Using LMDI Approach to Analyze Changes in Carbon Dioxide Emissions of China ’ s Logistics Industry

Purpose: China is confronting with tremendous pressure in carbon emission reduction. While logistics industry seriously relies on fossil fuel, and emits greenhouse gas, especially carbon dioxide. The aim of this article is to estimate the carbon dioxide emission in China’s logistics sector, and analyze the causes for the change of carbon dioxide emission, and identify the critical factors which mainly drive the change in carbon dioxide emissions of China’s logistics industry. Design/methodology/approach: The logarithmic mean Divisia index (LMDI) method has often been used to analyze decomposition of energy consumption and carbon emission due to its theoretical foundation, adaptability, ease of use and result interpretation. So we use the LMDI method to analyze the changes in carbon dioxide emission of China’s logistics industry in this paper. Findings: By analyzing carbon dioxide emission of China’s logistics, the results show that the carbon dioxide emission of logistics in China has increased by 21.5 times, from 45.1 million tons to 1014.1 million tons in the research period. The highway transport is the main contributor to carbon dioxide emission in logistics industry. The energy intensity and carbon dioxide emission factors contributed to the reduction of carbon dioxide emission in China’s logistics industry in overall study period.


Introduction
Many studies have suggested that the concentration of greenhouse gases (GHG) in the atmosphere has been increasing as a result of human activities (Loo & Li, 2012), and the high concentration of GHG has caused global warming which was measured by the increase of the Earth's average temperature (Chapman, 2007).It was reported that the average global surface temperature had increased by 0.74ºC over the last 100 years (2014), which was caused by the GHGin the atmosphere, due to the consumption of numerous fossil fuels.Intensive use of fossil fuels can be cited as the main reason of the significant increase in anthropogenic GHG that lead to climate change (Ipek-Tunç, Türüt-Aşık & Akbostanci, 2009).Carbon dioxide (CO2) was the most important composition and accounted for about 80% share of the greenhouse effect (Liao, Lu & Tseng, 2011).Figure 1   Confronted with global warming, CO2 emission as a main composition of GHG was widely paid attention and researched by most governments, scholars and enterprises in recent years.We can roughly summarize the researches as the following aspects.The first aspect primarily focused on the relationships between energy consumption, CO2 emission and economic activities in different countries and districts, such as China (Wang, Zhou & Zhou, 2011), Russia (Pao, Yu & Yang, 2011), India (Ghosh, 2010), Europe (Acaravcia & Ozturk, 2010), South Africa (Menyah & Wolde-Rufael, 2010), Turkey (Halicioglu, 2009), and so on.The research results showed there were different causal relationships in different countries.The second aspect was on the forecasting of CO2 emission.For example, Azadeh, Khakestani and Saberi (2009) forecasted the oil consumption and CO2 emission in Canada, United States, Japan and Australia during 1990-2005 by using a flexible fuzzy regression algorithm.Olsthoorn (2001) estimated CO2 emissions from international aviation from 1950 to 2050 through a regression model.
Finally, it was case study of CO2 emission decomposition.For example, Ipek-Tunç et al. (2009) identified the factors that contributed to the changes in CO2 emissions in agriculture, industry and services.Hammond and Norman (2012)  As we know, transport is one of main resources of CO2 emissions.Some research achievements were made.Steenhof, Woudsman and Sparling (2006) analyzed the change in GHG emissions produced by Canada's freight transport using a decomposition analysis framework.Liao et al. (2011) examined CO2 emissions of truck-only transportation using activity-based emission modeling and compared those with intermodal coastal shipping and truck movements.Fatumata and Lee (2009) compared the energy intensity and CO2 emission of truck freight in Australia, France, Japan, the United Kingdom and the United States from 1973 to the present, using a bottom-up approach relying on national data.Solís and Sheinbaum (2013) presented a disaggregation of the fuel consumption and its related CO2 emissions from passenger and freight road transport in Mexico.Zhou, Chung and Zhang (2013) studied CO2 emissions performance of the transport sector throughout China's 30 administrative regions using Data Envelopment Analysis (DEA) models with different return of scales.Kellner and Igl (2015) examined how the network carbon footprint of a real-world distribution system was affected by the logistics service provider network that was chosen to forward goods from production facilities to customers.Xu and Lin (2015) adopted provincial panel data from 2000 to 2012 and nonparametric additive regression models to examine the key influencing factors of CO2 emissions in the transport sector in China.
We can find that many research achievements about CO2 emissions sprang up in different fields from above listed references.However, few scholars researched on carbon dioxide emission in logistics industry.A small quantity of scholars researched CO2 of logistics.For example, Zając (2011) presented of the conception of counting the energy consumption of logistics warehouse systems.Tang, Wang, Yan and Hao (2014) examined the issue of cutting emissions by reducing shipment frequency within the framework of periodic inventory review system.Hammami, Nouira and Frein (2014) developed a deterministic optimization model that incorporates carbon emissions in a multi-echelon production-inventory model with lead time constraints.Logistics is a process of planning, implementing and controlling the efficient, costeffective flow and storage of raw materials, in-process inventory, finished goods and related information from point of origin to point of consumption for the purpose of conforming to customer requirement (Cooper, Lambert & Pagh, 1997).Logistics has played an extremely important role in economic growth in China (Zhang & Peng, 2009;Peng, 2011;Tan, 2003).On the other hand, it is a relatively energy-intensity industry, such as haul trucks, shipping and aircraft, which seriously rely on fossil fuel, and they emit greenhouse gas.Logistics activities accounted for roughly 5.5% share of global GHG emissions, around 90% of which came from transport, and the rest come from warehouses, load and unload (McKinnon, 2012).
There are close correlations between logistics and transportation.However, logistics presents some obvious differences with transportation.Transportation can be known as an important element of logistics.The generalized transportation includes passenger transport and freight transport, but logistics doesn't cover passenger transport.as follows.Section 2 mainly introduces the decomposition method of carbon dioxide emission and the source of data.Section 3 describes the statistical analysis of carbon dioxide emission in China's logistics industry, and presents the results of carbon dioxide emission decomposition.Section 4 presents the conclusions and some suggestions for sustainable logistics.

