Francis O’Sullivan and Richard Schmalensee
Rutgers Energy Institute, November 6, 2015
The latest in the MIT “Future of…” studies exploring the
roles of key energy technologies in a carbon-constrained future
Limiting climate risk to acceptable levels will require
drastic reductions in global carbon dioxide emissions
from electricity generation by mid-century.
This will be politically difficult unless the electric power sector can also
meet the needs of a growing global economy at reasonable cost.
Solar is about 1% of global generation; can it be scaled up by around 50x
by 2050 to play a major role in meeting future electricity demand?
If so, what policies would make this most likely?
2
The scale and distribution of the solar resource make it one of the few low
carbon technologies capable of meeting a substantial fraction of
worldwide electricity demand even with rapid economic growth.
Map showing global variations in average annual solar irradiance
With today’s technology, total U.S. electricity demand could be
met by solar covering 0.43% of the contiguous U.S.
Source: Map adapted from Albuisson, M., M. Lefevre, and L. Wald. Averaged Solar Radiation 1990-2004, Ecole des Mines de Paris. (2006).
3
Today we have two practical pathways for generating solar electricity, PV
and CSP – PV dominates contemporary solar electricity generation and it will
continue to do so for the foreseeable future
Solar photovoltaics (PV)
Concentrated solar power (CSP)
-
Mature:
~97% of global
solar capacity
-
Modular:
efficiency does not
depend on scale
-
Output responds
immediately to
changes in
insolation
-
Less mature,
more expensive
-
Capital costs
fall with scale
-
Needs clear skies
Dispatchable when
thermal storage is
added
4
The past half decade has borne witness to remarkable growth in the scale of
installed solar generation capacity – This year will see 65GW of new PV
capacity come online with 40 GW coming from the US, Japan and China alone
Cumulative global installed PV capacity
GW
250
65 GW of new PV
capacity in 2015
USA
China
200
Europe
ROW
150
100
50
0
2008
2009
2010
2011
2012
2013
2014
2015E
Global installed solar capacity will approach 250GW
by the end of 2015, a 12X expansion since 2008
Source: MIT Analysis, National Renewable Energy Laboratory, Lawrence Berkeley National Laboratory, Solar Energy Industry Association, European Photovoltaic Industry Association, IHS
5
The pathway for solar growth depends on the local market – In the US and
China, utility scale systems are the dominant growth vector, while in markets
like Japan distributed systems lead the way
Annual US PV capacity additions by system type
MW
Cumulative PV capacity by state (2014)
MW
Other
9000
Utility
8000
Commercial
7000
Residential
20000
New Mexico
Texas
16000
New York
6000
Hawaii
12000
5000
Nevada
Massachusetts
4000
8000
North Carolina
3000
New Jersey
2000
4000
Arizona
1000
California
0
0
2008 2009 2010 2011 2012 2013 2014 2015E
2014
In the US, close to 60% of all PV capacity is in the
form of utility-scale units
Source: MIT Analysis, National Renewable Energy Laboratory, Lawrence Berkeley National Laboratory, Solar Energy Industry Association, European Photovoltaic Industry Association, IHS
6
A large reduction in the cost of PV modules has been a key factor in the
recent growth of solar installations – These dynamics also mean that the
focus of solar economics is shifting to the balance-of-system (BOS)
Rapid declines in PV module prices
have been important drivers of
growth
Evolution of PV module & system prices
$/Wp
… but these declines may have slowed
RESIDENTIAL
PV System
… and BOS costs have declined much
less rapidly
UTILITY
PV system
Deployment support at federal,
state, and local levels has also
driven growth
… but federal subsidies are scheduled to
be drastically cut from 2017, and
state programs have not expanded
recently
… and there has been a backlash
against rooftop solar in some states
MODULE
Price Drop
~85%
BOS
MODULE
Source: MIT Analysis, National Renewable Energy Laboratory, Lawrence Berkeley National Laboratory, U.S. Department of Energy, Solar Energy Industry Association, Photon Consulting
LLC
7
With these lower costs, utility-scale PV is increasingly competitive in
regions with high quality solar resource like CA, even without subsidy –
But residential PV and CSP systems have notably higher costs
PV Systems
Levelized cost of electricity
$/MWh
350
300
CSP Systems*
ITC Subsidy Value
After Subsidy LCOE
331
Benchmark LCOE
for Natural Gas
Generation
287
250
Regional variation
200
192
Minimum LCOE
158
141
150
105
100
76
50
0
Gas Combined Cycle
CA
MA
Utility-Scale
PV
CA
MA
Residential-Scale
PV
* CSP LCOE numbers based on CA system having 11 hours and MA system having 8 hours of nameplate capacity storage
Source: MIT Analysis, U.S. Energy Information Administration
CA
MA
CSP
8
However, as PV penetration increases the average price a PV generator
receives will be suppressed significantly – For solar to succeed at very large
scales, its costs must be reduced substantially
Illustration of how the price a solar generator receives for its output can fall well
below the average market price as solar penetration increases
$/MWh
60
55
50
45
40
35
30
25
20
0
Source: MIT Analysis
6
12
18
24
Solar Penetration
(% Peak Demand)
30
36
9
Increasing solar penetration in Germany has already lead to this new pricing
paradigm in their power system – Large-scale solar generation has led to
shaving of peak prices in the Germany wholesale power market
At marginal
penetration the
realized peak
price is high
As penetration rises
the peak price is
suppressed
Source: MIT Future of Solar Study
10
In light of all this, what needs to be done now to make it more likely that
solar energy can play a major role in limiting climate change?
