understanding model estimates of the economic...
TRANSCRIPT
Understanding Model Estimates of the Economic Costs of Climate Policy
EPRI Modeling WorkshopMay 8, 2008
Tom WilsonSenior Program Manager
2© 2007 Electric Power Research Institute, Inc. All rights reserved.
U.S. Climate Policy Proposals Focused on Cutting Emissions Significantly Below Historic Levels
3© 2007 Electric Power Research Institute, Inc. All rights reserved.
Elements of Lieberman-Warner Climate Security Act of 2007
• Economy-wide Coverage – about 87% of 2005 emissions– Downstream on coal (units > 5,000 tons/yr)– Upstream on oil, gas, F-gases, N2O
• Targets for Covered Sectors (in MtCO2e)2012: 5,7752020: 4,9242030: 3,8602050: 1,732 (~ 70% below 2005 level)
• Cap-and-trade system– 22.5% auction phasing to 69.5% by 2030– Permits/auction revenues designated for a wide array of uses
• Provisions for limited use of offsets• Cost containment via a Carbon Market Efficiency Board
4© 2007 Electric Power Research Institute, Inc. All rights reserved.
And the Cost of Lieberman-Warner 2007 is ….(Draft)
0
50
100
150
200
250
300
350
400
450
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055
$/to
n C
O2e
CRA, banking
MIT, 15% offsets, CCS subs
MIT, no offsets, CCS subs
MIT, 15% offsets, no CCS subs
MIT, no offsets, no CCS subs
EPA, ADAGE core
EPA, ADAGE low intl
EPA, ADAGE constrained N&B
EPA, ADAGE constrained N,B,CCS
EPA, ADAGE constrained N,B,CCS,NG cartel
EPA, ADAGE alt ref
EPA, IGEM core
EPA, IGEM unlim
EPA, IGEM no offset
EPA, IGEM alt ref
CATF
ACCF, low
ACCF, high
EIA, core
EIA, high cost
EIA, lim alt
EIA, no intl offsets
EIA, lim/no intl
5© 2007 Electric Power Research Institute, Inc. All rights reserved.
Topics
• A Digression – A Perspective on Using Models• Models Here Today• Different Assumptions Often Make the Biggest Differences• Concluding Thoughts
6© 2007 Electric Power Research Institute, Inc. All rights reserved.
Part I:What Are Models Good For?
“It’s hard to make predictions, especially about the future”
-Yogi BerraFamous American philosopher (and baseball player)
7© 2007 Electric Power Research Institute, Inc. All rights reserved.
The Challenge of Modeling (Thanks to Jae Edmonds)
“I’m amused year after year by your high emissions scenarios – and happy to see others now chuckle at them too!”
– Note Jae got from a friend in 1983
Examine Jae and John’s Forecasts of 2000 CO2Emissions + Oil Prices
… Made in 1982
8© 2007 Electric Power Research Institute, Inc. All rights reserved.
Projected and Actual 2000 Global CO2Emissions
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
1975 1980 1985 1990 1995 2000
Mill
ions
of M
etric
Ton
s of
Car
bon
Per Y
ear
ForecastCDIAC Historical Global CO2 EmissionsExpon. (CDIAC Historical Global CO2 Emissions)
Source: Edmonds and Reilly -- Global Energy Production And Use To The Year 2050. Institute for Energy Analysis, Oak Ridge Universities, 1982
9© 2007 Electric Power Research Institute, Inc. All rights reserved.
Projected and Actual 2000 World Oil Prices
$0.00$10.00$20.00$30.00$40.00$50.00$60.00$70.00$80.00$90.00
$100.00
1975 1980 1985 1990 1995 2000
Cru
de O
il D
omes
tic F
irst P
urch
ase
Rea
l Pric
e 96
$$/
barr
el
Actual PriceER 1982 Scenario
Conventional Wisdom
Source: Edmonds and Reilly -- Global Energy Production And Use To The Year 2050. Institute for Energy Analysis, Oak Ridge Universities, 1982
10© 2007 Electric Power Research Institute, Inc. All rights reserved.
