mitigation potential in the US iron and steel industry between 1958 and 1994. They found a total cost-effective reduction potential of 3.8 GJ/t with a payback period of three years or less. Despite the wide use of this method, it has some limitations because it is focused only on the existing technology. Using the scenario analysis method, Ke et al. [11] studied the energy saving and CO2 emissions reduction potential in China’s cement industry. They estimated that the carbon reduction potential represents 3.2–4.4 gigatonnes during 2011–2030 and that the energy efficiency is the key policy measure to drive down carbon emissions intensity in the cement’s industry. Ari and Aydinalp Koksal [12] analyzed the reduction potential of carbon dioxide emission from the Turkish electricity sector. Based on four scenarios, they argued that a significant decrease in the amount of CO2 emissions from electricity generation can be achieved if the share of the fossil-fueled power plants is lowered. Later, Özer et al. [13] used the Long-range Energy Alternatives Planning system (LEAP) model to investigate mitigation potential of emissions in the electricity sector. They found that cumulative CO2 emission reduction between the BAU and Mitigation Scenarios from 2006 to 2030 is 903 million tons. He et al. [14] attempted to forecast the future CO2 emissions from fossil energy combustion in China from 2010 to 2020; similar to Li et al. [15] who focused their research on shanghai city (China). Henriques et al. [16] analyzed the carbon dioxide (CO2) emissions reduction potential from energy consumption in Brazilian industrial sector. Their conclusions showed that by improving energy efficiency, recycling and energy substitution, it will be possible to reduce carbon dioxide emissions by 43% in the industrial sector during 2010–2030 period. However, an important variable such as price was not considered. Far to be exhaustive, several methods and variables have been used to determine the carbon emissions change and reduction potential. However to the best of our knowledge, studies conducted on reduction potential of CO2 emissions in the Chinese industries are limited. Therefore, we extended the research on Chinese textile industry by conducting an in-depth study on mitigation potential of CO2 emissions. 3. Methodology and data sources 3.1. Co-integration approach The Johansen co-integration test has a property to determine the long term relationship among variables. It is employed to determine a long run relationship among carbon intensity, technology, energy price, labor productivity and energy substitution. Prior to the regression analysis, we verified that the variables are stationary in order to avoid spurious results. So, to determine whether these variables have the same integration order, the unit root tests were applied. The most common tests used are Augmented Dickey–Fuller (ADF) tests [17], Phillips–Perron (PP) tests [18] and Kwiatkowski–Phillips–Schmidt–Shin (KPSS) tests [19]. Eq. (1) is the Augmented Dickey–Fuller (ADF) test representation to test the unit root hypothesis.
Dzt ¼ b0 þ a0 t þ a1 zt1 þ
m X bi Dzt1 þ et i¼1
ð1Þ
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
With, zt representing the variable at time t; Dzt1 the zt1 zt2; et as the disturbance with a mean 0 and a variance r2; t the linear time trend and m is the lag order. The null hypothesis H0 is: a1 = 0 in Eq. (1), if a1 is significantly less than zero, the null hypothesis of a unit root is rejected. However, the ADF tests may lack efficiency when applied to a small sample size. The PP test uses the same models as the ADF test, but is remarkably insensitive to the heteroscedasticity and the autocorrelation of the residuals. Both ADF and PP tests require assumptions that tested sequence may contain a constant term and trend variables, the application has a certain degree of inconvenience. To strengthen it, KPSS test is also applied; which is more appropriate for small samples when it chooses a lower lag truncation parameter [20]. Therefore, we used all three test methods and compared their results. We found that the variables are stationary and consequently proceeded to test for cointegration. There are two most popular methods used to establish co-integration relationship: Engle and Granger [21] and Johansen and Juselius method [22,23]. The difference between these approaches is that the Engle–Granger procedure is mostly applied to a single equation co-integration test, while the second approach can detect the existence of co-integration among variables and also determine the number of co-integration vectors. As being the most convenient for this paper, we applied the Johansen–Juselius method trace test and maximum eigenvalue test to determine the number of cointegrating vectors in our model. 