Three main messages:
1. A long-term approach should be taken to technology
development
2. Preparation should be made for much greater
penetration of PV generation
3. Subsidies for solar deployment should be reformed
to improve their efficiency
11
Message 1:
A long-term approach should be taken to technology
development
What that means in practice:
Federal R&D spending should focus on emerging technologies with
the potential to deliver transformative cost reductions; the private
sector has the incentives and ability to improve those technologies
that are currently commercially marketed.
12
Wafer-based PV technologies and in particular crystalline silicon (c-Si)
dominate today’s solar market – In may respects this is a very attractive
technology but it has limitations
Current c-Si PV technology
ADVANTAGES
DISADVANTAGES
Efficient
Thick wafers
Reliable
Rigid and heavy
Robust and
Durable
Complex
manufacturing
Abundant
Non-toxic
c-Si PV technology is efficient and mature, but its intrinsic properties
may limit the potential for much further system cost reductions
13
With today’s c-Si PV technology balance of system (BOS) costs dominate
total system costs – Industry has the ability and incentive to reduce BOS
costs
Utility-Scale
PV
BOS now accounts
for 65% of utility-scale
system cost
2014 System cost build-up
$/W
2.00
1.80
Balance of System
1.00
0.05
0.40
0.65
0.30
0.40
0.00
Module
Residential-Scale
PV
Inverter & Other Engineering and
Hardware
Construction
Sales Tax
Margin and G&A
System Cost
2014 System cost build-up
$/W
3.25
Balance of System
3.00
0.74
0.05
BOS now accounts for
80% of residential-scale
system cost
0.56
2.00
0.35
1.00
0.90
0.65
0.00
Module
Source: MIT Analysis
Inverter, Other
Hardware &
Logistics
Installation
Labor
Customer
Acquisition &
PII
Sales Tax
Margin and
G&A
System Cost
14
Emerging thin-film technologies have the potential to lower both
module and BOS costs
Light & Flexible
High-throughput
Abundant
Kaltenbrunner, et al. 2012
Much more R&D needs to be done, and this is where federal
solar R&D should focus
(Current) Challenges
Low efficiency
Low stability
Unproven at scale
15
DOE solar R&D funding has increasingly focused on
areas other than core solar technology development
Breakdown of DOE’s Solar Energy Technology Office budget
400
$Millions
350
300
$241M
or
69%
250
Other
200
150
0
2010
$110M
or
31%
PV
2011
• grid integration
• enhanced manufacturing
competitiveness
• reduction of c-Si
BOS “soft costs”
CSP
100
50
Funding for work on
Current Technologies
addressing:
2012
2013
* 2016 SETO budget values are proposed not actual
Source: Department of Energy Annual Budget Justification statements
2014
2015
Funding for work
directly focused on
Advanced Solar
Technologies
2016*
16
To reduce CSP costs substantially, new high-temperature system
designs & materials must be developed and tested at pilot scale
CSP energy losses and opportunities
More efficient solar
collectors can convert
more of the incident
solar energy into thermal
energy
Source: MIT Analysis
Higher-temperature
power cycles can
convert more of the
absorbed thermal
energy into electricity
Reminder:
Storage is integral for CSP
in the form of stored heat
that can be used on demand
to produce electricity
17
Key Recommendations:
Technology Development
• Federal PV R&D should focus on transformative
technologies rather than on near-term reductions
in the cost of crystalline silicon systems.