The Magic of Countervailing Errors –Geographic
Year 2000 Global Primary Energy
0
20
40
60
80
100
120
ER 2000 scenario Est. 2000 historical
Exa
joul
es p
er Y
ear
US
Canada & Western Europe
EE&FSU
China, South & East Asis,& Africa
Japan, Australia & NZ
Latin America
China, South & East Asia,& Africa
Exa
joul
espe
r Yea
r
11© 2007 Electric Power Research Institute, Inc. All rights reserved.
The Magic of Countervailing Errors – Technology Choice
2000 Global Energy Consumption by Fuel (ER
Scenario)
32%
17%
31%
12%
8%
OilGasCoalNuclearHydro
`
2000 Global Energy Consumption by Fuel (EIA
Historical)
39%
23%
24%
6%
8%
OilGasCoalNuclearHydro
12© 2007 Electric Power Research Institute, Inc. All rights reserved.
Insights from This Example
• Important stuff happens, especially over 10, 20, or 40 year time horizons– Models can provide very useful insights … but are less
useful for providing “predictions”– Some things are easier to “predict” than others
• What could happen to change “predictions” dramatically?– Extended recession?– Rebound of the US$?– Technology accident?– Technology breakthrough?
• What modeling insights seem robust?
14© 2007 Electric Power Research Institute, Inc. All rights reserved.
Models to Be Discussed Today
• EIA NEMS=National Energy Modeling System• CATF NEMS• ACCF NEMS
• EPA ADAGE = Applied Dynamic Analysis of the Global EconomyIGEM = Intertemporal General Equilibrium ModelIPM = Integrated Planning ModelOthers
• MIT EPPA = Emissions Prediction and Policy Analysis Model• CRAI MRN-NEEM = Multi-Region National Model +
North American Electricity and Environment Model
15© 2007 Electric Power Research Institute, Inc. All rights reserved.
Selected Model Characteristics(Draft)
Geographic Time ElectricModel/Modeling Effort Scope Sectors Horizon Foresight Sector Detail–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––EIA/CATF/ACCF
• NEMS US-regional All, Process 2030 Recursive Plant type
EPA
• ADAGE Global(?) All, Trade1 2050 Perfect Production function
• IGEM US All, Trade1 2050 Perfect Production function
• IPM US-regional Electric 2025 Perfect Plant type/unit
MIT-EPPA Global All, Trade1 2050 Recursive Production function
CRAI- MRN-NEEM U.S.- regional All, Trade1 2050 Perfect Plant type/unit
1Based on IMPLAN state-level, SIC code trade data with energy trade adjusted to match EIA physical trade data and reference scenario calibrated to EIA-AEO projections.
16© 2007 Electric Power Research Institute, Inc. All rights reserved.
Part IIIDifferent Assumptions Often Make the Biggest Differences
17© 2007 Electric Power Research Institute, Inc. All rights reserved.
Recall the Cost of Lieberman-Warner 2007 is ….(Draft)
0
50
100
150
200
250
300
350
400
450
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055
$/to
n C
O2e
CRA, banking
MIT, 15% offsets, CCS subsMIT, no offsets, CCS subs
MIT, 15% offsets, no CCS subsMIT, no offsets, no CCS subs
EPA, ADAGE core
EPA, ADAGE low intlEPA, ADAGE constrained N&B
EPA, ADAGE constrained N,B,CCSEPA, ADAGE constrained N,B,CCS,NG cartel
EPA, ADAGE alt ref
EPA, IGEM coreEPA, IGEM unlim
EPA, IGEM no offsetEPA, IGEM alt ref
CATF
ACCF, lowACCF, high
EIA, coreEIA, high cost
EIA, lim alt
EIA, no intl offsetsEIA, lim/no intl
18© 2007 Electric Power Research Institute, Inc. All rights reserved.
Cost Estimates from Application of NEMS by EIA, CATF, and NAM/ACCF
0
50
100
150
200
250
300
2010 2015 2020 2025 2030 2035
$/to
n C
O2e
CATFACCF, low costACCF, high costEIA, coreEIA, high costEIA, limited alternativesEIA, no intl offsetsEIA, limited/no intl offsets
19© 2007 Electric Power Research Institute, Inc. All rights reserved.
