Energy 144 (2018) 232e242

Contents lists available at ScienceDirect

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

Modeling the point of use EROI and its implications for economic
growth in China
Jingxuan Feng a, Lianyong Feng a, *, Jianliang Wang a, Carey W. King b
a
School of Business Administration, China University of Petroleum (Beijing), Beijing 102249, China
b
Energy Institute, The University of Texas at Austin, 2304 Whitis Ave, C2400, Austin, TX 78712, USA

a r t i c l e i n f o a b s t r a c t

Article history: Energy return on energy invested (EROI) might be considered a measure of net energy, and most current
Received 16 June 2017 studies focus on standard EROI (EROIST). If the energy inputs required to obtain a fuel was extended from
Received in revised form a wellhead to the point of use, the energy delivered decreases, and the energy input of delivering the fuel
3 November 2017
increases. These factors combine to reduce the EROIST to what is referred to as the point of use EROI
Accepted 9 November 2017
Available online 11 November 2017
(EROIPOU). This study calculates the direct and indirect energy (embodied energy) inputs for energy
production sectors (including extraction, processing and delivery) by means of an Input-Output table to
calculate China's EROIPOU and the net energies from 1987 to 2012. Based on calculations in this study, the
Keywords:
Embodied energy
EROIPOU of China's energy production sector declined from 11.01:1 to 5.26:1 between 1987 and 2012. In
Point of use EROI 1987, the energy production sectors consumed 1 ton standard coal equivalent (TCE) energy inputs for
Net energy every 10.01 TCE of produce net energy. However, in 2012, this number declined to 4.25. Additionally, this
China'S economic growth study simulates and forecast economic Gross Domestic Product (GDP) trends in China using net energy
production function. The results reveal how declining EROIPOU for Chinese fossil fuels influence China's
GDP growth.
© 2017 Published by Elsevier Ltd.

1. Introduction energy [4]. Economic development is actually supported by net

Energy production sectors (including energy extraction,
processing and delivery) require energy inputs (also known as
energy costs) from themselves for energy outputs because a
significant amount of energy is needed for processing. An approach
that does not consider the costs of energy (including energy inputs)
cannot provide a proper assessment of energy consumed at the
final end (i.e., the point of use). A comprehensive evaluation should
focus not only on energy resources but also on both the energy
input that is part of the energy production process and the net
energy contribution to this process, particularly the impact of its
size on the economic system [1].
Energy return on investment (EROI) is a concept that originated
in ecology [2] and represents the ratio of the energy output to the
energy input during the process of energy production [3]. This
metric, which evaluates energy and its related problems from the
perspective of the “ratio of net energy”, transforms the gross energy
by positing that what is important to society is the available net
* Corresponding author.
E-mail address: fenglyenergy@163.com (L. Feng).
energy, as opposed to gross energy, because energy is used in the
production of energy itself. The EROI is significant in both theory
and reality because net energy supports economic development.
However, net energy is associated not only with extraction but also
with the entire energy industry (including the energy extraction,
processing and delivery sectors) that delivers useful energy to a
location at which it is used to generate economic goods and
services.
In recent years, significant research has been performed on
EROI, but these studies have tended to not be comparable with
one another because the difference between the boundaries is the
most important and most easily overlooked variable. Most
research on EROI involves calculations based on the extraction of
fossil energy; however, net energy is influenced not only by
extraction but also by the entire energy industry chain. For
example, when the EROI of extraction decreases, the EROI of the
entire energy industry chain may nonetheless be decreasing, sta-
ble or increasing based on the technological development status
(the determinants of EROI include a technological component and
a physical depletion component [5]). Hall further clarified the
boundaries used in EROI calculations into the following categories

