The efficiency improvement potential for coal, oil and electricity in ...

Energy xxx (2015) 1e11

Contents lists available at ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

The efficiency improvement potential for coal, oil and electricity in China's manufacturing sectors Ke Li a, c, Boqiang Lin a, b, * a

Collaborative Innovation Center for Energy Economics and Energy Policy, China Institute for Studies in Energy Policy, Xiamen University, Fujian 361005, China b Newhuadu Business School, Minjiang University, Fuzhou 350108, China c College of Mathematics & Computer, Hunan Normal University, Changsha 410081, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 January 2015 Received in revised form 15 March 2015 Accepted 11 April 2015 Available online xxx

This paper introduces an improved total-factor ESTR (energy-saving target ratio) index, which combines the sequence technique and the “energy direction” to a DEA (data envelopment analysis) model, in order to measure the possible energy saving potential of a manufacturing sector. Afterward, the energy saving potentials of four different energy carriers, namely coal, gasoline, diesel oil and electricity, for 27 manufacturing sectors during the period of 1998e2011 in China are calculated. The results and its policy implications are as follows: (1) the average ESTRs of coal, gasoline, diesel oil and electricity are 1.714%, 49.939%, 24.465% and 3.487% respectively. Hence, energy saving of manufacturing sectors should put more emphasis on gasoline and diesel oil. (2) The key sectors for gasoline saving is the energy-intensive sectors, while the key sectors for diesel oil saving is the equipment manufacturing sectors. (3) The manufacture of raw chemical materials and chemical products sector not only consumes a large amount of oil, but also has a low efficiency of oil usage. Therefore, it is the key sector for oil saving. (4) Manufacture of tobacco and manufacture of communication equipment, computers and other electronic equipment are the benchmark for the four major energy carriers of energy-saving ratios. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Energy saving potential Total-factor effects Data envelopment analysis Manufacturing sectors China

1. Introduction Along with China's economic growth, its energy consumption also increases rapidly. In 2011, China consumed about 348 million ton of coal equivalent (tce), which is about 2.56 times than that in 1998, with an average annual growth of 7.56%. Nowadays, China faces challenges to cope with increasing energy demand, supply restraints, huge environmental costs and backward technology. Hence, the government argues that it is urgent for China to take steps to improve energy efficiency and control the growth rate of energy consumption, thereby creating a resource-conservative and environment-friendly society. It can be said that the core of China's energy strategy is to safeguard national energy security by fully implementing energy-saving policies. At present, China has entered an accelerated period of the industrialization and urbanization. One of the features of this

* Corresponding author. Newhuadu Business School, Minjiang University, Fuzhou, Fujian, 350108, China. Tel.: þ86 5922186076; fax: þ86 5922186075 E-mail addresses: [email protected] (K. Li), [email protected], bqlin2004@ vip.sina.com (B. Lin).

period is that the energy-intensive manufacturing sectors, such as petrochemical, iron and steel, are playing key roles in economic growth, thus energy demand is increasing quickly. In 2011, the number of manufacturing enterprises accounts for 92.59% of the total industry enterprises, and its assets is about 75.96% of the total industry assets. The gross industrial output values, taxes and other charges on principal business and the employees of manufacturing sectors account for 86.94%, 82.85% and 87.86% of the whole industrial sectors1 respectively. At the same year (2011), the industry consumed about 70.82% of whole energy usage; while the manufacturing sectors consumed about 81.32% of industrial energy consumption. Therefore, the manufacturing sectors are the main energy consumers, and they are also the key sectors for implementing energy saving policies. Fig. 1 shows manufacturing sectors consume about 55%e60% of the total energy consumption. In terms of energy carriers, manufacturing sectors consume about 38% and 52% of the total coal and electricity consumption respectively. But their proportions 1 It includes (i) mining and quarrying sectors, (ii) manufacturing sectors, and (iii) electric power, gas and water production and supply sectors.

http://dx.doi.org/10.1016/j.energy.2015.04.013 0360-5442/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Li K, Lin B, The efficiency improvement potential for coal, oil and electricity in China's manufacturing sectors, Energy (2015), http://dx.doi.org/10.1016/j.energy.2015.04.013

2

K. Li, B. Lin / Energy xxx (2015) 1e11

70 60

%

50 40 30 20 10 0 1998

1999

2000

2001

energy

2002

2003

coal

2004

2005

gasoline

2006

2007

diesel oil

2008

2009

2010

2011

electricity

Fig. 1. The proportion of energy usage of manufacturing sectors in China's total energy usage (1998e2011).

100 90

80 70

electricity diesel oil gasoline coal

%

60 50 40 30 20

Measuring…

Electrical Eq.

Computers etc.

Special Mac.

Transport Eq.

General Mac.

Non-Ferrous…

Metal Products

Ferrous Press

Nonmetal Ma.

Fibers

Rubber and…

Medicine

Chemical

Cultural…

Paper

Printing

Furniture

Leather

Wood Proc.

Textile

Apparel

Tobacco

Beverage

Food Ma.

Food Proc.

0

Petroleum Pro.

10

Fig. 2. The average structural of four main energy carriers of China's manufacturing sectors during the period of 1998e2011. Note: The consumption of each energy carrier is transformed to 104 tce.

show downward and upward trends respectively. Gasoline and diesel oil consumption in manufacturing sectors account for a small proportion of the total consumption, and both show downward trends. Fig. 2 shows the average structural of four main energy carriers (coal, gasoline, diesel oil and electricity) of China's manufacturing sectors during the period of 1998e2011. According to it, there are 15 sectors' energy consumption are dominated by coal (it account for more than 50% of four main energy carriers), with the highest is petroleum pro.2(95.29%). In addition, four sectors, including cultural articles, metal products, computers etc. and measuring inst., are dominated by electricity. For energy-saving policy, the Chinese government has put the declining targets of energy intensity (it is defined as energy consumption per unit of Gross Domestic Product) in the national economic development plan, and has adopted the TRS (target responsibility system) for provinces, industries and even energyintensive units since 2006. Energy intensity has become a core index for China's energy saving policy, and the government expects sustainable development by keeping it at a downward trend. As a single factor index, changes in the national energy intensity can be attributed to changes in economic structural and energy

2 We use the abbreviated name in italic type to make the text clear and concise. The full name and its abbreviation of each manufacturing sector are provided in Appendix.

intensities of sectors [1]. Energy substitution will also lead to energy intensity changes. Accordingly, in the context of measures designing, the Chinese government pays more attention to technological progress to bring down the energy intensities of energyintensive sectors/products. Furthermore, the government also accelerates the adjustment and upgrades of the industrial structure in order to improve energy efficiency. It seems that the Chinese government faces two contradictory targets, namely to maintain economic growth and reduce the growth rate of energy consumption, or even further, to control total energy consumption. In this sense, optimizing the energy consumption structure is conducive to maintaining economic growth under the constraint of total energy consumption control. More importantly, the environmental deterioration is directly related with the massive usage of coal. Thus, it is not enough to only analyze the changes in energy intensity. According to Figs. 1 and 2, the Chinese government should pay more attention to energy-saving of the manufacturing sectors, but the sectors have different energy consumption structure. Hence, it is important and meaningful to investigate the energy-saving potential of different energy carriers across sectors. The remainder of this paper is organized as follows. Section 2 is literature review and the main contribution of the current study. Section 3 introduces an improved DEA (data envelopment analysis) model to compute the energy-saving targets. Section 4 cites the empirical results, and section 5 is discussion. The final section concludes the research findings and presents policy implications.



