Quantifying regional economical impacts of the CO2

intensity reduction target allocation in China

Da Zhang∗ Sebastian Rausch† Valerie Karplus‡ Tianyu Qi§ Xiliang Zhang¶

Version 2: May 4, 2012

Abstract

To address concerns about rising energy use and CO2 emissions, China’s leadership hasdefined national and disaggregated provincial energy and CO2 intensity targets under theTwelfth Five-Year Plan (2011-2015). An important consideration in this process has beenhow the reduction burden and associated costs of policy will be distributed across China’sprovinces. By developing and applying a static regional model of China’s economy (theChina Regional Energy Model (C-REM)) built on the 2007 regional input-output tables forChina and the GTAP 8 global data set, we perform a preliminary assessment of the impactof provincial-level CO2 emissions intensity targets under the Twelfth-Five Year Plan. Wefocus on the impacts of meeting these targets on CO2 emissions intensity, CO2 emissions,and economic welfare at the national and provincial level under two policy scenarios. Wefurther compare the impact of imposing a single national carbon intensity reduction targetto the impact of current provincially-disaggregated targets. We find that the single nationalcarbon intensity reduction target has less significant welfare loss at the national level, andsubstantial differences in the magnitude and sign of these impacts across provinces.

Keywords: Intensity target, targets allocation, climate policy, computable generalequilibrium modeling.JEL classification: C68, F18, Q54, R13.

∗Tsinghua University, Institute of Energy, Environment, and Economy, China, and Massachusetts Institute of Tech-nology, Joint Program on the Science and Policy of Global Change, Cambridge, MA, USA. Email: zhangda@mit.edu.†Massachusetts Institute of Technology, Joint Program on the Science and Policy of Global Change, Cambridge,

MA, USA. Email: rausch@mit.edu.‡Massachusetts Institute of Technology, Joint Program on the Science and Policy of Global Change, Cambridge,

MA, USA. Email: vkarplus@mit.edu.§Tsinghua University, Institute of Energy, Environment, and Economy, China, and Massachusetts Institute of Tech-

nology, Joint Program on the Science and Policy of Global Change, Cambridge, MA, USA. Email: qitianyu@mit.edu.¶Tsinghua University, Institute of Energy, Environment, and Economy, China. Email: zhang_xl@tsinghua.edu.cn.

1

1 Introduction

Over the past decade China’s policy has signaled strong intentions to reduce the country’s grow-

ing energy and CO2 emissions footprint. Sustained rapid growth in China has brought great

benefits but has also sharpened concerns about energy security, air quality, and the acceleration

of global climate change caused by human activities. Increasingly, China’s comprehensive Five-

Year Plans, which lay out the government’s priorities and program of work every five years, have

reflected these concerns. Most recently, China’s Twelfth Five-Year Plan (2011-2015) has, for the

first time, introduced a national target for reducing the nation’s carbon intensity by 17% over

the period 2011 to 2015, in line with the nation’s commitment at the 2009 Copenhagen Summit

to reduce its CO2 emissions intensity by 40-45% over the period 2005 to 2020. This national

carbon intensity target has been disaggregated at the provincial level (State Council, 2012).

While meeting these targets is mandatory, their existence does not by itself create incentives

for firms and households across China to reduce CO2 emissions intensity. To meet these short-

and medium-term policy targets, China’s policy makers have announced a range of programs to

support target attainment. These include an industrial energy efficiency mandate, targets for the

deployment of renewable and nuclear electricity generation, and reduced subsidies to China’s

energy-intensive, export-oriented sectors. Also in the early stages is a pilot cap-and-trade system

for CO2 emissions for a subset of China’s provinces. An energy cap is also under discussion.

Alongside economic growth and sustainable development, promoting inter-regional equity

remains among the top concerns of China’s policy makers. Answering the critical question of

which policy mechanism(s) will enable China to meet its targets at the least cost, without ex-

aggerating the welfare gap across China’s provinces, requires a modeling framework capable of

resolving policy impacts at the provincial level. In this paper, we first describe how we have

developed a new energy-economic computable general equilibrium model that disaggregates

China into 30 provinces connected by inter-provincial trade flows. The model includes signif-

icant detail in the energy system and quantifies energy-related CO2 emissions. We then apply

this new tool to perform an initial analysis of China’s CO2 intensity targets.