https://doi.org/10.1016/j.energy.2017.11.061
0360-5442/© 2017 Published by Elsevier Ltd.
 J. Feng et al. / Energy 144 (2018) 232e242
Fig. 1. Different boundaries used to calculate the different EROIs.
[6,7]: standard EROI (EROIST), point of use EROI (EROIPOU),
extended EROI (EROIEXT), and societal EROI (EROISOC). The calcu-
lation of the EROIST divides the energy output for a project, region
or country by the amount of energy used to generate that output
per year. The EROIPOU is more comprehensive than the EROI
because the EROIPOU includes the energy consumed in extracting,
processing and delivering the energy. The EROIEXT is an expanded
analysis that considers the energy required not only to extract but
also to consume the energy at the final end use of a unit of
energy.(see Fig. 1).
Hall [8] gives the following description:
“A reasonable question is what would be the EROI of a fuel at the
point where it is used, since there may be very different
efficiencies for different fuels between the source and the point
at which it is used. Unfortunately, such studies are rare, and we
must remember to always start with a source of energy from
nature.”
A lower EROI value indicates that a smaller ratio of net energy is
provided for the socio-economic system that is defined outside of
the boundary of the EROI analysis. Using a gross output of 100 units
of energy, if, for example, the EROI is 1.2:1, only 16.7% of the gross
output can be consumed [9]. Lower net energy describes numeri-
cally greater input and thus less net energy remaining to power
society after fossil energy extraction, processing and delivery.
Unsurprisingly, lower net energy negatively affects whether the
economy continues to grow. Researchers generally study the
connection between economic growth and the gross energy supply
as opposed to the net energy [10e13]. However, considering the
long-term structural scarcity in net energy supplies, the energy cost
of obtaining the energy is necessary to calculate the energy con-
straints on economic growth. Thus, net energy plays an important
role in spurring economic growth.
This study aims to build a model to calculate the EROIPOU from
embodied energy perspectives in China. Next, using a simulated
production function, predicts the effects of EROIPOU changes on
China's economic growth from 2016 to 2030 (flowchart see
Fig. 2).
2. Methods and data
2.1. Modeling point of use EROI with embodied energy analysis
2.1.1. The relationship between net energy and EROI for fossil fuels
The basic equation for EROI is as follows:
233
Energy Output Total Primary Energy Supply
EROI ¼ ¼
Energy Input Energy Required to Get That Energy
(1)
The EROI is defined as the total energy output divided by the
sum of all direct and indirect energy inputs, the energy output is the
energy returned to society and the economy, and the energy input
is the energy required to acquire that energy [6]. The energy input
includes direct and indirect energy, i.e., the energy “embodied” in
the energy production process or “embodied energy”. “Embodied”
indicates the sum of all goods and services (direct and indirect) that
are required to produce any goods or services considered as if the
goods and services are incorporated or “embodied” in the product
itself.
Consider the case of a single fuel in which this fuel is used by the
entire economic system. The baseline variables are F, G and C.
Variable F is the total primary energy supply by the entire economy.
G represents the final energy consumption at the point of use that
also could be defined as the net energy that is used at the final end.
C is the sum of all energy inputs in each energy production sector,
and this energy cannot be used by the non-energy portion of the
economic system (see Equation (2)).
G¼FC (2)
The definition of the EROI includes the energy inputs necessary
to invest in producing the net energy. Thus, the EROI can also
broadly be described as the ratio of the energy made available to
society through a certain process to the energy inputs necessary to
implement this process [14,15].
Notably, much research on EROI involves calculating the oil and
gas field extraction processes and does not include final use at the
end, which can lead to a problem, i.e., the EROI and net energy will
be overestimated. The EROI has primarily been used to estimate
changes in the net energy productivities of oil and gas fields
[16e18]. However, the energy inputs not only include oil and gas
extraction but also include separation, washing, coking, and
transport. Many studies on EROI find it difficult to calculate the net
energy at the point of use because the input data for fossil fuels are
difficult to collect.
A metric can be defined in this study that is calculated at
different boundaries if “delivered to society” as the total energy
consumption (i.e., the net energy plus what is used in the energy
production sectors) and assume that the energy “used in the
energy production sectors” is the sum of the energy inputs of the
energy extraction, processing and delivery sectors. Based on these
definitions, the ratio of the energy production sectors EROI (point of
use EROI) is
P
n P
n
Fi Fi
EROIPOU ¼ i¼1
¼ i¼1
(3)
Pn P
n P
n
Ci Fi  Gi
i¼1 i¼1 i¼1
where EROIPOU is the point of use EROI of all energy sectors, and the
P P
net energy from all primary energy is ni¼1 Gi . ni¼1 Fi is the total
energy used by the entire economy (including the energy used in
extraction, processing and delivery). The variable i represents the
P
different types of energy. ni¼1 Ci represents all the types of energy
inputs in the different processes throughout the entire energy
industry chain. Therefore, the equation of net energy can be derived
as follows:
234 J. Feng et al. / Energy 144 (2018) 232e242

Fig. 2. Flowchart for the modeling of the EROIPOU and the evaluation of its impact on economic growth.

