What Does It Cost To Build A Power Plant?: ILAR Working Paper #17 September 2012
What Does It Cost To Build A Power Plant?: ILAR Working Paper #17 September 2012
What Does It Cost To Build A Power Plant?: ILAR Working Paper #17 September 2012
(ILAR)
The Laboratory is part of School of International Relations and Pacific Studies at University of
California, San Diego. ILAR gratefully acknowledges anchor funding from the nonpartisan
Electric Power Research Institute, BP, plc, the Norwegian Research Foundation and from UC San
Diego’s School of International Relations and Pacific Studies.
Pacific Studies, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA
92093, USA
*
Corresponding e-mail address: farong@ucsd.edu. Tel: (01)-949-3356986. ILAR's research is funded by
EPRI, BP, UCSD, and the Norwegian Research Foundation and their support is gratefully acknowledged.
Page 1
Abstract
The massive size of the investment in the electric power sector—along with other important
consequences—hinge on the relative cost of different types of power plants. Surprisingly, within
the literature on energy modeling and forecasting there has been little attention to the factors
that determine these capital costs. Where scholars have studied these factors in detail they
have focused mainly on a particular subset of power plants. This essay offers a review of these
capex factors along with a case study on China. Our aim is to help the energy modeling
community take a closer look at pivotal assumptions in their models and suggest ways of
framing future scenarios for capital costs. We suggest that capital costs for new power plants is
likely to vary worldwide to a much greater degree than widely assumed—largely because labor
costs and various state-sponsored mechanism for risk management vary so much. We also
suggest that labor saving innovations, such as in robotics and through international trade in pre-
assembled components, may substantially lower costs even for plants that have long been
thought immune to such improvements, such as large coal and nuclear facilities.
1. Introduction
The central scenario in International Energy Agency (IEA)’s World Energy Outlook (WEO)
2011—the so-called New Policies (NPS) Scenario which assumes recent government policy
commitments are implemented in a cautious manner—envisions that world primary demand for
energy will increase by one-third between 2010 and 2035 (IEA, 2011). Electricity is expected to
grow faster than any other final form of energy. Of the total $38 trillion in global investment in
energy-supply infrastructure IEA projects will be needed, the power sector alone will claim
about half. Those investment needs vary, of course, with scenario. In IEA’s “450 Scenario”—a
scenario which prescribes strong policy actions to limit climate change— total investment is
Page 2
slightly lower but the share for electricity is higher as IEA’s model assumes there are advantages
to reducing world emissions through even more rapid electrification (IEA, 2011). While we focus
here on the IEA results, other large world energy models make similar projections.
Pivotal to these forecasts is the capital cost of building new power plants. While other
costs—notably fuel—along with plant performance play central roles in determining which
power plants and fuels will be selected, capital costs are central. For some technologies, such as
nuclear and hydro, nearly all of the levelized cost of electricity stems from construction costs
and related charges such as interest. In the last decade, capital costs have been rising much
faster than inflation in all developed countries, and that trend has led analysts to expect that
capital costs will be high into the future as well. Yet the experience in other countries—
including in the most capital intensive technologies such as nuclear power—suggests that this
trend is not universal. For example, a Joint report by the IEA and the OECD Nuclear Energy
Agency (NEA)—which examined the cost of building a wide array of types of power plants—
found that while costs vary substantially. Nuclear costs are as low as 1,560 USD/kWe (Korea)
and 1,750 USD/kWe (China), or about half the overnight costs in the United States and one-third
the likely cost in Belgium (IEA and NEA, 2010). Many analysts have explored the reasons for
these variations in costs, including whether low costs in one country might be “portable” to
other jurisdictions (e.g., Yang 2011 which includes a comparison of U.S. and Chinese nuclear
costs). Yet, for the most part, the results of detailed assessments of the driving factors of plant
costs are not accessible in the energy modeling community, and a close look at those analyses
suggests that the long-term future for capital costs could be quite different from the
The assumed level of capital cost for power plants has potentially profound implications for
policy. Capital (and fuel) costs play a central role in determining the structure of the energy
Page 3
system and thus on the kinds of externalities such as insecurity and environmental emissions
that have been a central focus of policy makers. There is a huge literature that looks closely at
how policy instruments could affect deployment of technology (e.g. Victor, 2002; Rai, et al.,
2010; Fischer et al., 2008; Nemet, 2012). The choice of those policy instruments and design of
the interventions depends, in part, on the existing and potential relative costs of different
attaining political support) required for policies that try to advance otherwise uneconomic
governments attempt to deploy renewable energy technologies that are more expensive than
fossil-based rivals (Victor and Vanosek, 2011). For decades there has been an active literature
on national systems of innovation; a few scholars have applied those concepts to innovation in
energy-related technologies in particular (Gallagher et al., 2012). For policy makers, one
implication is that efforts to enhance innovation in energy policy—a goal that many recent
studies have advocated (Holliday, et al, 2010; Victor, 2011; Pielke, 2011; Wessner and Wolff,
2012; Chu and Majumdar, 2012)—should look at the full array of policy factors that affect how
firms mobilize and deploy capital for fixed investments such as power plants.
This paper offers an in-depth review of major studies on capital costs of building new power
plants while exploring the major factors that determine these costs with an eye to how those
driving forces might change in the future. In addition, it offers a brief case study on how those
First, the capital cost of new power plants varies substantially across countries. Costs for
most power generating technologies including both fossil fuel plants (e.g., advanced coal) and
renewables (e.g., onshore wind) are much lower in some emerging Asian countries, particularly
China. The cost to build a supercritical power plant in China, for example, is possibly less than
Page 4
one third of building the similar one in the United States. The variation in costs for constructing
Second, on average the cost of building new plants in the advanced industrialized world has
risen three times faster than inflation over the last decade. The financial crisis imposed a short-
lived decline, but costs have rebounded quickly due to sustained economic growth in Asia that
has yielded higher commodity prices and shortages in supply of skilled labor and engineering
services. Over the last decade, major energy models have reflected this pattern by substantially
raising their assumed cost for the construction of new power plants. However, since the
patterns of the last decade are extremely unusual and likely a very poor guide for long-term
evolution of capital costs, it would be questionable whether the trend will continue into the
future or not, assuming real costs actually stabilize as the construction market reaches
equilibrium.