Decomposition Methods
There are several methods for decomposition analysis of energy consumption and CO2 emission.Especially, two famous decomposition methods, i.e. structural decomposition analysis (SDA) and index decomposition analysis (IDA), have been widely used, such as Wachsmann, Wood, Lenzen and Schaeffer (2009), Chang, Lewis and Lin (2008), Lise (2006), Akbostanc, Ipek-Tunç andTürüt-Aşık (2011).SDA was based on the input-output model, given to an analytical framework by Leontief (1966).Rose and Casler (1996) reviewed the development of SDA and its relationships to other methodologies, and presented the fundamental principles of alternative approaches.IDA was first used to study the impact of changes in product mix on industrial energy demand (Ang, Zhang & Choi, 1998).Ang (2004) comprehensively compared the two popular index decomposition analysis methods, namely the Laspeyres index decompostition and the Divisia index decomposition, and recommended the log mean Divisia index (LMDI) method for general use due to its theoretical foundation, adaptability, ease of use and result interpretation, along with some other desirable properties in the context of decomposition analysis (Liu, Fan, Wu & Wei, 2007).So we use the LMDI method to analyze the changes in CO2 emission of China's logistics industry in this paper.
According to the LMDI method firstly introduced by Ang, Zhang and Choi (1998), and the practical guide presented by Ang (2005), the changes in CO2 emissions from industry may be studied by quantifying the contributions from five different factors: activity effect, structure effect, intensity effect, fuel-mix effect and emission-factor effect.About 90% CO2 emissions came from transport activity in logistics sector (McKinnon, 2012), so we further to divide the intensity effect into transport intensity effect and energy intensity effect.So we can research the changes in CO2 emission from logistics industry by quantifying the contributions from the following six different factors: logistics activity (measured by logistics added-value), transport intensity, transport mode shift, energy intensity, fuel mix and CO2 emission factors.The general index decomposition analysis (IDA) identity may be written as shown in Equation ( 1).
The significance of symbols can be seen in Table 1.