2. Literature review The energy intensity index contains many structural factors, such as industrial structural and energy mix. Changes in energy intensity usually represent the results of economic structural changes, and do not reflect the substitution among various inputs, thereby possibly exaggerating the energy efficiency [2]. More importantly, the energy intensity index does not provide any information about the realizable improvement gap in energy efficiency [3], which can be overcome by using total-factor effects in energy efficiency measurement [4]. Early in 2000, Boyd and Pang [5] indicated that energy intensity declining (or energy efficiency improvement) depended on total factor productivity improvement. Hu and Wang [6] believed that the economic outputs were the results of all inputs (including energy), and energy intensity essentially reflected the outcome of changes in total factor productivity. Based on the total factor analysis framework, they firstly introduced the TFEE (total factor energy efficiency) index. Compared with energy intensity, TFEE reflects not only the substitution effects among factors, but also the comprehensive level of energy use under a given structure of production factors, and it also measures the energy efficiency improvement space. In recent years, the TFEE index has been widely accepted and is a popular method used to measure energy efficiency [7]. According to Hu and Wang [6], the TFEE is constructed by taking the ratio of the actual energy input to target energy input which is conducted through a DEA (data envelopment analysis) model. In other words, a DEA model constructs the efficient frontier of energy consumption, and the distance from the inefficient one to the corresponding frontier one is the value of TFEE. If the value of the TFEE is 1(unity), it means the DMU (decision-making unit) operates at the efficiency frontier of energy consumption; otherwise, it means that the DMU's energy efficiency has improvement potential. Hence, TFEE measures the relative value of energy efficiency. In essence, it presents the ratio of actual energy usage to target energy usage under a given technical level, thus it is closely related to energy-saving potential, whereas this information can not be given by the energy intensity index [3]. In addition, the DEA method, compared with SFA (stochastic frontier analysis), does not need to set a functional form between inputs and output, nor any price information [8]. It only needs the inputs and outputs to measure the energy efficiency, which gives DEA a significant advantage to calculate TFEE [9]. There are many literature that adopt the DEA method to evaluate the energy efficiency of China from various perspectives. By using a cross country panel data, Hu and Kao [10] adopted TFEE to analyze the ESTR (energy-saving target ratios) for 17 APEC (AsiaPacific Economic Cooperation) economies during 1991e2000, and found that China had the largest ESTR, it could save almost half its current usage. Zhang et al. [11] applied the DEA window to evaluate the TFEE of 23 developing countries in 1980e2005, and found China had the most rapid improvement. Wei et al. [12] found China's average TFEE ranked 147 among 156 countries. Some literature also estimated the TFEE across Chinese provinces. The pioneer study is Hu and Wang [6]. They analyzed the TFEE of China's 29 provinces in 1995e2002, and found that China could improve its energy efficiency without the cost of economic growth. Specially, central China had the worst energy efficiency and had the greatest potential in energy savings. Wei and Shen [13] analyzed the TFEE of China's 29 provinces over the period of 1995e2004, and found that most provinces' TFEE show “first increasing, and then declining”, and the turning point generally appeared around the year 2000. Wang et al. [14] measured energy efficiency of China's provinces by using the DEA window analysis,

3

and found the eastern China has the highest energy efficiency, while western China was the lowest. Wang et al. [15] utilized the global DEA method to explore China's energy efficiency, and found there was an overall declining trend for China's energy efficiency from 2001 to 2005. Taking the desirable GDP (Gross Domestic Product) and the undesirable CO2 and SO2 as outputs, Li and Hu [16] calculated the ETFEE (ecological total-factor energy efficiency) of China during the period of 2005e2009. Bian et al. [17] estimated the potential energy savings and potential CO2 emission reduction for 31 provinces by a non-radial DEA model. Zhang et al. [18] conducted an empirical analysis of regional ecological TFEE by incorporating CO2, SO2 emissions and the COD (chemical oxygen demand) of China during 2001e2010. Their results indicated that most provinces were not performing at high ecological energy efficiencies. Some researches analysis the energy efficiency of China's industrial sectors. Wang et al. [19] evaluated the TFEE of industrial sectors of 30 provinces in China. Pan et al. [20] applied the DEA method to analyze China's provincial industrial TFEE, and investigated its determinants using Tobit regressions. Zhao et al. [21] found the industrial sectors of the eastern provinces had higher TFEEs than that in other provinces. Wei et al. [22] investigated the energy efficiency of China's iron and steel sector in 1994e2003 based on DEA-Malmquist Index Decomposition. Lin et al. [23] evaluated the potential future energy efficiency gap of China's steel industry. He et al. [24] adopted DEA-Malmquist to evaluate the energy efficiency of China's iron and steel enterprises, and found the average efficiency was only 61.1%. Xue et al. [25] used an input-oriented DEA model to measure the energy efficiency of the construction industry from 2004 to 2009 with data in 26 provinces of China. Their results indicated only Guangdong province experienced effective energy consumption during 2004e2009. Zhou et al. [26] adopted an environmental DEA technology to measure the energy efficiency performance of China's transport sector. Although there are many studies calculate the energy efficiency by adopting the TFEE, few literature clearly analyze the improvement potential of energy-saving by the result of TFEE. Chang [3] established an indicator of the improvement potential in energy efficiency by measuring the difference between the target level of energy intensity that is suggested from a DEA model and the actual energy intensity, and then investigated the improvement potential for 27 EU (European Union) members. Further, Li and Lin [4] adopted an improved DEA model to re-estimates the TFEE, and then used it to calculated the improvement potential in energy intensity for China's provinces under the metafrontier and groupfrontier respectively. Lee et al. [27] calculated the efficiency of electricity, coal, and gasoline oil savings for 27 provinces in China during the period of 2000e2003. To our knowledge, the research of Lee et al. [27] is the first study about efficiency improvement potential that also simultaneously incorporates various carriers of energy, which inspires the current study. The contribution of this paper is twofold: (1) we calculate the DEA model based on the “energy direction” to measure the TFEE (total factor energy efficiency). For the calculation method of the DEA model, the sequence technique is adopted. Thus we introduce an improved ESTR (energy-saving target ratio) index, which was proposed by Lee et al. [27] and Chang [3]. Since the sequence technique avoids the “technical regress” over periods in efficiency evaluation, we believe the ESTR can be compared over years. (2) This paper takes various carriers of energy usage (coal consumption, gasoline consumption, diesel oil consumption and electricity consumption) into consideration, and we estimate the ESTR for different energy carriers in China's manufacturing sectors. Therefore, we can discuss the main energy carrier and the main manufacturing sector of the implementation of energy saving



energy inpust/output

4

A B

A' C

A'' B' D B''

O

output=1

other inputs/output Fig. 3. DEA representation of energy efficiency.

policies. These are greatly significant for policy making with respect to energy consumption control planning, as well as energy conservation and emissions abatement targets.

