Ongoing Macro-Stand Alone and CGE modelling approaches at UCL Energy Institute
1. Ongoing Macro-stand alone and CGE modelling
approaches at UCL Energy Institute
Dr. Matthew Winning
Energy Institute, University College London
ETSAP Workshop, Copenhagen
19 Nov 2014
2. Contents
1. UK Times Model (UKTM)
2. Macro Stand-Alone
3. Capital-energy substitution
4. UKTM-MSA scenarios, issues and results
5. GTAP-UCL
6. Conclusions and extensions
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
3. UKTM
Energy Systems model
Bottom-up with energy technology choice
Partial equilibrium, least-cost optimisation
Exogenous energy demands
New features
Non-CO2 greenhouse gases
Non-energy mitigation options
Energy storage and other energy infrastructures
New time slices (4 intra-day x 4 seasonal)
Updated industry and residential sectors
Development process
Transparency at the forefront of development (data, assumptions, structure is clear and traceable,
full replicability of results, comprehensive QA processes)
Full sectoral data update & 2010 base-year recalibration
User constraints categorized & explicit
UKTM will be fully open-source from January 2015
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
4. MARKAL-Macro in UK
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Strachan and Kannan (2008), Hybrid modelling of long-
term carbon reduction scenarios for the UK, Energy
Economics (30), 2947 – 2963
Strachan, Pye and Hughes (2008), The Role of
International Drivers on UK scenarios of a low-carbon
Society, Climate Policy, 8:sup1, 125 – 139
Results: loss of GDP in UK ranging from 0.3% to 1.5% by
2050
5. Motivation
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Q/ Is it worthwhile linking full-blown CGE model to
TIMES for overall impacts i.e. marginal benefit of using
TIMES-MSA vs CGE for overall economic impacts?
Q/ TIMES-MSA vs. CGE parameterisation of elasticity of
substitution between capital and energy
6. TIMES-MSA
(Kypreos and Lehtila, 2013)
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
• A single-agent, single-sector neoclassical optimal
growth general equilibrium model
• Cumulative and discounted utility maximisation of a
representative consumer-producer agent
• Production function is energy and a capital/labour
composite
• GDP comprises of consumption, investment and
energy system costs
• Explore MSA parameter sensitivities
• e.g. elasticity of substitution between
capital/labour composite and energy
• E.g. capital to GDP ratio in the UK
7. TIMES-MSA
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
UKTM
ENERGY SOURCES
TECHNOLOGY CHARACTERISTICS
ENVIRONMENTAL CONSTRAINTS
& POLICIES
TECHNOLOGY MIX
FUEL MIX
EMISSIONS SOURCES & LEVELS
FUEL & EMISSION MARGINAL COSTS
RANKING OF MITIGATION OPTIONS
MACRO
LABOUR
GDP
CONSUMPTION
CAPITAL INVESTMENT
USEFUL ENERGY
SERVICES
ENERGY
PAYMENTS
8. TIMES-MSA
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Parameter Description Default Value
TM_ARBM Arbitrary multiplier for the last period replication 1
TM_DEFVAL(item) Default values for regional Macro constants
TM_DEFVAL(ESC) 1.03
TM_DEPR(r) Depreciation rate (percentage) 5.00
TM_DMTOL(r) Lower bound factor for the demand variables 0.50
TM_ESUB(r) Elasticity of substitution 0.25
TM_GDP0(r) GDP in the first period
TM_GR(r,y) Projected annual GDP growth in per cent
TM_IVETOL(r) Investment and energy cost upper bound tolerance 0.50
TM_KGDP(r) Initial capital to GDP ratio 2.50
TM_KPVS(r) Initial capital value share in all production factors 0.25
TM_SCALE_CST Scaling factor for cost units 0.00
TM_SCALE_NRG Scaling factor for the demand units 1.00
TM_SCALE_UTIL Scaling factor for the utility function 0.00
TM_QFAC(r) Switch for market penetration penalty function * 0.00
9. TIMES-MSA
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Parameter Description Default Value
TM_ARBM Arbitrary multiplier for the last period replication 1
TM_DEFVAL(item) Default values for regional Macro constants
TM_DEFVAL(ESC) 1.03
TM_DEPR(r) Depreciation rate (percentage) 5.00
TM_DMTOL(r) Lower bound factor for the demand variables 0.50
TM_ESUB(r) Elasticity of substitution 0.25
TM_GDP0(r) GDP in the first period
TM_GR(r,y) Projected annual GDP growth in per cent
TM_IVETOL(r) Investment and energy cost upper bound tolerance 0.50
TM_KGDP(r) Initial capital to GDP ratio 2.50
TM_KPVS(r) Initial capital value share in all production factors 0.25
TM_SCALE_CST Scaling factor for cost units 0.00
TM_SCALE_NRG Scaling factor for the demand units 1.00
TM_SCALE_UTIL Scaling factor for the utility function 0.00
TM_QFAC(r) Switch for market penetration penalty function * 0.00
10. TIMES-MSA
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Parameter Description Default Value
TM_ARBM Arbitrary multiplier for the last period replication 1
