Energy Research & Social Science 46 (2018) 103–113

Contents lists available at ScienceDirect

Energy Research & Social Science

journal homepage: www.elsevier.com/locate/erss

Original research article

The politics of energy scenarios: Are International Energy Agency and other T
conservative projections hampering the renewable energy transition?
⁎
Gerry Carrington, Janet Stephenson
Centre for Sustainability, University of Otago, P.O. Box 56, Dunedin 9054, New Zealand

A R T I C LE I N FO A B S T R A C T

Keywords: Scenario-building assists commissioning organisations to understand the multiple forces that shape their future.
Solar PV Governments and investors use the scenario projections of authoritative organisations to help drive their
Scenario planning and decision-making. But what if scenarios consistently fail to represent a credibly established tech-
Growth projection nology trajectory, particularly for a topic as critical as the world's future energy systems? We examine solar PV
Energy transition
projections in 26 recent global energy scenarios, contrasting them with academic studies and other analyses, and
find that they all fail to account fully for technology developments and recognise plausible upper levels of solar
PV growth. Drilling deeper into the influential World Energy Outlook scenarios of the International Energy
Agency, which are amongst the more conservative of the 26 scenarios, we explore possible reasons for, and the
implications of, their projections for solar PV growth. We conclude that low scenario projections such as the
IEA's are likely to deter investments in innovation and development that would otherwise occur. If authoritative
scenarios fully acknowledge the possibility of continuing rapid growth of solar PV, investors, governments and
the energy sector will be encouraged to take a more optimistic view of the market potential, thus creating
positive feedback loops of belief, investment, and growth.

1. Introduction
If the expectations and ambitions of investors and policy makers for
the future of energy systems are shaped by scenarios that are un-
necessarily pessimistic about the potential rate of change, might this in
itself become a self-fulfilling prophesy, as suggested by the
International Renewable Energy Agency [1]? In this paper we explore
the proposition that estimates of the rate at which the global energy
system could plausibly become low-carbon have implications for in-
vestment decisions as well as for energy policy.
The predominant view in the academic literature is that energy
transitions are inevitably long protracted affairs [2]. On the other hand
Sovacool [3] suggests that the evidence does not consistently support
this position. A brisk debate [2,4–8] was stimulated when Sovacool [3]
posed the questions: “What does the mainstream academic literature
suggest about the time scale of energy transitions?” and “What does
some of the more recent empirical data related to transitions say, or
challenge, about the mainstream view?” (p202). In this paper we aim to
advance this debate by reviewing global energy scenarios and asso-
ciated literature on solar photovoltaic (solar PV) growth, and con-
sidering the implications for policy and investment decision-makers.
Solar PV is likely to be a key technology in a low-carbon energy system
because of its rapid historic growth and because solar energy has more
potential for electricity generation than other renewable resources [9].
Global solar PV capacity has increased by 34% per annum on average
over 57 years to 2017 (Fig. 1). We review studies which indicate that
continued growth at a similar rate until 2030 is a reasonable possibility.
We also explore the implications of scenarios that promulgate more
conservative projections.
Projections for the future growth of solar PV cover a wide range. At
one extreme Seba [10] argues that solar PV is an exponential tech-
nology for which historic growth rates will be maintained at least until
the 2030’s, as part of a wider technological transition in energy and
transport. If this eventuates, solar PV would provide more than half of
global electricity production by 2030. Although Seba is arguably the
most prominent advocate for a transition of this speed and scale, it is
also considered to be plausible by others in the research community
[11–16]. In this paper we will refer to an outcome in which solar PV
generates more than 50% of global electricity as a “Seba-like” transi-
tion.
Seba’s view contrasts with those of Kramer and Haigh [17], Höök
et al. [18], Bezdek and Wendling [19] and Smil [20]. Kramer and Haigh
[17] propose two “laws of energy technology deployment” based on the
concept of “materiality” – the time at which a new energy technology

⁎
Corresponding author.
E-mail addresses: gerald.carrington@otago.ac.nz (G. Carrington), janet.stephenson@otago.ac.nz (J. Stephenson).

https://doi.org/10.1016/j.erss.2018.07.011
Received 15 September 2017; Received in revised form 2 July 2018; Accepted 6 July 2018
Available online 17 July 2018
2214-6296/ © 2018 Elsevier Ltd. All rights reserved.
G. Carrington, J. Stephenson Energy Research & Social Science 46 (2018) 103–113

scenarios are not only shaped by their own world-views, but they also
shape the world-views of others.
In this paper we question the messages scenarios convey to others,
and examine “what they can reveal about current agendas, motivations,
and existing social orders” ([24]: 482). This is particularly significant
given that expectations “play a central role in mobilizing resources both
at the macro level, for example in national policy […] and at the meso
level of sectors and innovation networks” ([26]:286). Critically, they
also underpin the decisions of investors [28]. Consequently, scenarios
that promulgate a limited-scope vision may have the effect of con-
straining the “cognitive aspects” of investors ([29]:6).
To examine our title question, “Are International Energy Agency
and other conservative projections hampering the renewable energy
transition?” we consider three subsidiary questions: Is there evidence
that the International Energy Agency and other scenarios are con-
servative? What evidence is there that solar PV could continue to grow
at historic rates? From a public-good perspective, should global energy
scenarios explore rapid (Seba-like) solar PV uptake as a plausible fu-
Fig. 1. The open circles show the cumulative global solar PV peak generation
capacity (MWp) at the end of each year, 1960–2017. The dashed line is a fitted
ture?
exponential growth curve with a compound annual average growth rate of 34% The paper is structured as follows: in Section 2 we describe the
pa. methods used to analyse global energy scenarios and in Section 3 we
examine data on the deployment of solar PV in the past and consider
the implications of different growth rates in the future. We examine the
produces or consumes between 1% and 10% of world energy supply.
results of the scenario analysis in Section 4, compare them with other
Prior to materiality they claim that the rate of deployment of a new
evidence on the future growth of solar PV, and consider the implica-
energy technology may grow exponentially. But subsequently the rate
tions of conservative scenarios. In Section 5 we present our conclusions.
of growth reduces due to existing infrastructure lock-in, the time
needed to establish large-scale technology supply capacity, and the time
2. Methods
required to adapt to it. The position of Kramer and Haigh is broadly
consistent with what Sovacool [3] calls “the mainstream view that
We examined 26 well-known global energy scenarios, all but one
energy transitions all take time” (p205).
published between 2013 and 2017. The slightly earlier World Wide
To establish an understanding of expectations for solar PV growth
Fund for Nature scenario [30] was also included for contrast because it
we examine 26 global energy scenarios that are by well-recognised
projected a faster rate of reduction in global CO2 emissions than the
organisations, broadly read, accessible and widely referenced. We
others. Six of the scenarios are by fossil-fuel companies (Statoil 3; Shell
compare the scenario projections with analysis in the academic and
2, ExxonMobil 1). Nine are by the International Energy Agency, an
grey literatures, including reputable commercial reports. Throughout
intergovernmental agency established by the OECD, and five are by the
we accept the definition of the IPCC [21] that a scenario is “a coherent,
World Energy Council, a global network of energy organisations. Three
internally consistent and plausible description of a possible future state
are by Greenpeace. There is one scenario each by Carbon Tracker (a
of the world. It is not a forecast; rather, each scenario is one alternative
think tank based in the UK), the World Wide Fund for Nature and the
image of how the future can unfold.” Scenario development is under-
USA’s Energy Information Administration.
taken for many reasons including to systematically analyse possible
Table 1 lists the scenarios in order of increasing projected CO2
future pathways, to examine technology potentials, to elucidate stra-
emissions in 2040 (top to bottom), showing source details, the letter
tegic options and more generally “to capture the richness and range of
codes used to identify them in Fig. 4, and the projected global CO2
possibilities, stimulating decision makers to consider changes they
emissions due to energy use in 2040. We recognise that the scenarios
would otherwise ignore” ([22]:38). In this context we use the verb
may have been produced for a range of reasons and may use different
“project” and noun “projection” in the same way as the IEA [23]: a
methodologies, but their findings cover a similar range of topics and are
projection is an estimate of a possible future state under specific as-
thus empirically comparable. The Appendix A lists numerical data for
sumptions. It is not a prediction.
each scenario.
Energy scenarios produced by influential global organisations
As a basis for comparing the scenarios we used eight global in-
matter because they “affect what may actually happen. They are per-
dicators for 2040, the last year for many of the scenarios. These were
formative.” ([24]:465). They shape the beliefs and expectations of de-
selected as indicators of projected changes in the global energy system
cision-makers in their sector and beyond [25]. They have a pervasive
in response to global climate change: CO2 emissions due to fossil-fuel
influence on the unfolding energy transition because “[t]he language
combustion, total primary energy demand (TPED), natural gas use, oil
we use to describe transitions serves as more than a mere analytical
use, coal use, total electricity generation, solar PV generation and wind
tool—it can shape how energy system users, investors, operators,
generation. The projected quantity of CO2 emissions in 2040 was ex-
builders and financiers frame energy problems and also envision future
pressed in units of GT of CO2. TPED includes all energy sources before
pathways for change” ([8]:236). Scenarios establish expectations that
being converted into the form used by consumers. TPED and fuels are
can become “taken-for-granted assumptions of what is going to happen,
measured here in units of EJ (1018 J). Total electricity generation is the
thus falling into a deterministic lock-in, where future states become not
energy supplied as electricity which is measured in units of PWh (1015
just a promise but an inevitability” ([24]:469). Widely-shared ex-
Wh). Where the scenarios used other units, we applied the following
pectations and visions about the future can underpin unconscious
conversion factors, selected for consistency between the IEA [32] and
biases, or become hard-wired into analytical tools and models [26].
the WEC [35] scenarios: primary energy 0.0417 EJ/MTOE; natural gas
Organisations themselves may be constrained by their own beliefs and
0.0342 EJ/bcm; oil 1.95 EJ pa/mb per day; coal 0.0287 EJ/Mtce; heat
norms that support a particular energy culture [27] which makes it
1.055 EJ/quad; electricity 3.6 EJ/PWh.
difficult to examine a future that is beyond the scope of those beliefs.
Most of the scenarios provided the required data in the form of
For these reasons the projections that organisations declare in their
tables, but for the ExxonMobil [38] and WWF [30] scenarios some data