3.1.1. Stability Once the co-integration relationship is determined and the parameters are estimated, it is imperative to test for robustness of the model. After the error correction model has been built, the Pesaran and Pesaran [24] test is applied to the cumulative sum of recursive residuals (CUSUM) and the cumulative sum of square (CUSUMSQ) to examine the parameter stability. The model is qualified stable if the plots of CUSUM and CUSUMSQ statistics are confined within the 5% critical bounds of parameter stability. 3.1.2. Scenarios and variables We explored three carbon dioxide reduction scenarios. The feasibility of these scenarios is justified by the fact that for each variable, the future trend is based on the historical data from 1985 to 2010. The scenarios are: 1. The BAU (Business As Usual) scenario, which is based on the historical trend observed for each variable. According to this scenario, the future values of the variables will be similar to their respective average growth rate observed during the period 1985–2010. For The carbon dioxide reduction potential, we considered the BAU scenario as a base line for policy actions. 2. The optimum scenario represents the most committed scenario to reduce the carbon intensity. For this scenario, the future growth rate of each variable is considered to be the highest historical value observed since 2001. There are important growths observed during the period prior to 2001. However, it seems not realistic to expect that it will be possible to reach this anterior performance due to important transformations that occurred in several aspects of the Chinese economy, such as structural change since 2001 and the important transformation that occurred in the Chinese textile industry since the country joined the world trade organization (2001). For instance, exportation of textile and clothing product considerably increased during that period. The promotion of industrial upgrading and setting up structure regulation optimization have become priorities. Furthermore, reduction of carbon dioxide emissions and promotion of energy saving became major concerns.
783
3. The medium scenario is considered to be the average policy commitment. For this scenario, the future values of the variables will be based on the average growth of BAU scenario and the optimum scenario. The growth rate of variables is more important than in BAU scenario; meaning that the policy strength of this scenario is weaker than the optimum scenario but stronger than the BAU scenario. In this paper, four factors are integrated to detect their respective weights on the carbon intensity: Energy Price; it is expected to have an opposite trend to the carbon intensity [25]. This means that a growth of price leads to a reduction in energy consumption and carbon intensity. Because of lack of availability of data, raw material and fuel price index is considered [26]. Energy substitution; it is the proportion of non-fossil fuel energy consumed in the Chinese textile industry. An increase in share of energy consumption from non-fossil fuels is expected to reduce the carbon intensity and the CO2 emissions. This includes energy such as nuclear power, hydropower and renewable energy. This variable has an important impact on the carbon intensity [27,28]. Labor productivity; efficiency in labor productivity contributes to reduce the carbon intensity in textile industry. We assume that labor productivity is negatively correlated to the carbon intensity. It is represented by the industrial value added per capita. Increase in aggregate carbon intensity is driven primarily by an increase in activity [29]. Technology; it is considered as a key element resulting in carbon intensity reduction [30,31]. The Research and Development brings new techniques and methods in the production system which leads to reduced energy consumption and improvements of efficiency. In this paper, the technology is represented by the expenditure for science and technology as percentage of total revenue from sales. Furthermore, there are other variables which may have also impact on the trend of carbon intensity. The CO2 emissions trading scheme in China is an interesting factor which will have impact on the trend of carbon intensity. However, it is still at early development stage in China. It is confronted with the difficulty of the availability and accuracy of CO2 emissions data from participating manufacturers. There is also controversy about defining emission limits under diversified energy structures. Therefore we ignored this variable in our analysis. Adjustment for high added value products in the textile industry is also an important factor to drive down carbon intensity. The optimization of the product structure such as the design, brand value and apparel products may help to reduce the energy consumption, therefore to decrease the carbon emissions level. However, due to the lack of available data and the complexity of the Chinese textile industry, we do not consider this variable as key factor in this study.