This paper is organized as follows. In Section 2, we summarize previous studies and identify

the contribution of this work, as well as provide some background on China’s CO2 intensity

2

target and the allocation of reductions to each of China’s provinces. In Section 3, we describe

briefly how we have established the new model, including model structure, data preparation,

representation of inter-provincial trade, integration with the 2007 edition of the Global Trade

Analysis Project, and the incorporation of supplemental CO2 emissions accounting. In Section

4, we describe the results of the two policy scenarios. Section 5 discusses some preliminary

conclusions and topics for future investigation.

2 Background on CO2 Intensity in China and Policy Targets

2.1 Previous work

Our study builds on previous work that has used computable general equilibrium models to

understand how China’s economy will respond to energy and environmental policy. In partic-

ular we focus on the distributional effects of policy, using a new model that is disaggregated

at the provincial level. Previous research has employed single-region models of China to fo-

cus on the impacts of carbon mitigation measures (Cao, 2007; Wang, Wang and Chen, 2009).

Other analyses have used models with various levels of regional disaggregation to investigate

a wide range of economic issues (Horridge and Wittwer, 2008; Li, Shi and Wang, 2009; Wang

and Ezaki, 2006; Li and He, 2005; Xu and Li, 2008). Wei, Ni and Du (2011) estimate CO2

reduction potential and marginal abatement costs by province in a model using a distance func-

tion approach. Li and He (2010) are among the few to analyze carbon mitigation policy in a

regionally-disaggregated CGE model. However, these models are mostly based on older input-

output data (e.g. 2002 input-output data of China) and lack the flexibility to choose the desired

regional aggregation and do not include physical accounting in the energy sector. Moreover, as

they are not integrated with any global trade data set, these models treat China as a small or

large open economy, which can significantly affect the reliability of simulation results.

2.2 Description of the Twelfth Five-Year Plan CO2 Intensity Target Allocation

China’s primary policy initiative to reduce energy and CO2 emissions takes the form of intensity

targets, defined as the allowable energy consumption or emissions per unit of GDP. Prior to the

Twelfth Five-Year Plan (2011-2015), policy was focused on energy intensity. The Eleventh Five-

3

Year Plan (FYP) included the first target for energy intensity reduction of 20% nationwide. This

target was not formally allocated to provinces, although provinces made non-binding pledges

to undertake a certain level of reductions at the outset of the policy. At the conclusion of the

Eleventh FYP, China’s leaders officially declared that a 19.1% reduction in energy intensity was

achieved (Industrial Efficiency Policy Database, IEPD). Much of the reduction was thought to be

achieved by energy efficiency improvements in heavy industry (much of it said to be achieved

through an initiative called the 1,000 Enterprises Program) and the closure of small, inefficient

industrial and power generation facilities (Price, Wang and Yun, 2010; Price and et.al., 2011).

A CO2 intensity target was formally introduced for the first time under the Twelfth FYP, with

a target of 17% (State Council, 2012). Much of reduction in CO2 intensity over this period is ex-

pected to come from reductions in energy intensity (through further improvements in industrial

energy efficiency as well a shift in economic structure away from energy-intensive industries),

as well as the further introduction of low carbon electricity sources into China’s electric power

generation mix. For the first time, binding targets for CO2 emissions reductions were assigned

at the provincial level. These targets are shown in Table 1:

Table 1: CO2 Intensity reduction targets across provinces of mainland China.

Carbon intensity Provincesreduction target (%)

19.5 Guangdong19 Tianjin, Shanghai, Jiangsu, Zhejiang18 Beijing, Hebei, Liaoning, Shandong17.5 Fujian, Sichuan17 Shanxi, Jilin, Anhui, Jiangxi, Henan, Hubei, Hunan, Chongqing, Shannxi16.5 Yunan16 Neimenggu, Heilongjiang, Guangxi, Guizhou, Gansu, Ningxia11 Hainan, Xinjiang10 Xizang

A driving principle behind the allocation is to assign reduction burdens according to provin-

cial wealth, which is expected to ease pressure on less affluent regions or regions targeted for

accelerated development given significant heterogeneity across provinces (see Figures 1,2,3).