104
G. Carrington, J. Stephenson Energy Research & Social Science 46 (2018) 103–113

Table 1
The 26 scenarios, listed in order of projected annual CO2 emissions in 2040.
Scenario Short name Authorship End date ID code 2040 CO2 emissions (GT)

World Wide Fund for Nature & Ecofys The Energy Report 2011 [30] WWF 2011 International NGO 2050 A 7.0
Greenpeace Energy Outlook 2015 Advanced Energy [R]evolution [31] GrnPce’15 AdE[R] International NGO 2050 B 8.1
Greenpeace Energy Outlook 2015 Energy [R]evolution [31] GrnPce’15 E[R] International NGO 2050 C 12.0
IEA World Energy Outlook 2017 Sustainable Development Scenario [23] IEA WEO’17 SDS Global policy agency 2040 D 18.3
IEA World Energy Outlook 2016 450 [32] IEA WEO’16 450 Global policy agency 2040 E 18.4
IEA World Energy Outlook 2015 450 [33] IEA WEO’15 450 Global policy agency 2040 F 18.8
Statoil Energy Perspectives 2017 Renewal [34] Statoil’17 Renewl Fossil-fuel firm 2050 G 19.5
Carbon Tracker 2017 Strong_PV_EV_low [16] Carbon Trkr’17 Academic institute 2050 H 24.1
World Energy Council World Energy Scenarios 2016 Unfinished Symphony [35] WEC’16 Unf Smph Global policy network 2060 I 25.8
World Energy Council World Energy Scenarios 2013 Symphony [36] WEC’13 Smphny Global policy network 2050 J 26.0
Statoil Energy Perspectives 2017 Reform [34] Statoil’17 Reform Fossil-fuel firm 2050 K 32.1
World Energy Council World Energy Scenarios 2016 Modern Jazz [35] WEC’16 Mod Jazz Global policy network 2060 L 34.1
IEA World Energy Outlook 2017 New Policies [23] IEA WEO’17 NP Global policy agency 2040 M 35.7
IEA World Energy Outlook 2016 New Policies [32] IEA WEO’16 NP Global policy agency 2040 N 36.3
IEA World Energy Outlook 2015 New Policies [33] IEA WEO’15 NP Global policy agency 2040 O 36.7
Shell New Lens Scenarios Mountains 2013 [37] Shell’13 Mountns Fossil-fuel firm 2060 P 37.0
ExxonMobil The Outlook for Energy 2017 [38] ExxonMob’17 Fossil-fuel firm 2040 Q 37.0
World Energy Council World Energy Scenarios 2016 Hard Rock [35] WEC’16 Hard Rk Global policy network 2060 R 37.2
Statoil Energy Perspectives 2017 Rivalry [34] Statoil’17 Rivalry Fossil-fuel firm 2050 S 38.1
Shell New Lens Scenarios Oceans 2013 [37] Shell’13 Oceans Fossil-fuel firm 2060 T 41.0
IEA World Energy Outlook 2017 Current Policies [23] IEA WEO’17 CP Global policy agency 2040 U 42.7
EIA International Energy Outlook 2016 [39] EIA Outlook’16 National policy agency 2040 V 43.2
IEA World Energy Outlook 2016 Current Policies [32] IEA WEO’16 CP Global policy agency 2040 W 43.7
IEA World Energy Outlook 2015 Current Policies [33] IEA WEO’15 CP Global policy agency 2040 X 44.1
Greenpeace Energy Outlook 2015 Reference [31] GrnPce’15 Ref International NGO 2050 Y 44.3
World Energy Council World Energy Scenarios 2013 Jazz [36] WEC’13 Jazz Global policy network 2050 Z 45.3