3.2. Data sources We used data covering the period from 1985 to 2010 in this study. The data were mainly collected from China Statistical Yearbook [32], China Statistical Yearbook on Science and Technology [33], CEIC China Database [34], China Energy Statistical Yearbook [2] and were calculated based on 2005 constant prices. For stability of the data, we used the natural logarithm of the variables carbon intensity (LnCI), energy price (LnEP), energy substitution (LnES), labor productivity (LnLP) and technology (LnTD). The coefficients of the CO2 emissions of the various energy sources (coal, coke, crude oil, diesel, fuel oil, gasoline, kerosene and natural gas) were chosen based on the estimations made by the Intergovernmental Panel on Climate Change [35]. CO2 emissions were estimated by multiplying
784
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
Table 1 Unit root test. Series
LnCI LnEP LnES LnLP LnTD 4LnCI 4LnEP 4LnES 4LnLP 4LnTD
Table 2 Lag selection criteria.
ADF
PP
KPSS
Lag Log L
No trend
With trend
No trend
With trend
No trend
With trend
0.411 1.995 0.436 1.054 1.142 3.330b 3.693b 5.113a 2.489 4.584a
1.138 2.213 2.328 1.617 1.794 3.262c 3.959b 4.966a 2.446 4.530a
0.449 1.954 0.401 0.349 1.148 3.292b 2.674c 5.120a 2.445 4.592a
1.513 1.335 2.328 1.623 1.866 3.218 2.671 4.694a 2.446 4.537a
0.735a 0.619a 0.733a 0.735a 0.674a 0.121c 0.230c 0.114c 0.144c 0.097c
0.120b 0.165a 0.118c 0.115c 0.127b 0.114c 0.094c 0.144b 0.143b 0.074c
0 1 2
consumption of individual fuels by their CO2 emissions coefficients. Similar to Lin et al. [36], we used the following relation to determine the carbon dioxide reduction potential:
ð2Þ
CRPt is the carbon dioxide reduction potential in period t; CIBAUt is the carbon intensity from the business as usual scenario in period t. CISCENARIOt is the carbon intensity of period t for optimum and medium scenarios, respectively. VAt is the value added in textile industry for period t. 4. Empirical analysis 4.1. Unit root tests The Augmented Dickey–Fuller (ADF), Phillips–Perron (PP) and Kwiatkowski–Phillips–Schmidt–Shin (KPSS) tests were applied in levels and in first differences (Table 1).
ð0:07775Þ
AIC
SC
HQ
43.34087 NA 2.82e08 3.195072 2.949644 3.129960 156.6918 170.0265a 1.89e11a 10.55765a 9.085087a 10.16698a 177.3132 22.33975 3.72e11 10.19276 7.493056 9.476530
Table 3 Johansen co-integration test. Hypothesized No. of CE(s) b
None At most At most At most At most
1b 2b 3 4
Eigen value
Trace Statistic
0.05 Critical value
p-Valuesa
0.973500 0.746096 0.585954 0.429147 0.016030
148.5794 65.07531 33.54692 13.26603 0.371681
69.81889 47.85613 29.79707 15.49471 3.841466
0.0000 0.0006 0.0177 0.1054 0.5421
Max-Eigen
0.05
Statistic
Critical value
p-Valuesa
33.87687 27.58434 21.13162 14.26460 3.841466
0.0000 0.0147 0.0654 0.0814 0.5421
Hypothesized No. of CE(s) b
None At most At most At most At most
1b 2 3 4
Eigen value 0.973500 0.746096 0.585954 0.429147 0.016030
ð0:04059Þ
According to the results, the null hypothesis of a unit root could not be rejected in levels for all the variables tested. Exception is observed with the KPSS test, revealing that the null hypothesis could be rejected for the variables (Table 1). However, despite this inconsistency, it was safer and reasonable to accept the null hypothesis for the level series [37] and then to reject the null hypothesis at the first differences. 4.2. Co-integration analysis An unrestricted vector autoregressive [38] was built to determine the cointegration relation among variables (LnCI, LnES, LnEP, LnTD and LnLP). Using the Akaike information criterion (AIC), Schwarz information criterion (SC), sequential modified LR test statistic (LR), Final prediction error (FPE), and Hannan–Quinn information criterion (HQ), a lag of 1 was selected (Table 2).