Third, while there has been much scholarly attention to “learning,” three other factors also
play major roles in determining capital costs: the price of commodities (e.g., steel), labor, and a
country’s business and regulatory context. (This last factor includes environmental regulation,
subsidies and other incentives, and the broader context that how government and business
interact.) For mature technologies, labor costs are usually the single largest contributor to total
cost. In a nuclear plant built in a typical western country, labor is 90% of the total cost. For
Fourth, China has surged as a worldwide leader not only in building most of the world’s
energy efficient super-critical coal power plants but in renewable powers such as solar and wind
at a much faster speed and at much cheaper costs. The so-called “Chinese Price,” which for
mature technologies is about one half to one-third of western levels, notably reflects the
country’s large pool of low-cost highly skilled and semi-skilled labor, even with sharply rising
Page 5
labor costs in recent years. The country’s system of state capitalism has also reinforced the
factors that allow for low capital costs (and possibly hidden the real cost of projects that have
Fifth, it is an extremely challenging task to project future capex of building power plants and
there is an urgent need for a critical review of information and data available. Almost all
published capital cost estimates have been constantly revised upward overtime and these
adjustments are made not only on the nuclear power—a notable example of cost escalation—
but on almost all other power generating technologies including coal, wind, solar, etc. The
projection challenge is further reflected by the fact that the major widely-cited studies as well as
the energy models they use show a wide range of cost estimates for future capex of building
power plants and some of their projections are fundamentally different from each other on the
future cost trend—whether the costs will decrease, increase, or keep constant.
Last, the paper suggests the possibility that costs could come way down—partly from labor-
saving innovations in robotics and other fields, which have barely made a dent in the power
plant industry, and partly from pre-assembly of major components overseas, particularly from
Asian countries like China, which could reduce costs in the developed nations.
We proceed in four parts. First is an overview of the most recent literature on trends in the
costs of new power plants, drawn mainly from industry tracking services. Second is an effort to
decompose those costs and identify the major driving factors. Ideally that decomposition would
be quantitative; since most studies are not presented in a comparable fashion we, instead, offer
a qualitative assessment. Third is a case study on China, aimed at explaining the so-called
Page 6
2. General Trends in Power Plant Costs
institutions which examines capital costs of building new power plants. Table 1 lists the most
recent major studies we have examined. Our selection concentrated on studies that report
information across multiple technologies in ways that allow for comparisons across technologies
and markets. Thus we exclude, for the purpose of the present paper, the large number of
studies that focus on single technologies—such as the studies that concentrate on particular
renewable technologies and their improvement through “learning.” Notable examples include
Hultman et al. (2007), CBO (2008), Schlissel and Biewald (2008), Du and Parsons (2009), and NEI
(2011) with a focus on nuclear, Zhao et al (2008) on coal, Blanco (2009) and GWEC (2010) on
wind, IEA (2010) and EPIA (2011) on solar photovoltaic. Most of these studies we review focus
on a single country such as the study conducted by AATSE (2011) in Australia and another one
by EIA (2010) in the U.S., while a few of them compare capital costs across multiple countries. A
few are more comprehensive—covering all markets and all major technologies (e.g., IEA and
NEA, 2010; IEA, 2011; EIA, 2011). Although most of these studies are technically oriented, a few
also explore the political legal contexts that shape these technical factors. For example, Kalpan
(2008) explains how policy (including regulation) affects uncertainty (and thus cost) in power
plant costs. We have also selected studies that include a substantial focus on underlying driving
forces (e.g. Mott MacDonald, 2011), rather than simply reproducing cost estimates from other
sources.
Page 7
Table 1: Summary of Most Recent Literature on Capital Costs of Building Power Plants
Nov 2011 International Energy World Energy Outlook (WEO) annually includes both current
(IEA, 2011).
July 2011 IHS CERA IHS CERA Power Capital Costs Index (PCCI) and European
June 2011 Electric Power Provided an overview of near-term (2015) as well as longer
April 2011 U.S. Energy The Annual Energy Outlook (AEO). AEO 2011 includes
Information estimated costs for various types of power plants, starting with
March Australian Academy Reports current and projected levelized costs of electricity
2011 of Technological including capital costs, for a range of new power generating
Sciences and technologies (CCGT w/o CCS, SC w/o CCS, IGCC, wind, solar,
Page 8
Engineering (AATSE) geothermal, biomass steam, etc.) in Australia in 2020, 2030,
May 2011 Mott MacDonald Provides current and future costs of renewable and other low
March Melbourne Energy Reviews both current and future costs of three forms of
November U.S. Energy Summarizes current cost estimates for the generic utility-scale
Administration (EIA) for estimates of material and installation costs, indirect costs,
June 2010 Mott MacDonald Assesses current and future costs for the main large scale
MacDonald, 2010).
May 2010 Carter Reviews other studies of the cost of electricity from various
(Carter, 2010).
Agency (IEA) and the generation technologies for almost 200 plants in 21 countries
Nuclear Energy including coal and gas w/o CCS), nuclear, hydro, onshore and
Agency (NEA) offshore wind, biomass, solar, wave and tidal as well as CHP
Page 9
(IEA and OECD NEA, 2010).
February Lazard Ltd. Compares the levelized cost of energy for various conventional
2009).
November Kaplan Projects for 2015 costs for new fossil, nuclear, and renewable
While the existing literature—even the sample shown in table 1—is large and complex, most
First, across the advanced industrialized world—where there is reliable longitudinal tracking
of costs—capital costs have been rising rapidly for the last decade. The IHS Cambridge Energy
Research Associates (CERA) Power Capital Costs Index (PCCI) tracks and forecasts the costs
associated with the construction of a portfolio of 30 different power generation plants in North
America including coal, gas, wind and nuclear power plants, while the IHS CERA European Power
Capital Costs Index (EPCCI) tracks and forecasts the costs associated with the construction of a
portfolio of power generation plants in Europe (IHS, 2011). Both are indexed to year 2000 and
show that the costs of building new power plants have shown an overall upward trend—rising
much faster than GDP—despite several years of temporary decline during the most recent
worldwide financial crisis (see Figure 1). In particular, capital costs of building new nuclear
Page 10
power plants are soaring. The IHS PCCI without nuclear, for example, has a much flatter trend
Figure 1: North America and European Power Capital Cost Index (PCCI and EPCCI)
250
230
Overall PCCI
210
Overall PCCI without nuclear
Cost index (2000=100)
EPCCI
190
170
130
GDP deflator
110
90
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Data source: IHS (2011) and U.S. Bureau of Economic Analysis (2012)
The cost increases have been attributed to many factors such as global commodity prices—
notably steel, copper, and aluminum—and the demand for power generating equipment.
Shortages of construction services have also played a role, as have regulatory delays.
Second, there is huge cost variation across countries. For example, Figure 2 presents cross-
supercritical coal, nuclear pressurized water reactors, and onshore wind) drawn from the
systematic analysis of the IEA and the OECD Nuclear Energy Agency (NEA).
Page 11
Figure 2: 2010 Overnight Capex for Selected Power Generating Technologies (2010 USD/kWe)
Page 12
Note: Data is only available for ultra-supercritical coal in The Netherlands. Advanced nuclear here
includes different Generation III technologies including the European pressurized reactor (EPR), other
advanced pressurized water reactor designs as well as advanced boiling water reactor designs. Please be
noted that these numbers are the results of the work carried out in 2009 and focus on the expected plant-
level costs of baseload electricity generation by power plants. Real costs today may be higher than what
have been presented here. For example, overnight costs of a supercritical coal power plant in the United
States could be easily over $3,000/kW (EIA 2010), well above the number quoted in Figure 2.