  an accounting method that allows one to estimate the sum total of

X
n X
n
1 the energy necessary for an entire product lifecycle.
Gi ¼ Fi  1  (4)
EROIPOU Input-Output tables can be used when calculating the embodied
i¼1 i¼1
energy used by sectors. An Input-Output table is a matrix that dis-
Equations (3) and (4) link the net energy with the EROI. Mearns plays the manner in which industries or sectors interact. The present
[19] has discussed this relation and noted there is a negative study also proposes a calculation method to obtain consumption-
relationship between gross and net energy. The relation between based intermediate input data based on an Input-Output table. The
net and gross energy is called “The Net Energy Cliff”. Input-Output table includes the energy mining or extraction sectors
Measuring the net energy is the real application of the EROI. The and the delivery, separating, coking, processing and other energy
EROI calculation extends from the energy exploration stage to the production sectors. Therefore, the Input-Output table is used to
energy point of use and thus provides a complete estimate of net calculate the energy costs embodied in the energy sectors.
energy, which is significant because when the EROI of an energy- The Input-Output (IO) model is useful for analyzing the
producing process falls below approximately 10:1, the net energy economic relationships of linkages among the sectors of an
delivered to society from that process begins to decline sharply economy and was developed by Leontief [22] in the 1930s. In the
[6,19]. basic IO model, X represents the total output of an economy and can
be expressed as the sum of intermediate consumption (AX) and
final consumption (Y) as follows:
2.1.2. Modeling the point of use EROI for fossil fuels
The EROIPOU is defined as a metric of net energy from the AX þ Y ¼ X (5)
wellhead to the point of use. The EROIPOU can be difficult to
calculate because the complete and accurate energy costs or input where A is the technical coefficient matrix that can be expressed as
data are often hard to obtain because the boundary of the EROIPOU in Equation (6).
is bigger than that of the EROIST. Thus, a method or perspective to 2 3
obtain these data was needed. a11 a12 … a1n
Based on the basic Equation (1), the “energy required to get that 6 a21 a22 … a2n 7
A¼6
4 …
7 (6)
energy” includes direct and indirect energy inputs, i.e., the energy … … … 5
“embodied” in the energy production process. The “total primary an1 an2 … ann
energy supply” is the final energy consumption. The question is
In Equation (6), aij is the technical coefficient that represents
how the energy “embodied” in the energy production sectors can
how much sector j must directly consume of sector i's product to
be calculated.
produce a unit of the product of its sector, which can be calculated
Embodied energy analysis is the process of determining the
as in Equation (7):
energy required directly and indirectly to produce a specific
product or service [20] after considering that the energy was xij
incorporated or “embodied” in the product itself. Embodied energy aij ¼ (7)
xj
primarily refers to the indirect effects described by Bullard and
Herendeen [21]. Moreover, calculating embodied energy requires The solution to Equation (5) can be expressed as follows:
 J. Feng et al. / Energy 144 (2018) 232e242
2 3
B11 B12 … B1n
6 B21 B22 … B2n 7
X ¼ ðI  AÞ 1 6
Y ¼ BY ¼ 4 7Y
… … 1 … 5
Bn1 Bn2 … Bnn
The matrix B or ðI  AÞ1 is called the Leontief inverse matrix,
and it indicates the direct and indirect costs of sector i that are
necessary to produce a unit product of sector j.
Assume that yj is the final demand vector of the energy sector.
Therefore, the embodied energy of the energy sector can be
described as follows:
X
n
Embodied Energy of Energy Sectors ¼ Ej
j¼1
! where Y is the output in GDP; A(t) is the multi-factor productivity;
X
n X
n
¼ ei $Bij $yj
j¼1 i¼1
where ei is the direct energy consumption coefficient that repre-
sents the total energy consumption of sector i divided by total
output of sector i. The unit of ei is tons of standard coal equivalent
(TCE) per yuan, and the data are calculated at the producers' prices.
The Input-Output tables for China that are available for the ten
years of 1987, 1990, 1992, 1995, 1997, 2002, 2005, 2007, 2010 and
2012 [23] and that cover a total of 42 sectors. All Input-Output
tables used in this study were obtained from official national
institutes, and newer tables are not yet available. Therefore, the
study period is from 1987 to 2012.
The data associated with domestic energy consumption by
sector in China were obtained from the Chinese Energy Statistical
Yearbook [24]. The total energy production and consumption data
were obtained from the IEA data online.1 The energy data are given
as the TCEs. The energy consumption data by sector were extracted
from the Chinese Statistics Year Book [25]. The total energy
consumption by each sector can be converted into tons standard
coal equivalent based on the calorific values of the various energy
types.
The Input-Output tables were available for 33 industrial sectors
during the following years: 1987, 1990, 1992 and 1995. The tables
were available for 40 industrial sectors in 1997 and 42 industrial
sectors for the years 2002, 2005, 2007, 2010 and 2012. The energy
consumption data were available for 29 sectors. The economic
sectors were aggregated into 10 sectors to maintain dataset con-
sistency [26]. The similar sectors were merged in this study due to
the mismatch between the sectoral classifications in China's IO
tables and the sectoral classifications in energy consumption.
Similar classifications of energy consumption sectors have previ-
ously been for IO-related research in China [27,28]. Table A.1 in the
appendix presents the different sector classifications and the 10
aggregated sectors.
For analysis, sectors B, C and D are the energy production sectors
and include mining, separating and washing of coal, extraction of
petroleum and natural gas, coking coal and petroleum refining. The
sectors of the economic system are not isolated; for example, the
extraction of oil and gas not only requires the consumption of oil
and gas but also includes steel consumption from another sector.
The embodied energies used by sectors B, C, and D are the direct
and indirect energy inputs of the entire energy industry chain.
1
http://www.iea.org/statistics/statisticssearch/report/?year¼2014&cou
ntry¼CHINA&product¼Balances.
235
2.2. The impact of EROIPOU on economic growth
The establishment of the causal links between the EROIPOU and
economic growth is important in this study for interpreting the
(8)
impact of net energy on economic growth. The energy-economy
link is often omitted from economic growth formulations for
which the production function includes multi-factor productivity.
Energy is one of important inputs for the development of
industries and economic growth. The importance of energy as an
essential ingredient to life on earth is well established in the
physical sciences both in terms of biological energy and as a
primary enabler of industrialization [29]. If energy as an input
factor that includes capital and labor, the multi-factor production
function can be expressed as follows:
Y ¼ f （K; L; E） ¼ AðtÞK k Ll Eg (11)
(9) K is the capital; L is the labor; E is the energy; and g, k and l are
variables with value and the output elasticities of each factor input.
Neoclassical economics apply well to periods with large and
growing energy supplies and relatively low energy extraction costs.
However, economic theories can no longer regard resources as “free
gifts”; indeed, resources are clearly limited [30e32]. Assume gross
fossil fuel extraction does not sufficiently increase due to
environment constraints or fossil fuel production peaks, the
increases in direct and indirect energy inputs in energy production
will eventually lead to a reduction in net energy.
The impact of energy consumption on economic growth is
significant. Energy has become an essential part of production ac-
tivities. To highlight net energy's restrictions on the growth of the
economy, Nel and Cooper [29] built the following energy-type
production:
X 
Y ¼ A0 eat mi ðtÞEth;i  di ðt; Ei Þ (12)
i
where Y represents the economic output as reflected by the GDP,
and A0 eat represents the exponential growth factor. To fulfill the
need for economic growth, assume that capital and production
capacity are sufficient; thus, these factors will be resolved as
endogenous variables. a represents the growth index, t represents
the time in years, Eth;i represents the total energy content of the fuel
type “i”, and mi represents the effective factors or fuel efficiency of
the fuel type “i”. di ðEi ; tÞ represents the energy inputs of the energy
production sectors in t years, which denotes the energy inputs
required to obtain the direct and indirect energy commodities.
P
di ðEi ; tÞ corresponds to ni¼1 ci in Equation (3).
In Equation (12), mi ðtÞ can be represented by a logistic curve or s-
curve to account for long-term (outside the time scales of this
assessment) and past incremental efficiency improvements with
m0 , m1 , c and t0 as constants (see Equation (13)). This was necessary
because efficiency improvement is not only derived from technol-
ogy improvement, but also by changing the modes in which fuels
are used.
On the one hand, efficiency improvement (mi ðtÞ) represents
some technologies or programs used on the demand side, e.g.,
demand side management (DSM). DSM programs are designed to
influence the utility customers' energy utilization for load leveling
and optimization of the energy system [33]. DSM does not create
net energy but rather saves net energy on the demand side. DSM
Moreover, the extent to which DSM reduces the intensity of energy
consumption is limited because, sometimes, additional energy is
required to replace the existing equipment. For example, the results
of some studies have predicted a very large rebound effect on
236 J. Feng et al. / Energy 144 (2018) 232e242
Table 1
Parameters used in Equation (13).
m0 m1 c tj A0
Numerical values 0.1 0.7 0.05 1971 354.66
energy use in 2040 of 76%, i.e., a 10% improvement in energy
efficiency leads to only a 2.4% reduction in energy use [34]. To
consider the efficiency improvement effect, net energy and eco-
nomic growth more generally, the net energy saved by efficiency
improvement is considered in Equation (13) as mi ðtÞ. For the pa-
rameters used in mi ðtÞ in this study, see Table 1.
On the other hand, efficiency improvements are not derived
from technology improvements alone but also from changes in the
modes in which fuels are used [29]. For example, increasing the
percentage of natural gas combined-cycle power plants (which
have greater efficiencies than steam cycles and combustion
turbines alone) used in electricity generation can increase the
productivity of the economy in terms of delivering motive power
and driving electronic equipment [35].
If di ðEi ; tÞ represents the energy inputs for producing and
processing the total energy andEth;i represents the total energy
outputs, then mi ðtÞEth;i  di ðEi ; tÞ represents the net energy supply of
the economic system or the energy return. Based on Equation (1),
the net energy can be calculated with the EROIPOU and total energy
consumption, and by combining Equations (1) and (3), the net
energy production function incorporates net energy and the
EROIPOU:
 
