Chen 2020
Chen 2020
Chen 2020
a r t i c l e i n f o a b s t r a c t
Article history: In this paper, we comprehensively consider the environment and economic performance by employing
Received 23 November 2019 non-radial distance function to study the regional eco-efficiency of China’s Non-ferrous Metals Industry
Received in revised form (NMI) during 2000e2016, from the perspective of total-factor through Data Envelopment Analysis
18 June 2020
(DEA)-Malmquist analysis framework. The change of regional technology gap ratio and the source of
Accepted 20 July 2020
Available online 3 August 2020
regional eco-inefficiency is also explored. Moreover, this paper analyzes the evolution of regional eco-
efficiency and the main factors that contribute to changes in eco-efficiency. The results indicate that:
Handling editor. Bin Chen First, the eco-efficiency of China’s NMI has improved and has shown a certain degree of regional het-
erogeneity. Second, the main source of eco-inefficiency is management inefficiency in eastern China. The
Keywords: central, western and northeastern China all showed an increasing trend of technology gap inefficiency
Eco-efficiency and a declining trend of management inefficiency during the sample period. Third, eastern China is
China’S non-ferrous metals industry regarded as a technology leader. The improvement of eco-efficiency in the central and northeastern
Eco-inefficiency China is mainly due to the promotion of technological progress, while the western region relies on the
DEA
improvement of technical efficiency. The study suggests that strengthening scientific and technological
Malmquist
research, promoting inter-regional technical exchanges are particularly important for eco-efficiency
improvement and regional coordinated development in China’s NMI.
© 2020 Elsevier Ltd. All rights reserved.
https://doi.org/10.1016/j.jclepro.2020.123388
0959-6526/© 2020 Elsevier Ltd. All rights reserved.
2 X. Chen, B. Lin / Journal of Cleaner Production 277 (2020) 123388
Fig. 1. CO2 emissions and annual growth rate of China’s NMI during 2000e2016
Notes: The data were collected and calculated from the statistical yearbook.
degree of industrial relevance and play an important role in eco- The reasonable evaluation of eco-efficiency needs to consider
nomic construction, national defense construction, social devel- both economic performance and environmental performance. To
opment and other aspects.3 China is the largest producer and this end, in this paper, we comprehensively consider environ-
consumer of non-ferrous metals products, accounting for nearly mental and economic performance and establish an eco-efficiency
half of the world’s nonferrous metal production (Lin and Chen, indicator to measure regional ecological performance from the
2019). Non-ferrous metals are indispensable in the industrial sys- perspective of total-factor.
tem. However, the smelting and processing of non-ferrous metals We hereby raise the following research questions. (1) What is
are accompanied by high energy consumption. The energy con- China’s NMI eco-efficiency change trend at the national and
sumption of China’s NMI (155.58 Mtoe) even exceeds the national regional levels? (2) What is the technology gap of regional eco-
energy consumption of Italy (151.3 Mtoe) in 2016. Under the efficiency in China’s NMI? (3) What are the causes of eco-
background of green development, China’s NMI shoulders the re- inefficiency and how are regional differences reflected? (4) What
sponsibility of energy conservation, emission reduction and clean are the factors that affect the change of eco-efficiency in China’s
production. The research on the evolution of eco-efficiency can NMI and what are their regional heterogeneity?
provide a meaningful reference for the sustainable development of To achieve the above research objectives, we use the data of
China’s NMI. China’s NMI in 2000e2016 through the DEA-Malmquist frame-
China is facing the pressure of ecological environmental pro- work. Within this framework, we discuss the sources of regional
tection while ensuring rapid economic development (Song et al., eco-inefficiency and construct the dynamic evolution trend of
2018; Wang and Liu, 2018; Lin and Kuang, 2020). The improve- regional eco-efficiency through the establishment of contempora-
ment of energy-saving and efficiency (Ouyang et al., 2018), the neous, intertemporal, and global production technology sets. In
improvement of the ecological environment and the high-quality addition, the driving forces of technical efficiency, technological
development of the economy (Song et al., 2015) cannot be progress and technology leadership on the change of regional eco-
ignored in the process of ecological civilization construction. China efficiency are also discussed.
is working to establish a carbon market to manage GHG emissions, Eco-efficiency research is extremely important for both the in-
including the current pilot cap-and-trade system. ternational community and China. This paper has two contribu-
Due to the urgent needs of China’s emission reduction and tions. On one hand, this paper enriches the literature of applying
sustainable development, China’s NMI, a heavy industry, has also the DEA-Malmquist framework to study eco-efficiency. Non-radial
been included in the carbon emissions trading system (ETS) as a efficiency measure and Malmquist analysis give full play to their
pilot industry. The starting point of the establishment of carbon advantages in the construction and decomposition of dynamic eco-
market is to promote energy conservation and emission reduction, efficiency, and appropriately depict the evolution of eco-efficiency.
