Transcript Global CGE Model – Key Characteristics

Global and Country Specific CGE Models at the World
Bank for Climate Change Analysis
Govinda Timilsina
The World Bank, Washington, DC
Skopje, Macedonia
March 01, 2011
Presentation Outline
 Introduction
 Global CGE Model
 Data for Global CGE Model
 Single Country CGE Model
 Data for Single Country CGE Model
Introduction
•
•
•
Costs of climate change (impacts and mitigation) are carried out at
activity or sector level and such analysis does not account the
inter-sectoral linkages. Due to inter-linkages between productive
sectors; between economic agents and international trade, an
activity, if implemented at a large scale, could have economy wide
effects;
The impacts or costs measured at the activity or project level could
be significantly different from those measured at the economy-wide
levels
A GHG mitigation technology, for example, attractive from activity
or sectoral approach may not necessarily be attractive if its impacts
to the overall economy are accounted for (or the rankings of GHG
mitigation options could change)
Global CGE Model – Key Characteristics
• Multi-sector, multi-region, global recursive dynamic CGE model
• The model is flexible enough to accommodate new regions/countries
or sectors and is calibrated with GTAP database
• Nested CES and CET functional forms to represent production
behavior and land supply, respectively
• Nonhomothetic Constant Difference of Elasticities (CDE) function
form for households
• Detailed representation of land-use biofuel sectors
• Representation of bilateral and international trade
Global CGE Model Structure – Production Sector
Nested CES structure of the model for production sectors
CGE Model – Production Sector (Cont.)
Cost minimization formulation
𝐦𝐢𝐧
𝑽𝑨𝑬𝒊,𝒓,𝒕,𝒗 , 𝑵𝑫𝒊,𝒓,𝒕
𝑷𝑽𝑨𝒊,𝒓,𝒕,𝒗 𝑽𝑨𝑬𝒊,𝒓,𝒕,𝒗 +𝑷𝑵𝑫𝒊,𝒓,𝒕 𝑵𝑫𝒊,𝒓,𝒕
where VAE is the value added and energy bundle, ND is the non energy bundle.
PVA and PND are the prices of VAE and ND, respectively.
𝝈𝑽𝑨𝑬𝑵𝑫
−𝟏
𝒊,𝒓,𝒗
𝟏
𝒔. 𝒕.
𝜶𝑽𝑨𝑬
𝒊,𝒓,𝒗
𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗
𝝀𝑽𝑨𝑬
𝒊,𝒓,𝒕,𝒗 𝑽𝑨𝑬𝒊,𝒓,𝒕,𝒗
𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗
𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗
𝟏
𝜶𝑵𝑫
𝒊,𝒓,𝒗
+
𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗
𝝀𝑵𝑫
𝒊,𝒓,𝒕,𝒗 𝑵𝑫𝒊,𝒓,𝒕
𝝈𝑽𝑨𝑬𝑵𝑫
−𝟏 𝝈𝑽𝑨𝑬𝑵𝑫
−𝟏
𝒊,𝒓,𝒗
𝒊,𝒓,𝒗
𝑽𝑨𝑬𝑵𝑫
𝝈𝒊,𝒓,𝒗
≥ 𝑿𝒗𝒊,𝒓,𝒕,𝒗
𝒗
where X is gross output. VAE and ND correspond to the share parameters for VAE
and ND, respectively, and VAEND is the elasticity of substitution between VAE and
ND. VAE and ND are the productivity parameters that represent the state of the
technology. The indices i, r, v and t correspond to sector, country/region, capital
vintage and time, respectively.
CGE Model – Production Sector (Cont.)
• Derivation of demand and price variables
𝑽𝑨𝑬
𝑽𝑨𝑬𝒊,𝒓,𝒕,𝒗 = 𝜶𝑽𝑨𝑬
𝒊,𝒓,𝒗 𝝀𝒊,𝒓,𝒕,𝒗
𝝈𝑽𝑨𝑬𝑵𝑫
−𝟏
𝒊,𝒓,𝒗
𝑵𝑫
𝜶𝑵𝑫
𝒊,𝒓,𝒗 𝝀𝒊,𝒓,𝒕,𝒗
𝑵𝑫𝒊,𝒓,𝒕 =
𝑷𝑿𝒗𝐢,𝐫,𝐭,𝐯
𝑷𝑽𝑨𝒊,𝒓,𝒕,𝒗
𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗 −𝟏
𝒗
𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗
𝑷𝑿𝒗𝐢,𝐫,𝐭,𝐯
𝑷𝑵𝑫𝒊,𝒓,𝒕,𝒗
𝑿𝒗𝒊,𝒓,𝒕,𝒗
𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗
𝑿𝒗𝒊,𝒓,𝒕,𝒗
𝟏
𝑷𝑿𝒗𝐢,𝐫,𝐭,𝐯 = 𝜶𝑽𝑨𝑬
𝒊,𝒓,𝒗
𝑷𝑽𝑨𝒊,𝒓,𝒕,𝒗
𝝀𝑽𝑨𝑬
𝒊,𝒓,𝒕,𝒗
𝟏−𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗
+ 𝜶𝑵𝑫
𝒊,𝒓,𝒗
𝑷𝑵𝑫𝒊,𝒓,𝒕,𝒗
𝝀𝑵𝑫
𝒊,𝒓,𝒕,𝒗
𝟏−𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗
𝟏− 𝝈𝑽𝑨𝑬𝑵𝑫
𝒊,𝒓,𝒗
• In the similar manner, all demand and price variables were derived
Global CGE Model – Energy Sector
Figure 1 (c): Nested CES structure of the model for energy demand
Global CGE Model – Land Use
Figure 1 (b): Nested CET structure for land supply
Global CGE Model – Land Use (Cont.)