Key Assumptions IncludeUncertainties and Policy Levers
I will talk about three uncertainties• Reference Case• Technology Cost• Non-economic limits on technology deployment
Later speakers will address• Price Sensitivity• Other Energy Policies• Offsets• Technology Policies• Etc
20© 2007 Electric Power Research Institute, Inc. All rights reserved.
Policy Cost Depends on the Reference Case…Reference Cases Change Over Time
4000
4500
5000
5500
6000
6500
7000
7500
8000
8500
1985 1990 1995 2000 2005 2010 2015 2020 2025 2030
CO2-
e (M
MT)
AEO 1998
AEO 1999
AEO 2000
AEO 2001
AEO 2002
AEO 2003
AEO 2004
AEO 2005
AEO 2006
AEO 2007
AEO 2008
Actual
EIA Annual Energy Outlook Projected CO2 Emissions (AEO 1998 – AEO 2008)
21© 2007 Electric Power Research Institute, Inc. All rights reserved.
Energy Technology Costs Are Highly Uncertain:All Nuclear and CCS Costs are Engineering Estimates
Pleasant Prairie Chilled Ammonia, Post-Combustion Capture Demo
Late 2007 photo, courtesy of Alstom
22© 2007 Electric Power Research Institute, Inc. All rights reserved.
Energy Technology Costs Are Highly Uncertain:Are Today’s Costs a New “Plateau” or a “Bubble”
100
110
120
130
140
150
160
170
2000 2001 2002 2003 2004 2005 2006 2007 2008
Mid-Year
Inde
x Va
lue
GDP Deflator
EIA Generation Data –PC & IGCC
Chemical Engineering Plant Cost Index -CEPCI
Marshall & Swift Equipment Cost Index
Handy-Whitman –Electricity Utility Construction
RS Means Construction Cost Index
IHS-CERA Downstream Construction Cost Index -DCCI
Nelson Farrar Refinery Construction Inflation Index
Construction Cost Escalation Continues
(Comparison of Several Cost Indices, Mid-2000 = 100)
23© 2007 Electric Power Research Institute, Inc. All rights reserved.
Reported and Guesstimated Technology Cost Estimates for Analyses you Will See Today
Overnight Capital Cost (2008$/kW)
0500
100015002000250030003500400045005000
Nuclear Coal CCS
$/kW
EIA
CATF
ACCF
EPA
MIT
CRAI
Caveats: Converted to 2008$
CATF= AEO2007, EPA/IPM= AEO2005, CRAI= updated AEO2007
Some costs decline rapidly over time, e.g., CRAI Coal CCS to $3203/kW by 2050
24© 2007 Electric Power Research Institute, Inc. All rights reserved.
Technology Deployment is Even More Uncertain
PublicPerception
Cost
Leakage
Regulation
LegalLiability
25© 2007 Electric Power Research Institute, Inc. All rights reserved.
Cumulative Capacity Additions of Coal CCS, Nuclear and Renewables (preliminary)
Cumulative Capacity Additions by 2030
0
100
200
300
400
500
600
CRA EPAIPM
CATF ACCF,low
LIMIT
ACCF,highLIMIT
EIA,core
EIA,highcost
EIA,lim alt
EIA, nointl
offsets
EIA,lim/no
intl
GW
RenewNuclearCoal CCS
26© 2007 Electric Power Research Institute, Inc. All rights reserved.
Concluding Thoughts
• US climate policy analyses likely to be released frequently over the next few years
• All of the models here are credible• None are great predictors; but all can provide important
insights about policy• Differences in cost estimates appear to represent true
uncertainties (e.g., technology deployment) and effects of policy choices (e.g., use of offsets) more than differences in models
I would be more worried and feel less-informed if there were little diversity in the cost estimates
27© 2007 Electric Power Research Institute, Inc. All rights reserved.
Questions?
See EPRI Journal, Summer 2007
Contacts
Palo Alto –Tom Wilson – [email protected]