so the high emission characteristics of NMI is an important reason On the other hand, the contribution is reflected in the content of the
to be included in ETS. The CO2 emissions of China’s NMI have research. China’s NMI has a special international status. This paper
increased significantly in recent years, with an annual growth rate can provide a meaningful reference for the sustainable develop-
of 13.07%, increased from 98 million tons in 2000 to 701 million ment of regional development in China’s NMI by analyzing the
tons in 2016, as shown in Fig. 1. The Chinese government has causes of regional eco-inefficiency and the decomposition of eco-
strengthened its control over GHG emissions through various efficiency.
methods such as command-and-control regulation and carbon The rest of the paper is organized as follows; in the second part,
emissions trading. Achieving the improvement of eco-efficiency is the paper summarizes the literature related to eco-efficiency. The
the common aspiration of the people and the government (Shao third part introduces the methodology related to the construction
et al., 2016). of dynamic eco-efficiency and data sources. The fourth part reports
and analyses the results. The fifth part is the summary of this paper
and the policy recommendations.
3
Source: “Non-ferrous metal industry adjustment and revitalization plan”.
X. Chen, B. Lin / Journal of Cleaner Production 277 (2020) 123388 3
Table 1
Literatures relevant to eco-efficiency research (except China).
Arabi et al. (2014) Iran 52 Power plants. 2003 SBM-MLindex Input: Capital, Fuel.
e2010 Output: Generated power & SO2, NOx, CO2.
Arabi et al. (2016) Iran 52 power plants. 2003 DEA, ML index Input: Fuel, Capital.
e2010 Output: Generated Electricity, Operational Availability & Deviation from
Generation Plan, SO2.
Camarero et al. 22 OECD 1980 DEA Output: GDP & CO2, NOX, SOX.
(2013) countries e2008
Egilmez and Park U.S Manufacturing EIO-LCA, DEA Input: Water, Energy.
(2014) sectors. Output: Economic output & Carbon.
Gomez-Limon et al. Spanish 292 Andalusian olive 2010 DEA Output: Net income & Environmental pressures.
(2012) farmers.
Gomez-Calvet et al. European Union 1993 DEA Output: GDP & CO2e, SO2, NOx.
(2016) (EU-27) e2010
Korhonen and European 24 power plants. DEA Input: Total costs.
Luptacik (2004) country Output: Electricity generation & DUST, NOx, SO2.
Lorenzo-Toja et al. Spanish 113 Wastewater 2009 LCA, DEA Electricity use, Chemical consumption, Sludge production.
(2015) treatment plants. e2012
Martinez (2013) Swedish Service industries. 1993 DEA-Malmquist Input: Energy.
e2008 Output: & CO2 emission.
Monastyrenko European 15 electricity 2005 DEA, ML index Input: Installed production capacity, Total operational expenditures.
(2017) Producers. e2013 Output: Generated electricity & CO2.
Moutinho et al. 16 Latin America 1994 DEA Input: Energy, Population density, Labor productivity, Renewable energy
(2018) countries e2003 consumption, Gross Capital Formation productivity.
Output: GDP & CO2.
Munisamy and Iran 48 power plants. 2003 SBM, meta- Input: Effective capacity, Fuel consumption
Arabi (2015) e2010 frontier ML index Output: Generated energy & SO2, NOx, Cox.
Oggioni et al. 21 countries Cement industry. 2005 DEA Input: Capacity, Energy, Labor, Materials
(2011) e2008 Output: Cement & CO2.
Picazo-Tadeo et al. Spanish Olive-growing farms. 2010 DEA Input: Labor, Fertilizers, Pesticides, Energy, Contracted services, Fixed costs.
(2012) Output: Net income & Erosion, Pesticide risk, Energy, CO2 fixation.
Ramli et al. (2013) Malaysia Manufacturing sector. 2009 Scale DDF Input: Operating expenditure, Capital.
Output: Sales value & CO2.
Rashidi and Saen 19 OECD 2012 DEA Input: Labor, Precipitation average, Energy.
(2015) countries Output: GDP & CO2.
Robaina-Alves 26 European 2000 SFA Input: Fossil fuel, Renewable Energy, Capital, Labor.
et al. (2015) countries e2011 Output: GDP & GHG.
Suh et al. (2014) South Korea 272 firms in 16 2008 DEA Input: CO2, Energy.
industries. e2010 Output: Production, value-added &
Notes: DEA means data envelopment analysis. DDF means directional distance function. SFA means stochastic frontier analysis. SBM means slack-based measure. ML means
Malmquist-Luenberger. EIO-LCA means economic input-output life cycle assessment. In the variable column, the variable before & is the desired output, and the variable after
& is the undesired output.
Table 2
Literatures relevant to eco-efficiency research (China).
Fei and Lin (2017) China Agricultural sector 2001 DEA Input: Energy, Capital, Labor
e2012 Output: GDP & CO2.