m1 Eth;i
Y ¼ Eq:ð12Þ ¼ A0 eat （m0 þ ）Eth;i 
1 þ ectt0 EROIPOU
Each mi ðtÞ (i.e., the effective factor of the fuel) is time related.
With advances in technology and the optimization of the fuel
consumption structure, the effective factor follows the trend of a
logistic curve.
The known parameters of the historical data from 1987 to 2012
and used the first-order equation of the residuals to obtain the
unknown coefficients in Equation (13) based on the least squares
criterion. The parameters obtain from the simulation are presented
in Table 1.
With the above parameters, the historical GDP and project the
GDP was estimated using Equation (13). As Fig. 8 shows, the actual
GDP and the estimated historical GDP match well. However,
compared with other periods, the simulated GDP from 2005 to
2008 is relatively poor. Since 2005, new policies have been
constantly applied, and the country's instability and variability
make it difficult to accurately simulate GDP growth. The new pol-
icies increased subsidies and control over the health care sector,
accelerated the urbanization process and adopted a loose monetary
policy. China's economic growth since the reform has been very
rapid. China's GDP growth from 1978 to 2013 has been between
9.5% and approximately 11.5% a year.
3. Results and discussion
3.1. Net energy analysis of China's energy production sectors
3.1.1. EROIPOU in China from 1987 to 2012
The EROIPOU values for China between 1987 and 2012 can be
calculated. Fig. 3 presents the EROIPOU values of the Chinese energy
production sectors, which appeared to decline from 1987 to 2012.
The EROIPOU was as low as 5.26:1. However, compared to the values
a
0.03
Fig. 3. EROIPOU for the energy industry in China between 1987 and 2012.
determined in other studies, how low is this value? A comparison of
EROIPOU with the study results from Hu et al. (using the standard
EROI) and Chen et al. (using the emergy-EROI) [36e38] demon-
strated that the EROIPOU in this study was less than the EROIST but
nearly equal to the emergy-EROI because the emergy-EROI and
EROIPOU have greater borders. King et al. [39] used the ‘energy in-
dustry own use’ (EIOU) data from the International Energy Agency
(IEA) as an estimate of the direct energy consumed by the energy
industry to produce energy, e.g., the consumption of oil or gas to
produce the same. These studies do not yet include estimates of
(13) indirect energy inputs that represent the embodied energy in
capital and labor. Following this idea, data in this study obtain from
the IEA2 to calculate the total primary consumption/EIOU in China
from 1987 to 2012 (Fig. 4). The results illustrated that the EROIPOU in
China from 1987 to 2012 exhibited the pattern of change as the
EIOU in China. The EROIPOU is also lower than other EROIST esti-
mates in China. Fig. 4 reveals that, in the long term, the present
indicates that both EROIs exhibit downward trends. The question
that remains is what caused the EROIPOU to decline?
Fig. 5 indicates that the embodied energy used by the Chinese
energy sectors has been rising rapidly. In 1987, the energy pro-
duction sectors consumed 1 ton of TCE energy inputs for every
10.01 TCE of produce net energy. However, in 2012, this number
declined to 4.25. Although the growth of embodied energy used by
the energy production sectors caused the EROIPOU to decline, it is
important to understand which sector is the primary cause. The
decline of the EROIPOU directly leads to a slowing of the growth in
net energy. In 2012, China's total primary energy consumption was
4.29 billion TCE, and the net energy was only 3.47 billion TCE, which
means that approximately 0.82 billion TCE could not be utilized by
the economy and society.
Fig. 5 illustrates the ratio of indirect energy inputs to direct
energy inputs. This ratio revealed the degree of change in indirect
energy inputs relative to direct energy inputs. If this ratio is suffi-
ciently high, the indirect energy inputs require greater percentages
of energy inputs. Figs. 5 and 6 illustrate that the rates of embodied
energy used by sector D (the processing and supply of energy) are
growing far faster than those of the other sectors. The calculation
results further proved that the entire energy production chain,
2
http://www.iea.org/statistics/statisticssearch/report/?year¼2014&cou
ntry¼CHINA&product¼Balances.
 J. Feng et al. / Energy 144 (2018) 232e242 237

40

35

30

EROI From Present Study

25

20

15

10

0
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
EIOU (Total primary energy supply / Energy industry own use)

Emergy-EROI for Daqing oilfield in China by Chen et al. 2017

EROIst of Daqing oilfield in China by Chen et al. 2017

EROIst of Oil and NG in China by Hu et al 2011

EROIst of Coal in China by Hu et al 2011

EROIpou

Fig. 4. Comparison of the EROIPOU with the present EROI study for fossil fuels in China.