were obtained from figures. The Statoil [34] scenario provided pro-
jected energy use and CO2 emissions for 2030 and 2050 and we used
linear interpolation to obtain estimates for 2040. For the Carbon
Tracker scenario [16] the total primary energy demand was estimated
from the fossil-fuels demand plus the contributions from nuclear en-
ergy, biomass and variable renewables used in electricity production.
Here the average thermal efficiency for power generation was taken to
be 33% for biomass and 25% for nuclear. The captured energy ap-
proach was used to determine the contribution to primary energy due
to electricity generation by variable renewables and geothermal sources
[40].
Two projections we cite for comparison [41,42] provided data only
on the global peak solar PV production in 2040. Here we estimated the
global annual electrical generation using an average solar PV capacity
factor of 18.5%, equivalent to 1.62 PWh-pa/TWp, the value for 2040
obtained from the IEA Sustainable Development Scenario [23].
Where we needed to estimate the average compound growth rate for
a quantity, Q, over y-years in % pa, the equation ((Q y / Q0 )1/ y−1) × 100
was used, where Q0 and Qy are the values for Q at the end of year-0 and
year-y respectively.
In order to establish whether or not the mean values of indicators
for two groups of scenarios were significantly different we used the
T.TEST function [43] to estimate the probability (p) that the two
samples come from the same population. In applying this function we
used the two-tailed distribution and two-sample unequal variance op-
tions.
represents a fitted exponential growth curve (R2 = 0.99) with a com-
pound annual average growth rate of 34% pa. If this were to continue
until 2030, solar PV would then produce 20 PWh of electricity, over
half of projected global electricity generation in 2030, 30 −35 PWh
[23].
The logarithmic scale used for the cumulative solar PV capacity in
Fig. 1 masks past volatility in the growth rate. This is illustrated in
Fig. 2 by the 3-year and 15-year average annual growth rates. The 3-
year average growth rate for 2015–2017, 31% pa, is less than the
previous high in 2010, but is nevertheless close to the long-term
average, 34% pa. The 15-year average growth rate shows less volatility
than the 3-year data and has not dropped below 20% pa.
To illustrate the effect of different growth rates on PV generation in
the future, Fig. 3 shows global annual solar PV electricity generation at
three future dates plotted against the compound average annual growth
rate, starting in 2014. The use of the compound growth rate here is not
intended to suggest that future growth will be exponential. The figure
shows that, at the higher levels of uptake, a relatively small increase in
average growth rate can significantly increase the quantity of electricity

3. Solar PV growth

The increasing importance of solar PV in the global electricity

system was highlighted in the Renewables 2017 Global Status Report
[44]: “Solar PV saw record additions and, for the first time, accounted
for more additional capacity, net of decommissioning, than did any
other power generating technology” (p20). This increase is unsurprising
in the context of historic growth, illustrated in Fig. 1, which shows the Fig. 2. The open circles show the 3-year average annual growth rate (% pa) for
global cumulative solar peak power (MWp) from 1960 to 2017 with global solar PV peak generation capacity. This represents the running average
data from the Performance Curve Data Base [45] (1960–2003), the IEA for the three years prior to the end of each year, using the data shown in Fig. 1.
[46] (2004–2016) and Renewables Now [47] for 2017. The dashed line The open triangles show the corresponding 15-year average annual growth rate.