Hu et al. (2019) China 281 wastewater treatment 2016 SBM-DEA Input: Investment, Energy, Operating cost, Relative capacity load rate,
plants. Wastewater.
Output: Removal efficiency of COD, TN, NH3eN, TP.
Huang et al. China 30 provinces. 2000 GB-US-SBM model Input: Energy, Capital, Labor, Land.
(2014) e2010 Output: GDP & COD, Wastewater, Exhaust gas, SO2, dust, Solid waste, Smoke
dust.
Huang et al. China 30 provinces 2001 GB-US-SBM model Input: Energy, Capital, Labor, Land, Water.
(2018) e2014 Output: GDP & CO2, SO2, Wastewater, COD, AN, Dust.
Huang et al. China 273 cities. 2003 DEA Input: Capital, Labor, Land, Water, Electricity.
(2018) e2015 Output: GDP & SO2, CO2, Dust.
Li et al. (2017) China 262 cities. 2005 SBM -DEA Input: Labor, Capital.
e2012 Output: GDP & SO2.
Liu et al. (2013) China Water system. 2010 DEA Input: Capital, Workforce.
Output: GDP, Public green areas, etc.
Long et al. (2015) China Cement manufactures. 2005 SBM-ML index Input: Capital, Labor, Coal, Electricity, Clinker.
e2014 Output: Cement & CO2.
Long et al. (2017) China Cement manufacturers. 2005 DDF, SBM Same as (Long et al., 2015)
e2014
Wang et al. (2019) China 28 typical coal-mining 2007 SFA Input: Resource input, System management.
cities. e2014 Output: GDP & Wastewater, SO2, Smoke, Dust
Xing et al. (2018) China 26 sectors. 2012 EIO-LCA, DEA Input: Energy.
Output: Economic output & Water withdrawal, CO2, Hazardous waste,
Wastewater, Exhaust emission.
Yang et al. (2017) China 30 provinces. 2003 DEA Input: Capital, Labor, Material resource.
e2012 Output: GDP & CO2, Household refuse, Wastewater.
Yang and Yang China Agriculture, Forestry, etc. 1978 Ecological footprint Human consumption, Energy, Pollution emissions, Land.
(2019) e2016 model
Yang and Zhang China 30 provinces. 2003 Global DEA Input: Capital, Labor, Material resource.
(2018) e2014 Output: GDP & Solid waste, Household refuse, SO2, Soot and industrial dust,
Wastewater.
Yu et al. (2016) China Pulp and paper industry. 2000 DEA-SBM model, ML Input: Water.
e2013 index Output: Industrial output & Wastewater, COD, Ammonia nitrogen.
Yu et al. (2018) China Industrial sector. 2001 DEA Input: Capital, Labor, Land, Water, Energy.
e2015 Output: Industrial output & Industrial environmental pollutants.
Zhang et al. (2008) China 30 provincial industrial 2004 DEA Input: Water, Resource, Energy.
sector. Output: Value-added & COD, Nitrogen, SO2, Soot, Dust, Solid Wastes.
Zhang et al. (2017) China Industrial sector. 2005 Three-stage DEA Input: Capital, Labor, Energy.
e2013 Output: Gross industrial output value.
Notes: DEA means data envelopment analysis. DDF means directional distance function. SFA means stochastic frontier analysis. SBM means slack-based measure. ML means
Malmquist-Luenberger. EIO-LCA means economic input-output life cycle assessment. GB-US means global benchmark-undesirable output, super efficiency. In the variable
column, the variable before & is the desired output, and the variable after & is the undesired output.
!C n o
D ðK;L;E;Y;C;gÞ¼sup wT bC : ðK;L;E;Y;CÞþg:diag bC 2TRCh !G n
G
G
o
D ðK;L;E;Y;C;gÞ¼sup wT b : ðK;L;E;Y;CÞþg:diag b 2T G
(6)
(10)
The intertemporal production technology set is the envelope of
the contemporaneous production technology set, which is According to the NDDF defined by three types of production
expressed as Eq. (7): technology sets, the corresponding TEEI can be obtained by solving
the function, such as Eq. (11).
TRI h ¼ TRC1
h
∪TRC2
h
∪TRC3
h
∪:::∪TRCTh ; t ¼ 1; 2; 3; :::; T (7)
j
TEEIG ðK; L; E; Y; CÞjRh G t t tþ1 tþ1 TEEIG xtþ1 ; ytþ1
TGRRh ¼ (12) MNMEEI x ;y ;x ;y ¼ (16)
TEEII ðK; L; E; Y; CÞjRh TEEIG ðxt ; yt Þ
In Eq. (16), xrepresents input variables K; L; E and yrepresents
Because the global production technology set is the envelope of
output variables Y; C.
the intertemporal production technology set, that isT I 4 T G , so
This paper decomposes MNMEEI with reference to Oh and Lee
there’s alwaysTEEI G ðK; L; E; Y; CÞjRh TEEI I ðK; L; E; Y; CÞjRh , that’s 0 < (2010) to obtain three indicators: technical efficiency change (EC),
TGR 1. best-practice gap change (BPC) and technology gap rate change
The value TGR indicates the heterogeneity of the two production (TGC). The decomposition process is shown in Eq. (17).