Studies on individual countries and component technologies confirm these patterns. Prices
for the flue gas desulphurization (FGD) market, for example, had stabilized around $150/kW,
when after a few years of production, the Chinese manufactures had brought prices down to
$75/kW for product delivered in China (Mott MacDonald, 2011). The so-called “China price”—
that power plants can be built in China much cheaper— is strongly evident in figure 2 (which
also suggests that for nuclear and coal technologies the phenomenon should be called the
“China/Korea price”). A study utilizing Chinese data, for example, indicated that the cost of
constructing an IGCC plant in China would be almost half the cost of constructing an equivalent
Third, the few studies that have simultaneously looked at historical patterns while also
making projections point to the difficulty of serious forecasting. There is suggestive evidence
that much forecasting in this area is done retrospectively. Analysts use current costs to anchor
their expectations for long-term future costs rather than modeling capital costs as the result of
fundamental forces of supply and demand. This is evident, for example, at how major modeling
teams have adjusted the assumed future cost of generation technologies in response to short-
term variations in the real world based on current cost of those same technologies. All
Page 13
published capital cost estimates for building new power plants by utility company, investment
firms, and organizations have almost all been continually revised upward in recent years; what is
unclear, however, is whether those revisions reflect fundamental shifts in technology markets
(and thus likely to occur in the real world) or if they merely reflect variations in current
conditions that could have little bearing on the situation 5 or more (let along 50) years in the
One perspective is the updating of base capital costs. For this, we focus on the EIA’s Annual
Energy Outlook—not to criticize the report but, rather, because EIA reliably publishes its
estimates and revisions and thus analysts such as ourselves can make ready comparisons.
Figure 3 presents variations in EIA’s base year estimates of overnight capital costs for five
generation technologies over the period 2000 to 2010. Comparing Figure 3 with Figure 1
suggests that the base year estimates in EIA’s forecasts lag real world conditions by about 2
years. When looking back its past projections, EIA realized the need to significantly raising their
cost estimates. The EIA has assumed no real cost increases until 2008 in AEO 2008 (except for
fuel cell whose cost estimates were significantly raised as earlier as in AEO 2004), when it for the
first time significantly raised costs by at least 15 percent over AEO 2007 levels for all
conventional and renewable energy technologies. Overnight capital cost estimates of building
new nuclear power plants in 2011, for example, are 37% higher than those in 2010, increasing
from $3,901/kW to $5,339/kW and for advanced PC without CCS cost estimates in 2011 are 25
Page 14
Figure 3: EIA’s Base Overnight Capital Costs Estimates (2001-2010)
350
300
Constant Dollar index (2000=100)
Fuel cells
250
PC w/o CCS
200
IGCC w/o CCS
150
Advaned NGCC
100
Advanced nuclear
50
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
A second perspective comes from looking at the technology that is most capital intensive:
nuclear. Published capital cost estimates for nuclear by utility company, investment firms, and
organizations have all been continually revised upward. For example, in 2002 and 2003 the
Department of Energy and MIT envisioned that the cost of new nuclear plants would be
$1,200/kWe to $2,000 (MIT, 2003). At the time, the MIT number was striking for its high value.
At the end of that decade costs had more than doubled (see Figure 4). The same researchers
looking at the same phenomenon yielded massive changes in estimated costs over a period that
was about half the time needed for full planning, siting and construction of a nuclear plant.
Page 15
Figure 4: Recent Nuclear Construction Costs Estimates Summarized in Major Studies
$6,000
Overnight Cost
$5,000
Total Plant Cost
$4,000
$3,000
$2,000
$1,000
$0
DOE (2002) MIT (2003) Moody's Florida MIT (2009) DOE (2010) DOE (2011)
Investor Power &
Services Light (2007)
(2007)
different studies—something that the National Renewable Energy Laboratory (NREL) did in 2010
for six credible and widely discussed models: U.S. EIA, IEA, NREL, Pacific Northwest National
Laboratory (PNNL), ICF International, and Electric Power Research Institute (EPRI). Although the
models had nearly identical base years and aimed to explain the same underlying phenomena,
neither in base year nor in projected years was there any convergence around particular values
(see Figure 5). For example, the studies by U.S. EIA, ICF International, and PNNL all suggest the
future capex of coal-fired power plants is expected to continue declining, while the other two
Page 16
Figure 5: Comparison of Overnight Costs of Selected Technologies in the U.S. in Selected Studies
Page 17
3. Component Cost Analysis: What Are Major Capex Drivers?
Building new power plants is a complicated process affected by numerous factors. One is
definitions since “capital cost” is a general term. Capital costs of building power plants can be
defined at five levels: the Bare Erected Cost (BEC), the Engineering, Procurement and
Construction Cost (EPCC), the Total Plant Cost (TPC), the Total Overnight Capital (TOC), and the
Total As-Spent Capital (TASC). As shown on figure 6, the first four are “overnight” costs, which
are the focus of this paper and expressed in “base-year” dollars (for definitions of these terms,
please see the caption). By contrast, the last one, TASC, comprises the TOC plus both escalation
and interests on debts during the capital expenditure period and therefore, is expressed in
Figure 6: Base Plant Site Capital Cost Estimates for Selected Technologies (2010$/kW)
6,000
Owners' cost
3,000
TOC
TPC
2,000 EPCC
BEC
1,000
0
Advanced Supercritical Coal Onshore wind Advanced nuclear PWR
Page 18
BEC: refers to Bare Erected Cost which comprises the cost of process equipment, on-site facilities and
infrastructure that support the plant and the direct and indirect labor required for its construction and/or
installation.
EPCC: refers to the Engineering, Procurement and Construction (EPC) Cost which comprises the BEC plus
the cost of services provided by the EPC contractor. EPC Services refers to the cost of services provided by
the EPC contractor including detailed design, contractor permitting (i.e., those permits that individual
contractors must obtain to perform their scopes of work, as opposed to project permitting, which is not
TPC: refers to the Total Plant Cost which comprises the EPCC plus process and project contingencies.
TOC: refers to the Total Overnight Capital which comprises the TPC plus owner’s costs which include
preproduction (start-up) costs, inventory capital, and cost of securing financing but excluding interest
BEC normally accounts for the largest share of the total capex, ranging from 45 to 85%
percent of the total depending on power technologies. For onshore wind, for example, over 80%
of the cost is BEC, mainly referring to the cost of the wind turbine generator and its installation.
By contrast, for the nuclear power plant BEC accounts for less than a half of the total capex and
EPC services could be as high as one quarter of the total, reflecting the factor that a significant
share of the capex for the nuclear is labor. In fact, if excluding a significant share of labor input
already embodied within the equipment, the equipment costs could be as low as 10% for
nuclear (Mott MacDonald, 2011). In this paper we have used “capital cost” principally to mean
BEC or EPCC—a definition that corresponds closely to the concept of “overnight” capital cost.
The literature suggests that four factors dominate capital cost trends for most large power
plants. (There is not much literature with a focus on capex drivers of power plants. Most of
literature in Table 1 discusses non-policy factors including learning, materials, and labor.