• Land is split into 18 Agro-Ecological Zones (AEZs)
AEZ
Climate type
Humidity level
AEZ1
Arid
AEZ2
AEZ6
Dry semi-arid
Moist semi-arid
Sub-Humid
Humid
Humid > 300-days LGP*
AEZ7
Arid
AEZ8
AEZ12
Dry semi-arid
Moist semi-arid
Sub-Humid
Humid
Humid > 300-days LGP
AEZ13
Arid
AEZ14
Dry semi-arid
Moist semi-arid
Sub-Humid
Humid
Humid > 300-days LGP
AEZ3
Tropical
AEZ4
AEZ5
AEZ9
AEZ10
Temperate
AEZ11
AEZ15
AEZ16
AEZ17
AEZ18
Boreal
*LGP stands for Length of growing period
Model Dynamics and Closure
• Medium variant of UN population forecasts
• Per capita GDP growth is exogenous (World Bank projections)
• Resource prices (e.g., oil price forecasts) are exogenous
• Annual sector specific productivity growth (2.1% for agriculture, 1%
for service and 2% for manufacturing)
• Autonomous energy efficiency improvement (1% )
• Long-term sustainability  Government deficit and capital account
are fixed
Data & Parameters
• Data are coming from the GTAP (Global Trade Analysis Project)
database (Purdue University, Indiana)
• The database provides SAMs and international trade (bilateral flows,
trade barriers)
• Database version 7.1
– Year 2004
– 112 countries/regions
– 57 sectors
GTAP 7.1 – Geographic disaggregation
XNA
XEF
FIN
NOR SWE
EST
LVA
DNK
LTU
BLR
IRL GBR NLD
POL
BEL DEU
LUX CZE
UKR
AUT SVK XEE
FRA CHE SVNHUNROU
HRV
XER
BGR
ITA
GEO
ALB
ARM
AZE
PRTESP
GRC
TUR
CAN
USA
MLT
TUN
RUS
KAZ
XSU
KGZ
KOR
CYP
IRN
MAR
XNF
XEA
XSA
PAK
EGY
XWS
MEX
IND
XCB
XCB
GTM
XCA
NIC
SEN
NGA
VEN
XCF
ETH
HKG
PHL
PHL
MYS
SGP
UGA
ECU
XAC
TWN
LKA
XSM
COL
BGD
MMR
LAO
VNM
THA
KHM
XWF
XEC
CRI
PAN
JPN
CHN
MYS
IDN
TZA
XOC
XSE
BRA
PER
ZMB
MWI
MOZ
BOL
ZWE
PRY
MDG
MUS
XSC BWA
AUS
ZAF
CHL
ARG
URY
NZL
NZL
Orange – individual countries; red- combined a regions
GTAP 7.1 – Sectoral Disaggregation
Paddy rice
Coal
Wood. Prod.
Electricity
Wheat
Oil
Paper etc.
Gas Dist.
Oth. Cereals
Gas
Ref. oil etc.
Water
Veg. & Fruits
Oth. Minerals
Chemicals etc.
Construction
Oil seeds
Red meat
Oth. Min. Prod.
Trade
Sug. Cane & Beet
White meat
Ferr. Met.
Land trns.
Plant-based fibers
Veg. Oils
Oth. Met.
Sea trns.
Oth. Crops
Dairy Prod.
Met. Prod.
Air trns.
Beef etc.
Proc. Rice
Mot. Veh. & parts
Communication
Poultry, Pork, etc.
Ref. sugar
Oth. Trp. Eqpt.
Fin. Serv.
Raw milk
Oth. Food
Electronic Eqpt.
Insurance
Wool etc.
Bev. & Tob.
Oth. Mach. & Eqpt.
Oth. Bus. Serv.
Forestry
Textiles
Oth. Manu.
Recr. & Oth. Serv.
Fishing
Clothing
Publ. Serv.
Leath. Prod.