83.50406 31.52839 20.28089 12.89435 0.371681
a MacKinnon–Haug–Michelis probability values, MacKinnon–Haug Michelis (1999). b Rejection of the hypothesis at the 0.05 level.
Following this, we proceeded to the determination of the number of co-integration relations and estimation of the co-integration equation. The results of the rank tests for the logarithm of the variables are given in Table 3. The trace statistic and the maximum eigenvalue statistic revealed a unique cointegration equation among the logarithm of the variables. Thus, the co-integration equation can be formulated as:
LnCI ¼ 2:523464 0:224566 LnEP 0:248886 LnES 0:458779 LnLP 0:439590 LnTD ð0:03365Þ
FPE
LR: sequential modified LR test statistic (each test at 5% level). FPE: Final prediction error. AIC: Akaike information criterion. SC: Schwarz information criterion. HQ: Hannan–Quinn information criterion. a Lag order selected by the criterion.
Note: Eviews7 has been used for the tests. D Indicates series in first difference. The critical value for ADF with constant (and trend) at the 5% significance is 2.986 (3.603) and at the 1% significance is 3.724 (4.374). The critical value for KPSS with constant (and trend) at the 5% significance is 0.463 (0.146) and at the 1% significance is 0.739 (0.216). KPSS, Kwiatkowski et al. [19], refers to testing the null hypothesis of stationarity against the alternative of unit root. PP refers to Phillips and Perron [18] unit root test. a Significance at 1%. b Significance at 5%. c Significance at 10%.
CRPt ¼ ðCIBAUt CISCENARIOt Þ VAt
LR
ð0:03315Þ
ð3Þ
The numbers in parentheses are the resulting standard errors. Some comments can be formulated from Eq. (3). Eq. (3) indicates that there exists a long-term equilibrium relationship among variables carbon intensity, energy price, labor productivity, energy substitution and technology. (a) Eq. (3) confirms that as expected the variables energy price, labor productivity, energy substitution and technology are negatively correlated to the carbon intensity. In the long run, these variables contribute to considerably reduce the carbon intensity level. According coefficients values, a 1% increase in energy price, energy substitution, labor productivity and technology will result to a decline of carbon intensity by 0.22%, 0.24%, 0.45%, and 0.43% respectively.
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
785
Fig. 2. (a) Plot of cumulative sum of recursive residuals (CUSUM) and (b) the CUSUM of square (CUSUMSQ).
(a) The labor productivity and technology are the most influencing factors; meaning that these variables had predominant impact on the decrease of carbon intensity during the past 25 years in the Chinese textile industry. This implies that improving the technology and efficiency in labor productivity are the most important aspects driving down the CO2 intensity. 4.3. Test of stability We verified the stability of our model by employing the method proposed by Pesaran and Pesaran [24]. For a certainty in the estimations made, we applied the test of stability. The relevance of this test is to determine whether or not there are over changes in the estimated parameters; because in case of no consistency, the validity of conclusions may be biased. Fig. 2(a and b) illustrate plots of CUSUM and CUSUMSQ statistics of the stability tests. The curves are within the 5% critical bounds indicating that model and the parameters of Eq. (3) satisfied the stability test, and then valid. 4.4. Carbon intensity projections The error correction model applied reflects a short-term relationship among the variables, and the cointegration model indicated the existence of a long-term relationship. Therefore, we conducted analysis in order to predict the reduction potential of carbon dioxide emissions of the Chinese textile industry up to 2025. Similar approach to Lin et al. [39], we used the co-integration Eq. (3) to estimate the amount of carbon dioxide emissions that can be reduced from the textile industry. The first step for the estimation was to determine a measure of the prediction by substituting the historical data of the variables (CI, ES, EP, TD and LP) into the co-integration Eq. (3) in order to obtain fitted trend of carbon intensity for the period 1985–2010. Fig. 3 illustrates the average relative error for the fitted trend compared to the actual trend. Following this, we proceeded to the estimation for future values