Page 19
MacDonald (2011) presents a full list, but it only focus on low-carbon technologies and does not
rank all influential factors.) We start with the factor that is already most extensively analyzed in
through economics of scale. Since 2000, the learning curve or experience curve concept has
become an increasingly popular tool to analyze the cost development of electricity generation
OECD/IEA, 2000; McDonald and Schrattenholzer, 2001; Papineau, 2006; Jamasb and Köhler,
2008, etc.). The use of the learning curve is based on a deeper history of studying technological
change in the energy industry (Grubler, et al., 1999). Strictly speaking, “learning” applies only
from the moment that the first commercial demonstration projects are built. Learning rates—
the level of improvement for every doubling of installed capacity—have often been observed at
the rate of about 30% for immature technologies, slowing to 10% to 20% as the technology
matures and then flattens. At extremely large deployments the concept of “learning” may
become less relevant, not least because other factors—such as labor costs and materials—can
The concepts are familiar and summarized in figure 7. An emerging technology moves from
a concept whose uncertain concept through a costly first-of-a-kind demonstration into niche
markets and then more widespread deployment. As many units are built, the low costs become
more certain and buyers will face lower project risks. Currently, various power generation
technologies face their different phases of learning curves with different cost trends. Subcritical
coal (SCPC) plants, for example, are considered as a mature technology and their unit capital
costs are expected to stabilize. By contrast, advanced Gen III/III+ nuclear technologies are still in
Page 20
the phase of development or demonstration with highly uncertain cost estimates. Its unit
capital cost is expected to decline through learning as more units are built until the technology
matures and the cost trend line becomes flat. Because learning has been subjected to extensive
prior analysis (e.g. Grubler, et al., 1999; OECD/IEA, 2000; McDonald and Schrattenholzer, 2001;
Papineau, 2006; Jamasb and Köhler, 2008, Gallagher et al., 2012, etc.) we don’t examine it in
more detail here. We note that the exact role and wisdom of studying these phenomena as
“learning” remains controversial since “learning” is general term that can obscure lots of other
processes at work; moreever, learning must compete with other factors that affect performance,
Page 21
3.2 The commodity market
A second major factor that determines total cost for a plant is raw material (e.g., steel,
cement). For mature technologies—that is, technologies for which learning effects are small—
the cost of raw materials is widely believed to be the largest source of revision in assumed
capital costs. The last decade has seen dramatically increasing raw materials prices, tightly
linked to a rise in actual capital costs. These trends are seen in figure 8, which shows the overall
PCCI (reprinted from figure 1) along with the index for metal prices.
Figure 8: U.S. Commodity Prices and Power Plant Capital Cost over Years
250
Overall PCCI
200
Index (2000=100
150
50
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Data sources: IHS (2011) and U.S. Bureau of Economic Analysis (2012)
The cost increases in raw materials have been much faster than the growth in GDP in
mature economies (figure 8 shows the US). This decoupling is due to high global demand for
commodities (notably from China’s rapidly expanding economy) and lagging investment in new
Page 22
output capacity; higher production and transportation costs (in part owing to high fuel prices,
which are driven by the same underlying phenomena) along with a weakening U.S. dollar all play
However, it is highly unlikely that rising costs for raw materials will lead to a major rise in
the total capital cost of power plants. The direct linkage between rise in raw material prices and
total capital costs is weak. Mott MacDonald (2011)’s study, for example, indicates that for low-
carbon technologies the basic raw materials even at the peak of the market typically account for
This decadal rise in commodity prices has been referred by some analysts as a “super cycle”
and is, perhaps, immune to the normal forces that bring prices back to historical patterns
(Cuddington and Jerrett, 2008). In fact, a return to normalcy may be under way at this writing.
Following more than two years of strong growth and reaching a peak in early 2011, commodity
prices has recently declined on concerns about global macroeconomic and financial outlook and
slowing demand in emerging markets, notably in China (World Bank, 2012). According to World
Bank’s forecasts, despite of both downward and upward risks, commodity prices are most likely
to decline into the medium term for all metals with the exception of aluminum, which is
expected to rise and supported by higher costs for power and other inputs (World Bank, 2012).
Looking to the longer term, commodity markets are inherently volatile and cyclical with booms
and slumps (Cashin et al., 2012). Prices will depend not just on future demand from emerging
economies but also the prospects for expanded investment in supply as well as innovation in the
methods for extracting and processing existing materials and reduction in demand due to
lightweighting and switching to substitutes (Streifel, 2006). We are mindful that even for high
end uses of metals such as aluminum—which for decades had a prized position in some
industries such as aircraft manufacture—rivals exist. During the Second World War Howard
Page 23
Hughes built a troop airplane from wood because aluminum was in short supply. (It flew once
and was a financial disaster.) Today, about half the airframe of the Boeing 787 is carbon fiber,
plastics and other composite materials. (It appears to be a commercial success, and earlier
experience with composites on aircraft, such as in the tail of the Boeing 777, indicate that
composites don’t just replace aluminum and weigh less—they also require lower maintenance.)
Third, we look at labor costs. Labor comes in many forms--low-, semi- and highly-skilled—
and all are relevant for power plants. Figure 9 shows both U.S. and China’s labor cost index in
the manufacturing sector over the last decade. Between December 2000 and December 2011, in
the United States the general inflation rate (measured by the GDP deflator) increased about 27
percent and during the same period, the country’s cost of manufacturing labor increased about
31 percent, slightly higher than the rate of general inflation. By contrast, China’s labor cost has
increased dramatically, more than tripling within ten years with a double digit growth rate
annually.
Page 24
Figure 9: Labor Cost and Power Plant Capital Cost Over Years
400
350
China Labor Cost Index
(Manufacturing)
300
Index (2000=100
250
Overall PCCI
200
U.S. Labor
150 China GDP Deflator Cost Index
(Manufacturing)
100
U.S. GDP Deflator
50
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Data sources: IHS (2011), U.S. Bureau of Economic Analysis (2012), and National Bureau of
There are growing concerns about a gap between demand and supply of skilled construction
labor. In most developed countries, the average age of the current construction skilled
workforce is rising rapidly, and high attrition rates in construction are compounding the problem.