We investigate the impacts of alternative target allocation scenarios in an integrated modeling

framework, as it allows us to implement alternative policy designs and examine the economic

4

5000

10000

15000

20000

25000

30000

35000

Figure 1: GDP of China’s provinces in 2007 (100 million yuan).

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Figure 2: CO2 emission of China’s provinces in 2007 (100 million tons).

outcomes, as well as the impact on energy and CO2 emissions.

3 Modeling framework

3.1 Data

This study makes use of a comprehensive energy-economy data set that features a consistent

representation of energy markets in physical units as well as detailed accounts of regional pro-

duction and bilateral trade for the year 2007. The dat set merges detailed provincial-level data

for China with national economic and energy data for regions in the rest of the world. Social

accounting matrices (SAMs) in our hybrid data set are based on data from the Global Trade

Analysis Project (GTAP, 2012) and China’s National and the full set of China’s 2007 provin-

5

2

3

4

5

6

7

8

9

10

11

Figure 3: CO2 emission intensity of China’s provinces in 2007 (ton/10,000 yuan).

cial Input-Output Tables (National Infortion Center, 2011). Energy use and emissions data is

based on data from GTAP and China’s National and Provincial Energy Statistical Yearbook 2007

(National Statistics Bereau, 2008).

The GTAP 8 data set provides consistent global accounts of production, consumption, and bi-

lateral trade as well as consistent accounts of physical energy flows, energy prices and emissions

in the year 2007, and identifies 129 countries and regions and 57 commodities (GTAP, 2012).

The China’s National and Provincial Input-Output data specifies benchmark economic ac-

counts for the 30 China provinces (except for Tibet because of the small scale of its economic

activities). The data set consists of input-output tables for each province. Each table identifies

the forward and backward linkages associated with production of 42 commodities and existing

taxes. Based on these input-output tables, we established our SAM tables for each province after

some minor adjustments1 and updates2 for balancing. We applied the following least-squares

optimization problem to obtain the balanced SAM tables for each province p (see Table 2):

min{xpij}∑

i,j(xpij −Xpij)2 + PENALTY

∑i∈E or j∈E(xpij −Xpij)

2

s.t.∑

j xpij =∑

j xpji for all i

VOMpi ≤ VXMpi for all i

(1)

1 We set to zero any negative input or output values in the raw data.2 To improve the characterization of energy markets, we merged input-output data with data on physical energy

quantities and energy prices from China’s National and Provincial Energy Statistical Yearbook 2007.

6

where i and j represent rows or columns’ indices of the SAM table, and xpij is value of elements

of the SAM table for province p. E represents rows or columns related to energy sectors (energy

production, use and trade), and PENALTY is the penalty term associated with changing ele-

ments related to the energy sector. VOMpi and VXMpi are output and total outflows (domestic

outflows and international export) of sector i in province p.

The objective function minimizes the extent to which the value of SAM elements can be

altered, especially in the case of elements related to energy sectors given that we have already

modified the energy data to improve its quality. Constraints in the optimization problem force all

accounts in the SAM table to be balanced and require output of every sector to be greater than

the total outflow for each province to satisfy the Armington assumption (Armington, 1969).

Table 2: Structure of SAM tables for each province in China

A C F H G1 G2 T DX X I1 I2 M

A AC SAC CA CH G2D DER ER CS1 CS2 VDSTF FAH HF HG2 DHR HR

G1 G1G2 CG1SG2 G2G1 TR

T TADX DRC DRH

X RC RHI1 DP PSV1 G1SVI2 IC PSV2M MG

Note: AC - sector output; SA - sector subsidy; CA - intermediate use; CH - household consump-tion; G2D - local government; DER - domestic outflow; ER - export; CS1 - investment; CS2 -inventory addition; VDST - domestic transportation service use; FA - factor input; HF - Factorearning; HG2 - transfer from central government to household; DHR - domestic trade deficit;HR - international trade deficit; G1G2 - transfer from local government to central government;CG1S - Balancing term for central government; G2G1 - transfer from central government tolocal government; TR - tex revenue for local government; TA - production tax; DRC - domesticinflow; DRH - domestic trade surplus; RC - import; RH - international trade surplus; DP - capitaldepreciation; PSV1 - balancing term for investment; G1SV - balancing term for investment; IC -inventory deletion; PSV2 - balancing term for inventory; MG - domestic trade margin.