105
G. Carrington, J. Stephenson
Fig. 3. Global annual solar PV electricity production in 2020, 2030, 2040 for a
range of compound annual average growth rates from 2014. The thick shaded
line shows the range of projections for 2040 by the 26 global energy scenarios
examined. The vertical dashed line shows the historical solar PV average
growth rate.
generated.
Among the 26 global energy scenarios examined in this paper the
highest projected growth rate for global annual solar PV generation
(2014–2040) is 17% pa [37] with a corresponding projected global
electricity production by solar PV of 11.4 PWh in 2040. The lowest
projected growth rate is 5% pa in the WEC [36] Jazz scenario, with
0.7 PWh generated in 2040. The average for all the scenarios is 3.8 PWh
in 2040. The thick shaded line in Fig. 3 shows the range of projected
growth rates for the 26 global energy scenarios, which highlights the
difference between the historic growth rate (vertical dashed line, 34%
pa) and the scenario projected rates. It also shows that a continued
average growth rate at the historic level would result in 20 PWh gen-
erated globally by PV by 2030.
4. Results and discussion
4.1. Group averages
Here we examine the average values of the eight selected scenario
indicators in 2040 in order to illustrate progress towards a low-carbon
transition projected by scenario groups and identify group differences.
Compared with 2014, 10 of the 26 scenarios (A– J) project reduced CO2
emissions in 2040, while the other 16 (KeZ) project the same or
Fig. 4. Projected CO2 emissions and solar PV generation in 2040 for 26 scenarios. The symbol Δ shows actual data for the calendar year 2014.
106
Energy Research & Social Science 46 (2018) 103–113
increased emissions. In the following we denote these as the “de-
creasing emissions” and “increasing emissions” groups respectively.
Table 2 lists average projected values for the 2040 indicators with
data for all 26 scenarios, the decreasing emissions group, the increasing
emissions group, the six fossil-fuel firms (G, K, P, Q, S, T) and the three
IEA low emissions scenarios (D–F). The standard deviations in the
group averages are shown in parentheses. As a point of reference, the
table also lists actual indicator values for 2014 [32]. The table shows
that the standard deviations for the indicators of the decreasing emis-
sions group are mostly larger than those of the increasing emissions
group. The differences between the average indicators of these two
groups are significant (p < 5%) for all indicators except total elec-
tricity generation (p = 17%).
In relation to the fossil-fuel firms, the average indicators for CO2
emissions, TPED, natural gas, oil and coal use are significantly larger
than those for the decreasing emissions group (p < 5%). On the other
hand, the fossil-fuel firms’ projections for electricity generation and
solar PV are consistent with those for the decreasing emissions group
(p > 55%). Interestingly, among the 26 scenarios the highest level of
solar PV generation in 2040 (11.4 PWh) is projected in the Shell Oceans
scenario [37], which is in the increasing emissions group. This indicates
that Shell is prepared to entertain a wider range of solar PV futures than
other scenario organisations.
We note that the average annual growth rate for solar PV
(2014–2040) is 14% for the decreasing emissions group, 13% for the
fossil-fuel firms, 12% for all 26 scenarios and for the IEA low emission
group and 11% for the increasing emissions group. These projected
growth rates for solar PV are all much less than the historical long-term
average growth rate (34% pa). In its three low emission scenarios (D–F)
the IEA solar PV projections are closer to the increasing than the de-
creasing emissions group.
Considering the IEA low emission scenarios (D–F) further, most
average indicators are not significantly different from those for the rest
of the decreasing emissions group (p > 13%). Solar PV, on the other
hand, is different with p = 5%. By contrast, most average indicators for
the IEA low emission group are significantly different from those of the
increasing emissions group, with p < 0.1% for all except wind
(p = 3%) and solar PV (p = 41%). Thus the solar PV projections of the
IEA low emissions scenarios for 2040 are different from the IEA group’s
other indicators, being more strongly aligned with the increasing
emissions group than the decreasing emissions group. We conclude that
the solar PV projections of the IEA low emissions group are indeed
conservative.
G. Carrington, J. Stephenson
Table 2
Scenario group average projected CO2 emissions and energy production in 2040. The standard deviation for the scenarios in each group is shown in brackets and the
actual data for the calendar year 2014 are also shown.
Actual 2014 Scenario group
All scenarios
A-Z
CO2 (GT) 32.2 31 (12)
TPED 571 692 (135)
Fossil-fuels (EJ) Natural Gas 121 164 (44)
Oil 179 181 (60)
Coal 163 141 (65)
Electricity (PWh) Total gen. 23.8 39 (4)
Solar PV 0.2 4 (3)
Wind 0.7 5 (3)
4.2. Solar PV trends
Fig. 4 shows the projected global CO2 emissions as a function of
solar PV generation in 2040 for the 26 scenarios. All anticipate in-
creasing solar PV capacity compared with 2014 (Δ symbol), but there is
wide variation. There are two clusters of scenarios with projected
emissions equal to, or higher than, 2014 (32 GT). One cluster (UeZ),
with emissions above 42 GT in 2040, comprises mainly reference sce-
narios, such as the IEA Current Policies scenarios (U, W, X). For this
cluster the projected generation by solar PV in 2040 is in the range
0.7–2.2 PWh. The second cluster (KeS) has emissions between 32 and
38 GT in 2040 and includes the IEA New Policies scenarios (M, N, O).
For this group the projected range of solar PV generation in 2040 is
larger than the first, 1.2–3.7 PWh. The scenarios with decreasing CO2
emissions (A–J) tend to project higher, but more variable, levels of solar
PV generation than those with increasing emissions. Yet even for these
scenarios the projected solar PV generation (2.2–10.7 PWh) is much
lower than the level (20 PWh) that would indicate a Seba-like transition
by 2040.
The figure shows that there is a broad correlation between CO2
emissions and solar PV generation, although the Shell Oceans (T) pro-
jection is an outlier. This indicates there is general agreement across the
scenarios that solar PV is likely to have a role in a decarbonised world,
but there is no agreement about its scale in 2040 or about its influence
on CO2 emissions.
Fig. 4 shows that successive IEA low emissions scenarios – 2015 (F),
2016 (E) and 2017 (D) – project increasing levels of solar PV generation
in 2040. We recognise that the policy assumptions used in previous IEA
450 scenarios are not all the same as those in the IEA SDS (Sustainable
Development Scenario) [23], but many are in common [48]. In 2016
and 2017 the increases in the solar PV projections for the IEA low
emissions scenarios were 52% and 64% (respectively) of the previous
year’s projection. These increases are much higher than the current
short-term growth rate for solar PV generation (31% pa) and they may
well be a response to that growth, but we do not have independent
evidence to support this suggestion.
Fig. 4 highlights our point (Section 4.1) that the IEA’s low emission
projections (D–F) for solar PV generation in 2040 are low compared
with the other decreasing emissions scenarios. In addition the IEA [23]
SDS 2040 projection (5.3 PWh) is lower than the IRENA Remap pro-
jection [41], which is for 7.5 PWh of global solar PV generation in
2040, based on a global average capacity factor of 18.5% (Section 2).
Bloomberg New Energy Finance forecast for 2040 [42] indicates a si-
milar level of global solar PV generation (7.9 PWh), again based on a
global average capacity factor of 18.5%. These data support our con-
clusion that the three IEA low emission scenarios (D–F) provide con-
servative projections for solar PV generation in 2040.
Energy Research & Social Science 46 (2018) 103–113
Decreasing emissions Increasing emissions Fossil-fuel firms IEA 450 & SDS
A-J K-Z G,K,P,Q,S,T D,E,F
18 (7) 39 (4) 34 (8) 19 (1)
561 (120) 773 (59) 729 (104) 614 (24)
124 (44) 188 (21) 171 (37) 140 (4)
122 (56) 218 (19) 205 (31) 143 (1)
69 (27) 187 (32) 157 (46) 86 (15)
37 (5) 40 (3) 38 (4) 35 (1)
6 (3) 3 (3) 5 (4) 4 (2)
7 (4) 3 (1) 4 (2) 6 (1)
4.3. Is continuing rapid solar PV growth plausible?
It is increasingly accepted that solar PV is likely to be the least cost
source of electricity, and the dominant global generation technology, in
the long-term [9,11,12]. But how quickly might this occur? Some short-
term projections [49,50] foresee solar PV growth of 20% pa until at
least 2020. Seba [10], on the other hand, maintains that rapid solar PV
growth is likely to continue at historic rates: “Can solar keep growing at
this exponential rate for another ten or twenty years? The answer is that
the solar growth rate could actually accelerate. As a rule, when a
technology product achieves critical mass (the point of no return), its
market growth actually accelerates” (p38). Motyka and Sanborn [51]
concur: “When costs fall below a certain threshold and adoption in-
creases, reducing costs and boosting demand further in a virtuous cycle;
a “tipping point” may be reached, beyond which adoption will accel-
erate” (p31).
A number of factors have enabled solar PV generation costs to fall in
the past: improved efficiency through technological innovation [52,53],
knowledge spill-overs from other technologies [54], and reduced cost of
raw materials and increased investment in the industry [53]. Wood-
house et al. [55] examine the potential for combinations of efficiency
gains, degradation rate, reliability, lifetime, and other factors to influ-
ence the levelised cost of solar PV generation and identify a pathway for
reducing this by 70% compared with the 2015 benchmark for utility
scale systems. Such reductions would open new markets which enable
scaled-up fabrication and further cost reductions as a function of vo-
lume manufactured. In the opinion of Nemet [54] “This sequence from
R&D to market formation to local learning-by-doing for the historically
dominant crystallized silicon PV may now be repeated for emerging
alternative solar technologies including thin film and organic PV” (p2).
To gain an understanding of the dynamics of solar PV growth,
Needleman et al. [15] used bottom-up cost modelling to determine
trajectories for the global capacity of solar PV in terms of technological,
economic and market parameters, up to 2030. Where cost reduction is
achieved mainly by increased efficiency, they conclude that this “re-
sults in faster growth and a cumulative installed capacity in 2030 of
11.2 TW” (p2126). At an average capacity factor of 18.5% (Section 2),
the annual global electricity generation by solar PV would then be
approximately 18 PWh, corresponding to an average growth rate
(2014–2030) of 33% pa. They consider that there are “clear and re-
dundant pathways” to achieve these improvements.
This assessment is endorsed by Haegel et al. [13] who used data
from Jones-Albertus et al. [56] to make a bottom-up techno-economic
analysis of opportunities for solar PV cost reduction. They conclude
“The PV industry is on a trajectory to reach at least three TW of cu-
mulative PV installations by 2030” (p143). These authors also conclude
that it would be feasible for the installed global capacity to reach
10 TW. Annual solar PV generation would then be approximately
16 PWh and the corresponding average growth rate for solar PV
(2014–2030) would be 32% pa.
107
G. Carrington, J. Stephenson
As a reference point for these projections for 2030, total global
electricity generation for 2030 projected by the IEA SDS scenario [23]
is 30.5 PWh. On this basis, if solar PV generation were to reach
16−18 PWh in 2030, as suggested by the authors above, this would