Demand for such workers is expected to outpace supply over the next decade (DOE, 2006), and
within the power industry there are perennial concerns about how to address problems in the
aging workforce. In developing countries, even China with the world’s largest labor force, there
Population and Labor Economics at the Chinese Academy of Social Sciences (CASS), China
already reached a “Lewisian turning point,” where rural surplus labor is depleted to such a level
that continuing industrialization cannot be supported cheaply (Cai, 2007). A 2011 report
Page 25
published by the U.S. Bureau of Labor Statistics (BLS) stated that from 2002 to 2008, hourly
labor costs in the manufacturing sector in the United States increased by 19%, while the
Increased labor costs have a direct impact on capital costs. A large contributor is the cost of
construction (EPC) services. Prices for EPC services have risen sharply in the last decade, and a
backlog of projects at the four major EPC firms—Fluor Corporation, Bechtel Corporation, The
Shaw Group Inc., and Tyco International Ltd—has grown. Between 2005 and 2006 alone the
backlog rose one-third (from $4.1 billion to $5.6 billion), although that pipeline of projects
includes a wide array of EPC activities—in addition to power plants there are roads, port
facilities and water infrastructure (Chupka and Gregory, 2007). Over the same period, this
severe tightening in the EPC market affected EPC costs for particular power plants. A filing by
Oklahoma Gas & Electric Company (OG&E) indicated, for example, that the estimate for EPC
services for a new coal power plant had increased by more than 50% during the nine month
period in 2006, contributing to an overall 10% increase in the expected capital cost of the plant
One major source of uncertainty in future labor costs concerns the prospects for labor-
saving innovations, such as from robotics. For decades, robots have become a strategic
technology for industries like agriculture (if considering mechanization as an earlier version of
for these industries are to lower unit costs of production and to improve product quality
(Siciliano and Khatib, 2008). However, robotics has yet made a significant dent in the utility
industry. The only partial exception is in nuclear power plants where robots have been used to
reduce radiation exposure to human workers and also to reduce plant downtime (Moore, 1985;
Page 26
Kim, 2010; Sugisaka, 2011). The wider application of robotics in the sector could be
revolutionary. It is reported, for example, new robots are building solar power plants in the U.S.
(V-Kraftwerker, 2012). Since building open area solar parks is mostly manual work and identical
processes are repeated hundreds of thousands of times, the new robots are able to massively
Another possible huge cost-savings may come from the import of pre-assembled major
components overseas. In effect, pre-assembly allows markets where labor costs are high to take
advantage of the lower costs in other manufacturing countries; as pre-assembly expands it can
be applied not just to low-value labor operations but also highly skilled activities (Mahmud,
2008). It is reported, for example, that a $7.2 billion repair of the San Francisco-Oakland Bay
Bridge could save at least 400 million by importing pre-assembled bridge decks from China (Pyle,
2011). Looking to the future could see a larger role for major component imports. Over a
longer time horizon, a shift to modular nuclear reactors could allow for international trade—
rather than assembly in place—of the most expensive elements of nuclear plants.
Regulation, in some settings, may also be a major factor in the cost of building new power
plants. However, in many integrated assessment models major regulations are resolved
independently and treated, in effect, as a form of technological change. For example, models
typically treat base coal plants differently from plants that include carbon capture and storage
technologies; some models also treat other pollution control technologies (e.g., scrubbers and
SCR) separately. A few studies have looked at the scope for changes in such technologies—such
as learning by doing and other forms of innovation in scrubbers (Taylor, et al. 2005).
Due to tightened and changing health and safety compliance requirements, for example,
very few coal-fired power plants have been built in the US in the past decade and the share of
Page 27
newly built coal-fired plants among the total has declined for years (see Figure 10). (Looking to
the future, those regulations along with low prices for natural gas suggest that little if any new
coal-fired capacity with be built.) Since 1990, more than 80% of new capacity has been natural
gas-fired, which are cost-competitive with coal and emit no SO2, no mercury, no other
hazardous air pollutants, and only about half of the warming emissions of coal-fired units (EIA,
2011). Recently, U.S. EPA is proposing a series of regulations on coal-fired power plants. As of a
result, a large number of old plants will have to retrofit or even retire. Some of these retired
units will be replaced by new capacity, of which some will be coal-fired, but most replacements
are likely to be natural gas combined cycle units (McCarthy and Copeland, 2011).
Figure 10: U.S. Power Plant Capacity, by Type and Year It Entered Service
60,000
40,000
coal (MW)
30,000
20,000
10,000
Other types of power plants also face regulatory interventions—notably nuclear power,
where regulatory burdens and growing public concern led to higher costs and longer delays. The
Page 28
time from project initiation to ground breaking, for example, was increased from 16 months in
1967, to 32 months in 1972, to 54 months in 1980 (Cohen, 1990). In the aftermath of Three
Mile Island no new plants were initiated until a new regulatory framework was crafted in the
late 1990s and a license to build two reactors was just recently approved by the U.S. Nuclear
Regulatory Commission, which is the first in over 30 years (Hargreaves, 2012). Even with this
new framework, the construction and operation of a nuclear power plant in the U.S. is subject to
many veto points. Siting in other countries has been much easier. In Korea and China—the two
countries that are most active builders (and lowest cost providers) of nuclear plants today see
Policy makers intervene in markets with many tools beyond just regulation, and their
interventions don’t always raise costs. A wide array of policies aimed at supporting novel
with learning) to lower costs. For example, the German Feed-In Tariff, part of the German
Renewable Energy Sources Act of 2000 (Ernaeuerbare-Energien-Gesetz, EEG), has been the most
important driving force in that country’s expansion of renewable energy in the electricity sector
(Frondel et al., 2008). In the U.S., an important policy driver for renewable energy has been
production and investment tax credits (along with cash payments in some circumstances).
Government has also played a role in lowering the risks associated with investing in (and thus
liability. Another is government-backed financing, which lowers borrowing costs. The 2009 U.S.
economic stimulus, for example, provided more than $2.5 billion to finance clean-energy loan
guarantees—funds that were leveraged into about $16 billion in investment projects (Synder,
2012). Still another model is direct financing—something that is easier, perhaps, in countries
Page 29
where governments own the firms that deploy new technologies and have control over banks
and other sources of capital. We return to this topic in our case study on China.
A perennial challenge in the deployment of new technology is managing the risks associated
with deploying pre-competitive and unproven technologies; while many different forms of risk
management techniques have emerged, most share the common feature that the government
assumes (or assigns, such as to consumers or to regulated enterprises) the extra risk associated