MNMEEIG xt ; yt ; xtþ1 ; ytþ1
( )
TEEIG xtþ1 ; ytþ1 TEEIC xtþ1 ; ytþ1 TEEIC xt ; yt TEEIG xtþ1 ; ytþ1
¼ ¼
TEEIG ðxt ; yt Þ TEEI C ðxt ; yt Þ TEEIC xtþ1 ; ytþ1 TEEI G ðxt ; yt Þ
( ) ( )
TEEItþ1 xtþ1 ; ytþ1 TEEIC xt ; yt TEEII xtþ1 ; ytþ1 TEEII xt ; yt TEEIG xtþ1 ; ytþ1
¼
TEEIt ðxt ; yt Þ TEEI tC xtþ1 ; ytþ1 TEEII ðxt ; yt Þ TEEI I xtþ1 ; ytþ1 TEEIG ðxt ; yt Þ
. .
TEEIC xtþ1 ; ytþ1 TEEII xtþ1 ; ytþ1 TEEIC xtþ1 ; ytþ1 TEEIG xtþ1 ; ytþ1 TEEII xtþ1 ; ytþ1
¼ . .
TEEIC ðxt ; yt Þ TEEII xtþ1 ; ytþ1 TEEIC ðxt ; yt Þ TEEI G ðxt ; yt Þ TEEII ðxt ; yt Þ
technologies and the difference of eco-efficiency between the two Technical efficiency change (EC) represents the movement of
production technologies. The closer TGR is to 0, the greater the eco-efficiency in the contemporaneous production technology set
heterogeneity of eco-efficiency between the group and the global and represents the change of technical efficiency (TE) related to
production technology. The closer TGR is to 1, the closer the group is eco-efficiency. BPC represents the movement of eco-efficiency from
to the global eco-efficiency frontier. the contemporaneous production technology to the intertemporal
Due to the heterogeneity between the group and the global production technology, and represents the best-practice gap (BPG)
production technology, the difference in eco-efficiency between the change related to the eco-efficiency. This index is equivalent to
two technology sets is obtained. To this end, with reference to Chiu technological progress. TGC represents the movement of the TGR
et al. (2012); Wang et al. (2013), this paper measures the Eco- that is, the movement of the intertemporal production technology
inefficiency (EIN) and decomposes it into technology gap eco- to the global production technology, representing the changes in
inefficiency (TGEIN) and management eco-inefficiency (GMEIN), as the technology leadership of the group. The values of EC, BPCand
shown in Eq. (13). TGEIN represents the eco-inefficiency caused by TGC are divided into three cases: <1, ¼1 and >1. When the value < 1,
the gap between group and global production technology, as in Eq. the performance of the indicators is worse than that of the previous
(14). GMEIN refers to the eco-inefficiency caused by management period. When the value ¼ 1, the performance of the indicators is
failure in the same group with similar technology level, as in Eq. (15). unchanged compared with that of the previous period. When the
value > 1, the performance of the indicators is better than that of
j j j the previous period.
EINRh ¼ TGEINRh þ GMEINRh (13)
3.2. Data
j j j
TGEINRh ¼ TEEII ðK; L; E; Y; CÞRh 1 TGRRh (14)
This paper evaluates the eco-efficiency of China’s NMI by using
j I j the data from 2000 to 2016. Data of Hong Kong, Taiwan and Macao
GMEINRh ¼ 1 TEEI ðK; L; E; Y; CÞRh (15)
are not available and some of the data of Hainan and Tibet are
Referring to Malmquist productivity index and decomposition missing, we, therefore, selected the data of 29 provinces in China.
framework established by Caves et al. (1982); Fare et al. (1994); Oh This paper also divides China into four regions (eastern, central,
and Lee (2010); this paper constructs a dynamic eco-efficiency in- western and northeastern)4 to study the heterogeneity of regional
dex, the Meta-frontier Non-radial Malmquist Eco-efficiency Index eco-efficiency. It should be noted here that we have linearly
(MNMEEI), expressed as Eq. (16). Where, meta-frontier refers to the interpolated and replaced the obvious outliers existing in the
frontier of global production technology for all decision-making
units, which to be a function that envelops the deterministic
4
components of the group frontiers (O’Donnell et al., 2008). In this Eastern China: Beijing, Tianjin, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian,
Shandong and Guangdong. Central China: Shanxi, Anhui, Jiangxi, Henan, Hubei and
paper, group frontier refers to the technology frontier of four re- Hunan. Western China: Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou,
gions in China. And meta-frontier refers to China’s technology Yunnan, Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang. Northeast China: Heilongjiang,
frontier, which is equivalent to the global technology frontier. Jilin and Liaoning.