provide 52–59% of projected global electricity production. The energy
produced would represent 10–11% of the corresponding projected
TPED for 2030, 577 EJ [23].
In a separate study, Mayer et al. [14] reported three scenarios for
the growth of global solar PV, using a bottom-up approach to estimate
market development, and a top-down cross-check with projected future
electricity demand. Following expert workshops, a solar PV “break-
through” scenario was also developed, to assess the impact of an “ex-
treme” market. They note that “[t]his scenario was considered as most
realistic by most attending industry experts” (p21). Under the break-
through scenario, global solar PV generation would be 23 PWh pa in
2040, 59% of the average projected global electricity production in
2040 (Table 2) and 12% of average projected TPED (Table 2). The
corresponding average growth rate for solar PV generation
(2014–2040) would be 20% pa.
The three studies [13–15] indicate that Seba-like rates of solar PV
growth are plausible in the medium term. The projected global solar PV
capacity in these three studies would generate 10–12% of anticipated
global primary energy demand, suggesting that growth may be influ-
enced by the materiality condition of Kramer & Haigh [17]. On the
other hand, as Motyka and Sanborn [51] note: “solar adoption may no
longer be subject to the previous ‘laws’ of energy technology adoption.”
(p23). In particular, it is likely that solar PV growth will be supported
by technology convergence in which “the disruptors (solar, electric
vehicles, and autonomous cars) complement and accelerate one an-
other’s adoption.” ([10]:8).
Fig. 5, a modified form of Fig. 3, highlights the considerable dif-
ference between the solar PV growth rates found to be plausible in these
three studies (symbols ◻,○, Δ) and the range of projections of the 26
scenarios (thick shaded line). The diamond symbols (◇) show the 2040
solar PV projections of the IEA 450 and SDS scenarios [23,32,33]. These
projections are relatively low compared with the other scenarios, as
noted elsewhere.
Overall there is evidence that, under favourable conditions, global
solar PV capacity could grow at close to the historic rate (34% pa) until
2030 or at an average rate of 20% pa until 2040. If this were to happen
solar PV would have a major role in global electricity production in the
decade 2030–2040. These growth rates are much larger than those
projected by the 26 scenarios we examined, especially so for the IEA
projections. We conclude that a Seba-like level of solar PV generation
by 2030 or 2040 cannot be rejected on the basis of technical, economic
or market feasibility. None of the 26 scenarios contain projections ap-
proaching this scale of growth.
108
Energy Research & Social Science 46 (2018) 103–113
4.4. Finance and risks
To illustrate the financial implications of a low-carbon transition,
we cite scenario estimates by Channell et al. [57]. They find that global
expenditure on energy (capital and fuel) in the 25 years up to 2040 will
be in the order of US$190 trillion both for a low-carbon “Action” sce-
nario and for business-as-usual, on a non-discounted basis. Under the
Action scenario some of the US$6 – 9 trillion, spent each year on en-
ergy, shifts from fossil-fuels to other sectors. The low-carbon option is
projected to cost a total of US$0.8 trillion more up to 2025 but become
cheaper after 2035, with a net saving of $1.8 trillion by 2040. These
sums do not take business-as-usual climate change liabilities into ac-
count, estimated by Channell et al. [57] to be US$20 – 72 trillion by
2060, undiscounted.
The Action scenario of Channell et al. [57] does not consider rapid
solar PV growth, but it illustrates broadly the scale of investment
needed to maintain Seba-like growth rates, both in solar PV technology
and in the integration of solar PV with the existing electricity infra-
structure. Here we consider if investment on this scale is plausible,
bearing in mind that investment decisions are shaped as much by
subjective matters as by the objective qualities of a product or service:
this includes behavioural biases such as familiarity with the product, a
belief in past performance of investments, and anchoring in a particular
piece of influential information [28]. Such psychological influences
mean that investors tend to default to the status quo unless they are
strongly motivated to change.
Nevertheless, there is evidence of growing interest by businesses
and investors in low-carbon investments, arising from a dual motivation
of perceived risks and opportunities. Perceived risks include the threat
to businesses of climate change itself, the possibility of businesses
failing to reduce their own emissions, and the risk of devalued invest-
ments in fossil-fuel related portfolios [58,59]. There are also risks in
how a Seba-like transition could bring about disruptive change to the
energy industry, whereby long-term sector players may face competi-
tion from new business models, resulting in stranded assets and reduced
profit margins [12,16,60].
The views of long-term institutional investors are of interest because
of their particular concerns and financial resources. Members of the
Institutional Investors Group on Climate Change and the Principles for
Responsible Investment manage some €13 trillion and US$60 trillion
respectively, the assets of long-term investors such as pension funds,
trusts, charitable foundations, family wealth offices, insurance capital
and sovereign wealth funds [61]. Such investors are especially con-
cerned about the long-term risks that climate change poses to invest-
ment returns [57,61]. In this context it is financially rational to invest in
assets that support a rapid low-carbon transition when the long-term
implications are taken into account [62], for a rapid energy transition
“is the only way of avoiding the potentially large risk to portfolio value
Fig. 5. The symbols (◻,○, Δ) show plausible rates for solar PV
generation in 2030 and 2040, and corresponding growth rates
from 2014, determined by three studies, compared with the 2040
projections by 26 global scenarios (thick shaded line). The dia-
mond symbols (◇) indicate the projected solar PV generation in
2040, and growth rates from 2014, for the IEA 450 and SDS sce-
narios (2015–2017). The vertical dashed line represents the his-
torical solar PV average growth rate.
G. Carrington, J. Stephenson
if the economy is vulnerable to climate change” ([63]:18). Mercer [64]
describes the emergence of “future makers”, investors that seek “to
influence which scenario comes to pass.” (p78) by “doing whatever
they reasonably can to bring a rapid energy transition about” ([65]:1).
Shah and Booream-Phelps [66] see solar “becoming increasingly
mainstream as it passes cost competitiveness with traditional forms of
generation” (p80). Stephen Byrd claims that “Numerous key markets
have reached an inflection point where renewables will have become
the cheapest form of new power generation by 2020, a dynamic we see
spreading to nearly every country we cover” [67]. These cost factors
“could continue to shift capital allocation away from fossil fuels”
([68]:2). The rapid developments in these and other smart technologies
and systems are also in themselves attracting new businesses and
business models to cater for emerging markets for technologies and
services.
This combination of factors is attracting new investors to renewable
energy [44,69], but the much higher level of investment required to
achieve a low-carbon transition will be challenging to achieve. The
dynamics of investment is driven by feedback loops that are strongly
founded in the beliefs of investors and fund managers, many of whom
do not give sufficient weight to long-term investment value drivers
[65,70]. If a technology is believed to have technical and market po-
tential, and if the investment vehicles are of sufficient quality, invest-
ment is likely to be forthcoming. At scale this would enable technical
advancement, volume growth, lower costs and thus increased demand
for solar PV, a positive feedback loop that could support continuing
Seba-like growth. On the other hand, if investors and fund-managers
have little confidence in the technical potential or the size of the market
there are likely to be lower levels of interest in developing the product
and increasing production, resulting in lower levels of uptake.
In this situation global energy forecasts and scenarios are likely to
be influential [24]. They play an important role as ‘anchoring’ docu-
ments that influence investors’ beliefs in the potential of different en-
ergy technologies, and the risks and opportunities they or their sectors
face from different energy pathways [71]. The International Energy
Agency’s annual World Energy Outlook, “widely regarded as the gold
standard of energy analysis” [72], is generally accepted as an author-
itative analysis of global energy trends, so is worth looking at in some
detail.
4.5. Scenarios and expectations
In Sections 4.1 and 4.2 we showed that the projected rates of solar
PV generation for 2040 in the IEA’s low emissions scenarios are con-
servative relative to the other decreasing emissions scenarios, none of
which project a Seba-like rate of uptake. Its most optimistic projection
of 5.3 PWh of solar PV by 2040 [23] is low compared with the
16–23 PWh of solar generation by 2030–2040 considered feasible by
the studies discussed in Section 4.3. We note that none of the IEA re-
ports [23,32,33,46,49,73,74] refer to solar PV as an “exponential
technology”, a central theme of Seba [10], although the IEA solar
technology roadmap acknowledges exponential growth drivers: “Rapid
deployment and falling costs have each been driving the other”
([73]:1).
Not surprisingly the solar PV projections in the IEA World Energy
Outlook have attracted critical comment [11,25,75–78]. The Executive
Director of the IEA has responded: “Some colleagues and friends in the
renewables industry have at times criticised the projections of future
renewables energy supply in our main scenario as too conservative.
They may indeed turn out to be too conservative; I sincerely hope that
they do. But they rest squarely on the foundation of officially declared
policy intentions. More can and should be done, as we demonstrate
clearly in our other scenarios that require a more rapid pace of dec-
arbonisation” ([32]:3).
But Sussams et al. [25] argue that there are other issues involved.
They suggest that the IEA’s low solar PV projections arise from the
109
Energy Research & Social Science 46 (2018) 103–113
difficulty of including technology advances in the IEA’s World Energy
Model: “policy developments, such as the INDCs, are easier than tech-
nological developments for energy modellers to translate into changes
in the energy mix […] The advancement of new technologies tends to
be less predictable and more tumultuous and is therefore often under-
weighted in forecasts. […] So far, these factors have tended to mean
that the rate of growth can outperform predictions within a few years”
(p49). This argument is supported by Creutzig et al. [9] who show that
many conservative solar PV projections are affected by underlying bias
in their energy models: “support policies, public support, rapid tech-
nological learning, and underperforming technological competitors
explain the more rapid development of PV compared to model projec-
tions in the past” (p4).
Thus a key weakness of the World Energy Outlook is that its models
for technological advancement are limited. It has no analysis of tech-
nology convergence in which complementary technologies (e.g. solar
PV, electric and autonomous vehicles) combine to accelerate the tran-
sition to renewables. There is no consideration of emerging technolo-
gies, such as perovskite solar cells [79,80] or solid state batteries [81],
which have the potential to exert a critical influence on the cost of solar