Dwellings
Regional and sector decomposition
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Paddy rice
Sugar (cane & beet)
Vegetables, fruit
Wheat
Corn
Other cereal grains
Oilseeds
Livestock
Sugar Ethanol
Corn Ethanol
Grains Ethanol
Biodiesel
Processed food
Forestry
Coal
Crude oil
Natural gas
Other mining
Gasoline
Diesel
Refined oil
Chemicals
Other manufacturing
Electricity
Gas distribution
Construction
Transport services
Other services
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Australia and New Zealand
Japan
Canada
United States
France
Germany
Italy
Spain
UK
Rest of EU & EFTA
China
Indonesia
Malaysia
Thailand
Rest of East Asia & Pacific
India
Rest of South Asia
Argentina
Brazil
Rest of LAC
Russia
Rest of ECA
MENA
South Africa
Rest of Sub-Saharan Africa
•
Computational limitations require
aggregation of countries/regions
and sectors
(GTAP: 112 regions & 57 sectors
or 112* 57 = 6,384 equations for 1
variable only defined on 2 dimensions)
•
Focus on main countries/regions
producer of biofuels
•
Keep as much detail as possible
for agriculture (especially biofuel
feedstocks) and for energy
sectors
Key Elasticity Parameters
• Elasticity parameters other than related to biofuels and land-use are
from Burniaux and Chateau (2010); van der Werf (2008); Timilsina
and Shrestha (2006); Ma et al (2010); Jarrett and Torres (1987) and
Narayanan and Walmsley (2008).
• Elasticity of substitution between biofuels and fossil fuels
• Existing studies (e.g., Birur et al. 2007) - 2.0 based on historical data
• Increase the value from 1.2 (2004) to 3.0 (2020) to reflect expansion of flexfuel vehicles
• Elasticity parameters for land-use module:
• A high value (18) between AEZ (based on literature)
• CET elasticity values -- -0.2, -0.5 and -1.0, respectively for top, middle and bottom nests
(Choi, 2004; Hertel et al. 2008)
Single Country Model: Key Features
• Multisector, SAM based general equilibrium model for a country
(Thailand, Brazil, Nigeria and Morocco);
• It has two regions: the country and the rest of the world (but
assumption small open economy)
• Number of sectors are flexible based on policy questions to be
analyzed (for example, in Thailand 187 sectors are aggregated to 21
sectors)
• Deep nested structures for representing the behavior of production
and household sectors;
Single Country Model: Difference from other models
• Has detailed representation of the energy sectors and commodities
(e.g., coal, crude oil, natural gas, fuel wood, petroleum refinery, gas
processing and electricity generation);
• The electricity sector is further divided into seven sub-sectors: hydro;
coal-, oil- and gas- fired steam turbine; oil- and gas- fired combined
cycle; and diesel fired internal combustion engine;
• Refined petroleum products are divided into three category: gasoline,
diesel and others
• Land use and biofuels are explicitly represented to allows modeling of
GHG mitigation options in the land use change and forestry sector
Single Country Model: Production Structure
(Excluding transport, agriculture and forestry sectors)
Output
Materials
Value Added & Energy
Labor
Capital & Energy
Capital
Energy
Electricity
Non-electricity
Liquid fuel
(Petroleum)
Coal
Gasoline
Diesel
Gas
Others
Single Country Model: Production Structure
(transport sector)
Liquid fuels
Other petroleum products
Biofuels-gasoline-diesel
Ethanol & gasoline
Ethanol
Gasoline
Biodiesel & diesel
Biodiesel
Diesel
Single Country Model: Production Structure
(Agriculture and forestry sectors)
Output
Materials
Land
Value Added & Energy
Labor
Capital & Energy
Capital
Energy
Electricity
Non-electricity
Coal
Gasoline
Liquid fuel
(Petroleum)
Diesel
Gas
Others
Single Country Model: Land Supply
Total land
supply
Protected
forests
Unprotected
Forest lands
Land for
sugar crop
Other lands
Crop lands
Land for
soybean
Pasture land
Land for other crops
Single Country Model: Electricity Supply
Electricity
Hydro & Renewable
Thermal
Combined cycle/gas
turbine
Steam Turbine
Coal
Oil
Gas
Gas
Oil
Modeling Challenges
– CGE models, normally do not have technology level details of
production sectors, especially when a production sector is an
aggregate of several sub-sectors, which in turn are aggregate of
several technologies (e.g., food & beverage sector, chemical sector)
– Since CGE models are based on database of a base year (SAM),
modeling a technology which does not exist in the base year is
difficult (although there might be some tricks)
– Since MAC curve is a product of a separate models/modules outside
the CGE model, there exits always a danger of inconsistencies on
assumptions on the common economic variables (e.g., GDP growth,
price assumptions, etc.)
– Precise estimation cost of climate change impacts is difficult if not
impossible.
THANK YOU
Govinda R. Timilsina
Sr. Research Economist (Climate Change & Clean Energy)
Development Research Group
The World Bank
1818 H Street, NW
Washington, DC 20433, USA
Room: MC3-451
Mail Drop: MC3-300
Tel: 1 202 473 2767
Fax: 1 202 522 1151
E-mail: [email protected]