3. Method and data 3.1. Method How TFEE measures the energy saving potential is illustrated by Fig. 3. The efficient frontier under the input-oriented is line output ¼ 1, point C and point D are on the frontier and their values of TFEE are 1 (unity), or their energy usages are efficient, thereby their ESTRs are zero. Point B is the actual input set, while point B0 is the projected point on the frontier, so point B can improve its efficiency by reducing the radial adjustment BB0 . While for point A, its energy-saving target is the summation amount of AA0 and A0 C. AA0 is the radial adjustment, while A0 C is the slack adjustment. However, the above results are the minimum of all inputs under a given output. In other words, it can be named as the total factor energy efficiency, or total factor capital efficiency, etc. To be more precise, we believe the calculate results of the above method are the total factor productivity. So this TFEE under the traditional DEA method has been criticized. Stern [28] introduces the “energy direction” to measure TFEE, which is simply the ratio of energy use to the minimum technically feasible energy input, ceteris paribus (or assuming other inputs are efficiency). In Fig. 3, the energy-saving target of point A is AA00 , while point B is BB00 , ceteris paribus. Using a DEA method to model the above analysis. Suppose with the K DMUs (decision-making units, manufacturing sectors in this study), each uses m kinds of non-energy inputs x ¼ ðx1 ; x2 ; /; xm Þ2ℝþ and h kinds of energy inputs m e ¼ ðe1 ; e2 ; /; eh Þ2ℝþ to produce s kinds of outputs h þ y ¼ ðy1 ; y2 ; /; ys Þ2ℝs . The envelopment of the kth DMU can be derived from the following linear programming problem:

Minq;l q s:t: yk þ Yl 0; xk Xl 0; qek El 0; l0

(1)

(1) calculates the score of energy use efficiency for the kth DMU. According to Farrell [29], the value of unity means the DMU locates on the efficient frontier, or achieving the minimum energy inputs by given outputs and non-energy inputs. Additionally, Eq. (1) may observe “technical regress”, which means the best practice frontier under the DEA model may move backwards over periods. The reason is that the year-specific frontier is calculated by outputs and inputs data for each year in the conventional DEA model. Thus, the technology of previous periods may become unfeasible in the present or in subsequent periods. This contradicts the logical thinking of the notion that “technology has been mastered will not forget”, or technology remains unchanged to a certain extent because technological progress is closely associated with the accumulation of knowledge and experience [30]. This implies the energy efficiency calculated by Eq. (1) may not be comparable over periods. To overcome the “technical regress”, the previous literature adopt three different methods to improve the DEA model, namely global technology [31], the windows analysis [14,32] and the sequential technique [30]. The first one uses all observations to construct a single efficient frontier, and then uses it to evaluate contemporaneous DMUs; the windows analysis constructs frontier by using observations of three or more consecutive years (determined by the width of window). However, the frontier constructed by these two methods may still be infeasible for the earlier observations, so it cannot completely avoid “technical regress”. Hence, this study adopts the sequential technique, which takes all current and past observations to construct the frontier, and uses it to evaluate the current DMUs [4,33]. Therefore, the envelopment of the kth DMU at time t can be calculated by linear programming problem as follow:

Minq;l q s:t:

ytk þ

t X

Y t lt 0;

t¼1

xtk

t X

X t lt 0;

t¼1 t X

qetk

(2)

Et lt 0;

t¼1 t

l 0

Based on TFEE, Hu and Wang [6], Hu and Kao [10], Lee et al. [27] and Chang [3] introduced a total-factor ESTR (energy-saving target ratio) index to measure how large potential of a DMU's energy saving,

ESTR ¼

EST AEC TEC ¼ AEC AEC

(3)

where EST is the energy-saving target, which is calculated by the difference between AEC (actual energy use) and the TEC (target energy use). While TEC is calculated by a DEA model,

TEC ¼ AEC q

(4)

Using Eqs. (2)e(4), ESTR is (1 q). Obviously, the higher ESTR is, the lower the energy efficiency (TFEE) will be. If energy use is efficient, then q is unity, and ESTR is zero. 3.2. Data

where q is a scalar representing the efficiency score for the kth DMU; l is an K 1 vector of constants, and is larger than zero meaning the efficient frontier is convexity; Y, X, and E represent an (s K) matrix of the outputs, a (m K) matrix of the non-energy inputs and an (h K) matrix of energy inputs, respectively. Eq.

This study analyses the energy efficiency and energy-saving potential of 27 two-digit manufacturing sectors during the period of 1998e2011. We excluded the sectors “artwork and other manufacturing” and “recycling and disposal of waste”, because the



observations of some variables are missing. Also, this paper combines the sectors “rubber” and “plastics” as “rubber and plastic” because of the changes in the statistical range. Therefore, the specific sectors can be seen in Table 1. For the statistic scope perspective, the Chinese government reported the data of industrial enterprises on the basic of at- and above-township level before 1998, however since 1998 it on the basis of state-owned and non-state-owned industrial enterprises above designated size (the annual revenue from principal business over 5 million Yuan). In consideration of the data available and the consistency of the statistic scope, the current study takes the study period as the 1998e2011.3 In this paper, we use six inputs and one output to calculate the TFEE and the energy-saving potential. These inputs are capital stock, number of employed persons, and coal consumption, gasoline oil consumption, diesel oil consumption and electricity consumption.4 The only one output is the gross industrial output. All variables are explained as below. (1) Capital stock. It is estimated by using the perpetual inventory approach because it can not be obtained directly, namely

Kit ¼ Iit þ ð1 dit Þ Ki;t1 ;

5

Tables 1 and 2 are summary statistics of inputs and outputs. The largest value of gross industrial output is observed by computers etc. in 2011; while the smallest value is observed by furniture in 1998. Petroleum pro. in 2011, nonmetal ma. in the years of 2002 and 2001, and ferrous press in 2011 are observed the largest consumption of coal, gasoline, diesel oil and electricity respectively. In order to analysis the energy consumption features of different manufacturing sectors, the 27 two-digit manufacturing sectors are integrated into three categories based on the classification of Chenery et al. [36] and Wang and Chen [37]. The third category dominated by equipment manufacturing has the largest output, but its energy consumption is far less than that of the second category. The coal consumption of the third category is even less than the first category, which mainly consists of light industry sectors. The second category, which is comprised of many energy-intensive sectors, such as petroleum, raw chemical material, medicine, chemical fiber, rubber, plastic, non-metal, and metal manufacturing sectors, is the main energy usage body. Its average coal consumption is about 10.22 and 21.97 times than that of the first and the third category respectively. 4. Results