PV collectors, the balance of system costs and grid integration costs. In
this context the IEA notes, without elaboration, that technology ad-
vances have “the potential to alter fundamentally the future orientation
of the energy system” ([23]:57).
We suggest that the IEA may in time respond purposefully to these
difficulties as the energy transition progresses and the focus of the
energy sector moves from fossil-fuels, on which the IEA has built its
reputation, to renewable technologies. The IEA will be motivated to
strengthen its analysis of technologies and learning rates in order to
retain its authoritative reputation. If this occurs the IEA’s solar PV
projections would be likely to exhibit a much wider range than hitherto
and provide more detailed analysis of the uncertainties that surround
them. In the short term, however, the IEA’s solar PV projections are
likely to be moderated by its current mandate: “to promote energy se-
curity amongst its member countries through collective response to
physical disruptions in oil supply, and provide authoritative research
and analysis on ways to ensure reliable, affordable and clean energy for
its 29 member countries and beyond” ([32]:3). In particular, to avoid
disruption in energy markets the IEA is concerned about the investment
needed to maintain the supply of fossil-fuels during the energy transi-
tion [32]. Thus it presently has incentives to issue conservative sce-
narios in order to sustain confidence in fossil-fuels and contain investor
expectations for variable renewables.
Another factor that inhibits the IEA’s capacity to project change in
the global energy system is the political character of the World Energy
Outlook. As noted by Gaede and Meadowcroft [82]: “… the Outlook is
inherently less objective in that it presents a ‘fictional’ (scenario-based)
perspective on the future with the aim of inducing policymakers to
action. The basic structure and intent of these scenarios are strongly
influenced by what the IEA member countries want out of an autono-
mous, global perspective on the future for energy and what they will
countenance” (p619). Member countries are all members of the OECD
and largely consist of more developed nations, for many of which a
significant proportion of GDP is still derived from fossil fuels (e.g.
Australia, USA, Canada).
A further factor is that, until recently, markets for solar PV have
largely depended on a supportive policy environment, with various
forms of subsidy required to encourage uptake in many nations.
Scenarios models developed to be attuned to policy settings, such as the
IEA’s, have been unprepared for emerging non-policy drivers of PV
growth. The most important of these drivers is arguably the ongoing
reductions in the costs of production to a point where retail grid parity
has already been reached in a number of countries, so that consumer-
side subsidies will increasingly be unnecessary to make it financially
attractive to invest in solar PV [83]. Less visible forms of support have
also stimulated growth in manufacturing, such as China’s facilitation of
G. Carrington, J. Stephenson
solar PV entrepreneurs with cheap capital, land, and “institutional ad-
vantages” ([84]:787).
The IEA’s current technology conservatism has consequences. One is
its influence on other models and modellers. The World Energy Outlook
has a strong discursive power amongst energy agencies, and as noted
earlier is self-identified “as the gold standard for energy analysis”. A
result may be that energy modellers are unwilling to be out of step with
the consensus narrative, and thus do not stray far from the dominating
influence of IEA modelling assumptions. Another consequence is that it
is “likely to have fed through to energy industry thinking more broadly”
([25]:6). Thus “low scenarios might become self-fulfilling prophecies”
([1]:71), as anchoring documents that may constrain investor con-
fidence in renewables while maintaining confidence in fossil-fuels. A
scenario-constructed belief that any transition to renewables will be
slow is also likely to delay planning for the possibility of a rapid Seba-
like transformation, as noted by Creutzig et al. [9]: “… the consistently
underestimated potential of solar energy — if continued — has im-
plications for the future as decision-makers might treat PV too re-
luctantly. Specifically, policymakers might fail to address the integra-
tion challenge and insufficiently plan for adequate grid and storage
infrastructure” (p6).
Thus conservative scenarios for variable renewables, such as solar
PV, increase the risk of a disruptive transition. But variable renewables
at scale need not be disruptive if planned for ahead of time and man-
aged well [51,85]. The importance of contingency planning for a rapid
transition is noted by Sussams and Leaton [16] who observe that “[i]n
such a scenario of significant change, the mass stranding of downstream
fossil fuel assets is highly likely” (p17). To mitigate this risk Sussams
et al. [25] urge energy modellers to “entertain a range of options that
deviate from their current view of the world” (p7). Similarly Metayer
et al. [76] advise policymakers “to consider the expansion of renew-
ables well beyond the WEO projections in their energy policies in order
to avoid stranded investments in future” (p3220).
More generally, conservative scenarios risk limiting the scope of
change that can be envisaged by sector decision-makers and weaken the
ambition of global climate change mitigation efforts [86]. The sub-
jective barriers that Baker and Ricciardi [28] describe as constraining
the ability to investors to adjust to a changing future may be equally
applicable to those working in the policy world. In addition, the pro-
mulgation of a conservative energy culture [27] risks reversing the
investment feedback loop by depressing investors’ confidence in the
future growth potential of solar PV, making large-scale investment in
this and other variable renewables less likely, and thus slow the rate of
transition.
In Section 1 we noted that energy scenarios produced by influential
global organisations matter because they are shaped by, and convey,
beliefs and expectations about the present and future state of the world.
Is it evidence of ‘group think’ that all of the 26 scenarios examined here
cluster around the same ‘possibility space’ for solar PV uptake occupied
by the IEA, and that none stretch out into the higher uptake possibilities
discussed in Section 4.3 and illustrated in Fig. 5? Further research is
needed to explore this question, and to examine routes through which
institutional norms, biases or other subjectivities might be built into
models like the IEA’s World Energy Model.
To summarise, we have found that the IEA is not considering Seba-
like scenarios for solar PV and related technologies, despite evidence
that they are plausible. One reason is that the IEA makes insufficient
allowance for ongoing technology advancement, the emergence of new
technologies and technology convergence. While this could be an out-
come of the modelling parameters, institutional decisions are also in-
volved. We suggest the IEA may be influenced in its scenario modelling
by its commitment to avoiding disruption to the status quo, as well as
by its culture, its mandate and possibly by its member countries. The
IEA, whether consciously or not, is creating market signals about slow
uptake of solar PV that will tend to reduce investment in variable re-
newables and delay the renewable energy transition. This also signals to
110
Energy Research & Social Science 46 (2018) 103–113
governments and utilities that they do not yet need to prepare for high
levels of renewable penetration in the electricity grid. Contrary to its
mandate, this strongly suggests that the IEA’s conservative solar PV
projections are lowering the ambition for climate change mitigation
and raising the risk of a disruptive unplanned transition. We suggest
that the IEA should take responsibility for the discursive power of the
scenarios that it produces, and explore a wider range of plausible fu-
tures.
5. Conclusions
In considering the title question “Are International Energy Agency
and other conservative projections hampering the renewable energy
transition?” we posed three subsidiary questions: “Is there evidence
that the International Energy Agency scenarios are conservative? What
evidence is there that solar PV could continue to grow at historic rates?
From a public-good perspective, should global energy scenarios explore
rapid (Seba-like) solar PV uptake as a plausible future?”
Regarding the first question, we found that the solar PV projections
for 2040 in the IEA 450 and SDS energy scenarios are more aligned with
scenarios that project increasing rather than decreasing CO2 emissions.
Yet by comparison, the other energy indicators for the IEA 450 and SDS
scenarios are consistent with those of the decreasing emissions group
but not the increasing emissions group. This and other evidence leads
us to conclude that the solar PV projections for 2040 in the IEA 450 and
SDS energy scenarios are indeed conservative.
In relation to the second question, none of the 26 global energy
scenarios incorporated continued high growth rates for solar PV despite
plausible evidence that solar PV could generate over 50% of projected
global electricity generation by 2030 or 2040, if supported by appro-
priate policies and investments. Foreseeable technology advances,
technological learning and technology convergence increase the like-
lihood of such growth. In addition, long-term investors are motivated to
commit their large financial resources to promote such a rapid transi-
tion, in order to reduce their exposure to climate change risks. There is
therefore credible evidence that global solar PV could continue to grow
at close to historic rates until 2030 at least.
Regarding the third question, we conclude that unless authoritative
global energy scenarios explore the rapid deployment of solar PV as a
plausible future the expectations of policy makers and investors of what
is feasible will be constrained. It is likely this will depress the level of
investment in innovation, development and deployment of solar PV. It
may also lead to a level of unpreparedness amongst policy and industry
communities for the scale and scope of change, should solar PV growth
continue at current rates. Thus, from a public good perspective, au-
thoritative global energy scenarios should explore the possibility of
rapid (Seba-like) deployment of solar PV.
Of special interest are the influential IEA scenarios which have
conservative projections for solar PV growth and offer little analysis of
emerging energy technologies. Despite some hiking in the past few
editions of the World Energy Outlook, the IEA scenarios still fail to
consider a future in which the growth of solar PV (and other variable
renewables) could create a rapid transition supported by technological
advancement. This signals to governments and utilities that they do not
yet need to prepare for high levels of variable renewables, and to in-
vestors that growth possibilities are limited. Thus, in relation to our title
question, we conclude that the current conservative solar PV projec-
tions by the International Energy Agency and others are indeed likely to
hamper the renewable energy transition.
We have examined the implications of the IEA’s scenarios in some
detail because of their influential role in the energy sector, but our
conclusions apply equally to other conservative scenarios. We were
particularly interested to find that fossil-fuel firms, which could be
assumed to have reasons to project low renewables growth rates, were
among the scenario organisations that have explored the highest levels
of solar PV generation. This indicates that these firms are well aware of
G. Carrington, J. Stephenson Energy Research & Social Science 46 (2018) 103–113

