Computable general equilibrium (CGE) modelling

Download Report

Transcript Computable general equilibrium (CGE) modelling

AMOS Energy CGE Modelling
Karen Turner
Department of Economics and Fraser of Allander Institute
University of Strathclyde
ESRC Grant Ref: RES-061-25-0010
FAI Brown Bag Seminar
Overview
•
•
•
•
•
Policy issues
Model requirements
Computable general equilibrium modelling
Development of the AMOSENVI framework
Current and future research
Policy issues
• Impacts of changes in economic conditions and/or policy on
environmental/sustainability indicators
• Sustainability – global and local concerns
• UK – regional and national focus
• Supply-side issues
• Energy and/or carbon taxation
• Resource productivity, energy efficiency
• Changes in technology
• Demand-side issues
• Nature and structure of energy demand and use
• Elasticity of energy demand
Requirements for energy-economyenvironment modelling
• Multi-sectoral modelling
• Different energy-use and pollution generation
characteristics of different production and
consumption activities
• System-wide
• Interaction between different production and
consumption activities
Input-output accounting and modelling
• Input-output accounts
• Snapshot of economic activity
• Multi-sectoral, economy-wide
• Single entry book-keeping
• Regional and national accounts
• Multiplier analysis
• Input-output models
• Assumptions
• Simple, transparent form of general equilibrium modelling
• Popular with policymakers
• But restrictive – demand or supply, prices or quantities,
universal Leontief technology
Computable general equilibrium (CGE)
modelling (1)
• IO database
• Extended to social accounting matrix (SAM)
• Labour supply and demand, capital stocks,
investment demands
• Key: analytical work with IO/SAM – structural
characteristics
• Theoretical origins: Walrasian general equilibrium
theory
• In practice: wide variety of approaches, assumptions
and focus
Computable general equilibrium (CGE)
modelling (2)
• Common features:
• Both supply and demand matter
• Prices and quantities modelled simultaneously
• Data inform about structure
• Modelling based on theory and observation of
behavioural relationships, market interactions
• Involves making assumptions
• Should be transparent and subject to sensitivity
analysis
Example - AMOS
• Originally developed as A micro and macroeconomic Model Of
Scotland
• 3-sector, single region
• Now N-sector, single and inter-regional
• Particular focus on regional labour markets
• Recent developments – energy-environment, linked demographic
model
AMOSENVI (1)
• Environmental impact version of AMOS
• N-sector – identify sectors with distinct energy supply/use and/or
pollution generation characteristics
• Initial version – emphasis on pollution generation – Leontief
ouput-pollution coefficients
– Captures changes in pollution due to scale and composition
effects
– But not due to input substitution and technology effects
• ESRC funded project ‘Modelling the impacts of sustainability
policies in Scotland’
– Link pollution generation to energy input use
– Introduction KLEM production structure – substitution
between energy and other inputs
AMOSENVI (2)
• 3 transactor groups – households, firms, government
• 25 commodities and activities (5 energy supply)
• Two exogenous external transactors (RUK and ROW)
• Commodity markets taken to be competitive
• Scotland modelled as a small open economy
AMOSENVI (3)
•
Assume cost minimisation in production
•
Multi-level production functions
•
Four major components of final demand:
•
Government expenditure – exogenous or endogenous
•
Consumption a linear homogenous function of real disposable income
•
Exports (and imports) determined via an Armington link (relative price
sensitive)
•
Investment
AMOSENVI (4)
Single Scottish labour market
Perfect sectoral mobility
Wages subject to a regional bargained real wage function
Labour and capital stocks updated between periods
Capital – investment equals depreciation plus fraction gap between
desired and actual capital stocks
Population updating via migration function
AMOSENVI (5)
• KLEM production structure
• Input- and output-CO2 coefficients
• Calibrated on 1999 Scottish SAM
• Scottish IO tables with estimated electricity
disaggregation
• Pilot region-specific sectoral CO2 accounts
Sectoral breakdown of the 1999 Scottish 25 sector extended (KLEM) AMOSENVI model
IOC
1
AGRICULTURE
1
2
FORESTRY PLANTING AND LOGGING
3
FISHING
3.1
4
FISH FARMING
3.2
5
Other mining and quarrying
6,7
6
Oil and gas extraction
7
Mfr food, drink and tobacco
8
Mfr textiles and clothing
21 to 30
9
Mfr chemicals etc
36 to 45
10
Mfr metal and non-metal goods
46 to 61
11
Mfr transport and other machinery, electrical and inst eng
12
Other manufacturing
13
Water
87
14
Construction
88
15
Distribution
89 to 92
16
Transport
93 to 97
17
Communications, finance and business
18
R&D
19
Education
20
Public and other services
2.1, 2.2
5
8 to 20
62 to 80