We then construct another least-square optimization problem to balance all the SAM tables

for each province simultaneously to ensure that the domestic trade flows for each sector in China

are balanced. Prior to this optimization, we adjust the international trade data of each province

7

to satisfy the constraint that the total international trade flows for all provinces are consistent

with totals reported for China’s international trade in GTAP for each sector. We then fix the

international trade within each sector in each province to reconcile trade differences between

China’s data and the GTAP data set.

min{xpij}∑

p,i,j(xpij −Xpij)2 + PENALTY

∑i∈E or j∈E(xpij −Xpij)

2

s.t.∑

j xpij =∑

j xpji for all p, i

VOMpi ≤ VXMpi for all p, i∑p VDXMpi =

∑p VDIMpi for all i

(2)

This optimization problem is similar to previous one. VDXMpi and VDIMpi are domestic exports

and imports, respectively, from sector i for province p.

Using the balanced provincial SAM data, bilateral inter-provincial trade data is estimated

using the least-squares approach under the assumption that the import source composition of

each sector is the same as the source composition of the total imports for each province. Bilateral

province-to-country trade flows are also estimated by disaggregating China’s bilateral interna-

tional trade data in GTAP according to each province’s value share in China’s import/export

flows by sector.

For this study, we aggregate the data set to 30 provinces in China and to three regions in

the rest of the world (the United States, the European Union and other European countries,

and the Rest of World), and into 26 commodity groups (see Table 3). However, we maintain

the flexibility to aggregate the regions as desired for other studies. Our commodity aggregation

identifies six energy sectors and 20 non-energy composites. The mapping of GTAP commodities

and sectors identified in our study is provided in Table 3. Primary factors in the data set include

labor and capital. Labor and capital earnings represent gross earnings denominated in 2007 US

dollars.

3.2 The numerical model

Our modeling framework draws on a multi-commodity, multi-region static numerical general

equilibrium model of the world economy with sub-national detail for China’s economy. The key

features of the model are outlined below.

8

Table 3: Regions, commodity classifications and mappings in the model.

Region Abbreviation GTAP commodity Aggregatedcommodity

Beijing BEJ Paddy rice AGRTianjin TAJ Wheat AGRHebei HEB Cereal grains AGRShanxi SHX Vegetables, fruit, nuts AGRNeimenggu NMG Oil seeds AGRLiaoning LIN Sugar cane, sugar beet AGRJilin JIL Plant-based fibers AGRHeilongjiang HEL Crop AGRShanghai SHH Bovine cattle, sheep and goats, horses AGRJiangsu JSU Animal products AGRZhejiang ZHJ Raw milk AGRAnhui ANH Wool, silk-worm cocoons AGRFujian FUJ Forestry AGRJiangxi JXI Fishing AGRShandong SHD Coal COLHenan HEN Oil CRUHubei HUB Gas GASHunan HUN Minerals OMNGuangdong GUD Bovine meat products AGRGuangxi GUX Meat products AGRHainan HAI Vegetable oils and fats AGRChongqing CHQ Dairy products AGRSichuan SIC Processed rice AGRGuizhou GZH Sugar AGRYunnan YUN Food products AGRShanxi SHX Beverages and tobacco products B_TShannxi SHA Textiles TEXGansu GAN Wearing apparel CLOQinghai QIH Leather products CLONingxia NIX Wood products LUMXinjiang XIN Paper products, publishing PPP

Petroleum, coal products OILUnited States USA Chemical, rubber, plastic products CRPEurope Union and EUR Mineral products NMMother European countries Ferrous metals MSPRest of world ROW Metals MSP

Metal products FMPMotor vehicles and parts TMETransport equipment TMEElectronic equipment ELQMachinery equipment OMEManufactures OMFElectricity ELEGas manufacture, distribution GDTWater WTRConstruction CONTrade TRDTransport TRPWater transport TRPAir transport TRPCommunication OTHFinancial services OTHInsurance OTHBusiness services OTHRecreational and other services OTHPublic Administration, defense, OTHeducation, healthDwellings OTH

9

3.2.1 Production and behaviour of consumer and government

For each industry (i = 1, . . . , I, i = j) in each region (r = 1, . . . , R) gross output (Yir) is

produced using inputs of labor (Lir), capital (Kir) and produced intermediate inputs (Xjir):3

Yir = Fir(Lir,Kir;X1ir, . . . , XIir) . (3)

We employ constant-elasticity-of-substitution (CES) functions to characterize the production

technologies. All industries are characterized by constant returns to scale and are traded in

perfectly competitive markets. Nesting structures for each type of production system are de-

picted in Figure 4.