(5) 4.1. Saving ratios for coal consumption

where I is new investment values at the price level in 2000 (measured in 100 million Yuan), and d is the depreciation rate. The relative data in Eq. (5) can be calculated using the equations as follow [34,35], depreciation values(t) ¼ accumulated depreciation (t)- accumulated depreciation (t-1) depreciation rate(t) ¼ depreciation values(t)/original value of fixed assets(t-1) new investment values at the current price (t) ¼ original value of fixed assets(t)- original value of fixed assets(t-1) new investment values at the price level in 2000 (t) ¼ new investment values at the current price (t)/investment in fixed assets price index (2000 ¼ 1) t and (t-1) stand for the current year and the previous year. In addition, the average net values of fixed assets at the current price in the year of 1998 are transferred to the constant price at the level in 2000, and they are regarded as the original capital stock in Eq. (5). So far, the data of capital stock at constant 2000 prices can be calculated by Eq. (5). The above data is collected from the China Industry Economy Statistical Yearbook. (2) Labor. It refers to the annual number of employed persons (measured in 10,000 person), and collected from the China Statistical Yearbook. (3) Energy use. It is given by coal consumption, gasoline oil consumption, diesel oil consumption and electricity consumption of manufacturing sectors, and the data is from China Energy Statistical Yearbook. The former three is measured by 104 ton, while the electricity consumption is measured by 100 million kwh. (4) Gross industrial output. It is measured in 100 million Yuan and at 2000 prices, and the data is collected from China Statistical Yearbooks.

3 China conducted a general economic survey for the year of 2012, whereas the data has not been published. 4 In order to analyze the energy-saving targets in-depth, we choose four different energy carriers, namely coal, gasoline oil, diesel oil and electricity. The reason for choosing these four carriers is because each sector consumes a considerable amount of them. While for other energy carriers, for example coke and crude oil, some sectors have no consumption data in China Energy Statistical Yearbook.

Fig. 4 depicts the target savings and its ratios of coal consumption during 1998e2011 for manufacturing industries in China. Overall, coal consumption of manufacturing industries is much efficient. During the period of 1998e2011, the average coal saving target ratio (ESTR_COAL) of 27 manufacturing sectors is 1.714%. If all sectors achieve the best efficient coal usages, it could save approximately 42.783 million tons of coal. In addition, ESTR_COAL is fluctuation at a low level during 1998e2002, and then jumps to 1.717% in 2003. In 2004e2010, ESTR_COAL increases rapidly from 1.561% to 3.055%, which indicates that the coal efficiency is in a downward trend during this period, especially in 2008e2010. In 2011, ESTR_COAL declines to 2.060%. From the manufacturing category perspective, the third category has the highest ESTR_COAL, and its average value is 3.040%, followed by the first category and the second category (1.445% and 1.188% respectively). However, due to the huge consumption of the second category, its accumulated amount of coal-savings under the efficient usage is 21.166 million tons, which is much larger than the first and the third category (10.591 million tons and 11.026 million tons respectively). From the sectors perspective, there are eight sectors that experienced higher ESTR_COAL than the average ESTR_COAL of 27 manufacturing sectors (1.714%), while the other 19 sectors witnessed the opposite results (Table 3). Two sectors, namely tobacco and computers etc. are observed the efficient coal usage, and their average ESTR_COALs are close to zero. According to Table 3, seven sectors, namely food proc., textile, chemical, mrubber and plastics, ferrous press, general mac. and transport eq. are observed large coal-saving targets, each accounts for more than 5% of the total coal-saving targets or is larger than 2.139(¼42.783*5%) million tons. In addition, the printing sector has the largest average value of ESTR_COAL (6.472%), but its coal-saving target amount is 459,230 tons5 because of its small coal consumption (during 1998e2011, it cumulatively consumed 7,435,800 tons of coal). While for sector chemical, although its average ESTR_COAL is as small as 0.339%, it

5 It is calculated by the sum of each year's energy savings, which is the result of multiplying each year's coal saving ratios by its coal consumption quantity.


6


Table 1 Summary statistics of inputs and outputs by sectors. Sectors

Typea

Index

Gross industrial output

Capital stock

Labor

Coal

Gasoline

Diesel oil

Electricity

Food Proc.

I

Food Ma.

I

Beverage

I

Tobacco

I

Textile

I

Apparel

I

Leather

I

Wood Proc.

I

Furniture

I

Paper

I

Printing

I

Cultural Articles

I

Petroleum Pro.

II

Chemical

II

Medicine

II

Fibers

II

Rubber and Plastic

II

Nonmetal Ma.

II

Ferrous Press

II

Non-Ferrous Pr.

II

Metal Products

II

General Mac.

III

Special Mac.

III

Transport Eq.

III

Electrical Eq.

III

Computers etc.

III

Measuring Inst.

III

Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.

10773.94 7609.01 4407.32 3248.76 4146.85 2872.89 2956.22 1514.51 13402.46 8043.41 5932.06 3772.26 3636.71 2272.83 2754.13 2468.86 1730.27 1437.97 4802.40 3366.49 1837.49 1202.37 1561.89 894.48 7590.38 3311.86 16939.35 12330.00 5398.26 4092.55 2645.76 1512.56 8239.28 5732.37 12344.72 9821.53 16899.86 12064.91 7824.22 6012.20 7889.31 5824.71 14699.05 12713.39 8462.91 7229.86 24429.40 21521.60 17447.62 14045.38 38743.23 31543.38 3398.35 2636.34 4828.48 5185.86 9530.13 8739.67 17863.43 20654.35

2438.78 1673.42 1249.71 666.92 1379.24 491.38 711.44 145.32 3591.93 1318.58 984.08 484.20 494.68 239.55 627.86 389.26 335.04 226.13 2053.60 1039.24 680.37 261.91 293.80 125.01 3406.20 1770.29 7186.80 4061.38 1649.40 875.45 1066.35 224.54 2287.97 1189.17 4812.85 2643.63 8386.87 5003.39 2902.38 1999.20 1657.05 1037.81 3003.42 1968.78 1886.71 1121.33 5058.34 3077.78 2968.72 2084.39 4767.47 3055.60 576.20 324.90 1236.71 1206.52 3706.21 3480.16 3043.48 2624.30