• Federal PV R&D efforts should focus on new and
emerging thin-film PV technologies that use
environmentally benign, Earth-abundant materials
and that are compatible with low-cost manufacturing
and lower BOS costs.
• Federal CSP R&D efforts should focus on new
materials and system designs, and should establish
a program to test new designs in pilot-scale facilities,
akin to those common in the chemical industry.
18
Message 2:
Preparation should be made for much greater
penetration of PV generation
What that means in practice:
Given that c-Si PV will likely be the dominant solar technology for
many decades to come and very large-scale reliance on PV will pose
much more serious challenges than have been encountered to date, it
is necessary to focus on developing both the technical and
market/policy solutions needed to mitigate these challenges
19
Higher levels of PV penetration yield a number of challenges for the grid
operation including capacity and ramping requirements – These issues
can be mitigated to various degrees by storage
Simulated net demand for non-PV generation at different levels of PV penetration
ERCOT (Texas) typical summer day
ELECTRICITY
DEMAND
PEAK NON-PV
GENERATION
INCREASED RAMPING
RATE REQUIRED
24 hour day
20
Diurnal and seasonal changes in PV output are predictable, but PV
output varies with the weather, which is imperfectly predictable
Hourly solar radiation at Golden, Colorado during 2012
• At high levels of reliance on PV, large-scale storage with various capacities
(e.g., minutes, hours, days, …) and response speeds will likely be necessary.
• Apart from pumped (hydro) storage, which is economical but difficult to site,
large cost reductions in storage are necessary for widespread deployment.
Source: NREL
21
Distributed PV can help lower line losses, but as penetration grows those
savings are generally outweighed by investments needed to maintain
power quality
Average total costs with increased distributed PV penetration under
different assumptions about design standards & generation mix
Source: MIT Analysis
22
Net metering subsidizes residential PV more than utility-scale PV at the
expense of other customers – This has already produced conflict
System after A becomes a net solar seller
System before A installs solar
Network cost paid by customer per kWh
Network cost paid to customer A per kWh
Energy cost paid by customer per kWh
Energy cost paid to net-metered customer per kWh
Additional network cost paid by customers without solar
Utility Rate
$/kWh
Utility Rate
$/kWh
Higher retail price
with cost shifted
Retail price
including
network
costs
Wholesale
energy price
Wholesale
energy price
A
B
C
Utility Customers
…N
A
B
C
…N
Utility Customers
- When A sells power, she gets the retail price, while utilityscale sellers get the wholesale price, often much lower
- When A stops covering any network costs, the retail rate
must go up so the other customers cover those costs –
plus the network cost paid to A!
Net-metered rate
paid to Customer A
23
Key Recommendations:
Grid integration
• R&D aimed at developing low-cost, scalable energy
storage technologies is a crucial part of a strategy to
achieve economic PV deployment at large scale.
• Utilities, regulators, and stakeholders should develop
and deploy fair pricing systems that allocate distribution
network costs to all users of the network—including
distributed solar generators.
24
Message 3:
Subsidies for solar deployment should be reformed to
improve their efficiency
What that means in practice:
There is a good case for continuing to subsidize the deployment of
solar generation, but today taxpayers and utility ratepayers are paying
considerably more per kilowatt-hour of solar generation than they
could be. Appropriate reforming of today’s subsidy mechanisms will
ensure greater solar deployment per dollar of subsidy investment
25
Federal, state, & local governments
subsidize the deployment of solar technologies through an array
of tax credits, regulatory requirements, and direct subsidies
• These subsidies help lay the foundation for a major solar scale-up
by building experience with manufacturing & deployment and
overcoming institutional barriers
• Particularly in the absence of a nation-wide price on carbon
emissions, the US should continue to subsidize solar deployment
• The main federal solar subsidies are accelerated depreciation and
a 30% investment tax credit (ITC) for businesses and individuals
who own a solar system.