Cij
Carbon dioxide emission resulting from fuel j in ith transport mode The change of the aggregate carbon dioxide emission in logistics industry from the base year 0 to the target year t, denoted by Dtot or Etot, can be decomposed to six affect factors as (2) (3) According to the multiplicative and additive decomposition method recommended by Ang (2005), each term of Equations ( 2) and ( 3) can be calculated applying the LMDI method, which were shown in Appendix A-Table A1.

Management of Data
There are no compiled data of logistics industry in the related Chinese yearbooks, but we can find a statistical index in China Statistical Yearbooks and China Transport Yearbooks, which included the related statistical data about transport, storage and communications, we can regard the index as a proximate proxy of logistics industry (Liu & Li, 2007;Zhang & Yu, 2012).
In this paper, we use the "top-down" method to calculate CO2 emissions in logistics industry, which estimates CO2 emissions on basis of the total amount of fuel consumption.CO2 emissions of the ith transport mode in year t can be calculated by multiplying fuel consumption of one fuel type used in the ith transport mode and the CO2 emission factor of the fuel type j.
If different fuels are used in one transport mode, we should summarize the CO2 emission of different fuel types in the mode (Loo & Li, 2012).For example, the fuel types include coal, diesel and electricity in railway transport mode.The main energy fuels are gasoline and diesel in highway transport mode, and they are jet kerosene and aviation gasoline in the aviation transport.For water transport, the primary fuel is diesel.Carbon emission coefficients of fuels have changed with an updating in level of fuels.Because of the relatively short study period, these changes of the emission coefficients can be ignored when we analyze the macro changes in CO2 emission, we assume that the CO2 coefficients of coal, kerosene, gasoline and diesel are constant.However, CO2 coefficient of electricity is continuously changing, due to the fuel mix used and technological improvements in the generation of electricity.We use the net standard coal consumption rate (g/kw.h) to describe the energy consumptions which were used to generate one kilowatt hour.We can obtain the data from China Energy Statistical Yearbook.So, the CO2 emission coefficient is restricted to the impact of changes in electricity emission rate, which is calculated on basis of the individual fuels used in power generation (Liu, Fan, Wu & Wei, 2007;Wang, Zhang & Zhou, 2011).From Figure 3, we can conclude that the highway transport was a main contributor to CO2 emission, it accounted for 76.8% (or 778.47 million tons CO2) share of total CO2 emission in 2010.The reasons were that the demand for highway transport was increasing in recent years due to its convenience and flexibility, and the highway transport mainly relied on the fossil energy.Next to highway transport, waterway transport ranked the second, which accounted for 18.1% (or 183.37million tons CO2) of total CO2 emission in 2010, and it is noticed that the statistical data of waterway transport included ocean transport in China Statistical Yearbook.

Total CO2 Emission in Chinese Logistics Industry
Railway transport ranked the third with reduction from 32.3% share in 1980 to 2.53% in 2010.
The main reason is the replacement of the coal-fired steam locomotives with diesel/electric locomotives (Wang, Zhang & Zhou, 2011).The freight aviation transport had small proportion and ranked the fourth, which only accounted for 1.69% share of total CO2 mode, for example, the Just-In-Time (JIT) requests to reduce the inventory as far as possible, which made promptness of delivery become an important factor.In order to satisfy customer's requirements, more and more suppliers choose the aviation transport due to its quick speed.

Intensity of CO2 Emission in Logistics Industry
The CO2 emission intensity of logistics can be defined as the CO2 emission per unit of logistics output, which can be measured by the added value of logistics and the turnover volume of

Decomposition Results of CO2 Emission from Chinese Logistics Industry
In this paper, we used logarithmic mean Divisia index (LMDI) method to explore the multiplicative and additive decomposition of CO2 emission in China's logistics industry.The decomposition results are based on Equations ( 2) and ( 3) and the LMDI formula in Appendix A-Table A1, as shown in Table 3 and Table 4. Figures 6-7 showed the radar charts for multiplicative decomposition and the bar charts for additive decomposition of CO2 emission in logistics industry by using the numerical results presented in Tables 3-4, respectively.However, the energy intensity and the factors that influenced CO2 emission contributed to the reduction of CO2 emission in overall research period, which factors had multiplicative indexes less than one and additive indexes below zero, as shown in Figure 6.