241.14 74.75 124.69 31.39 104.55 16.26 22.17 3.48 568.10 64.57 330.66 93.87 198.36 66.77 86.14 35.58 65.51 33.88 131.97 15.75 67.51 10.19 97.50 26.24 74.79 12.54 375.11 55.40 128.25 27.67 42.63 3.78 262.85 77.17 446.76 50.95 290.66 32.90 138.77 34.87 233.96 69.02 375.66 94.87 246.62 52.27 390.52 100.52 375.87 144.95 437.18 233.15 87.16 26.79 169.86 153.01 221.53 137.28 318.83 172.03

1612.00 128.58 934.22 172.38 818.08 50.03 141.97 39.74 2210.24 446.53 200.02 43.55 92.61 14.04 378.19 79.51 42.90 12.49 3092.57 965.15 53.11 13.66 20.60 2.68 18618.45 8961.91 12103.98 2723.22 684.54 46.57 821.13 109.80 680.65 134.77 18251.11 4787.97 19315.75 6553.32 2745.51 1506.69 319.81 38.63 451.02 82.23 504.77 90.92 862.28 52.89 247.51 98.62 134.03 42.52 31.97 9.92 799.71 1012.85 8171.21 9180.12 371.93 284.35

26.35 9.33 11.43 2.96 9.26 1.48 13.51 15.72 28.14 8.95 10.78 3.35 5.66 1.56 5.22 2.18 4.28 1.83 11.98 3.31 6.96 0.99 3.37 0.69 23.76 9.17 48.24 6.00 10.48 1.84 2.54 1.39 24.34 4.31 41.51 11.98 26.96 10.19 9.59 2.39 22.81 5.22 32.98 10.86 26.14 5.72 33.08 11.17 23.71 6.23 12.66 3.77 4.62 1.50 11.41 9.80 23.36 15.50 22.20 12.72

46.06 10.00 22.15 5.67 13.49 3.27 4.91 1.21 44.94 6.46 24.64 9.03 15.88 3.55 10.74 4.22 7.14 4.33 27.09 5.67 8.95 3.39 13.93 3.45 61.35 22.95 139.06 32.62 9.88 4.22 9.33 1.96 51.59 7.88 285.26 38.16 89.68 18.29 54.89 10.13 51.32 13.52 50.78 17.31 26.90 12.61 71.94 26.76 43.86 17.76 50.73 15.93 10.95 3.12 19.99 14.24 83.60 83.06 42.53 25.56

271.46 107.26 129.63 37.18 86.82 29.62 35.76 7.18 792.11 383.74 92.28 42.06 54.34 24.01 105.52 75.56 24.20 12.51 373.17 126.18 62.41 26.46 37.20 13.49 370.60 122.85 2084.51 843.25 144.90 52.39 234.30 49.01 481.62 258.66 1466.00 715.57 2592.25 1453.88 1661.52 996.41 501.29 283.48 349.44 184.04 181.88 91.54 392.42 213.88 262.50 167.24 344.52 220.49 47.05 22.62 172.07 244.30 1059.67 1096.82 262.97 200.10

The first category The second category The third category

a According to [36], manufacturing sectors can be divided into three categories. As the industrialization process advances, the manufacturing industry gradually transfers to the third category, i.e., high-end manufacturing. Based on their theory, Ref. [37] gave the category classification of China's manufacturing sectors.

would have saved about 5,514,970 tons of coal if it achieves efficient coal use because of its huge consumption (169.456 million tons). 4.2. Saving ratios for gasoline consumption Fig. 5 indicates that gasoline consumption of China's manufacturing sectors is inefficient, and its target saving ratios

(ESTR_GASO) are always higher than 30% and maintain in an upward trend. It is even higher than 60% during 2008e2011. In other words, if manufacturing achieved the best efficient usage of gasoline, 30%e60% of its consumption would be saved. Overall, the average ESTR_GASO during 1998e2011 is 49.939%. The second category has the highest average ESTR_GASO (59.828%), followed by the first category and the third category (45.277% and 44.430% respectively).



7

Table 2 Summary statistics of inputs and outputs. Variable (unit)

Mean

Std. Dev.

Median

Minimum

Maximum

Gross industrial output (100 million Yuan) Capital stock (100 million Yuan) Labor (10,000 person) Coal consumption (104 ton) Gasoline consumption (104 ton) Diesel oil consumption (104 ton) Electricity consumption (104 ton)

9292.35 2461.38 220.19 3161.81 17.79 46.20 488.14

12507.02 2727.37 162.58 6404.86 13.81 57.42 774.02

4738.70 1416.44 174.91 637.92 13.32 29.77 198.12

278.04 97.74 18.61 15.07 0.72 2.08 9.40

100082.49 17722.19 819.48 34087.24 63.29 330.41 5248.27

500

3.5

450

3.0

400

2.5

300

2.0

250

%

10,000 ton

350

1.5

200 150

1.0

100

0.5

50

0.0

0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

EST

ESTR

Fig. 4. The average coal savings target and its ratios for manufacturing industries of China (1998e2011).

80 70 60

%

50 40 30 20 10 0 1998

1999 2000 All industries

2001

2002 2003 Type I industries

2004

2005 2006 2007 Type II industries

2008 2009 2010 Type III industries

2011

Fig. 5. The average gasoline saving target ratios for manufacturing industries of China (1998e2011).

In 1999e2002, the ESTR_GASO of the third category declines from 36.381% to 13.819%. However, since 2003, the ESTR_GASOs of special mac., transport eq. and electrical eq. have increased rapidly. Especially for special mac., whose ESTR_GASO has increased from 7.478% in 2002 to 43.048% in 2003. This induces the ESTR_GASO of the third category increased quickly in 2003. In Table 3, there are 17 sectors that experienced higher values of ESTR_GASO than the average value of ESTR_GASO of 27 manufacturing sectors (the value is 49.939%). Nonmetal ma. is observed the highest average value of ESTR_GASO, namely 88.098%. 4.3. Saving ratios for diesel oil consumption Fig. 6 shows the target savings and its ratios of diesel oil consumption during 1998e2011 for manufacturing industries of China.

The average diesel oil saving target ratio (ESTR_DIES) of 27 manufacturing sectors is 24.465%. During the sample period, ESTR_DIES shows an upward trend and has maintained above 30% since 2009. From the manufacturing category perspective, the third category has the highest average ESTR_DIES (27.799%), followed by the first category and the second category (24.838% and 21.745% respectively). Different from ESTR_COAL and ESTR_GASO, the ESTR_DIES of the third category is higher than the first category and the second category in most years (except for the periods of 2001e2002 and 2009e2010), which indicates that the equipment manufacturing sectors have relatively low diesel oil efficiency. The third category consumed about 35.724 million tons of diesel oil, accounting for about 20.455% of total manufacturing consumption. If this category achieved the most efficient usage mode, it would be saved about 10.108 million tons of diesel oil.


8


40 35 30

%

25 20 15 10 5 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 All industries

Type I industries

Type II industries

Type III industries

Fig. 6. The average diesel oil saving target ratios for manufacturing industries of China (1998e2011).