500000 140000

450000
120000
Total Primary Energy and Net Energy

400000
Embodied Energy, Unit: 10,000 TCE
350000 100000
Unit: 10,000 TCE

300000
80000
250000
60000
200000

150000 40000
100000
20000
50000

0 0
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012

Indirect Energy Embodied in Extrac on of Petroleum and Natural Gas

Direct Energy Embodied in Extrac on of Petroleum and Natural Gas
Indirect Energy Embodied in Mining and Washing of Coal
Direct Energy Embodied in Mining and Washing of Coal
Indirect Energy Embodied in Processing and Supply of Energy
Direct Energy Embodied in Processing and Supply of Energy
Total Primary Energy Supply
Net Energy Supply

Fig. 5. The gaps between the net energy and total primary energy consumption in China from 1987 to 2012 (unit: 10,000 TCE).
238 J. Feng et al. / Energy 144 (2018) 232e242

2.5
Mining and Washing of Coal
Extrac on of Petroleum and Natural Gas 2.210

Indirect Energy Inputs / Direct Energy Inputs

Processing and Supply of of Energy
2

1.727
1.607 1.581 1.616
1.5
1.290
1.229 1.218 1.234

1 0.977

0.5 0.455
0.327 0.359
0.260 0.264 0.254
0.167 0.197 0.171 0.189
0.119 0.150
0.121 0.141 0.110
0.088
0.083 0.069 0.087 0.093
0
1987 1990 1992 1995 1997 2002 2005 2007 2010 2012
Fig. 6. Ratio of indirect energy inputs to direct energy inputs for energy production sectors in China.

especially sectors D, cannot be ignored in the EROI calculation 1987 and 2012, a large proportion of the energy flow was from
when calculating net energy in the energy industry. sector G to D. Furthermore, sector G consumed a large amount of
energy from sectors B, C, and D, which means that sector D
3.1.2. Embodied energies among the energy production sectors consumed a large amount of indirect energy. Simultaneously, from
Table 2 and Fig. 5 present the direct and indirect embodied 1987 to 2012, all the energy production sectors exhibited signifi-
energies used by 3 energy production sectors. The largest and cant increases in embodied energy consumption.
fastest-growing consumer of embodied energy in China is sector D,
which is the processing and supply of energy.
3.2. Forecast of China's GDP growth considering the declining
Fig. 7 illustrates the energy flows among the sectors of energy
EROIPOU
production in China. According to the widths of the flows, the
amount of flow between sectors can be discerned. The directions of
3.2.1. Setting up the future scenarios
the flows indicate the roles of the sectors in the energy network.
If the energy EROIPOU continues to decline, energy inputs in the
Links of variable thickness can represent the extent or magnitude of
energy production sectors will continue to increase, which will
the relationships between the elements. The data for the links with
result in the energy industry being able to provide a net energy of
variable thicknesses come from an aggregated Input-Output table.
zero. Eventually, net energy limitations will be reflected in the
From the directions of the flows, in 2012, sector D was dependent
weakening of the economy and in the depletion of resources.
on sector G. Moreover, sectors B and D were relatively self-energy-
Therefore, the foundations of economic theories each have their
sufficient in 2012 (e.g., they used their own inputs), but in 1987,
their self-energy-sufficiency was almost negligible.
By comparing the flows of different sectors, the result shows
that sector D is the sector that consumed the largest quantity of
embodied energy and played an important role in the energy
production chain; moreover, this trend is still strengthening. In

Table 2
Embodied energy used by the Chinese energy sectors (10,000 TCE).

Year Mining and Extraction of Processing and

Washing of Coal Petroleum and Supply of Energy
(sector code: B) Natural Gas (sector (sector code: D)
code: C)

Direct Indirect Direct Indirect Direct Indirect

1987 3433.0 302.9 1643.0 136.6 2138.0 2626.8

1990 4159.0 495.6 1930.0 133.0 2508.0 4030.6
1992 4427.2 664.7 2106.5 255.8 3188.1 3882.4
1995 5499.8 475.8 2812.6 397.2 5908.6 5775.3
1997 5791.2 966.5 3564.9 331.4 7815.6 9644.5
2002 5132.8 1009.0 4680.6 515.6 8612.8 13615.7
2005 6711.7 1744.5 3763.4 643.8 12507.4 16138.8
2007 8269.9 2179.7 3651.3 1195.3 13449.3 21738.1
2010 10574.4 4809.6 4057.6 1030.2 16694.8 28834.7
2012 12339.1 4430.9 3807.9 718.1 18809.1 41571.6
Fig. 7. Energy flows among the energy production and other sectors.

140000
120000
Billion CNY Constant LCU
100000
80000
60000
40000
20000
Fig. 8. Results of the forecast of China's GDP growth (Local Current Unit, LCU) with its confidence band.
appropriate time of applicability and periodicity. If we neglect the
background of the net energy theory, the neoclassic economic
theory that promotes the growth of the economy may impede
subsequent growth [40].
The first step to use a net energy production function (see
Equation (4)) to forecast the future GDP is to forecast the TPES. The
projections of the TPES from the International Energy Agency (IEA)
were used in this study.
The new policy scenario of the World Energy Outlook 2016
(WEO2016) was published by the IEA and incorporates existing
energy policies and an assessment of the results that are likely to
stem from the implementation of announced intentions, including
those in the climate pledges submitted to the United Nations
Climate Change Conference in 2015. The climate pledges, known as
Nationally Determined Contributions (NDCs), are the building
blocks of the Paris Agreement and provide a rich and authoritative
source of guidance for this scenario. The NDCs have been carefully
and individually assessed in this edition of the WEO2016. Where
policies exist to support the NDCs, the implementation measures
are clearly defined, and the effects are reflected in the new policies
scenario. Additionally, this scenario considers some of the policies
that are being implemented in further efforts by the Chinese gov-
ernment to reform the economic and energy structure, for example,
efforts to achieve peak CO2 emissions by approximately 2030 and to
expand the use of natural gas and renewable energy [41].
Additionally, the EROIPOU trends from 1987 to 2030 presented in
Fig. 3 were used to support the decline of the EROIPOU in China. The
function and parameters of the curve-fitting model are also pre-
sented in Fig. 3.
3.2.2. The effect of the decline of the EROIPOU of fossil fuels on
economic growth
Based on the new policies scenario from the WEO2016 and the
declining EROIPOU scenarios presented in this study, trends for the
GDP from 2016 to 2030 as presented in Fig. 8 for net energies. The
EROIPOU trend line in Fig. 3 was used to create basic forecasts.
Although the actual GDP generally trends upward, its growth rate
over five-year periods is in a downward trend, as we demonstrate
in Table 3.
J. Feng et al. / Energy 144 (2018) 232e242 239
95% Confidence Band of GDP Growth
Actual GDP (constant LCU)
Simulated GDP with decline EROIpou
Comparisons of the results of this study with those of other
studies reveal that the GDP growth rates forecast in this study from
2016 to 2030 are lower than the predictions of the World Bank
Global Economic Prospects (6.3% in 2019) [42] and the IMF World
Economic Outlook Database (5.9% in 2020, 5.7% in 2022) [43],
which includes the U.S. Energy Information Administration (EIA)
IEO-2016 (5.3% in 2020,4.2% in 2030, reference case) [44] and the
International Energy Agency WEO-2016 (5.8% in 2021) [41].
Because fossil fuel constraints and the energy inputs of growth in
energy production, GDP forecast for 2016 to 2030 in this study is
slower than the results predicted by other researchers. One of the
primary reasons for these differences is that the World Bank, IMF,
EIA and IEA have not considered the decline of the EROIPOU for fossil
fuels in China.
However, for the entire energy industry, the EROIPOU can be
increased by adding a high EROI energy ratio because a high EROI
energy uses fewer inputs to obtain the identical (or additional)
output. Increasing the proportion of gas used in electricity gener-
ation thereby increases its productivity in the economy in terms of
delivering power and driving electronic equipment [45].
However, the effects of fossil energy growth inputs on economic
growth cannot be ignored. Inevitably, rising energy inputs will
reduce the net energy supply if energy continues to be dominated
by fossil energy, which will affect economic growth. Additionally, as
the energy EROI weakens, the net energy supply will be further
reduced such that the decline in the EROI will substantially impact
economic growth.
4. Conclusion
The net energy from fossil fuels is the real contribution to the
social economy that is related to the GDP. However, the declining
Table 3
Forecast of the rate of change of the actual GDP in China from 2016 to 2030.
2016e2020 2021e2025 2025e2030
EROIPOU 4.88 4.55 4.27
GDP growth 4.51% 4.33% 3.99%
240 J. Feng et al. / Energy 144 (2018) 232e242