4. Case Study: Why Can China Build Power Plants Much Cheaper?
Asian nations, particular China, are leading the way in building most of world’s new power
plants at a much faster speed and at much cheaper costs. Today, for example, most of the
world’s most efficient (supercritical and ultra-supercritical) coal-fired power plants, are being
built in China (Nalbandian, 2008). The cost to build an ultra-supercritical power plant in China is
less than one third of building the similar one in the United States (see figure 2). Onshore wind
turbines are built in China for half the cost typical in the west, and compared with some high
cost countries (e.g., Switzerland) the Chinese cost is a lot lower (figure 2). Although famous for
its massive coal fleet, China overtook the United States in 2010 as the world leader in
accumulative wind power installed capacity with a total capacity of 62.4 GW by the end of 2011,
more than one fourth of the global wind capacity (World Wind Energy Association, 2012). China
China (and Taiwan’s) market share has sharply risen as they have driven down costs and prices
Page 30
Figure 11: Average Retail Module Price Sold in the U.S. and China & Taiwan Market Share of
6 60%
Retail Moduel Price Index ($)
4 40%
Price Per Watt Peak
3 30%
2 20%
1 10%
0 0%
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
An explanation for the very low capital costs in China is found by looking at all four of the
Learning requires building and experience. Some of the lower cost in China may simply
reflect huge economies of scale—and thus rapid riding down learning curves—that come from a
large domestic market and (where domestic demand is relatively small, such as in solar) an
export oriented manufacturing system. China’s large wind capacity has grown to its current 62
GW from just 1.3 GW in 2005. In nuclear power, 25 out of a world total of 61 are under
construction in China—by contrast, just one is under construction in the United States (World
Page 31
Nuclear Association, 2012). The total Chinese nuclear capacity (40 GW in 2011) is expected to
Of the four factors we have considered, the only one that appears not to play a major role in
China is commodity prices. Perhaps a decade (or more) earlier, central administration of the
Chinese economy could help explain why producers faced often lower prices for strategic
gradual transition to a market economy, China implemented a dual price system Since February
1985, permitting producer goods exchange at two different prices: a state-set price, for centrally
rationed supplies, and a higher free-market price (Naughton, 1996). Today, except for only a few,
almost all commodity prices in China conform to international prices (Streifel, 2006). In fact,
China has become the most important player in international commodity markets and
international prices will depend very much on the pace of demand in China and its status as a
The heart of “the China Price”—in general, which means 30% to 50% lower cost for
manufacturing products when compared with the U.S. or anywhere else—is widely believed to
be the country’s much cheaper labor costs (Engardio et al., 2004). While there has been
extensive analysis of the “China price” for manufacturing, there has been a few studies looking
at differences in large-scale engineering projects (e.g. Yang, 2011), where the “the China Price”
we argue is not only benefited from its lower labor costs, but more critically promoted by its
regulatory environment.
manufacturing were roughly 6 percent of those in the United States, slightly surpassing those of
some developing countries like the Philippines, but still significantly lagged behind those of
other countries like Brazil (see Figure 12). The average manufacturing wage in China in 2010 is
Page 32
still only about $2 an hour (CBS, 2011), compared with $35 in the United States (U.S. Bureau of
Labor Statistics, 2011), though the wage is expected to much higher up to about $5 in the
China’s hourly cost advantage, while still significant, is shrinking rapidly. The real wages for
manufacturing workers in China have grown nearly 12% per year (see Figure 10). Harder to
parse are the trends for highly skilled professionals where there is a large pool of workers who
can work in conditions where their western peers will not (e.g.,12-hour days, on weekends,
willing to live in sheds close to projects, etc.). Today, China produces far more number of first
university degrees in natural science and engineering (NS&E), compared to the United States
and also the largest number of doctoral degrees in engineering, rising four-fold from
approximately 4,000 in 2001 to over 15,000 in 2008, compared to 8,000 in the United States and
Page 33
Figure 12: Index of Hourly Compensation Cost in Manufacturing, Selected Countries, 2010
Index (U.S.=100)
140
120
100
80
60
40
20
0
United China Phillippines Brazil Korea, Japan Germany
States Republic of
Note: Direct comparisons are difficult due to differences in statistical techniques and the
Source: Data for China is from China Labor Statistics Yearbook 2011. Data for other
favorable for large construction projects. Its system of “state capitalism”—which is a political
system that “tries to meld the powers of the state with the powers of capitalism, depends on
government to pick winners and promote economic growth, and also uses capitalist tools such
as listing state-owned companies on the stock market and embracing globalization (The
Economist, 2012)”—are particularly well suited to large, financially risky engineering activities
Page 34
First, industry is dominated by state-owned enterprises (SOEs), which account for 80% of
the value of the stock market in China. Energy and utilities, in particular, are the two industrial
sectors that SOEs have the most influential footprints (The Economist, 2012). As in many other
settings where SOEs operate with soft budget constraints, well-managed SOEs (which are the
norm in China) are often better able to manage risks (and shift them to government) while
Second is the role of state-controlled finance, notably from Chinese policy banks that have
largely financed the buildup of new power plants through almost-zero interest debt, which
provides a significant cost advantage over those builders in economic systems where the state
plays a more distant role. Cheap or free public finance does not necessarily affect overnight
costs—such as bare erected cost (see Figure 6)—but it can lower risks that are inherent in such
projects and reduce costs for some forms of procurement. The most notable example is the
Chinese Development Bank (CDB), one of the three Chinese policy banks. Originally operated as
an arm of the Chinese central government, CDB is now a joint stock company with limited
liability that reports to China's national cabinet on certain policy issues and it is the only bank in
China whose governor is a full minister. This allows the Chinese government to direct funds to
activities that the government is keen to support—including infrastructure, which has garnered
massive investment as the central economy has tried to promote economic growth and has
been worried that adequate electric power supply could be an impediment to growth. It is
reported that CDB had $687.8 billion in loans on its books at the end of 2010, more than twice
as much as the World Bank (Bloomberg, 2011a). The CDB raises most of its money via long-term
bonds and with an implicit state guarantee, which makes it easier, cheaper, and a lot less risky
for utility builders to obtain financing. Such advantages are not unique to China, and they often
arise when state-backed firms raise debt—such as state-backed oil companies (e.g., Pemex)
Page 35
(Victor et al., 2012) or state-backed Tennessee Valley Authority, which recently raised long-term
One strategic focal point of CDB nowadays is to finance emerging industries, which includes
solar power and other forms of clean energy. It was reported, for example, that in 2010 alone
CDB reached an agreement on providing $30 billion in low-cost loans and credits to the top five
Chinese solar panel manufacturers, with $8.9 billion to LDK Solar alone, the largest Chinese solar
player (Mercom Capital Group, 2011). It is believed that the huge almost-zero interest debt
funded by Chinese policy banks at least help Chinese solar companies achieve economies of
scale, drop the price of solar at an astonishing pace, and therefore dominate the solar market
(Lacey, 2011). This advantage, which is rooted in the relationship between the Chinese state
and its state-dominated banking system, is also the source of growing trade conflicts—notably
Third, the Chinese state—as with essentially all governments that are funding new energy
sources—has offered massive subsidies. It has offered reduced corporate income tax (CIT),
significant reduction in value added tax (VAT), other tax incentives, and feed-in tariffs and
subsidies to operators of renewable energy projects to compensate for their costs. A reduced
CIT rate of 15%, for example, is given for advanced and advanced and new technology
enterprises including solar and wind, compared to a regular rate of 33%. Chinese wind farms are
also eligible for a 50% VAT rebate and VAT has been reduced from 17% to 8.5% (Zhang et al.,
2011). In July 2011, a national feed-in tariff for solar projects, 1.00 or 1.15 RMB/kWh (about
$0.15 to $0.18 per kWh) depending on the construction time, was announced to form its
domestic solar market (NDRC, 2011). The feed-in tariff for wind projects now ranges from 0.51
resources (Zhang et al., 2011). To stimulate its distributed/rooftop solar PV, for example, the
Page 36
China government launched in 2009 the "Solar Roofs Plan" and the "Golden Sun Project." As the
end of 2010, the "Solar Roofs Plan" approved 210 projects and covered 181 MW of approved PV
capacity with a total cost to the Chinese government of RMB 2.4 billion (or roughly 380 million)
and the "Golden Sun Project" approved 314 projects involving over 630 MW of installed capacity
with a total construction costs associated with the projects to be roughly RMB 20 billion (or
A small fraction of the subsidies that have benefitted China’s industry have come from
overseas—notably through the Clean Development Mechanism (CDM) that allows polluters in
the EU to offset carbon emissions at home by investing in clean energy abroad. It is estimated
that about one in five of China’s renewable energy projects have been the beneficiaries of the
Europe’s cap and trade program (Kraeme, 2012). As of March 6, 2012, China had 1,843 CDM
projects registered at the UN, about half of the world total, and expected average annual carbon
credits from the registered projects reach 336 million with a value ranging from 3 to 10 billion
euro depending on price of carbon credits. About 80 percent of these CDM projects involve
hydro and wind power projects (UNFCCC, 2012). It is estimated that emissions reduction
revenues make up 20 percent of their income wind and small hydropower projects (Lan, 2012).