of carbon intensity up to 2025. For the BAU scenario, the carbon intensity was estimated based on the average annual growth rate of the variables during the study period 1985–2010; which are TD = 7%, ES = 5%, EP = 5% and LP = 13%. Therefore, the forecasts of carbon intensity are 0.86 t CO2/10,000 yuan by 2015, 0.5 t CO2/10,000 yuan by 2020 and 0.29 t CO2/10,000 yuan by 2025. This finding implies that, based on the current trends of CO2 emissions, energy consumption and value added in the BAU case, the carbon intensity in the Chinese textile industry will decline to 0.29 t CO2/10,000 yuan by 2025; equivalent to 1/5 of the level observed in 2010 (1.49 t CO2/
10,000 yuan). It is therefore possible to determine the carbon intensity trend under different scenarios and the related reduction potential of CO2 emissions. 4.5. Future carbon reduction potential from scenarios The reduction potential of CO2 emissions is based on three scenarios; the business as usual, medium and optimum scenarios. The BAU scenario as described above is policy driven. The optimum and the medium scenarios are formulated to speed up the CO2 emissions reduction. The optimum scenario is the most active one in term of carbon dioxide mitigation policy. According to that scenario, the future growth rate of each variable influencing the carbon intensity will be similar to the highest historical value observed since 2001; which were ES = 11% in 2002, EP = 7% in 2010, LP = 15% in 2009 and TD = 29% in 2007. For the medium scenario, the growth rates of variables are the average of values between the BAU and active scenarios (Table 4). Consequently, we determined the future trends of the carbon intensity by substituting the values of the variables into the cointegration Eq. (3), (Table 5). Table 5 shows that the carbon intensity in the China’s textile industry will considerably decline during the period 2010–2025. The decrease rate differs and is more important when evolving from the BAU to the medium and to the optimum scenarios. In the BAU scenario case, the carbon intensity in the Chinese the textile industry will drop from 1.49 t CO2/10,000 yuan (the level in 2010) to 0.29 t CO2/10,000 yuan by 2025. For the other two scenarios, the carbon intensity will decrease to 0.12 t CO2/10,000 yuan with the medium scenario and 0.06 t CO2/10,000 yuan for the optimum scenario by 2025. Accordingly, Eq. (2) is used to determine the reduction potential of carbon dioxide emissions by scenarios. The previsions from the
Fig. 3. Curves for actual, fitted and forecasted trends of carbon intensity.
786
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787
Table 4 CO2 emissions reduction scenarios. Variables
BAU (%)
Medium scenario (%)
Optimum scenario (%)
EP ES LP TD
5 5 13 7
6 8 14 18
7 11 15 29
Table 5 Carbon intensity forecasts by scenarios (in t CO2/10,000 yuan). Scenarios
2015
2020
2025
BAU Medium scenario Optimum scenario
0.86 0.65 0.5
0.5 0.28 0.17
0.29 0.12 0.06
Fig. 4. Reduction potential of carbon dioxide emissions in the Chinese textile industry.
‘‘12th five-year plan for textile industry’’, the average annual growth rate of value added is predicted to be 8% from 2011 to 2015. Considering its historical trend, the value added of the Chinese textile industry 1.2 trillion yuan by 2020 and 1.8 trillion yuan by 2025. To estimate the CO2 emissions reduction potential, two cases are built: Case A is the change in carbon intensity from the basis scenario (BAU scenario) to the optimum scenario. Case B is the change in the carbon intensity from the BAU scenario to the medium Scenario. Using the expected future values of carbon intensity from scenarios and the forecasted industrial value added, the potential reduction in carbon dioxide emissions are estimated for the period of 2011–2025, as illustrated in Fig. 4. If the CO2 emissions in the Chinese textile industry change from the BAU scenario to medium scenario (Case B), the quantity of carbon dioxide reduction will be approximately 16.4 million tons CO2 by 2015, 25.8 million tons CO2 by 2020, and will increase to 30.5 million tons CO2 by 2025. If more ambitious measures are implemented, meaning that the CO2 emissions change from the BAU scenario to the optimum scenario (Case A), the reduction potential will be more important. It is estimated to be about 28.1 million tons CO2 by 2015, 39.7 million tons CO2 by 2020 and will be 43 million tons CO2 by 2025.