X. Chen, B. Lin / Journal of Cleaner Production 277 (2020) 123388 7
Table 3
Descriptive statistic for regional variables.
Region China Eastern China Central China Western China Northeast China
Variables Unit N N N N N
Average growth rate Average growth rate Average growth rate Average growth rate Average growth rate
(%) (%) (%) (%) (%)
Input: L 10 thousand 493 7.21 2.96 153 6.89 4.87 102 11.08 3.79 187 6.08 1.67 51 4.58 1.67
Input: E 10 thousand tce 493 523.1 13.39 153 212.1 13.66 102 781.8 10.61 187 723.2 15.58 51 204.9 15.58
Input: K 100 million RMB 493 174.1 13.78 153 161.8 14.85 102 257.7 13.94 187 161 13.94 51 91.54 13.94
output: Y 100 million RMB 493 552.1 18.20 153 736.0 18.87 102 778.5 18.94 187 373.2 17.84 51 204 17.84
output: C 10 thousand tons 493 1224 13.07 153 498.8 13.09 102 1849 10.54 187 1677 15.24 51 485.9 15.24
sample. data (base period ¼ 2000) by using the industrial product ex-
The variables selected in this paper are energy, capital and labor factory price index (29 provinces in 2000e2016) from the China
(as input variables), gross industrial output (as desirable output) Urban (City) Life and Price Yearbook, in units of RMB 100 million.
and CO2 (as undesirable output) of China’s NMI. Table 3 is a
descriptive statistic for regional variables. Besides, the annual data Kt ¼ It þ ð1 dt Þ Kt1 (18)
of variables are listed in the appendix, as shown in Table A .
E: Energy consumption is obtained by multiplying the physical
quantity of six major energy categories of coal, coke, gasoline,
4. Results and discussions
diesel, fuel oil and electricity, by corresponding standard coal
conversion coefficient. The data comes from the statistical year-
4.1. Regional TGR
books of the provinces and the unit is 10,000 tce. The energy
conversion coefficient comes from the China Energy Statistical
Fig. 2 shows the trend of TGR in China’s NMI during 2000e2016,
Yearbook (CESY).
obtained according to Eq. (12). It is the change in the ratio of static
C: CO2 emissions are obtained by multiplying the six major
eco-efficiency under the intertemporal and global production
energy categories by corresponding CO2 coefficient, with a unit of
technology set.
10,000 tons. The CO2 conversion coefficient5 is calculated according
First, the average TGR is a downward trend during the sample
to the method reported by IPCC2006. This paper calculates the
period, which represented the enlargement of the heterogeneity of
proportion of CO2 emissions of various energy varieties at the na-
the group’s eco-efficiency, and the group’s technology level moves
tional level, as shown in Table B in the Appendix. The CO2 emissions
away from the global technology frontier. Second, the TGR ¼ 1 in
of the energy varieties considered in this paper account for
the eastern region indicates that its group technology is the global
93.7e95% of the total emissions of China’s NMI, which shows that
technology frontier, and the technical level in eastern China is
the six energy varieties are representative.
advanced, which is consistent with China’s actual situation.
K: The perpetual inventory method is used to calculate the
Furthermore, the TGR in the central, western and northeastern
capital stock, as shown in Eq. (18), whereKt1 andKt represent the
China showed a downward trend. This trend indicates that the
capital stock of period t 1 and t respectively. It represents the
distance between the group eco-efficiency frontier and global eco-
investment in the period t. dt denotes the depreciation rate of the
efficiency frontier is enlarged. The economic development level and
period t. We followed Lin and Chen (2019), to obtain the depreci-
population density of these three regions are relatively small
ation rate of the industry. The data needed for the calculation of
compared to eastern China. The negative externalities of NMI on
capital stock comes from the CSY and China’s Industrial Economics
eco-efficiency are lower in these regions than in eastern China.
Statistical Yearbook (CIESY). The net value of fixed assets in 2000 is
selected as the capital stock in the base period (2000). Besides, the
capital stock is adjusted to the comparable data (base 4.2. Eco-inefficiency decomposition
period ¼ 2000) by using the fixed asset investment price index, and
the unit is 100 million RMB. The data for this index in different There are two sources of eco-inefficiency in China’s NMI. They
provinces (2000e2016) is derived from the CSY. are technology gap eco-inefficiency (TGEIN) and management eco-
L: The number of employees in CIESY is selected as the proxy inefficiency (GMEIN), which are calculated by Eq. (13,14,15) . The
variable of labor. Due to the lack of data in 2012, this paper uses the change of regional eco-inefficiency is shown in Fig. 3a3b3c3d3e.
linear interpolation of data in 2011 and 2013 to obtain data for On one hand, the average eco-inefficiency trend of China’s NMI
2012. The unit is 10000. is declining. This shows the improvement of eco-efficiency on the
Y: The gross industrial output value comes from CISEY. The data national average. Regionally, the eco-inefficiency decline trend is
for 2000e2011 is directly available, and the data for 2012e2016 is obvious in the eastern and central China, while the eco-inefficiency
constructed through the industrial sales value of 2012e2016. in the other two regions has not decreased significantly.