Six sectors, including food proc., food ma., textile, rubber and plastics, general mac. and special mac. show high values of ESTR_DIES (higher than 40%). The highest one is observed by special mac., and its values is 52.876%.

4.4. Saving ratios for electricity consumption Electricity is the most important energy carrier for the manufacturing production process. Our results show that electricity usage of manufacturing sectors is efficient. In Fig. 7, the electricity saving target ratios (ESTR_ELEC) of 27 manufacturing sectors is below than 4.5%, and the average value is 3.487%. More importantly, ESTR_ELEC shows a downward trend; in 2011 it becomes 2.571%, which declined by 1.850% compared with its highest value (4.422%, which was observed in 1999). This result suggests that the electricity efficiency of manufacturing sectors has improved. From the manufacturing category perspective, the third category has the highest average ESTR_ELEC (4.788%), followed by the first category and the second category (3.761% and 2.254% respectively). Despite the second category consuming a large amount of electricity, its electricity usage is the most efficient among the three categories. In particular, the ESTR_ELEC of the second category declined from 3.365% to 1.706% during the period of 1999e2008. In 2009, because of the expansion investment policy created against

the financial crisis which began in 2008, the ESTR_ELEC increased a little, to 2.554%. But it declined during 2010e2011. The ESTR_ELEC of the third category shows a similar trend of the ESTR of other energy carriers that quickly declined in 1999e2002 but increased in 2002e2003. The main reason is that the ESTR_ELECs of special mac. and electrical eq. experienced a rapid increase in 2003. According to Table 3, the average values of ESTR_ELEC of 10 sectors are higher than 5%, enhancing their electricity efficiencies will conducive to electricity saving. 5. Discussion Table 3 indicates that two sectors, namely tobacco and computers etc. do not have any saving potential of different energy carriers, which implies that their energy usage are efficient during the research period (1998e2011). Although some sectors including petroleum pro., chemical, nonmetal ma. and ferrous press consume more than 10% of total coal consumption of manufacturing sectors, their values of ESTR_COAL are small, which implies that coal-intensive consumption sectors have a relative high efficiency of coal usage. The chemical sector consumes about 10.043% of total gasoline of manufacturing sectors during 1998e2011, but its average value of ESTR_GASO is as high as 80.562%. Although this result does not mean this sector has such value of energy-saving potential in

8 7 6

%

5 4 3 2 1 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 All industries

Type I industries

Type II industries

Type III industries

Fig. 7. The average electricity saving target ratios (unit %) for manufacturing industries of China (1998e2011).



9

Table 3 The average annual saving target ratios for manufacturing sectors in China during 1998e2011. ID

Sector

1 Food Proc. 2 Food Ma. 3 Beverage 4 Tobacco 5 Textile 6 Apparel 7 Leather 8 Wood Proc. 9 Furniture 10 Paper 11 Printing 12 Cultural Articles 13 Petroleum Pro. 14 Chemical 15 Medicine 16 Fibers 17 Rubber and Plastic 18 Nonmetal Ma. 19 Ferrous Press 20 Non-Ferrous Pr. 21 Metal Products 22 General Mac. 23 Special Mac. 24 Transport Eq. 25 Electrical Eq. 26 Computers etc. 27 Measuring Inst. The first category The second category The third category

ESTR

Consumption ratio

Coal

Gasoline oil

Diesel oil

Electricity

Coal

Gasoline oil

Diesel oil

Electricity

1.158 0.961 0.656 0.000 0.854 2.858 1.799 0.812 1.404 0.354 6.472 0.017 0.018 0.339 0.368 0.059 3.080 0.230 0.132 0.296 6.174 6.049 2.718 2.807 5.967 0.000 0.697 1.445 1.188 3.040

73.539 77.541 60.916 0.000 67.647 51.777 20.651 54.075 10.823 80.620 45.605 0.133 13.259 80.562 24.980 24.783 84.761 88.098 71.856 65.835 84.317 77.571 60.395 69.670 53.681 0.000 5.261 45.277 59.828 44.430

42.721 41.074 38.213 0.000 40.193 21.152 10.800 26.303 5.616 35.741 36.215 0.021 13.521 29.022 19.731 5.072 40.450 12.802 24.922 12.115 38.070 51.132 52.876 32.816 28.185 0.000 1.785 24.838 21.745 27.799

7.647 7.019 6.114 0.000 2.964 5.516 1.852 2.914 1.631 2.943 6.516 0.013 0.996 2.135 1.429 0.182 5.347 3.387 1.296 0.581 4.938 8.077 8.255 6.495 5.039 0.000 0.864 3.761 2.254 4.788

1.888 1.094 0.958 0.166 2.589 0.234 0.108 0.443 0.050 3.623 0.062 0.024 21.809 14.178 0.802 0.962 0.797 21.379 22.626 3.216 0.375 0.528 0.591 1.010 0.290 0.157 0.037 11.241 86.145 2.614

5.485 2.379 1.928 2.813 5.858 2.245 1.178 1.087 0.891 2.493 1.449 0.702 4.946 10.043 2.181 0.528 5.067 8.641 5.612 1.997 4.749 6.866 5.442 6.887 4.935 2.635 0.962 28.508 43.765 27.727

3.692 1.776 1.082 0.394 3.603 1.975 1.273 0.861 0.572 2.171 0.718 1.117 4.918 11.147 0.792 0.748 4.136 22.867 7.189 4.400 4.114 4.071 2.156 5.767 3.516 4.067 0.878 19.234 60.312 20.455

2.060 0.984 0.659 0.271 6.010 0.700 0.412 0.801 0.184 2.831 0.474 0.282 2.812 15.816 1.099 1.778 3.654 11.123 19.669 12.607 3.803 2.651 1.380 2.977 1.992 2.614 0.357 15.667 72.361 11.972

practice because the difference of production technology and process between the efficient sector and the chemical sector, it does imply that there is a huge irrational consumption and a wasteful usage of gasoline in this sector, hence there is a huge space for gasoline-saving. For diesel oil, chemical and nonmetal ma. consume more than 10% of total diesel oil of manufacturing sectors during 1998e2011, and their values of ESTR_DIES are 29.022% and 12.802% respectively. The former is higher, while the latter is lower than the average value of ESTR_DIES of 27 manufacturing sectors (24.465%). Sectors of chemical, nonmetal ma., ferrous press and non-ferrous pr. are electricity-intensive. Each consumes more than 10% of total electricity of manufacturing sectors during 1998e2011, and their average values of ESTR_ELEC are 2.135%, 3.387%, 1.296% and 0.581% respectively. All of them are below the average value of all samples (3.487%). So, we can conclude that the electricity-intensive consumption sectors are efficient in the usage of electricity. The results show that the manufacturing sectors are efficiently using coal and electricity, when corresponding with the energysaving target ratios of coal and electricity that are small. On the contrary, the usage efficiencies of gasoline and diesel oil are small, so their energy-saving target ratios are large. In conclusion, the key strategy of energy saving in manufacturing sectors is to improve efficiency of oil (including gasoline and diesel oil), thereby dramatically reducing its saving ratios. Furthermore, the key sector for oil saving is the chemical sector. 6. Conclusions and policy implications One of the core goals of China's energy conversation policy is achieving energy savings as much as possible. Manufacturing industries consume about 55%e60% of the total energy consumption, so it is the key sector of energy savings. Electricity, coal, gasoline oil