• At the end of 2016 the business ITC is scheduled to be cut to 10%,
and the individual ITC is scheduled to expire
• Such a drastic cut in federal support would be unwise
• Federal, state, and local support of solar deployment should be
reformed to enhance the efficiency of these programs
26
Solar developers are generally not capable of monetizing the ITC without use
of the tax equity market – Having to partner with tax equity investors is costly and
reduces the effectiveness of the entire subsidy mechanism
Levelized cost of electricity
$/MWh
350
ITC subsidy cost per
kWh
-
The current solar ITC subsidy regime
means that more expensive systems
receive higher subsidies
-
Generation from residential systems
can receive 2X or more subsidy per
kWh than from utility-scale systems
-
Not only that, firms that build and
own residential solar systems can
calculate ITC and depreciation based
on the present value of systems’
income, which in markets with little
competition may be well above the
actual investment cost.
300
After ITC electricity
LCOE
250
107
200
72
150
57
100
180
37
50
101
120
68
0
CA
MA
Utility-Scale PV
Source: MIT Analysis
CA
MA
Residential-Scale PV
27
The 24 state-level RPSs that require utilities to buy solar electricity
from distributed generators are a major driver of solar deployment
All RPS programs are different; almost all restrict generator location; many states
have multiple solar support policies; some localities do also
Source: dsireusa.org
28
Key Recommendations:
Deployment policy
• Particularly in the absence of a nationwide price on
carbon emissions, drastic cuts in federal support for
solar technology deployment would be unwise.
• Policies to support solar deployment should reward
generation, not investment; should not provide greater
subsidies to residential generators than to utility-scale
generators; and should avoid the use of tax credits.
• State RPS programs should be replaced by a uniform
national program. If this is not possible, states should
remove restrictions on out-of-state siting of eligible
solar generation.
29
Working Draft
Thank You
Last Modified 4/28/2010 8:16:26 AM GMT Standard Time
Printed 4/28/2010 8:08:33 AM GMT Standard Time
Study Participants
Study Chair
RICHARD SCHMALENSEE
Howard W. Johnson Professor of Economics and Management
John C. Head III Dean (Emeritus)
Sloan School of Management, MIT
Study Co-Chair
VLADMIR BULOVIĆ
Fariborz Maseeh (1990) Professor of Emerging Technology
Associate Dean for Innovation
Electrical Engineering and Computer Science, MIT
Study Group
ROBERT ARMSTRONG
Chevron Professor, Department of Chemical Engineering, MIT
Director, MIT Energy Initiative
CARLOS BATTLE
Visiting Scholar, MIT Energy Initiative
Associate Professor, Institute for Research in Technology
Comillas Pontifical University
PATRICK BROWN
PhD Candidate, Department of Physics, MIT
JOHN DEUTCH
Institute Professor, Department of Chemistry, MIT
HENRY JACOBY
Professor (Emeritus), Sloan School of Management, MIT
ROBERT JAFFE
Morningstar Professor of Science, Department of
Physics, MIT
JOEL JEAN
PhD Candidate, Department of Electrical Engineering
and Computer Science, MIT
RAANAN MILLER
Associate Director, MIT Energy Initiative
Executive Director, Solar Energy Study
FRANCIS O’SULLIVAN
Senior Lecturer, Sloan School of Management, MIT
Director, Research and Analysis, MIT Energy Initiative
JOHN PARSONS
Senior Lecturer, Sloan School of Management, MIT
JOSE IGNACIO PĖREZ-ARRIAGA
Professor, Institute for Research in Technology
Comillas Pontifical University
Visiting Professor, Engineering Systems Division, MIT
NAVID SEIFKAR
Research Engineer, MIT Energy Initiative
ROBERT STONER
Deputy Director for Science and Technology, MIT
Energy Initiative
Director, Tata Center for Technology and Design, MIT
CLAUDIO VERGARA
Postdoctoral Associate, MIT Energy Initiative
31
Thin-film PV technologies promise lower BOS costs due to their
format that can eliminate heavy glass substrates, … but, unlike
c-Si, materials availability and high-temperature processing will
limit the scale-up of today’s commercial thin-film PV
more than 35
years of current
production
required by 2050
1400
years
6 years
Te, In, Ga, and Se
are now produced only as
by-products from the
production of other metals.
Substantial increases in
production volumes of these
materials would likely require
primary production with
unknown technologies.