EROIPOU will reduce the net energy according to the proportion of Acknowledgements
total fossil fuel consumption. This study aimed to build a model to
calculate the EROIPOU in China from embodied energy perspectives. This study was supported by the National Natural Science
Then, simulated net energy production function was used to fore- Foundation of China (501100001809) (grant No. 71373285/
cast the effects of the declining EROIPOU on China's economic 71503264/71303258), the National Social Science Foundation of
growth from 2016 to 2030. China (grant No. 11&ZD164/13&ZD159) and the Humanities and
The EROI method focuses on physical energy and reveals the Social Sciences Foundation of the Ministry of Education of China
importance of net energy in calculating the outputs and inputs. This (501100002338) (Grant No. 15YJC630121).
study calculated the indirect and direct energies (including the Additionally, we thank Carey King for his many helpful
embodied energies) used by the energy production sectors with an comments.
Input-Output table. The mathematical relationships among the
EROIPOU, net energy and total primary energy consumption is core
in this study. Then simulated and predicted economic GDP trends in Nomenclature/list of symbols
China using a net energy production function. The results indicate
the following: CNY China Yuan
DSM Demand Side Management
1) Fossil fuels has long processing and supply chain convert pri- EIA Energy Information Administration
mary energy source into secondary energy. Excessive embodied EIOU Energy Industry's Own Use
energy consumption by the fuel processing and supply energy EROI Energy Return on Investment
sector causes the EROIPOU to decline. The embodied energies EROIPOU Point of use EROI
from the processing and supply of energy are larger than those EROIST Standard EROI
of the fossil fuel extraction and mining stages. Clearly, the in- EROIEXT Extended EROI
direct energy cost of energy processing (after extraction) cannot FFC Full Fuel Cycle
be ignored. GDP Gross Domestic Product
2) The slowing economic growth enlightens our sense of the GHG Greenhouse Gas
considerable importance of energy inputs or the costs in the IEA International Energy Agency
energy production sectors. The impact of net energy on China's IEO International Energy Outlook
economy is significant, as determined by the ratio of energy cost LCU Local Current Unit
in production to energy produced. Over the long term, TCE Tons Standard Coal Equivalent
decreasing energy quality and increasing energy costs have NDC Nationally Determined Contributions
significant and increasing influences on economic growth. TPES Total Primary Energy Supply
China's energy-related policy-makers needs strengthen the USD United States Dollar
deep policy analysis regarding the relationship between GDP WEO World Energy Outlook
and net energy. For example, a policy that aims to decrease the A technical coefficient matrix [dimensionless]
embodied energy in equipment and buildings is necessary to A(t) multi-factor productivity [dimensionless]
control energy costs. B Leontief inverse matrix [dimensionless]
C sum of all energy inputs in each energy production sector
Net energy also concerns human welfare and a sustainable [10,000 TCE]
environment, which should be part of a nation's policy emphasis. c constants in logistic curve of efficiency improvements
Present-day human welfare cannot simply be measured by GDP [dimensionless]
alone [46]. With the declining EROI values of traditional major fossil di (Ei,t) energy inputs of the energy production sectors in t years
fuel energy sources in China, China's fossil fuels industry may be [dimensionless]
moving toward the “net energy cliff” [5]. Alternatively, decreasing E energy input [10,000 TCE]
the net energy will increase the capital and technological costs, and ei direct energy consumption coefficient of sector I
China will experience less growth. Consequently, the negative [dimensionless]
impact on welfare improvement in developing countries such as Ei embodied energy of sector i [10,000 TCE]
China will be greater than that in developed countries. Eth,i total energy content of the fuel type “i” [10,000 TCE]
With rapid economic growth, human welfare and quality of life F total primary energy supply [10,00 TCE]
have been prioritized by the Chinese government. A previous study G final energy consumption at the point of use [10,000 TCE]
[5] found that societal well-being and the per capita net energy i different types of energy
available to society appear to be linked to quality of life based on K capital input
empirical analyses relating societal EROI to indices of quality of life. k output elasticities of capital [dimensionless]
Therefore, it is necessary to review our work from a Chinese L labor input
perspective that is informed by other analyses, including those of l output elasticities of labor [dimensionless]
Chinese societal well-being, net energy availability and EROI, to m0 constants in logistic curve of efficiency improvements
develop China's welfare and quality of life with rational net energy [dimensionless]
use and distribution. m1 constants in logistic curve of efficiency improvements
Therefore, China should consider energy efficiency policies and [dimensionless]
industrial restructuring policies, among others, to eliminate or mi(t) efficiency improvements [dimensionless]
mitigate excessive natural resource exploitation in the environ- X total output of an economy matrix [dimensionless]
ment. In the future, we expect more developed technology to ease Y final consumption matrix [dimensionless]
fossil fuel depletion and the development of alternative energies. a constants in logistic curve of efficiency improvements
Moreover, for substitutes, a new evaluation method to compre- [dimensionless]
hensively consider energy, the economy, and the environment aij technical coefficient j consumed i [dimensionless]
must urgently be developed. g output elasticities of energy [dimensionless]
 J. Feng et al. / Energy 144 (2018) 232e242 241

Appendix A

Table A1
Sector Classifications of the Chinese Economic Sectors.