determined by considering the historical values of the variables, energy price, energy substitution, labor productivity and technology. Moreover, the future values of the carbon intensity have been estimated based on three scenarios. The results showed that the carbon intensity will decrease from 1.4 t CO2/10,000 yuan in 2010 to 0.29 t CO2/10,000 yuan for the BAU scenario, 0.12 t CO2/ 10,000 yuan for the medium scenario and 0.06 t CO2/10,000 yuan for the optimum scenario by 2025. Furthermore, the results showed that the carbon dioxide emissions will decrease by 16.4, 25.8 and 30.5 million tons CO2 by 2015, 2020 and 2025 respectively. If more ambitious measures are implemented, the reduction potential is estimated to be 43 million tons CO2 by 2025; equivalent to a 60% reduction in CO2 compared to the 2010 emissions level. The coefficients of elasticities from the cointegration equation revealed that all the variables statistically have significant impact on the evolution of the carbon intensity trend. The significance of these coefficients provides orientations of policies to be implemented to reach the targets of carbon intensity reduction and carbon dioxide emissions mitigation. We suggest that the Chinese government continues to invest a lot into research and development to come out with environmentally-friendly cutting-edge technologies in the production and distribution of power to sustain the robust economic growth. It is also necessary for the Chinese government to actively promote the technology transfer by facilitating international cooperation among and industries. Wang et al. [40] suggested that the technology upgrading can simultaneously improve economic growth, accelerate the energy efficiency and also reduce the carbon intensity per unit of output. Improving technology will also improve the productivity, which is also a determinant for the carbon intensity. Similar to our findings, the technology and labor productivity are the key elements to focus on in order to mitigate the carbon dioxide emissions in the Chinese textile industry. We also found that energy substitution is an important factor in the efforts to cut down emissions. Coal has been and will continue to be for the coming years the main source of energy in Chinese industries including the textile. Naturally, a progressive substitution of coal by other less- polluting energy sources such as renewable energy, natural gas and nuclear power will considerably reduce the CO2 emissions. The price of fossil fuel has been controlled in the past by the Chinese government. A progressive and complete release of control on energy price will have interesting impact on the long term. Increase in energy cost will lead to more efficiency and energy conservative measures at the firms’ level; which will in return contribute to decrease the emissions. The government should establish a proportional tax policy on fossil fuel consumption, subsidies and tax relieve in order to promote the interest for less polluting alternative energy sources and influence on the energy demand. As results, there will be an important decrease in carbon intensity in the short-run and long run, which will significantly cut down the CO2 emissions. Acknowledgment The paper is supported by Newhuadu Business School Research Fund, the China Sustainable Energy Program (G-1305-18257), National Social Science Foundation of China (Grant No.12 & ZD059), and Ministry of Education (Grant No. 10GBJ013). References
5. Conclusions and policy suggestions We estimated the reduction potential of carbon dioxide emissions in the Chinese textile industry by employing a co-integration approach. In effect, the carbon intensity (CO2/value added) was
[1] EIA, U.S. Energy Information Administration.