Further, this paper adjusts the desirable output to the comparable On the other hand, the regional changes in TGEIN and GMEIN
are obvious. The eco-inefficiency in eastern China is mainly due to
management eco-inefficiency, and a downward trend. The eastern
region itself is at the technology frontier of China’s eco-efficiency. In
5
When calculating the CO2 conversion coefficient of electricity, we considered the future, continuing to strengthen the management efficiency of
the proportion of thermal power in the power generation structure. Thermal power the NMI is the key to improving the eco-efficiency in eastern China.
is the main source of CO2 at the power generation end. Therefore, when calculating
the CO2 emission coefficient of electricity, we have taken into account the change in
In addition, the central, western and northeastern China all
the proportion of thermal power in each year. This paper uses the same electricity showed an increasing trend of technology gap inefficiency and a
CO2 emission coefficient for all provinces. declining trend of management eco-inefficiency during the sample
8 X. Chen, B. Lin / Journal of Cleaner Production 277 (2020) 123388
period. In the production of NMI, although there are improvements also necessary to continue to improve management efficiency.
in management efficiency, we need to pay attention to the fact that The dynamic eco-efficiency of China’s NMI, MNMEEI, was
the trend of technology gap inefficiency has been rising, similar to decomposed to obtain three indicators, EC, BPC and TGC. EC mea-
the previous discussion on the downward trend of the TGR in these sures the movement of eco-efficiency under the contemporaneous
regions. technology set, indicating the technical efficiency change of China’s
NMI, shown in Fig. 5.
4.3. Dynamic eco-efficiency and its decomposition in China’s NMI On the one hand, the national average EC has an obvious upward
trend, and the Multiplicative EC increased by 882% during the
Further, this paper constructs the dynamic index, MNMEEI, of sample period. The arithmetic mean growth rate is 15.9%. On the
eco-efficiency according to Eq. (16, 17), and decomposes the index other hand, the technical efficiency of the western region has
by Malmquist to get three factors affecting the improvement of eco- increased most obviously (17.48% on average), followed by the
efficiency in China’s NMI. eastern region (16.29% on average), both of which are higher than
Fig. 4 and Table 4 shows the dynamic trend of regional eco- the national average. The central region is lower than the national
efficiency of China’s NMI. The eco-efficiency of each region has average but is on the rise (6.49% on average). However, the tech-
increased relative to the base period, but the trend tells us that the nical efficiency of the northeastern region has declined (- 0.06% on
increase of eco-efficiency in 2000e2016 shows obvious regional average).
differences. The eco-efficiency of China’s NMI increased by 161.4% BPC measures the movement of eco-efficiency from the
during the sample period (2000e2016), with an annual average of contemporaneous technology set to the intertemporal technology
6.19%. set, indicating technological progress. The change of regional BPC is
The change of eco-efficiency in eastern China is higher than the shown in Fig. 6.
national average and other regions. During the sample period, the On the one hand, the multiplicative BPC of China’s NMI is on the
cumulative increase is 350.4%, and the annual growth rate is 9.86%. rise, with a multiplicative increase of 1400% during the sample
The technology level of production and strict environmental period and an average annual growth of 18.46%. On the other hand,
regulation (Song et al., 2018) in the eastern region contributed to the BPC in each region has been improved. BPC in the eastern re-
keeping its eco-efficiency ahead of the rest of China. gion increased most significantly (21.96% on average), which is
The annual growth rate of eco-efficiency in central China (6.11%) higher than the national average. Next is the northeastern region
is close to the national average, and it began to rise around 2008. (17.04% on average), then the central region (14.05% on average),
Benefited by the Chinese government’s support for regional coor- and finally the western region (7.5% on average). These three re-
dinated development (Zhang et al., 2017), the support of relevant gions are lower than the national average.
policies has promoted the improvement of production technology Here, we combine the results of EC with BPC for analysis. The
in central China and the improvement of eco-efficiency. effects of EC and BPC on the eco-efficiency of eastern China are both
Although the eco-efficiency of the western region and the obvious. However, it is noteworthy that EC and BPC have played a
northeast region has improved, it is lower than the national average differentiated role in the promotion of eco-efficiency in the other
and fluctuates greatly. The enlargement of the technology gap in- three regions of China. The improvement of eco-efficiency in the
efficiency inhibits the improvement of eco-efficiency. The central and northeastern China is mainly due to the promotion of
improvement of production technology is particularly urgent for technological progress, while the western region relies on the
the improvement of eco-efficiency in these two regions, and it is improvement of technical efficiency. The main reason may be that
X. Chen, B. Lin / Journal of Cleaner Production 277 (2020) 123388 9
the central and northeastern China is closer to the eastern than the
western, and it is most convenient to receive technology transfer
from the eastern region.