and diesel oil are the four major energy inputs during industrial production. Hence, investigating the efficient improvement potential according to the feasible production frontier is important and meaningful for China's policy-makers. Although Chinese government has introduced the declining targets of energy intensity to enhance energy efficiency, it can not reflect the room for improvement in energy efficiency. In this paper, the concept of room for improvement in energy intensity is based on the difference between the actual energy intensity and the target level of energy intensity, which is calculated on the basic of the total factor analysis framework or through the DEA approach. Because the best practice frontier calculated by the previous DEA model assumes all inputs can shrink to the minimum ones under a given output, which may distort the energy efficiency. In the current study, we calculate the DEA model based on the “energy direction”. Furthermore, in order to overcome the “technical regress”, the sequence technique is adopted to measure the TFEE (total factor energy efficiency). Finally, we introduce an improved total-factor ESTR (energy-saving target ratio) index, which shows energy efficiency improvement potential of the DMU. In our opinion, the combination of the above methods has an obvious novelty, and it can be widely used in energy efficiency evaluation and some related issues. In the context of application, we compute the ESTRs of 27 manufacturing sectors during the period of 1998e2011. A single output (the gross industrial output) and six inputs (real capital stock, labor, coal consumption, gasoline consumption, diesel oil consumption and electricity consumption) are used to the aboveimproved DEA model to construct the year-specific frontier. The efficiency scores and target values of these four energy carriers (coal, gasoline, diesel oil and electricity) for each sector during the period of 1998e2011 are hence obtained by comparing to the best practice frontier in that year.


10


The results show that the average energy saving target ratios of coal, gasoline, diesel oil and electricity are respectively 1.714%, 49.939%, 24.465% and 3.487%. In other words, the manufacturing sectors are efficiently using coal and electricity, but usage in gasoline and diesel oil remains inefficient. The third manufacturing category, which is comprised of many equipment manufacturing sectors, has observed the highest saving potential (or the lowest usage efficiency) of coal, diesel oil and electricity, and the efficient energy-saving ratios are 3.04%, 27.799% and 4.788% respectively. On the other hand, the second category, which mainly includes energy-intensive sectors, has the highest saving potential of gasoline, and the value is 59.828%. Our results suggest that the key carriers of energy saving is oil (including gasoline and diesel oil), and the key sectors for gasoline saving is the second category of manufacturing sectors, while the key sectors for diesel oil saving is the third category. From sectors perspective, tobacco and computers etc. are found to have the efficient usage of four major energy carriers in 1998e2011. Coal-intensive and electricity-intensive sectors, such as chemical and ferrous press are efficiently using coal and electricity. The chemical sector consumes about 10.043% and 11.147% of total gasoline and diesel oil of manufacturing sectors during 1998e2011 respectively. However, the target saving ratios of gasoline and diesel oil for this sector are much high. Although the results do not means the really energy-saving potential because of various production technologies and processes across sectors, high values of the target saving ratios do imply a huge irrational consumption or a wasteful usage of energy, hence a huge space for energy-saving. In this sense, the chemical sector is a key sector for oil saving. Besides the above conclusions and implications, this study also provides the following suggestions about China's energy policies. First, similar to Chang [3] and Li and Lin [4], we also conclude that energy efficiency improvement cannot only be judged from a decline in energy intensity, but rather it should be examined whether the room for improvement in terms of energy intensity decreases. Unfortunately, we find that the energy saving target ratios do not show downward trends. Actually, this result indicates that the energy efficiencies of manufacturing sectors show a divergence picture, or the distance between the inefficient sectors and the best practice sectors have widened. In the long term, this trend may be harmful to the energy saving potential of China. Since technology progress plays a critical role in energy efficiency improvement [38], we suggest that it is urgent to promoting technology spillovers among sectors, thus to impel the energy saving target ratios in a downward trends. Second, although the manufacturing sectors have a relative high efficient level of coal usage, it does not mean there is no significance of coal-usage efficiency improvement. Coal is the main energy carriers for China, which is determined by several factors, such as the resource endowment of “abundant coal, scarce oil and gas”, the process of industrialization and urbanization, and the price competitiveness of coal. However, the massive usage of coal induces a huge environmental price for China's development. Coal consumption, both its magnitude and proportion in the primary energy use structure must be reduced as much as possible to solve the environmental problem, for example the problem of haze and its cost on human health. Due to the difficult of coal substitution, we suggest the development of clean coal technology to decline the cost of coal usage in the near future. In addition, it is worth noting that our results are found under the current technology. We believe that the coal saving potential will become huge as technology progresses. In other words, technological progress is the main driven for coal saving. Third, the government should pay more attention to the energy savings of equipment manufacturing or the third category of the

manufacturing sectors. A traditional viewpoint indicates that the energy policy should pay high attention to the energy-intensive sectors or the second category of the manufacturing sectors. The equipment manufacturing or the third category of the manufacturing sectors, due to its a relative low energy consumption, the government encourages their development while ignoring their energy savings. In practice, it is an important way of implementation of energy saving by substituting the second category for the third category. But as shown in Figs. 6 and 7, there are huge spaces of gasoline and diesel oil savings for the third category. We believe technical progress of fuel usage, as well as reforming the fuel price and levying the energy tax will be conducive to oil savings. Acknowledgment The paper is supported by Newhuadu Business School Research Fund, Ministry of Education (Grant no. 10JBG013), Social Science Foundation (Grant No. 14JZD031), National Natural Science Foundation of China (no.71173074) and China Postdoctoral Science Foundation (no. 2014M560527). Appendix

Table The list of abbreviation for each manufacturing industry. ID

Sector

Abbreviation

1 2 3 4 5 6

Processing of Food from Agricultural Products Manufacture of Foods Manufacture of Beverages Manufacture of Tobacco Manufacture of Textile Manufacture of Textile Wearing Apparel, Footware and Caps Manufacture of Leather, Fur, Feather and Related Products Processing of Timber, Manufacture of Wood, Bamboo, Rattan, Palm and Straw Products Manufacture of Furniture Manufacture of Paper and Paper Products Printing, Reproduction of Recording Media Manufacture of Articles For Culture, Education and Sport Activities Processing of Petroleum, Coking, Processing of Nuclear Fuel Manufacture of Raw Chemical Materials and Chemical Products Manufacture of Medicines Manufacture of Chemical Fibers Manufacture of Rubber and Plastics Products Manufacture of Non-metallic Mineral Products Smelting and Pressing of Ferrous Metals Smelting and Pressing of Non-ferrous Metals Manufacture of Metal Products Manufacture of General Purpose Machinery Manufacture of Special Purpose Machinery Manufacture of Transport Equipment Manufacture of Electrical Machinery and Equipment Manufacture of Communication Equipment, Computers and Other Electronic Equipment Manufacture of Measuring Instruments and Machinery for Cultural Activity and Office Work

Food Proc. Food Ma. Beverage Tobacco Textile Apparel

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Leather Wood Proc. Furniture Paper Printing Cultural Articles Petroleum Pro. Chemical Medicine Fibers Rubber and Plastic Nonmetal Ma. Ferrous Press Non-Ferrous Pr. Metal Products General Mac. Special Mac. Transport Eq. Electrical Eq. Computers etc. Measuring Inst.