B. Lin, M. Moubarak / Applied Energy 113 (2014) 781–787 [4] Xu J-H et al. Energy consumption and CO2 emissions in China’s cement industry: a perspective from LMDI decomposition analysis. Energy Policy 2012;50:821–32. [5] Wang Y, Liang S. Carbon dioxide mitigation target of China in 2020 and key economic sectors. Energy Policy 2013;58:90–6. [6] Hatzigeorgiou E, Polatidis H, Haralambopoulos D. CO2 emissions in Greece for 1990–2002: a decomposition analysis and comparison of results using the Arithmetic Mean Divisia Index and Logarithmic Mean Divisia Index techniques. Energy 2008;33(3):492–9. [7] Guo X-D et al. Evaluation of potential reductions in carbon emissions in Chinese provinces based on environmental DEA. Energy Policy 2011;39(5):2352–60. [8] Choi Y, Zhang N, Zhou P. Efficiency and abatement costs of energy-related CO2 emissions in China: a slacks-based efficiency measure. Appl Energy 2012;98:198–208. [9] Hasanbeigi A, Menke C, Price L. The CO2 abatement cost curve for the Thailand cement industry. J Cleaner Prod 2010;18(15):1509–18. [10] Worrell E, Price L, Martin N. Energy efficiency and carbon dioxide emissions reduction opportunities in the US iron and steel sector. Energy 2001;26(5):513–36. [11] Ke J et al. Potential energy savings and CO2 emissions reduction of China’s cement industry. Energy Policy 2012;45:739–51. [12] Ari I, Aydinalp Koksal M. Carbon dioxide emission from the Turkish electricity sector and its mitigation options. Energy Policy 2011;39(10):6120–35. _ S. The scenario analysis on CO2 emission mitigation [13] Özer B, Görgün E, Incecik potential in the Turkish electricity sector: 2006–2030. Energy 2013;49:395–403. [14] He J, Deng J, Su M. CO2 emission from China’s energy sector and strategy for its control. Energy 2010;35(11):4494–8. [15] Li L et al. Energy demand and carbon emissions under different development scenarios for Shanghai, China. Energy Policy 2010;38(9):4797–807. [16] Henriques Jr MF, Dantas F, Schaeffer R. Potential for reduction of CO2 emissions and a low-carbon scenario for the Brazilian industrial sector. Energy Policy 2010;38(4):1946–61. [17] Dickey DA, Fuller WA. Distribution of the estimators for autoregressive time series with a unit root. J Am Stat Assoc 1979;74(366):427–31. [18] Phillips PCB, Perron P. Testing for a unit root in time series regression. Biometrica 1988;75(2):335–46. [19] Kwiatkowski D et al. Testing the null hypothesis of stationarity against the alternative of a unit root: how sure are we that economic time series have a unit root? J Econom 1992;54(1–3):159–78. [20] Sabuhoro JB, Larue B. The market efficiency hypothesis: the case of coffee and cocoa futures. Agr Econ 1997;16(3):171–84. [21] Engle RF, Granger CWJ. Cointegration and error correction: representation, estimation and testing. Econometrica 1987;55(2):251–76.
787
[22] Johansen S. Likelihood-based inference in cointegrated vector autoregressive models. Oxford: Oxford University Press; 1995. [23] Johansen S, Juselius K. Maximum likelihood estimation and inferences on cointegration with applications to the demand for money. Oxford Bull Econ Stat 1990;52(2):169–210. [24] Pesaran M, Pesaran B. Working with microfit 4.0: interactive econometric analysis. Oxford University Press; 1997. [25] Enevoldsen MK, Ryelund AV, Andersen MS. Decoupling of industrial energy consumption and CO2-emissions in energy-intensive industries in Scandinavia. Energy Econ 2007;29(4):665–92. [26] Lin BQ. Electricity demand in the People’s Republic of China: investment requirement and environmental impact. Asian Development Bank; 2003. Working Paper (No. 37). [27] Greening LA, Davis WB, Schipper L. Decomposition of aggregate carbon intensity for the manufacturing sector: comparison of declining trends from 10 OECD countries for the period 1971–1991. Energy Econ 1998;20(1):43–65. [28] Nag B, Parikh J. Indicators of carbon emission intensity from commercial energy use in India. Energy Econ 2000;22(4):441–61. [29] Greening LA, Ting M, Davis WB. Decomposition of aggregate carbon intensity for freight: trends from 10 OECD countries for the period 1971–1993. Energy Econ 1999;21(4):331–61. [30] Graus W, Worrell E. Methods for calculating CO2 intensity of power generation and consumption: a global perspective. Energy Policy 2011;39(2):613–27. [31] Cai W, Wang C, Chen J. Revisiting CO2 mitigation potential and costs in China’s electricity sector. Energy Policy 2010;38(8):4209–13. [32] China Statistical Yearbook, N.B.o.S., People’s Republic of China, China Statistics Press, Beijing; 1990–2011. [33] China statistical yearbook on science and technology, N.B.o.S., People’s Republic of China, China Statistics Press, Beijing; 1990–2011. [34] CEIC China Database; 2013.