Sector Sector Sectoral Sector Name

Code Code

A Agriculture, Hunting, Forestry and Fishing 01 Agriculture, Hunting, Forestry and Fishing
B Mining and Washing of Coal 02 Mining and Washing of Coal
C Extraction of Petroleum and Natural Gas 03 Extraction of Petroleum and Natural Gas
D Processing and Supply of Energy 04 Processing of Petroleum, Coking and Processing of Nuclear Fuel
05 Production and Distribution of Gas
E Production and Supply of Electric Power and Heat Power 06 Production and Supply of Electric Power and Heat Power
F Other Manufacturing 07 Mining of Metal Ores
08 Mining and Processing of Nonmetal Ores and Other Ores
09 Food and Tobacco
10 Manufacture of Textiles
11 Textiles, Apparel, Footwear, Caps, Fur, Leather
12 Processing of Timber and Manufacture of Furniture
13 Study making, Printing and Manufacture of Articles for Culture, Education and Sports
14 Chemicals Industry
15 Nonmetallic Mineral Products
16 Smelting and Rolling of Metals
17 Metal Products
18 General Machinery Purpose and Special Purpose Machinery
19 Transport Equipment
20 Electrical Machinery and Equipment
21 Communication, Computer and Other Electronic Equipment
22 Measuring Instrument and Machinery for Cultural Activity and Office Work
23 Other Manufacture
24 Scrap and Waste
25 Production and Distribution of Water
G Construction 26 Construction
H Transport, Storage and Post 27 Transport, Storage
28 Post
I Wholesale and Retail Trade, Hotels and Catering Services 30 Wholesale and Retail Trade,
31 Hotels and Catering Services
J Other Service Industries 29 Information Transmission, Computer Service and Software
32 Financial Intermediation
33 Real Estate
34 Leasing and Business Services
35 Research and Experimental Development
36 Comprehensive Technical Services
37 Water Conservancy, Environment and Public Facilities
38 Services to Households and Other Services
39 Education
40 Health, Social Security and Social Welfare
41 Culture, Sports and Entertainment
42 Public Management and Social Organization

Table A2
Energy Consumption by Chinese Economic Sectors (10000 TCE).

Years Agriculture, Mining Extraction of Processing Production and Other Construction Transport, Wholesale and Retail Other Total
Hunting, and Petroleum and and Supply Supply of Electric Manufacturing Storage Trade, Hotels and Service
Forestry and Washing Natural Gas of Energy Power and Heat and Post Catering Services Industries
Fishing of Coal Power

1987 4471.0 3433.0 1643.0 2138.0 2969.0 48609.0 1260.0 4126.0 907.0 17076.0 86632.0
1990 4852.0 4159.0 1930.0 2508.0 3867.0 55113.0 1213.0 4541.0 1247.0 19273.0 98703.0
1992 4982.6 4427.2 2106.5 3188.1 4504.1 59074.0 1237.3 4805.4 1401.2 19471.2 105197.6
1995 5505.1 5499.8 2812.6 5908.6 7052.7 74918.2 1334.5 5862.9 2017.8 20263.8 131176.0
1997 5905.4 5791.2 3564.9 7815.6 10076.4 72832.3 1179.0 7543.1 2394.4 21070.8 137799.0
2002 6612.5 4591.2 4550.3 15321.7 11807.0 80160.0 2544.0 11171.0 3520.0 24407.0 164684.7
2005 7978.4 6711.7 3763.4 12507.4 15826.8 119990.3 3411.1 16629.2 5031.1 32832.7 224682.0
2007 8244.6 7170.8 3677.5 13792.9 18474.6 146249.8 4031.4 20643.4 5962.1 37335.8 265582.9
2010 6477.3 10574.4 4057.6 16694.8 22584.1 177137.6 5309.3 26068.5 6826.8 49208.8 324939.1
2012 6784.4 12339.1 3807.9 18809.1 23809.2 192582.2 6167.4 31524.7 8545.9 57362.2 361732.0
242 J. Feng et al. / Energy 144 (2018) 232e242