TGC measures the intertemporal movement of the TGR, indi-
cating the improvement of technology leadership, as shown in
Fig. 7.
According to the national average trend, the Multiplicative TGC
decreased by 32% and the annual average decreased by 2.34%,
indicating the decline of the national average technology leader-
ship. However, from a regional perspective, the annual TGC of the
eastern region is 1, which shows that the eastern region is in the
leading position of China’s NMI. However, the TGC in the central,
western and northeastern regions decreased by 2.32%, 5.03% and
1.61%, respectively. This enlightens us that the three regions should
strengthen the R&D and introduction of advanced technologies
related to eco-efficiency improvement, hence, promote the coor-
dinated regional development of China’s NMI.
4.4. Discussions
The results of regional TGR show that the eastern region is the
leader in eco-efficient technologies for the NMI in China. Eastern
region has national advanced technology and management (Wang
et al. (2013)). In the study of China’s energy efficiency, Wang et al.
(2013) also found that eastern China is at the frontier of national
technology. The decrease of TGR indicates the expansion of the gap
between regional and eco-efficiency best technologies. This re-
minds us that we should focus on economically underdeveloped
regions (Yang and Zhang, 2018), especially the western and
northeastern regions, and strengthen the related technologies
(Zhang et al., 2017) to improve eco-efficiency to narrow the gap
with the global technology frontier.
The decomposition results of eco-inefficiency found that in
addition to technical level, the improvement of management effi-
ciency is also the key to solving eco-inefficiency. Improve the
management level, strengthen the efficiency of technology appli-
cation and raise awareness of energy saving (Wang et al. (2013)) to
resolve the eco-inefficiency of China’s NMI. The rise of regional
TGEIN and the decline of TGR both reveal that the regional average
technology level has fallen relative to the optimal technology level
(the eastern region).
In the DEA-Malmquist framework, the trend of regional eco-
efficiency is obtained. The results show that the eco-efficiency of
China’s NMI keeps rising during the sample period, and the
regional performance was different, and the improvement in the
eastern region was particularly significant. In a similar study using
the DEA framework, China’s energy efficiency (Hu and Wang,
2006), environmental efficiency (Wang et al., 2013) and sectoral
efficiency (such as Fei and Lin, 2016) has been significantly
improved, and the improvement of the eastern region is the most
noticeable. The difference of regional regulation intensity (Bi et al.,
2014), technical level (Fei and Lin, 2017) and management level
(Huang et al., 2014) results in the difference of regional efficiency
improvement.
The three driving factors obtained by decomposing MNMEEI
have regional heterogeneity, which means that their contributions
to regional eco-efficiency change are different (Yu et al., 2016). and
10 X. Chen, B. Lin / Journal of Cleaner Production 277 (2020) 123388
particularly important, Shao et al., 2016 also believes that these two
Table 4 factors are crucial for improving TFP. The investment in advanced
Regional arithmetic mean growth rate of indexes during 2000e2016 (%). technology and the development of clean energy are also effective
Region MNMEEI EC BPC TGC for the NMI to achieve cleaner production (Wang and Zhao, 2017).
Eastern China 9.86 16.29 21.96 0.00
Central China 6.11 6.49 14.05 2.32 5. Conclusions and policy recommendations
Western China 2.74 17.48 7.50 5.03
Northeast China 2.30 0.06 17.04 1.61 5.1. Conclusions
China Average 6.19 15.35 18.46 2.34
this paper are summarized as follows. technological progress have played an obvious role in the
First, the eastern region has the highest technology level in improvement of eco-efficiency in eastern China. Moreover, the
China. The frontier of the eco-efficiency of China’s NMI in the eastern region is at the frontier of technology leadership. Based on
eastern region is the global frontier. The TGR related to eco- the analysis of this paper, the improvement of eco-efficiency in the
efficiency is declining in the other three regions, which directly central and northeastern China is mainly due to the promotion of
indicates that the technology gap is widening relative to the global technological progress, while the western region relies on the
technology frontier, which is worthy of caution. The average eco- improvement of technical efficiency.
inefficiency trend of China’s NMI is declining. The source of eco-
inefficiency in eastern China is mainly management inefficiency.
The central, western and northeastern China all showed an 5.2. Policy recommendations
increasing trend of technology gap inefficiency and a declining
trend of management inefficiency during the sample period. This According to the above conclusions, to achieve environmental
shows that the development and introduction of advanced tech- sustainability and improve the regional eco-efficiency, this paper
nologies are particularly important to reduce the technology gap puts forward the corresponding policy recommendations for
inefficiency and improve the eco-efficiency. improving the eco-efficiency.