the potential for continuing rapid solar PV growth.
Variable renewables need not be disruptive provided the required
adaptations are anticipated appropriately. But contingency planning for
high levels of renewables will be hampered unless influential global
energy scenarios explore the possibility of higher, but feasible, growth
rates. Indicating the possibility of rapid uptake will encourage investors
and governments to lift their ambitions, promote investment and pre-
pare for an early transition. We suggest that the following
recommendations, if adopted, would help to ensure a smooth transition:
(1) Organisations with energy interests should prepare for the possibi-
lity of an early transition to large scale variable renewables by ex-
ploring a wider range of scenarios related to solar PV growth and take
account of foreseeable advancements in solar PV and related technol-
ogies; (2) The primary mandates of the IEA should be reviewed to help
it respond more effectively to the challenges of the approaching energy
transition.

Appendix A

Scenario Label CO2 (GT) Fossil-fuels (EJ) Electricity (PWh)

TPED Nat Gas Oil Coal Total gen. Solar PV Wind

2014 actual Δ 32.2 571 121 179 163 23.8 0.19 1.72
WWF 2011 A 7.0 276 32 20 24 28.8 4.7 7.1
G-peace 2015 Adv E [R]ev B 8.1 482 82 46 33 46.1 9.4 15.5
G-peace 2015 E [R]ev C 12.0 485 98 67 50 42.8 7.1 11.3
IEA-WEO 2017 SDS D 18.3 587 144 142 73 35.9 5.3 7.0
IEA-WEO 2016 450 E 18.4 620 137 143 82 34.1 3.2 6.1
IEA-WEO 2015 450 F 18.8 634 139 144 102 33.9 2.2 5.1
Statoil 2017 Renewal G 19.5 565 117 152 77 36.6 5.4 7.1
Carbon Tracker 2017 H 24.1 675 173 150 63 38.1 10.7 2.1
WEC 2016 Unfin Symph I 25.8 633 152 183 86 35.5 3.8 4.9
WEC 2013 Symphony J 26.0 654 170 170 101 39.6 5.8 3.0
Statoil 2017 Reform K 32.1 680 154 202 142 39.7 3.6 5.5
WEC 2016 Mod Jazz L 34.1 694 188 206 129 37.7 2.7 4.3
IEA-WEO 2017 NP M 35.7 733 181 205 161 39.3 3.2 4.3
IEA-WEO 2016 NP N 36.3 745 178 202 170 39.0 2.1 3.9
IEA-WEO 2015 NP O 36.7 748 176 202 181 39.4 1.5 3.6
Shell Mountains 2013 P 37.0 822 226 189 191 36.4 3.7 2.3
Exxon-Mob 2017 Q 37.0 739 188 236 151 33.6 1.7 3.3
WEC 2016 Hard Rock R 37.2 709 159 216 169 35.6 1.3 2.9
Statoil 2017 Rivalry S 38.1 713 156 229 181 37.6 2.6 3.3
Shell Oceans 2013 T 41.0 856 187 222 202 44.2 11.4 4.0
IEA-WEO 2017 CP U 42.7 805 195 238 207 42.3 2.2 3.4
EIA Outlook 2016 V 43.2 860 223 260 190 36.5 1.0 2.5
IEA-WEO 2016 CP W 43.7 819 195 228 218 42.5 1.5 3.1
IEA-WEO 2015 CP X 44.1 819 192 228 230 43.1 1.1 2.8
G-peace 2015 Reference Y 44.3 803 193 202 240 44.0 0.8 2.6
WEC 2013 Jazz Z 45.3 825 216 231 224 44.5 0.7 3.1

Scenario global projected energy indicators for 2040. TPED is total primary energy demand. The data for the calendar year 2014 is from the IEA
World Energy Outlook 2016 [32].