Change of Carbon Dioxide Emission in Different Periods
We divided the study period into multi-stages in order to examine the factors that caused the change in CO2 emission of logistics well.Chinese administration departments take five years as a plan cycle when they make macro policies.We suppose that there are no obvious differences in two continuous five-year plans.So we divide the research period into three sub-stages, i.e. 1980-1990, 1990-2000 and 2000-2010  The change in CO2 emission in different sub-stages (in turn, 1980-1990, 1990-2000, 2000-2010, accounted for 6.3%, 8.3%, 85.4% share of the total CO2 emission change, respectively, with the increasing of 61.27, 80.35, 827.43 million tons carbon emission in the study period.
The CO2emission of China's logistic industry had greatly increased in the period of 2000-2010.
Especially, in the stage of 2005-2010, it accounted for 73% share of total CO2 emission in research period, with the increasing of 707.6 million tons carbon emission.Global economy was in recession in that time, Chinese government endeavored to pull domestic demand by investing plenty of money in public service and infrastructure to deal with world financial crisis in 2008.China's GDP ranked the second place, which is next to America and exceeded Japan in the world in 2010.As previously mentioned, economic growth would make the demand for logistics service increase, and it enhanced the energy consumption and CO2 emission.
From Figures 8-9, we can conclude that the changes in CO2 emission and the influence factors have extreme similarities between the stages of 1980-1990 and 1990-2000.For example, logistics activity was the leading factor that caused the increase of CO2 emission, with 85.5% and 114.4% share of the total change in the two periods, respectively; the energy intensity and the emission factor had positive roles in reducing the CO2 emission due to the updating and innovation of logistics technologies and logistics equipment.

Conclusions
Carbon dioxide emission has a tight link with economic growth and energy consumption.As an emerging and important industry in China, the logistics industry is promoting economic growth; on the other hand, it consumed a great deal of energy and emitted plenty of CO2, and it was the main reason for global warming.In order to examine the change in CO2 emission of China's logistics industry, we firstly calculated the carbon dioxide emission of logistics industry in the period of 1980-2010; and further analyzed the factors that influenced the changes in CO2 emission of logistics industry by using logarithmic mean Divisia index (LMDI) method.We can draw some conclusions from the present study as follows: 1.

Policy Implications
Facing the rapid growth in demand for logistics services due to economic growth, it's difficult to decrease total CO2 emission in logistics industry, because there are tight relations between economic growth, logistics demand, energy growth and carbon emission.However, we can take some strategies to improve energy efficiency and reduce CO2 emission per unit of energy consumption, which can offset the increase of total CO2emission, such as: (1) Optimize social logistics system and update logistics facility and equipment.Logistics is an extremely complex economy activity.It is significant to optimize the social logistics system for energy conservation and emission reduction.Nevertheless, logistics has been paid attention in recent years, and the current logistics system is unreasonable in China, such as unscientific layout of logistics network, underdeveloped traffic condition, outdated warehouse, and so on.The traffic congestion and reduce energy consumption and emission (Liao et al., 2011).ICT was generally utilized in truck transport in developed countries; however, few operators use it in China.So it is necessary to introduce advanced information communication technology.(4) Increase investment on designing high fuel efficiency and low emission engine.Efficient engine design and fuel economy standard play a great role in cutting fuel demand, which can improve trucks fuel economy (Schipper, 2009).However, because of the limited technology resource, China's transport carrier emission standard is still lower than international emission standard.
shows the top five CO2 emission countries in the world.China has exceed America and become the first source of CO2 emission since 2007, with 8320.96 million tons carbon (MTC) in 2010, which accounted for 26.2% share of the total CO2 emission in the world (2014).As a signatory of the United Nations Framework Convention on Climate Change (UNFCCC), China approved the Kyoto Protocol in 2002, and promised that carbon emissions per unit of GDP would be reduced by 40%-45% in 2020 than 2005 in Copenhagen world climate meeting in 2009.It is greatly significant for policy makers to analyze the change in dioxide carbon emission and find the critical factors to achieve the goal of emission reduction.