Gross output iσklem

M1 · · · Mj · · · MJKLEσeva

Energyσenoe

ELE Non-ELEσen

COL GAS CRU OIL GDT

Capital-Laborσva

K L

Figure 4: Structure of production for all other industries except fossil fuels and OIL, ELE, GDT

In each region r, preferences of the representative consumers are represented by a CES

utility function of consumption goods (Ci) and investment (I):

Ur = min [g(C1r, . . . , CIr), g(I1r, . . . , IIr)] (4)

3 For simplicity, we abstract from the various tax rates that are used in the model. The model includes ad valoremoutput taxes and import tariffs.

10

The function g(·) is a CES composite of all goods.

In each region, a single government entity approximates government activities at both cen-

tral and local levels.

3.2.2 Supplies of final goods and intra-China and international trade

All intermediate and final consumption goods are differentiated following the Armington as-

sumption. For each demand class, the total supply of good i is a CES composite of a domestically

produced variety and an imported variety, as follows:

Xir =[ψz ZDρ

Diir + ξz ZMρDi

ir

]1/ρDi

(5)

Cir =[ψc CDρ

Diir + ξc CMρDi

ir

]1/ρDi

(6)

Iir =[ψi IDρ

Diir + ξi IMρDi

ir

]1/ρDi

(7)

Gir =[ψg GDρ

Diir + ξg GMρDi

ir

]1/ρDi

(8)

where Z, C, I, and G are inter-industry demand, consumer demand, investment demand,

and government demand of good i, respectively; and ZD, ZM, CD, CM, ID, IM, GD, GM are

domestic and imported components of each demand class, respectively. The ψ’s and ξ’s are the

CES share coefficients. The Armington substitution elasticities between domestic and imported

varieties in these composites is σDi = 1/(1− ρDi ).

The domestic and imported varieties are represented by nested CES functions. We replicate

a border effect within our Armington import specification by assuming that goods produced

within the country are closer substitutes than goods from international sources. We include

separate import specifications for China provinces (indexed by p = 1, . . . , P ) and international

regions (indexed by t = 1, . . . , T ). The nesting structure of the Armington composites are

depicted in Figures 5 and 6.

11

Dip +Mip

σDi

Local and domestic (Di)σDUi

LocalDomesticσPUi

yi1p · · · yip′p · · · yiPp

Foreign (Mip)σMi

yi1p · · · yitp · · · yiTp

Figure 5: Aggregation of local, domestic, and foreign varieties of good i for China province p.

Dir +Mir

σDi

DomesticForeign (Mit)

σMi

ChinaσRUi

yi1t · · · yipt · · · yiP t

yi1t · · · yit′t · · · yiT t

Figure 6: Aggregation of domestic and foreign varieties of good i for international region t.

3.2.3 Equilibrium, model closures, and model solution

Consumption, labor supply, and savings result from the decisions of the representative household

in each region that maximize its utility subject to a budget constraint that consumption equals

income. Given input prices gross of taxes, firms maximize profits subject to the technology

constraints. Minimizing input costs for a unit value of output yields unit cost indices (marginal

costs), pYir. Firms operate in perfectly competitive markets and maximize their profit by selling

their products at a price equal to these marginal costs. The main activities of the government

sector in each region are purchasing goods and services, income transfers, and raising revenues

through taxes. Market clearance equations for factors and all the goods in both domestic markets

12

and import markets are also imposed as a constraint on the model solution.

Numerically, the equilibrium is formulated as a mixed complementarity problem (MCP)

(Mathiesen, 1985;Rutherford, 1995). Our complementarity-based solution approach comprises

two classes of equilibrium conditions: zero profit and market clearance conditions. The former

condition determines a vector of activity levels and the latter determines a vector of prices.