We have argued that the assumed cost of building power plants is a variable that deserves
much more attention by analysts working with tools such as integrated assessment models.
Typically, such models assume that capital costs will be largely fixed except for new technologies
(often renewables) where costs will decline with experience. Yet our review shows that costs
for existing, mature technologies are, in fact, highly variable around the world and it is plausible
that that variability will remain high far into the future.
Page 37
While there has been much scholarly attention to “learning,” three other factors also play
major roles in determining capital costs: the price of commodities (e.g., steel), labor, and a
country’s business and regulatory context. (This last factor includes environmental regulation,
subsidies and other incentives, and the broader context that how government and business
interact.) The price of commodity materials, though widely believed to be important, only plays
a minimal role in the overall costs. For mature technologies, labor costs are usually the single
largest contributor to total cost. In a nuclear plant built in a typical western country, labor is 90%
of the total cost. Along with labor costs, regulatory contexts probably largely explain the large
differences across countries. Yet regulatory and policy issues are often ignored or hard to
include in coarse models that are designed to examine engineering rather than institutions.
Studies such as the one here have many implications for policy. One is that some of the
most consequential policies are those that, traditionally, have not been a central focus of energy
policy analysts. For example, given the central role of labor costs it is plausible that changes in
the structure of labor markets—such as initiatives to make those markets more competitive and
to allow migration of skilled labor—could have a much larger impact on future costs of energy
technologies than most of the other factors we commonly study. Similarly, policies that affect
the ability of firms to manage risks associated with large capital projects—for example, the
presence of soft budget constraints and state-backed financing typical of many state-owned
enterprises, including the large energy firms in China—can have a massive impact on the
This essay suggests that large-scale energy models might devote more attention to the
existing and plausible future variability in capital costs. For example, they might adopt scenarios
that explore the possibility of large changes in costs over time—even in mature technologies—
due to changes in labor markets, industrial organization and regulation. While the rise in labor
Page 38
costs points to a future where capital costs for plants is rising, it is possible that costs could
come way down—partly from labor-saving innovations in robotics and other fields, which have
barely made a dent in the power plant industry, and partly from pre-assembly of major
components overseas, particularly from low cost Asian countries like China and Korea.
Page 39
References
1. AATSE (Australian Academy of Technological Sciences and Engineering), 2011. New Power
Cost Comparisons: Levelized Cost of Electricity for a Range of New Power Generating
2. Holliday, C., Augustine, N. R., Burns, U., Doerr, J., Gates, B., Immelt, J., and SolsoAEIC, T.,
2010. A Business Plan for America’s Energy Future. American Energy Innovation Council,
Washington D.C.
3. Banister, J. and Cook, G., 2011. China’s employment and compensation costs in
4. Benkard, C. L., 2000. Learning and forgetting: The dynamics of aircraft production. American
5. Blanco, M.I., 2009. The economics of wind energy. Renewable and Sustainable Energy
6. Bloomberg, 2011. Financing China Costs Poised to Rise with CDB Losing Sovereign-Debt
Status. May 2.
7. Cai, F., 2007.Green Book of Population and Labor: The Coming Lewisian Turning Point and
8. Carter, L.D., 2010. Prospecting for Power: The cost of meeting increases in electricity
9. Cashin, P., McDermottb, C.J., and Scott, A., 2012. Booms and Slumps in World Commodity
10. CBO (Congressional Budget Office), 2008. Nuclear Power’s Role in Generating Electricity.
Congressional Budget Office, the Congress of the United States: Washington D.C.
Page 40
11. Chu, S. and Majumdar, A., 2012. Opportunities and challenges for a sustainable energy
12. Chupka, W.M. and Gregory, B., 2007. Rising Utility Construction Costs: Sources and Impacts.
13. Cohen, L. B., 1990. The Nuclear Energy Option. Plenum Press.
14. Cuddington, J. and Jerrett, D., 2008. IMF Staff Papers. 55, 41–565.
15. DOE (Department of Energy), 2002. Annual Energy Outlook 2002. Washington D.C.
16. DOE, 2006. Workforce Trends in the Electric Utility Industry: A Report to the United States
19. DOE NREL (National Renewable Energy Laboratory), 2010. Cost and Performance
6A20-48595.
20. DOE NETL (National Energy Technology Laboratory), 2011. Cost estimation methodologies
21. Du, Y. and Parsons, E.J., 2009. Update on the Cost of Nuclear Power. Massachusetts Institute
22. EIA (Energy Information Administration), 2010. Updated Capital Cost Estimates for
23. EIA, 2011. Today in Energy: Age of electric power generators vary widely. June 16.
Washington D.C.
24. Engardio, P., Roberts, D., and Bremner, B., 2004. The China Price. Bloomberg BusinessWeek.
December 6.
Page 41
25. EPIA (European Photovoltaic Industry Association), 2011. Solar Generation 6: Solar
26. EPRI (Electric Power Research Institute), 2011. Program on Technology Innovation:
27. Fischer, C. and Newell, R. G., 2008. Environmental and technology policies for climate
28. Flessner, D., 2012. TVA selling bonds at record low rates. August 8th. Timesfreepress.
29. Frondela, M., Rittera, N., and Schmidta, M.C., 2008. Germany's solar cell promotion: Dark
30. Grubler, A., Nakicenovic, N., and Victor, D.G., 1999. Dynamics of energy technologies and
31. Gallagher, K.S., Grubler, A., Kuhl, L., Nemet, G.F. and Wilson, C., 2012. The Energy
32. GWEC (Global Wind Energy Council), 2010. Global Wind Energy Outlook 2010.
33. Hargreaves, S., 2012. First new nuclear reactors OK'd in over 30 years. February 9. CNN.
34. Hultman, N.E., Koomey, J.G, and Kammen, D., 2007. What History Can Teach Us about the
Future Costs of U.S. Nuclear Power. Environmental Science & Technology, APRIL 1, 2088-
2093.