31 to 34, 81 to 84
98 to 107, 109 to 114
108
116
115, 117 to 123
ENERGY
21
COAL (EXTRACTION)
22
OIL (REFINING & DISTR OIL AND NUCLEAR)
35
23
GAS
86
ELECTRICITY
85
24
Renewable (hydro and wind)
25
Non-renewable (coal, nuke and gas)
4
Figure1. Production structure of each sector i in the 25 sector/commodity AMOSENVI KLEM
framework
GROSS OUTPUT
INTERMEDIATES
ROW composite
VALUE-ADDED
UK composite
RUK composite
LABOUR
CAPITAL
LOCAL composite
ENERGY composite
NON-ENERGY composite
Non-energy comm. j = 1……………………………….20
ELECTRICITY
composite
RENEWABLE
(comm j=24)
NON-ELECTRICITY
composite
NON-RENEWABLE
(comm j=25)
OIL
(comm j=22)
NON-OIL
composite
COAL
(comm j=21)
GAS
(comm j=23)
Application of AMOSENVI
• Debate in literature: “rebound” and “backfire” effects
• Policy concern – House of Lords (2005) report on energy efficiency
• Khazzoom-Brookes Postulate (KBP)
– Jevons (1865) – “confusion of ideas” regarding productive use of
fuel and diminished consumption – increase utility, impact on
implicit prices
• Will an increase in efficiency of energy use lead to increased or reduced
consumption of energy? E.g. increase energy efficiency by 5%
• No rebound or backfire – reduce energy consumption by 5%
• Rebound – reduce energy consumption by less than 5%
• Backfire – increase energy consumption
• Efficiency effects vs substitution, income and output effects
• General equilibrium effects
Case study: Scotland
• Rebound and backfire effects
• Difference in direction of effects over short- and long-run
• Key parameters governing the extent of rebound and backfire:
• Elasticities of substitution in production
• Energy and non-energy intermediates
• Value-added and intermediates
• RUK Export demand elasticity
• Target of shock
• Implications of variations in sectoral energy supply and use
characteristics
Figure 4. Impact of a 5% increase in energy efficiency in all sectors on energy indicator
variables
2.0
1.5
0.5
51
49
47
45
43
41
39
37
35
33
31
29
27
25
23
21
19
17
15
13
9
7
5
11
-0.5
3
0.0
1
% change from base
1.0
-1.0
-1.5
-2.0
-2.5
Period (year)
Y/m (1) - GDP £/tonnes oil equiv non-elec
Energy consumption (elec gigaw att hours)
GDP
Y/m (2) - GDP £/gigaw att hours elec
Share of electricty generation from renew able sources
Key results of sensitivity analysis (long run impacts)
Central case
Electricity rebound effect (%)
Non-electricity energy rebound effect (%)
123.0
116.1
Elasticity of substitution
Energy/Non-energy
Value-added/intermediate
Low 0.1
High 0.7
Low 0.1
High 0.7
109.5
150.3
120.2
128.5
106.7
135.1
113.8
120.7
Central case
Electricity rebound effect (%)
Non-electricity energy rebound effect (%)
123.0
116.1
Armington export
Electricity (RUK)
Low 2.0
87.0
84.5
Varying target of shock
Energy supply Non-energy
(ES)
supply (NES)
209.2
57.6
179.9
72.6
Heavier users Heavier users Heavier users gas
gas and oil gas and oil (ES) and oil (NES)
65.3
79.6
60.4
75.2
70.3
76.0
Case study - UK
• Scottish results driven by fact Scotland is a net exporter in
electricity to RUK
• Also, nature of shock (somewhat blunt, and targeting Scotland
only, in a single region model)
• DEFRA project with UKENVI model (UK national economy)
• No backfire
• Elasticities of substitution dominate rather then export
elasticities (particularly energy/non-energy)
• Time pattern different – rebound bigger in short-run than in
long-run
Key conclusion
•
Extent of rebound and backfire effects is always and everywhere
an empirical issue
•
A large number of parameters potentially important for influencing
general equilibrium effects
•
Elasticities of substitution in production important
•
But other characteristics such as
• Openness of the economy in question
• Elasticity of supply of other inputs
• Energy intensity of individual production sectors and final
consumption activities
• Elasticity of substitution between commodities in consumption
• Income elasticities of demand for commodities
Current and future research
•
EPSRC projects (ongoing)
• Inter-regional CGE framework for UK
• More sophisticated modelling of electricity production and markets
• Extension of policy applications – e.g. changes in technology,
carbon/energy taxation
•
ESRC 1st grant project (start October 07)
• Analysis of factors governing energy rebound effects in the UK economy
• Introduce different hierarchical production and consumption activities
• Identification different household consumption groups and consumption
activities
• Focus on key production and consumption activities
• Econometric estimation of key functions and parameters
• Examination of a wider range of policy scenarios
• Systematic programme of simulation, sensitivity and scenario analysis
• Contribute to evidence base on extent to which economy-wide rebound
effects are likely to occur in UK, and nature of effects
• Development of energy-economy-environment CGE modelling framework
for application to a wider range of regional and national economies, and
policy issues