We formulate the problem in GAMS and use the mathematical programming system MPSGE

(Rutherford, 1999) and the PATH solver (Dirkse and Ferris, 1995) to solve for non-negative

prices and quantities.

3.3 Scenarios

To perform an initial investigation using this new model, we design two scenarios to compare

impacts of different CO2 intensity target schemes in China. In the first scenario, Regional Targets

(RT), we first set different provincial CO2 intensity reduction targets which are in line with the

Twelfth FYP’s target allocation (see Table 1). In the second scenario, National Target (NT), we

impose a single CO2 intensity reduction target at the national level that is equivalent to the

national carbon intensity reduction achieved in Scenario RT. We implement both policies as an

endogenous tax on CO2 embodied in energy used across the range of economic activities. The

tax is adjusted until the CO2 intensity target is achieved. The revenue of the tax in each province

goes back to the household of this province.

4 Modeling results

After imposing the provincial-level targets in Scenario RT, we find that the total carbon intensity

reduction achieved at the national level is 18.525%, and so we impose this level as the national

reduction target in Scenario NT. In both scenarios, welfare loss is modest at the national level,

4.23% in Scenario RT and 3.76% in Scenario NT (see Figure 4). Slightly more welfare loss occurs

at the national level under Scenario RT, the provincial allocation scheme (0.49% reduction

relative to Scenario NT), which is consistent with the fact that the equilibrium allocation is

more constrained due to the provincial-level reductions required. Interestingly, CO2 emissions

are actually reduced slightly more than CO2 intensity at the national level (see Figure 4).

13

-70 -50 -30 -10 10 30 50 70

SHX

GAN

ZHJ

LIA

HUN

SHD

SHA

JSU

SIC

HEN

GUD

GXI

JIL

FUJ

ANH

BEJ

HEB

SHH

YUN

NXA

TAJ

GZH

XIN

NMG

HLJ

CHQ

HUB

HAI

JXI

QIH

CHN

% Change - Consumption

(1) Provincial Target

(2) National Target

Figure 7: Regional welfare change.

Comparing the CO2 intensity reduction undertaken across provinces in each of the two

scenarios reveals some significant differences (see Figure 4). Under the national target, several

provinces that had relatively low targets in Scenario RT end up contributing significantly more to

overall abatement (in particular Hainan, Qinghai, and Xinjiang), suggesting that these provinces

offer abatement opportunities at lower cost. By contrast, provinces that faced tough provincial

targets in Scenario RT contribute less to overall abatement under the national target (see for in-

stance Shannxi, Gansu, and Jiangxi). This result suggests that the Scenario RT target allocation

is demanding large reductions from provinces where abatement is relatively expensive, while

bypassing opportunities to make reductions inexpensively in other provinces.

The modest welfare loss at the national level also masks large variation in the welfare im-

pacts across provinces under both scenarios (see Figure 4). Indeed, many of the provinces with

low reduction burdens under the provincial allocation (Scenario RT) experience large consump-

tion gains, relative to other provinces which are hit very hard by the policy. Some provinces

experience very large welfare increases, e.g. Qinghai and Jiangxi, and some provinces expe-

rience large welfare decreases, e.g. Shanxi (more than 50% in Scenario RT). In general, the

pattern of welfare change is quite similar in both scenarios. However, the consumption loss or

gain incurred in each province is almost always larger in absolute magnitude under the provin-

cial target allocation (Scenario RT) relative to the national target allocation (Scenario NT),

14

-35

-30

-25

-20

-15

-10

-5

0

5

10

AN

H

BEJ

CH

Q

FUJ

GA

N

GU

D

GX

I

GZH

HA

I

HEB

HEN

HLJ

HU

B

HU

N

JIL

JSU

JXI

LIA

NM

G

NX

A

QIH

SHA

SHD

SHH

SHX

SIC

TAJ

XIN

YUN

ZHJ

CH

N

% C

han

ge -

Car

bo

n E

mis

sio

ns

(1) Provincial Target

(2) National Target

Figure 8: Regional carbon emissions reductions.

suggesting that the provincial-level target allocation is not achieving a cost-effective allocation

that could have significant welfare consequences in some provinces.