35. IEA (International Energy Agency), 2000. Experience Curve for Energy Technology Policy.
36. IEA (International Energy Agency), 2007. World Energy Outlook 2007. OECD/IEA, Paris.
37. IEA, 2010. Technology Roadmap - Solar Photovoltaic Energy, International Energy Agency.
Page 42
38. IEA and NEA (Nuclear Energy Agency), 2010. Projected Costs of Generating Electricity: 2010
40. IHS, 2011. Power Plant Construction Costs: Cost Pressures Returning. Available from: <
http://press.ihs.com/press-release/energy-power/power-plant-construction-costs-cost-
pressures-returning>.
41. Jamasb, T. and Köhler, J., 2008. Learning curves for energy technology: a critical assessment.
In ed. Grubb. M. et al. Delivering a Low Carbon Electricity System: Technologies, Economics
42. Kaplan, S., 2008. Power Plants: Characteristics and Costs. CRS Report for Congress.
43. Kee, E., 2010. Global Nuclear Power Developments - Asia Leads The Way. Nuclear Energy
44. Kim, S., 2010. Application of robotics for the nuclear power plants in Korea. 2010 1st
International Conference on Applied Robotics for the Power Industry (CARPI), 5-7 Oct. 2010.
1-5.
45. Kraeme, S., 2012. China, Beneficiary of EU’s Cap & Trade, Now Shirks its Responsibilities.
46. Lacey, S., 2011. Are the Chinese Using Predatory Pricing to Knock America Out of Solar
47. Lan, L., 2012. Buyers default on carbon credits. China Daily. February 9.
48. Lazard Ltd., 2009. Levelized Cost of Energy Analysis - Version 3.0. New York, NY: Lazard Ltd.
49. Mahmud, K. 2008. Outsource in the Construction Industry. Fraser Milner Casgrain LLP.
50. McDonald, A. and Schrattenholzer, L., 2001. Learning Rates for Energy Technologies, Energy
Page 43
51. Nemet, G. F., 2012. Subsidies for New Technologies and Knowledge Spillovers from Learning
52. Mercom Capital Group, 2011. Loans and Credit Agreements involving Chinese Banks to
credit-agreements-involving-chinese-banks-to-chinese-solar-companies-since-jan-2010.
53. Mints, P. and Donnelly, J., 2011. Photovoltaic Manufacturer Shipments, Capacity and
55. MIT, 2009. Update on the Cost of Nuclear Power. Boston, MA.
56. Moore, T., 1985. Robots for nuclear power plants. IAEA Bulletin. Autumn. 31-38.
57. Mott MacDonald, 2010. UK Electricity Generation Costs Update. Brighton, United Kingdom.
58. Mott MacDonald, 2011. Costs of low-carbon generation technologies. Brighton, United
Kingdom.
59. Nalbandian, H., 2008. Performance and risks of advanced pulverised coal plant. IEA Clean
Coal Center.
60. National Bureau of Statistics of China, 2012. China Labour Statistical Yearbook 2011. Beijing,
China.
61. Naughton, B., 1996. Growing Out of the Plan: Chinese Economic Reform, 1978-1993.
62. NDRC (National Development and Reform Commission), 2011. The Notice on Improving the
http://www.ndrc.gov.cn/zcfb/zcfbtz/2011tz/t20110801_426501.htm
Page 44
63. NEI (Nuclear Energy Institute), 2011. The Cost of New Generating Capacity in Perspective.
Washington D.C.
64. Nordhaus, W.D., 2009. The perils of the learning model for modeling endogenous
technological change. National Bureau of Economic Research Working Paper Series, No.
14638.
65. NPD Solarbuzz, 2012. Module Pricing: Retail Price Summary - March 2012 Update. Available
at http://www.solarbuzz.com/node/3184.
66. NSF (National Science Foundation), 2012. Science and Engineering Indicators 2012. National
67. Papineau, M., 2006. An Economic Perspective on Experience Curves and Dynamic
68. Pielke, R., 2011. The Climate Fix: What Scientists and Politicians Won't Tell You About Global
69. Powell, B., 2011. The End of Cheap Labor in China. The Time Magazine. June 29.
70. PV-Kraftwerker, 2012. New Robot builds solar power plants / PV-Kraftwerker introduces a
71. Pyle, R. 2011. Bridge Comes to San Francisco With a Made-in-China Label. June 25. New York
Times.
72. Rai, V., Victor, D. G., Thurber, M.C., 2010, Carbon Capture and Storage at Scale: Lessons
from the growth of analogous energy technologies, Energy Policy, 38, 4089-4098.
73. Schlissel, D. and Biewald, B., 2008. Nuclear Power Plant Construction Costs. Synaps Energy
74. Siciliano, B. and Khatib, O. (Eds.), 2008. Springer Handbook of Robotics. Springer.
Page 45
75. Streifel, S., 2006. Impact of China and India on global commodity markets focus on metals
76. Sugisaka, M., 2011. Working robots for nuclear power plant disasters. 2011 Proceedings of
the 5th IEEE International Conference on Digital Ecosystems and Technologies Conference
77. Synder, J., 2012. New Clean-Energy Loan Guarantees Planned by Obama Administration.
78. Taylor, M.; Rubin, E., and Hounshell, D., 2005. Control of SO2 emissions from power plants: A
case of induced technological innovation in the U.S. Technological Forecasting and Social
79. The Economist, 2012. The Special Report on State Capitalism Jan, 21st.
81. U.S. Bureau of Economic Analysis, 2012. Interactive Data Tables on National Economic Data.
Washington, D.C.
82. Victor, D.G., 2002. Innovation and performance in the electric power industry. In: Steil, B.,
83. Victor, D.G. and Yanosek, K., 2011, The Crisis in Clean Energy: Stark Realities of the
84. Victor, D.G., 2011, Global Warming Gridlock: Creating More Effective Strategies for
85. Victor, D.G., Hults, D, and Thurber, M., 2012 (Eds.). Oil and Governance: State-Owned
Enterprises and the World Energy Supply. Cambridge University Press, Cambridge.
Page 46
86. Wang, S., 2011. Distribution PV and Incentives in China. High Penetration PV Systems into
87. Wessner, C. W. and Wolff, A. Wm., 2012 (Eds.). Rising to the Challenge: U.S. Innovation
Policy for Global Economy. The National Academies Press. Washington D.C.
88. World Bank, 2012. Global Commodity Market Outlook. Washington D.C.
89. World Nuclear Association, 2012. World Nuclear Power Reactors & Uranium Requirements.
90. World Wind Energy Association, 2012. World Market recovers and sets a new record: 42 GW
http://www.wwindea.org/home/index.php.
91. Yang, C., 2011. A comparison of the nuclear options for greenhouse gas mitigation in China
92. Zhang, D., Zhang, X., He, J. and Chai, Q., 2011. Offshore wind energy development in China:
Current status and future perspective. Renewable and Sustainable Energy Reviews 15,
4673– 4684.
93. Zhao, L., Xiao, Y., Gallagher, K., Wang, B., Xu, X., 2008. Technical, environmental, and
Page 47