Second, the average trend of the eco-efficiency of China’s NMI is Firstly, the technology gap and management inefficiency among
improving both nationally and regionally, but regional heteroge- regions of China’s NMI need to be focused on. To promote the green
neity is obvious. From the decomposition results of the three in- development of NMI, it is particularly important to adjust the
dicators, they have different effects on the improvement of production structure, technology progress and improve the man-
industrial eco-efficiency. Improvements in technical efficiency and agement level. Eastern China is at the frontier of technology lead-
ership and has the best eco-efficiency improvement in China’s NMI,
12 X. Chen, B. Lin / Journal of Cleaner Production 277 (2020) 123388
while the technology gap between the other three regions and considered (Arabi et al., 2016).
eastern China has increased year by year, resulting in the expansion This paper has potential limitations. On the one hand, this paper
of eco-inefficiency. To this end, research related to eco-efficiency considers six major energy categories as energy inputs of NMI. The
improvement should be strengthened, and technology exchange reason is that most provinces only officially reported main energy
and technology transfer between regions should be promoted. The consumption varieties (there are differences in energy varieties
production of NMI should pay more attention to the renewal of reported by all provinces). To ensure the comparability of energy
production capacity rather than expansion (Li et al., 2018; Song consumption among provinces, only six major energy varieties
et al., 2018), and the high-level replacement of inefficient tech- were selected without considering other kinds of energy. On the
nologies to achieve green production capacity update (Song et al., other hand, we regard CO2 emissions as the undesirable output of
2018). Technological progress in the central, western and north- NMI in this paper. In the actual situation, there will be other
east China can narrow the gap with the global technology frontier, emissions such as wastewater, SO2 and so on. Due to the availability
thereby improving the overall eco-efficiency level of China’s NMI. of data, we have no way to obtain these relevant emissions. Based
Secondly, implementing stricter environmental regulation pol- on our discussion, the current research results can also provide
icies can promote enterprises to make more effective use of pro- some reference for the study of the eco-efficiency in NMI.
duction technology and strengthen environmental management. Furthermore, if relevant emissions data such as wastewater and
Reasonable regional regulations can promote the innovation effect solid waste can be obtained in the future, the results of this paper
of enterprises and enhance the industry competition effect (Song will be further improved.
et al., 2018; Lin and Chen, 2020). The improvement of the man-
agement level reduces the regional eco-inefficiency to varying de-
CRediT authorship contribution statement
grees. This paper supports the further improvement of
management efficiency to promote resource conservation and
Xing Chen: Methodology, Software, Data curation, Writing -
emission reduction (Wang et al., 2016) to reduce eco-inefficiency.
original draft. Boqiang Lin: Conceptualization, Methodology,
Thirdly, the central and the northeast regions are close to
Software, Data curation, Writing - original draft.
eastern China, and the geographical advantages of the convenience
of technology learning and talent introduction should be fully
utilized. For the western region, technological innovation is needed Declaration of competing interest
while improving technical efficiency.
Chongqing has incorporated NMI into the carbon emission The authors declare that they have no known competing
trading pilot to promote low-carbon development. The nationwide financial interests or personal relationships that could have
carbon emission trading pilot is worth promoting (Lin and Jia, appeared to influence the work reported in this paper.
2019), because under the condition of the scarcity of total emis-
sions, market participants can be more motivated to tap the po- Acknowledgements
tential of green development (Lin and Jia, 2019), thereby promoting
the improvement of eco-efficiency. The paper is supported by Report Series from Ministry of Edu-
In brief, eco-efficiency is a comprehensive measure of economic, cation of China (No. 10JBG013).
resource and environmental performance. The improvement of
eco-efficiency is related to the sustainability of human develop-
ment. Economic development is important, but improvements in Appendix A
eco-efficiency and regional coordinated development are also
critical. Besides, ecological carrying capacity must also be
Table A
Annual data of variables
Units Labor (L) Energy (E) Capital stock (K) Output (Y) CO2 emissions (C)
10 thousand 10 thousand tce 100 million RMB 100 million RMB 10 thousand tons
Table B
Proportion of CO2 emissions of various energy types in China’s NMI
% Coal Coke Gasoline Diesel oil Fuel oil Electric- Coke oven gas Blast Converter Other Other coking Crude oil
ity furnace gas gas products
2000 20.26 6.34 0.45 1.61 1.60 63.47 0.02 e e 0.64 0.35 0.02
2005 14.65 5.43 0.14 1.10 1.35 71.53 0.28 e e 0.75 0.41 4.83*103
2010 10.48 4.94 0.15 0.75 0.83 77.17 0.30 e 1.8*103 0.50 0.06 0.01
2016 9.08 2.74 0.06 0.38 0.21 82.53 0.14 0.01 e 0.90 0.12 0.01
Keros- Lubricants White Bitumen Petroleum LPG Other petroleum Natural LNG Heat Other energy This paper
ene spirit asphalt coke products gas considered