In most provinces in both scenarios, CO2 emissions reductions are bigger than CO2 intensity

reductions in percentage terms because the adjustments to production required to meet the

intensity constraint are costly and have the net effect of reducing consumption. In provinces

where these costs are significantly offset by reductions in energy-related costs per unit of output

and allow for increases in consumption, the CO2 emissions reduction is smaller than the CO2

intensity reduction.

For China as a whole, as well as for the USA, Europe, and the Rest of World, the welfare

change is fairly small (less than 1%) in both scenarios (See Table 4). However, this change may

be significant given that policy is not directly imposed in these regions, which have witness slow

economic growth in recent years, particularly when compared to growth trends in China.

5 Conclusions

This paper has described a new provincial-level CGE model of China and applied it to a pre-

liminary assessment of China’s CO2 intensity target. The main goal of this analysis was to

compare the impacts of two CO2 intensity target allocation scenarios: one policy scenario that

matches China’s Twelfth FYP targets imposed at the provincial level (Scenario RT), and one

15

-35

-30

-25

-20

-15

-10

-5

0

5

10

AN

H

BEJ

CH

Q

FUJ

GA

N

GU

D

GX

I

GZH

HA

I

HEB

HEN

HLJ

HU

B

HU

N

JIL

JSU

JXI

LIA

NM

G

NX

A

QIH

SHA

SHD

SHH

SHX

SIC

TAJ

XIN

YUN

ZHJ

CH

N

% C

han

ge -

Car

bo

n In

ten

sity

(1) Provincial Target

(2) National Target

Figure 9: Regional carbon intensity reduction.

Table 4: Results for China, USA, Europe and rest of world in the two scenarios in percentageterms

carbon intensity change carbon emission change welfare change

China (Scenario 1) -18.53 -22.05 -4.23China (Scenario 2) -18.53 -21.96 -3.76USA (Scenario 1) -0.12 0.07 0.00USA (Scenario 2) -0.13 0.07 -0.01Europe (Scenario 1) -0.06 0.19 0.03Europe (Scenario 2) -0.07 0.17 0.03Rest of world (Scenario 1) 0.12 0.34 0.02Rest of world (Scenario 2) 0.10 0.32 0.02

policy scenario in which China faces a single national target that achieves an equivalent na-

tional intensity reduction. While we find that the single national carbon intensity reduction

target results in less significant consumption loss at the national level (3.76%) than current

provincially-disaggregated targets (4.23%), we also find great disparities in the regional im-

pacts. Given that regional impacts are an important consideration in the formulation of national

energy and climate policy, insights offered by this preliminary study and future work are likely

to be highly relevant for the policy process in China.

In particular, our results suggest that assigning provincial targets may miss cost-effective

opportunities to reduce emissions in less-constrained provinces, while demanding more costly

16

reductions from highly-constrained provinces. Assigning the appropriate intensity target level

to each province is a difficult task, and it is very difficult in advance to perform an exhaustive

comparison of abatement costs across provinces, not least because it requires knowledge of these

costs (which are often proprietary, difficult to estimate, or otherwise unavailable). A national

target does this more effectively because it creates incentives to undertake reductions where

they are most cost effective, independent of where they are located in China. However, we note

that the challenges of implementing a national intensity target may be significant in practice,

as provincial governments are currently held accountable for target implementation, and it is

less clear how this responsibility would be assigned (and achievement verified) under a national

target.

Our model helps to make equity and efficiency trade-offs clear, and to inform the current

discussion about ways to improve the cost effectiveness and address the distributional impacts

of climate and energy policies in China. Our model results provide some first insights into the

impact of reducing energy intensity in China in a static regional energy-economic modeling

framework. An important caveat is that we assume in our model that China’s economy is char-

acterized by perfectly competitive markets, which may have important implications for welfare

loss. We also note that in practice, the government may take countermeasures to mitigate large

welfare losses in individual provinces, reducing the serious distributional inequalities predicted

by our model results. Our future work will involve developing more technology and other de-

tails in the model as well as introducing inter-temporal dynamics to conduct further analysis of

China’s energy and climate policy proposals.