TM_DEFVAL(item) Default values for regional Macro constants
TM_DEFVAL(ESC) 1.03
TM_DEPR(r) Depreciation rate (percentage) 5.00
TM_DMTOL(r) Lower bound factor for the demand variables 0.50
TM_ESUB(r) Elasticity of substitution 0.25
TM_GDP0(r) GDP in the first period
TM_GR(r,y) Projected annual GDP growth in per cent
TM_IVETOL(r) Investment and energy cost upper bound tolerance 0.50
TM_KGDP(r) Initial capital to GDP ratio 2.50
TM_KPVS(r) Initial capital value share in all production factors 0.25
TM_SCALE_CST Scaling factor for cost units 0.00
TM_SCALE_NRG Scaling factor for the demand units 1.00
TM_SCALE_UTIL Scaling factor for the utility function 0.00
TM_QFAC(r) Switch for market penetration penalty function * 0.00
11. Capital-energy substitution
Both cross-price substitution and Morishima elasticity of
substitution
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
SR LR SR LR
Morishima 0.41 – 0.44 1.05 – 1.07 0.14 – 0.17 0.77 – 0.80
Cross-price 0.22 – 0.38 0.36 – 0.52 0.17 – 0.34 0.31 – 0.48
North America Europe
(M.J. Koeste et al, 2008)
12. Capital-energy substitution
Both cross-price substitution and Morishima elasticity of
substitution
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
SR LR SR LR
Morishima 0.41 – 0.44 1.05 – 1.07 0.14 – 0.17 0.77 – 0.80
Cross-price 0.22 – 0.38 0.36 – 0.52 0.17 – 0.34 0.31 – 0.48
North America Europe
(M.J. Koeste et al, 2008)
13. UKTM-MSA results
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Total Energy
5400
5600
5800
6000
6200
6400
6600
6800
7000
UKTM_BASE UKTM_BASE_MSA UKTM_LowGHG_Int2 UKTM_LowGHG_MSA_Int2
Total energy 2040 (PJ)
14. UKTM-MSA results
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Base MSA
ALL BIO ALL COALS ALL ELECTRICITY
ALL GAS ALL HYDROGEN ALL MANFUELS
ALL OIL PRODUCTS ALL OTHER RNW
LowGHG-MSA
ALL BIO ALL COALS ALL ELECTRICITY ALL GAS
ALL HYDROGEN ALL MANFUELS ALL OIL PRODUCTS ALL OTHER RNW
Final energy by fuel
15. UKTM-MSA results
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Base
FUEL TECHS AGR FUEL TECHS ELC
FUEL TECHS HYG FUEL TECHS INDUS
FUEL TECHS PRC FUEL TECHS RES
FUEL TECHS SERV FUEL TECHS TRA
FUEL TECHS UPSTREAMLowGHG
Base MSA
LowGHG-
MSA
16. UKTM-MSA results
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Total Emissions
-100000
0
100000
200000
300000
400000
500000
600000
700000
BASE MSA_035 LowGHG_MSA_Int2
GHG emissions 2040 (kT)
EMIS GHG UPS
EMIS GHG TRA
EMIS GHG SER
EMIS GHG RES
EMIS GHG PRC
EMIS GHG NEU
EMIS GHG IND
EMIS GHG HYG
EMIS GHG ELC
EMIS GHG AGR
17. UKTM-MSA results
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Total System Costs
8500000
8800000
9100000
9400000
9700000
10000000
10300000
Total System Cost
18. UKTM-MSA results
Scenario 2020 2030 2040
LowGHG 0.29 - 0.67 - 0.73
LowGHG (esub 0.35) 0.21 - 0.68 - 0.72
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
GDP loss from baseline
19. UKTM-MSA results
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Macro variables: Consumption (£bn)
2020 2025 2030 2035 2040
UKTM_BASE_MSA 1458 1646 1866 2147 2607
UKTM_BASE_MSA_
035 1461 1649 1868 2148 2606
UKTM_LowGHG_M
SA 1463 1646 1861 2136 2589
UKTM_LowGHG_M
SA_035 1463 1648 1861 2137 2590
20. UKTM-MSA results
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Macro variables
300.00
350.00
400.00
450.00
500.00
2020 2025 2030 2035 2040
Energy System costs
UKTM_BASE_MSA_2050 UKTM_BASE_MSA_035 UKTM_LowGHG_MSA_2050_Int2 UKTM_LowGHG_Int2_MSA_035
21. UKTM-MSA results
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Macro variables
280.00
300.00
320.00
340.00
360.00
380.00
400.00
420.00
2015 2020 2025 2030 2035
Investment (£bn)
UKTM_BASE_MSA_2050 UKTM_BASE_MSA_035 UKTM_LowGHG_MSA_2050_Int2 UKTM_LowGHG_Int2_MSA_035
22. UKTM-MSA issues
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
• NLP solver MINOS vs. CONOPT
• Lower bound on demand required
• Demand marginals volatility – further analysis
required
23. GTAP-UCL
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
• Based on GTAP-E production structure
• GTAP8 database based on Social Accounting
Matrix and trade data
• 129 regions, 57 economic sectors, 5 factors of
production
• MCP formulation solved in GAMS
• Currently standardised methodology of the
disaggregation of electricity sector as part of
ADVANCE FP7 project
24. Conclusions and future work
Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
• Comparison of capital/energy substitution in
MSA vs. CGE models
• Linking UKTM with GTAP-UCL
• Application of MSA to European Times Model
• WholeSEM – link GTAP-UCL and/or UKTM-
MSA with UK land-water models such as
FORESEER at Cambridge
25. Ongoing Macro-stand alone and CGE modelling approaches at UCL Energy Institute
IEA-ETSAP Copenhagen November 2014
Thank you for listening
m.winning@ucl.ac.uk
Questions please!