Figure 1 .
Figure 1.Top five of total CO2 emission in the world (U.S. Energy Information Administration, EIA, 2012) researched on carbon emissions from UK manufacturing between 1990 and 2007.Sheinbaum, Ozawa and Castillo (2010) analyzed energy and CO2 emission trends of Mexico's iron and steel industry during the period 1970-2006 using Log mean Divisia index; Schmitz, Kamiński, Scalet and Soria (2011) represented a detailed analysis of CO2 emissions and energy consumption of European glass industry.Xu, Tobias and Eichhammer (2012) analyzed the change of energy consumption and CO2 emissions in China's cement industry and its driving factors over the period 1990-2009.
. The paper mainly contributes to reflect changes in carbon dioxide emission of China's logistics from a more extensive perspective, on the basis of transportation, storage, distribution, packaging, et al.It is different with the previous researches which were just based on transportation data, such as passenger transportation, or freight transportation, or the sum of the two.The main purpose of this paper is to: (1) estimate the CO2 emission in China's logistics sector; (2) analyze the causes for the change of CO2 emission; (3) identify the critical factors which mainly drive the change in CO2 emissions of logistics sector in China.The remainder of the paper is organized

C
Engineering and Management -http://dx.doi.org/10.3926/jiem.Total carbon dioxide emission in study period Y Logistics activity, measured by logistics added-value Q Turnover volume of freight transport Qi Freight turnover volume of the ith transport mode Eij Consumption of fuel j in ith transport mode follows: logistics activity effect (denoted by Dact or Eact), transport intensity effect (denoted by Dintt or Eintt), transport mode shift effect (denoted by Dstr or Estr), energy intensity effect (denoted by Dinte or Einte), fuel mix effect (denoted by Dmix or Emix), CO2 emission factors effect (denoted by Demf or Eemf), as shown in the form of Equation (2) and the additive form of Equation (3).

Figure 2
Figure 2 presents the trend of total CO2 emission in China's logistics industry from 1980 to 2010.It's obvious that the CO2 emission has a continuously raising trend in the study period.It has increased by 21.5 times from 45.1 million tons CO2 in 1980 to 1014.1 million tons CO2 in 2010, with the average growth rates was 11.7%, which accounted for 12.2% share of total CO2

Figure 2 .
Figure 2. Total CO2 emission in Chinese logistics industry

Figure 3 .
Figure 3. Percentage of CO2 emission in different transport modes freight transport.So we analyze the CO2 emission from two aspects, i.e.CO2 emission per added value of logistics (million tons CO2 per billion Yuan) and CO2 emission per turnover volume of each transport mode (million tons CO2 per billion km-tons).As shown in Figure 4, it presents the changing trend of CO2 emission of logistics in China.In general, the CO2 emission of logistics tends to flat in the period of 1980-2006.However, there were dramatic changes in intensity of CO2 emission between 2006 and 2010, with increasing from 3.06 million tons CO2 per billion Yuan in 2006 to 8.74 million tons CO2 per billion Yuan in 2010 (constant price in 1980).There are some reasons for the obvious changes in intensity of CO2 emission, for example, Chinese government invested vast funds in infrastructure for economic growth from 2006 to 2010, which boosted the increase of logistics demand rapidly, and it exceeded the technological updating speed for energy conservation and emission reduction.On the other hand, the transport mode obviously changed from lower energy consumption mode to higher energy consumption mode in 2006-2010.The turnover volume ratio of railways to total turnover volume of freight transport decreased from 24.71% in 2006 to 19.49% in 2010; the ratio of highway transport increased from 10.98% in 2006 to 30.59%, especially in 2008, the turnover volume of highway transport increased by 1.89 times than last year, from 1135.5 billion ton-km to 3286.8 billion ton-km.So the economic growth and energy mix are important factors for the changing in intensity of CO2 emission.