6 Acknowledgements

We acknowledge the support of the Ministry of Science and Technology of China through the

Institute for Energy, Environment, and Economy at Tsinghua University, and the support of

the scholarship from Graduate School of Tsinghua University, which is supporting Zhang Da’s

doctoral research as a visiting scholar at the Massachusetts Institute of Technology. We further

acknowledge the support of Eni S.p.A., ICF International, and Shell, initial founding sponsors of

the China Energy and Climate Project, for supporting this model development work. We would

17

further like to thank Dr. John Reilly for helpful comments and discussion.

18

References

Armington, Paul. 1969. “A Theory of Demand for Products Distinguished by Place of Produc-

tion.” International Monetary Fund Staff Papers, 16: 159–76.

Cao, J. 2007. “A Theory of Demand for Products Distinguished by Place of Production.” Harvard

University, Cambridge USA.

Dirkse, S. P., and M. C. Ferris. 1995. “The PATH Solver: a Non-monontone Stabilization Scheme

for Mixed Complementarity Problems.” Optimization Methods and Software, 5: 123–156.

GTAP, (Global Trade Analysis Project). 2012. Global Trade, Assistance, and Production: The

GTAP 8 Data Base. Center for Global Trade Analysis, Purdue University.

Horridge, Mark, and Glyn Wittwer. 2008. “SinoTERM, a multi-regional CGE model of China.”

China Economic Review, 4: 628âAS634.

Industrial Efficiency Policy Database(IEPD). 2012. “Intensity Target of the 11th Five Year

Plan.” http://iepd.iipnetwork.org/country/china.

Li, Na, Min-jun Shi, and Fei Wang. 2009. “Roles of Regional Differences and Linkages on

Chinese Regional Policy Effect in CGE Analysis.” Systems Engineering - Theory and Practice,

10: 35–44.

Li, S., and J. He. 2010. “Regional disparity and Mitigation cost for carbon policy in China:

Assessment based on multi-regional CGE model.” 2010 GTAP Annual Conference.

Li, S.T., and J.W. He. 2005. “A Three-regional Computable General Equilibrium (CGE) model

for China.” The 15th International Input-Output Conference.

Mathiesen, Lars. 1985. “Computation of Economic Equilibria by a Sequence of Linear Comple-

mentarity Problems.” Mathematical Programming Study, 23: 144–162.

National Infortion Center. 2011. “2007 National/Provincial Input-Output Tables.” National

Information Center, Beijing, China.

19

National Statistics Bereau. 2008. “2007 China Energy Statistical Yearbook.” National Statistics

Bereau, Beijing, China.

Price, L., and et.al. 2011. “Assessment of China’s Energy-Saving and Emission-Reduction Ac-

complishments and Opportunities During the 11th Five Year Plan.” Energy Policy, 4: 2165–

2178.

Price, L., X. Wang, and J Yun. 2010. “The Challenge of Reducing Energy Consumption of the

Top-1000 Largest Industrial Enterprises in China.” Energy Policy, 11: 6485–6498.

Rutherford, Thomas F. 1995. “Extension of GAMS for Complementarity Problems arising in

Applied Economics.” Journal of Economic Dynamics and Control, 19(8): 1299–1324.

Rutherford, Thomas F. 1999. “Applied General Equilibrium Modeling with MPSGE as a GAMS

Subsystem: an Overview of the Modeling Framework and Syntax.” Computational Economics,

14: 1–46.

State Council, China. 2012. “Inform on Issuing the Scheme of Greenhouse Gas Emission Control

During the 12 Five-Year Plan by State Council.” State Council, Beijing, China.

Wang, F.and S.H. Guo, and M. Ezaki. 2006. “Labor Migration and Regional Development in

China: A Regional CGE Analysis.” China Economic Quarterly, 4: 1067–1090.

Wang, Ke, Can Wang, and Jining Chen. 2009. “Analysis of the economic impact of different

Chinese climate policy options based on a CGE model incorporating endogenous technological

change.” Energy Policy, 8: 2930–2940.

Wei, C., J. Ni, and L Du. 2011. “Regional allocation of carbon dioxide abatement in China.”

China Economic Review.

Xu, Z.Y., and S.T. Li. 2008. “The Effect of Inter-regional Migration on economic Growth and

Regional Disparity.” The Journal of Quantitative and Technical Economics, 2: 38–52.

20