References [1] Han Z, Wei Y, Fan Y. Research on change features of Chinese energy intensity and economic structure. J Appl Stat Manag 2004;23(1):1e6 [in Chinese)]. [2] Wang B, Zhang JH, Zhang H. Total-factor energy efficiency and influencing factors across provinces in China in the presence of environmental regulation. Econ Rev 2011;4:31e43 [in Chinese)].


K. Li, B. Lin / Energy xxx (2015) 1e11 [3] Chang M-C. Energy intensity, target level of energy intensity, and room for improvement in energy intensity: an application to the study of regions in the EU. Energy Policy 2014;67:648e55. [4] Li K, Lin B. The improvement gap in energy intensity: analysis of China's thirty provincial regions using the improved DEA model. Energy 2015;84:589e99. [5] Boyd GA, Pang JX. Estimating the linkage between energy effciency and productivity. Energy Policy 2000;28:289e96. [6] Hu J-L, Wang S-C. Total-factor energy efficiency of regions in China. Energy Policy 2006;34(17):3206e17. [7] Chang M-C. A comment on the calculation of the total-factor energy efficiency (TFEE) index. Energy Policy 2013;53:500e4. [8] Zhou P, Ang BW, Poh KL. Measuring environmental performance under different environmental DEA technologies. Energy Econ 2008;30(1):1e14. [9] Zhou P, Ang BW, Poh KL. A survey of data envelopment analysis in energy and environmental studies. Eur J Oper Res 2008;189:1e18. [10] Hu J-L, Kao C-H. Efficient energy-saving targets for APEC economies. Energy Policy 2007;35:373e82. [11] Zhang X-P, Cheng X-M, Yuan J-H, Gao X-J. Total-factor energy efficiency in developing countries. Energy Policy 2011;39(2):644e50. [12] Wei C, Ni J, Sheng M. China's energy inefficiency: a cross-country comparison. Soc Sci J 2011;48:478e88. [13] Wei C, Shen MH. Energy efficiency and its influence factors: empirical research based on DEA model. Manag World 2007;8:66e76 [in Chinese)]. [14] Wang K, Yu S, Zhang W. China's regional energy and environmental efficiency: a DEA window analysis based dynamic evaluation. Math Comput Model 2013;58(5e6):1117e27. [15] Wang Z, Feng C, Zhang B. An empirical analysis of China's energy efficiency from both static and dynamic perspectives. Energy 2014;74:322e30. [16] Li L-B, Hu J-L. Ecological total-factor energy efficiency of regions in China. Energy Policy 2012;46:216e24. [17] Bian Y, He P, Xu H. Estimation of potential energy saving and carbon dioxide emission reduction in China based on an extended non-radial DEA approach. Energy Policy 2013;63:962e71. [18] Zhang C, Kong F, Yu Y. Measuring ecological total-factor energy efficiency incorporating regional heterogeneities in China. Ecol Indic 2015;51:165e72. [19] Wang Z-H, Zeng H-L, Wei Y-M, Zhang Y-X. Regional total factor energy efficiency: an empirical analysis of industrial sector in China. Appl Energy 2012;97:115e23. [20] Pan H, Zhang H, Zhang X. China's provincial industrial energy efficiency and its determinants. Math Comput Model 2013;58:1032e9.

11

[21] Zhao X, Yang R, Ma Q. China's total factor energy efficiency of provincial industrial sectors. Energy 2014;65:52e61. [22] Wei Y-M, Liao H, Fan Y. An empirical analysis of energy efficiency in China's iron and steel sector. Energy 2007;32:2262e70. [23] Lin B, Wu Y, Zhang L. Estimates of the potential for energy conservation in the Chinese steel industry. Energy Policy 2011;39(6):3680e9. [24] He F, Zhang Q, Lei J, Fu W, Xu X. Energy efficiency and productivity change of China's iron and steel industry: accounting for undesirable outputs. Energy Policy 2013;54:204e13. [25] Xue X, Wu H, Zhang X, Dai J, Su C. Measuring energy consumption efficiency of the construction industry: the case of China. J Clean Prod 2014. http:// dx.doi.org/10.1016/j.jclepro.2014.04.082. [26] Zhou G, Chung W, Zhang Y. Measuring energy efficiency performance of China's transport sector: a data envelopment analysis approach. Expert Syst Appl 2014;41(2):709e22. [27] Lee Y-C, Hu J-L, Kao C-H. Efficient saving targets of electricity and energy for regions in China. Int J Electr Power Energy Syst 2011;33(6):1211e9. [28] Stern DI. Modeling international trends in energy efficiency. Energy Econ 2012;34(6):2200e8. [29] Farrell MJ. The measurement of productive efficiency. J R Stat Soc 1957;120(3):253e90. [30] Shestalova V. Sequential Malmquist indices of productivity growth: an application to OECD industrial activities. J Prod Anal 2003;19(2):211e26. [31] Oh D-h. A global Malmquist-Luenberger productivity index. J Prod Anal 2010;34:183e97. [32] Wang K, Wei Y-M, Zhang X. A comparative analysis of China's regional energy and emission performance: which is the better way to deal with undesirable outputs? Energy Policy 2012;46:574e84. [33] Li K, Lin B. Metafroniter energy efficiency with CO2 emissions and its convergence analysis for China. Energy Econ 2015;48:230e41. [34] Chen S. Reconstruction of sub-industrial statistical data in China (1980-2008). China Econ Q 2011;10(4):735e76 [in Chinese)]. [35] Chen S. The evaluation indicator of ecological development transition in China's regional economy. Ecol Indic 2015:42e52. [36] Chenery H, Robinson S, Syrquin M. Industrialization and growth: a comparative study. Oxford University Press; 1986. [37] Wang YZ, Chen JL. Development report of Yangtze River Delta in 2005. Shanghai: Shanghai People's Publishing House; 2005 [in Chinese]. [38] Li K, Lin B. The nonlinear impacts of industrial structure on China's energy intensity. Energy 2014;69:258e65.