Figure 5 .
Figure 5. CO2 emission intensity in different transport modes

Figure 6 .
Figure 6.Radar chart for CO2 emission decomposition in 1980-2010 . Additionally, taking account of rapid economic growth and policy adjustment in recent ten years, we further divided the period 2000-2010 into 2000-2005 and 2005-2010.Figures 8 -9 showed the radar charts for multiplicative decomposition and the bar charts for additive decomposition of CO2emission in logistics industry in different sub-stages, respectively.

Figure 8 .
Figure 8. Radar chart for CO2 emission decomposition in multi-stages However, the critical factors to changing in carbon emission were transformed in recent ten years.For example, logistics activity, transport intensity and fuel mix had adverse effect on the change in carbon emission from 2000 to 2005; herein, transport intensity was the first primary contributor to the increase of carbon emission.Transport structure, energy intensity and emission factor had positive role in decreasing carbon emission.From 2005 to 2010, all decomposition factors leaded to the increase of CO2 emission, except for emission factor.Transport structure played the most important role in increasing CO2 emission with 358 million tons CO2 emission, which accounted for 50.6% share of total CO2 emission.The effect resulted from the change of transport model from lower energy consumption model such as railway transport to higher energy consumption model such as highway and aviation transport.Transport mode shift mainly depended on customer's demand for transportation service, such as cost, security, convenience, promptness, flexibility, and so on.
administration departments should further to optimize social logistics system by relocating and constructing gather and distribution centers, junction stations, and building modern warehouse for improving logistics operation efficiency, reducing energy consumption and emission.(2) Encourage transport mode to shift from high emission to low carbon mode.The research result showed that the transport mode has critical effect on the change of CO2 emission.Railway transport and waterway transport have low energy intensity and strong transport capacity; however, due to the flexibility and convenience of highway transport and aviation transport's promptness, the share of freight turnover volumes present an increasing trend in recent years in China.Policy makers can encourage people to choose lower energy consumption transport by making transport pricing policy.(3) Introduce advanced information communication technology.Information communication technology (ICT) helps operators to make efficient route plan and schedule, vehicle loading, driving time and travel distance, and it can avoid Scientific and technological administrations should increase investment and organize R&D team to design high fuel efficiency and low emission engine.(5) Improve electricity generation and fuel refining technology.The development of energy technology has a little positive effect in decreasing the energy consumption and carbon emission in recent years in China, but it's very limited.Energy and environment administration division should make effort to improve the electricity generation and fuel refining technology for reducing the carbon coefficients.For example, the power stations can adopt new energy technologies, such as nuclear power, wind power, solar power and biomass power to generate electricity.In addition, high quality fuel can release more energy and less emission, and the petroleum companies should explore and introduce new fuel refining technology to improve the purity of fuel oil.

Table 2 .
The energy consumption data of different freight transport modes can be obtained and calculated from China Statistical Yearbook, Yearbook of China Transportation and Communication, Compile of China Aviation Statistics in different stages.Emission factor data came from IPCC Guidelines for National Greenhouse Gas Inventories (2006), as shown in Table 2. CO2 emission factor by type of transport fuel Journal of Industrial Engineering and Management -http://dx.doi.org/10.3926/jiem.1443 emission, however, its growth rate was the most obvious, it increased by 89.4 times, from 0.19 million tons CO 2 in 1980 to 17.12 million tons CO2 in 2010.It may be explained by the changing of production -847-

Table 3 .
Multiplication decomposition of CO2 emission in logistics industry

Table 4 .
Additive decomposition of CO2 emission in logistics industry

3.3.1. Total Change in Carbon Dioxide Emission Between 1980 and 2010
Figures 6-7 describe the results of the multiplicative and additive decomposition analyses of logistics industry in China.According to the analysis, we can conclude that the total CO2 emission had changed greatly, with the accumulated increasing of 969.1 million tons CO2 emission from 1980 to 2010, and logistics activity, transport intensity, transport mode shift and fuel mix are important factors in increasing the CO2 emission of logistics sector in China.