References 1453448.html.
[24] National Bureau of Statistics of China (NBSC). China energy statistical year-
re JH. The end of cheap oil. Sci Am 1998;278:78e83. book (1986-2014). Beijing: China Statistics; 2015.
[1] Campbell CJ, Laherre
[25] National Bureau of Statistics of China (NBSC). Chinese statistics year book
[2] Cleveland CJ. Biophysical economics: historical perspective and current
(1987e2012). Beijing: China Statistics; 2012.
research trends. Ecol Model 1987;38:47e73.
[26] Zhu L, Yong G, Lindner S, Zhao H, Fujita T, Guan D. Embodied energy use in
[3] Hall CAS. Migration and metabolism in a temperate stream ecosystem. Ecol-
China's industrial sectors. Energ Policy 2012;49:751e8.
ogy 1972;53:585e604.
[27] Chen G, Zhang B. Greenhouse gas emissions in China 2007:inventory and
[4] Felix E, Tilley DR. Integrated energy, environmental and financial analysis of
inputeoutput analysis. Energ Policy 2010;38:6180e93.
ethanol production from cellulosic switchgrass. Energy 2009;34:410e36.
[28] Chen Q, Kang C, Xia Q, Guan D. Primary exploration on low-carbon technology
[5] Lambert JG, Hall CAS, Balogh S, Gupta A, Arnold M. Energy, EROI and quality of
road map of China's power sector. Energy 2011;36:1500e12.
life. Energ Policy 2014;64:153e67.
[29] Nel WP, Cooper CJ. Implications of fossil fuel constraints on economic growth
[6] Hall CAS, Balogh S, Murphy DJR. What is the minimum EROI that a sustainable
and global warming. Energ Policy 2009;37:166e80.
society must have? Energies 2009;2:25e47.
[30] Dale M, Krumdieck S, Bodger P. Net energy yield from production of con-
[7] Hall CAS, Lambert JG, Balogh SB. EROI of different fuels and the implications
ventional oil. Energ Policy 2011;39:7095e102.
for society. Energ Policy 2014;64:141e52.
[31] Dale M, Krumdieck S, Bodger P. Global energy modelling d a biophysical
[8] Hall CAS. Energy return on investment: a unifying principle for biology, eco-
approach (GEMBA) part 2: methodology. Ecol Econ 2012;73:158e67.
nomics, and sustainability. Cham: Springer International Publishing; 2017.
[32] King CW, Hall CA. Relating financial and energy return on investment. Sus-
[9] Murphy DJ, Hall CAS, Dale M, Cleveland CJ. Order from chaos: a preliminary
tainability 2011;3:1810e32.
protocol for determining the EROI of fuels. Sustainability 2011;3:1888e907.
[33] Ardakani FJ, Ardehali MM. Novel effects of demand side management data on
[10] Nathwani JS, Siddall E, Lind NC. Energy for 300 years. Waterloo, Canada:
accuracy of electrical energy consumption modeling and long-term fore-
Institute for Risk Research, University of Waterloo; 1992.
casting. Energ Convers Manage 2014;78:745e52.
[11] Dincer I. Energy and GDP analysis of OECD countries. Energy Convers Manag
[34] Wei T, Liu Y. Estimation of global rebound effect caused by energy efficiency
1997;38:685e96.
improvement. Energ Econ 2017;66:27e34.
[12] Messner S, Schrattenholzer L. MessageeMacro: linking an energy supply
[35] Adams FG, Miovic P. On relative fuel efficiency and the output elasticity of
model with a macroeconomic module and solving it iteratively. Energy
energy consumption in Western Europe. J Ind Econ 1968;17:41e56.
2000;25:267e82.
[36] Hu Y, Feng L, Hall CAS, Tian D. Analysis of the energy return on investment
[13] Ohler A, Fetters I. The causal relationship between renewable electricity
(EROI) of the huge daqing oil field in China. Sustainability 2011;3:2323e38.
generation and GDP growth: a study of energy sources. Energ Econ 2014;43:
[37] Hu Y, Hall CAS, Wang J, Feng L, Poisson A. Energy Return on Investment (EROI)
125e39.
of China's conventional fossil fuels: historical and future trends. Energy
[14] Cleveland CJ, O'Connor PA. Energy return on investment (EROI) of oil shale.
2013;54:352e64.
Sustainability 2011;3:2307e22.
[38] Chen Y, Feng L, Wang J. Energy-based energy return on investment method
[15] Lundin J. EROI of crystalline silicon photovoltaics: variations under different
and its application in Daqing oilfield. Resour Sci 2016;38(12):2270e82 (in
assumptions regarding manufacturing energy inputs and energy output.
Chinese).
Uppsala Universitet; 2013.
[39] King C, Maxwell JP, Donovan A. Comparing world economic and net energy
[16] Cleveland CJ. Net energy from the extraction of oil and gas in the United
metrics, part 2: total economy expenditure perspective. Energies 2015;8:
States. Energy 2005;30:769e82.
12975e96.
[17] Gagnon N, Hall CAS, Brinker L. A preliminary investigation of energy return on
[40] Hall CAS, Klitgaard KA. Energy and the wealth of nations. New York: Springer;
energy investment for global oil and gas production. Energies 2009;2:
2012.
490e503.
[41] International Energy Agency. World energy outlook 2016. 2016. http://www.
[18] Gately M. The EROI of U.S. offshore energy extraction: a net energy analysis of
worldenergyoutlook.org/publications/weo-2016/ [Accessed November 2016].
the Gulf of Mexico. Ecol Econ 2007;63:355e64.
[42] World Bank. China 2030: building a modern, harmonious, and creative high-
[19] Mearns E. The global energy crises and its role in the pending collapse of the
income society. Washington, DC: World Bank; 2012. http://documents.
global economy. Aberdeen, Scotland: Royal Society of Chemists; 2008.
worldbank.org/curated/en/781101468239669951/pdf/
October 29th, 2008.
762990PUB0china0Box374372B00PUBLIC0.pdf.
[20] International Federation of Institutes for Advanced Study (IFIAS). Energy
[43] International Monetary Fund (IMF). World economic outlook 2016 database.
analysis workshop report no. 6. Stockholm, Sweden: International Federation
2016. http://www.imf.org/external/pubs/ft/weo/2017/01/weodata/index.
of Institutes for Advanced Study; 1974.
aspx.
[21] Bullard CW, Herendeen RA. The energy cost of goods and services. Energ
[44] Energy Information Administration of U.S. (EIA). International energy outlook
Policy 1975;3:268e78.
with projections to 2040. 2016. https://www.osti.gov/scitech/servlets/purl/
[22] Leontief W. Quantitative input-output relations in the economic system of the
1296780.
U.S. Reviews of economies and statistics. 1936. p. 18.
[45] Nel Christopher WP, Cooper J. Implications of fossil fuel constraints on eco-
[23] National bureau of statistics of China (NBSC). Input-output tables of China,
nomic growth and global warming. Energ Policy 2009;37:166e80.
1987; 1990; 1995; 1997; 2002, 2007, 2012. Beijing: China Statistics; 2012.
[46] van den Bergh JC. The GDP paradox. J Econ Psychol 2009;30:117e35.
http://www.stats.gov.cn/ztjc/tjzdgg/trccxh/zlxz/trccb/201701/t20170113_