Source: MIT Analysis
COMMERCIAL
THIN FILM PV
32
There is a promising set of emerging thin-film PV
technologies that are not materials-constrained and
that can be developed at near room-temperature
EMERGING Thin-Film PV
Material Sets
at most 3 years
of current
production
required by 2050
Source: MIT Analysis
COMMERCIAL
THIN FILM PV
EMERGING
THIN FILM PV
33
The PV system cost reductions that have been achieved have not necessarily
been passes along to US consumers – In the utility sector, pricing tends to be
competitive, while “value pricing” is a prominent feature of the residential market
PV Pricing Mechanisms
Utility-scale PV – ~1MW and above
Residential-scale PV – up to 10kW
-
Utilities driving market by need to meet RPS
targets
-
Emerging awareness and demand among
homeowners
-
Strong competition among developers to
secure PPAs
-
Installers developing innovative business
models reducing upfront costs to owners
-
Pricing strongly linked to underlying cost
base
-
“Value Pricing” linking solar prices to local
utility rates
34
Today, utility sector PPAs are being signed in the $40-50/MWh range, this
is at or below what today’s capex costs can allow – This is evidence of
operators being confident they can take out further cost
Average PPA prices
$/MWh
200
ERCOT
Southwest
California
Northwest
MISO
SPP
Southeast
300MW
150
100MW
100
50
0
2008
2009
2010
2011
Sources: Bloomberg NEF, “U.S. PPA Market Outlook.” 07/08/15. GTM/SEIA, “US SMI Q1 2015.”
2012
2013
2014
2015
35
Price formation in the residential sector differs from market-to-market and is
often linked to regulated utility rates – Consumer willingness to pay can lead to
a decoupling of solar price from underlying cost
Reported price in
immature market
Reported
price in
competitive
market
$4.50/W
ITC: $1.35/W
$3.25/W
ITC: $0.98/W
Unsubsidized
Costs - Gross
Price to
Consumer
Federal Subsidy
Competitive Market
Source: MIT Team Analysis
$2.27/W
Net Price to
Consumer
WTP: $3.15/W
Net Consumer Federal Subsidy Gross Price to
Willingness to
Consumer
Pay
Immature or Uncompetitive Market
36
One of the most important factors in the growth of solar in the residential
market has been the rise of the “third party owned” business model – High
capital cost and tax appetite, two key barriers to US residential solar penetration
have been eliminated
Average system price by ownership type
$/Wp
8.00
AZ
CA, Host-owned
7.50
CA, 3rd-party
MA, Host-owned
7.00
MA, 3rd-party
MD
6.50
NY, Host-owned
NY, 3rd-party
6.00
5.50
$5.25
5.00
4.50
$4.15
4.00
3.50
3.00
Q4
'10
Q1
'11
Q2
'11
Q3
'11
Q4
'11
Q1
'12
Source: California Solar Initiative and other state reporting systems
Q2
'12
Q3
'12
Q4
'12
Q1
'13
Q2
'13
Q3
'13
Q4
'13
Q1
'14
Q2
'14
Q3
'14
Q4
'14
Q1
'15
Q2
'15
37
The success of the third-party owned model is rooted in the ability to “value
price” solar power relative to incumbent utility supplied power
Power Price
¢/kWh
Range of future utility
prices: PU, t
PU, 0
PPV, 0
Predefined future PV lease
or PPA price: PPV, t
0
1
2
3…
…N
Years
-
Third party solar ownership, either via leases or PPA structures is allowed allowed in at least 22
states today
-
The third party model makes residential solar very affordable and in most major markets it entirely
dominates installations – In CA more than 75% of new installations are third party
-
Third party solar developers are explicit in viewing themselves as competing directly with utilities
Source: MIT Team Analysis
38
Cost-basis calculation for ITC purposes is an area where the third-party model
causes issues
Allowable methods for establishing the solar ITC cost basis:
-
The cost method is the most straightforward and is based on the assumption that an informed
purchaser will pay no more for a system than the cost of replacing it.
-
The market method relies on data from recent sales of comparable systems.
-
The income method estimates FMV based on the cash flows generated by the system.
How the ITC cost basis is established based on the “income method”
Source: MIT Team Analysis
39
In many contemporary U.S. residential solar markets, allowing the ITC cost
basis be established via the “income method” amplifies the subsidy by 50%
or more – In highly competitive markets this amplification would be eliminated
Subsidies:
ITC:
$0.98/W
MACRS: $0.26/W
$3.25/W
Unsubsidized
Cost
$3.00/W
Lease PV
Cost Method
Source: MIT Team Analysis
$4.24/W
Subsidies:
ITC:
$1.45/W
MACRS: $0.39/W
$4.84/W
$3.00/W
Subsidy PV Total Income
PV
Lease PV
Subsidy PV Total Income
PV
Income Method
40