References
[1] IRENA, REmap: Roadmap for a Renewable Energy Future, 2016 edition,
International Renewable Energy Agency (IRENA), Abu Dhabi, 2016 (Accessed 30
June 2017), http://www.irena.org/DocumentDownloads/Publications/IRENA_
REmap_2016_edition_report.pdf.
[2] V. Smil, Examining energy transitions: a dozen insights based on performance,
Energy Res. Soc. Sci. 22 (2016) 194–197.
[3] B.K. Sovacool, How long will it take? Conceptualizing the temporal dynamics of
energy transitions, Energy Res. Soc. Sci. 13 (2016) 202–215.
[4] P.S. Bromley, Extraordinary interventions: toward a framework for rapid transition
and deep emission reductions in the energy space, Energy Res. Soc. Sci. 22 (2016)
165–171.
[5] R. Fouquet, Historical energy transitions: speed, prices and system transformation,
Energy Res. Soc. Sci. 22 (2016) 7–12.
[6] A. Grubler, C. Wilson, G. Nemet, Apples, oranges, and consistent comparisons of the
temporal dynamics of energy transitions, Energy Res. Soc. Sci. 22 (2016) 18–25.
[7] F. Kern, K. Rogge, The pace of governed energy transitions: agency, international
dynamics and the global Paris agreement accelerating decarbonisation processes?
Energy Res. Soc. Sci. 22 (2016) 13–17.
[8] B.K. Sovacool, F.W. Geels, Further reflections on the temporality of energy transi-
tions: a response to critics, Energy Res. Soc. Sci. 22 (2016) 232–237.
[9] F. Creutzig, P. Agoston, J.C. Goldschmidt, G. Luderer, G. Nemet, R.C. Pietzcker, The
underestimated potential of solar energy to mitigate climate change, Nat. Energy 2
(2017) 17140.
[10] T. Seba, Clean Disruption of Energy and Transportation. How Silicon Valley Will
Make Oil, Nuclear, Natural Gas, Coal, Electric Utilities and Conventional Cars
Obsolete by 2030, Clean Planet Ventures, Silicon Valley, California, Kindle Edition,
(2014).
[11] C. Breyer, D. Bogdanov, A. Gulagi, A. Aghahosseini, L.S.N.S. Barbosa, O. Koskinen,
M. Barasa, U. Caldera, S. Afanasyeva, M. Child, J. Farfan, P. Vainikka, On the role of
solar photovoltaics in global energy transition scenarios, Prog. Photovolt. Res. Appl.
25 (2017) 727–745.
[12] J. Green, P. Newman, Disruptive innovation, stranded assets and forecasting: the
rise and rise of renewable energy, J. Sustain. Financ. Invest. 7 (2017) 169–187.
[13] N.M. Haegel, R. Margolis, T. Buonassisi, D. Feldman, A. Froitzheim, R. Garabedian,
M. Green, S. Glunz, H.-M. Henning, B. Holder, I. Kaizuka, B. Kroposki,
K. Matsubara, S. Niki, K. Sakurai, R.A. Schindler, W. Tumas, E.R. Weber, G. Wilson,
M. Woodhouse, S. Kurtz, Terawatt-scale photovoltaics: trajectories and challenges,
Science 356 (6334) (2017) 141–143.
[14] J.N. Mayer, S. Philipps, N.S. Hussein, T. Schlegl, C. Senkpiel, Mara Marthe Kleiner
(Ed.), Current and Future Cost of Photovoltaics. Long-Term Scenarios for Market
Development, System Prices and LCOE of Utility-Scale PV Systems, Fraunhofer-
Institute for Solar Energy Systems, Freiburg, 2015.
[15] D.B. Needleman, J.R. Poindexter, R.C. Kurchin, I.M. Peters, G. Wilson,

111
G. Carrington, J. Stephenson
T. Buonassisi, Economically sustainable scaling of photovoltaics to meet climate
targets, Energy Environ. Sci. 9 (2016) 2122–2129.
[16] L. Sussams, J. Leaton, Expect the Unexpected. The Disruptive Power of Low-Carbon
Technology, Carbon Tracker Initiative, Grantham Institute, Imperial College,
London, 2017 (Accessed 6 June 2017), http://www.carbontracker.org/wp-content/
uploads/2017/02/Expect-the-Unexpected_CTI_Imperial.pdf.
[17] G.K. Kramer, M. Haigh, No quick switch to low-carbon energy, Nature 462 (2009)
568–569.
[18] M. Höök, J. Li, K. Johansson, S. Snowden, Growth rates of global energy systems
and future outlooks, Nat. Resour. Res. 21 (1) (2012) 23–41.
[19] R.H. Bezdek, R.M. Wendling, A half century of long-range energy forecasts: errors
made, lessons learned, and implications for forecasting, J. Fusion Energy 21 (3/4)
(2002) 155–172.
[20] V. Smil, Perils of long-range energy forecasting: reflections on looking far ahead,
Technol. Forecast. Soc. Change 65 (2000) 251–264.
[21] IPCC, Definition of Terms, (2008) http://www.ipcc-data.org/guidelines/pages/
definitions.html.
[22] P.J.H. Schoemaker, Scenario planning: a tool for strategic thinking, Sloan Manage.
Rev. 36 (1995) 25–40.
[23] IEA, World Energy Outlook 2017, International Energy Agency, Paris, 2017.
[24] K. Konrad, H. van Lente, C. Groves, C. Selin, Performing and governing the future in
science and technology, in: U. Felt, R. Fouché, C.A. Miller, L. Smith-Doerr (Eds.), Ch
16 in The Handbook of Science and Technology Studies, 4th edn., MIT Press,
Cambridge, MA, 2017, pp. 465–493.
[25] L. Sussams, J. Leaton, T. Drew, Lost in Transition: How the Energy Sector is Missing
Potential Demand Destruction, Carbon Tracker Initiative, Imperial College, London,
2015 (Accessed 6 June 2017), http://www.carbontracker.org/wp-content/
uploads/2015/10/Lost-in-transition_Clean_Draft.pdf.
[26] M. Borup, N. Brown, K. Konrad, H. Van Lente, The sociology of expectations in
science and technology, Technol. Anal. Strateg. Manag. 18 (3/4) (2006) 285–298.
[27] J. Stephenson, Sustainability cultures and energy research: an actor-centred inter-
pretation of cultural theory, Energy Res. Soc. Sci. 44 (2018) 242–249.
[28] H.K. Baker, V. Ricciardi (Eds.), Investor Behavior – The Psychology of Financial
Planning and Investing, John Wiley & Sons, Hoboken, NJ, 2014.
[29] R. Wüstenhagen, E. Menichetti, Strategic choices for renewable energy investment:
conceptual framework and opportunities for further research, Energy Policy 40
(2012) 1–10.
[30] WWF, The Energy Report: 100% Renewable Energy by 2050, WWF – World Wide
Fund for Nature, Gland, 2011 (Accessed 6 June 2017), https://c402277.ssl.cf1.
rackcdn.com/publications/384/files/original/The_Energy_Report.pdf?
1345748859.
[31] Greenpeace, Energy [R]evolution: A Sustainable World Energy Outlook 2015. 100%
Renewable Energy for All, Greenpeace International, 2015 (Accessed 6 June 2017),
http://www.greenpeace.org/international/Global/international/publications/
climate/2015/Energy-Revolution-2015-Full.pdf.
[32] IEA, World Energy Outlook 2016, International Energy Agency, Paris, 2016.
[33] IEA, World Energy Outlook 2015, International Energy Agency, Paris, 2015.
[34] Statoil, Energy Perspectives 2017: Long-Term Macro and Market Outlook, Statoil
ASA, Stavanger, 2017 (Accessed 20 December 2017), https://www.statoil.com/
content/dam/statoil/documents/energy-perspectives/energy-perspectives-2017-
v2.pdf.
[35] WEC, World Energy Scenarios 2016: The Grand Transition, World Energy Council,
London, 2016 (Accessed 6 June 2017), https://www.worldenergy.org/wp-content/
uploads/2016/10/World-Energy-Scenarios-2016_Full-report.pdf.
[36] WEC, World Energy Scenarios: Composing Energy Futures to 2050, World Energy
Council, London, 2013 (Accessed 6 June 2017), https://www.worldenergy.org/wp-
content/uploads/2013/09/World-Energy-Scenarios_Composing-energy-futures-to-
2050_Full-report.pdf.
[37] Shell, New Lens Scenarios. A Shift in Perspective for a World in Transition, Shell
International BV, 2013 (Accessed 6 June 2017), http://www.shell.com/energy-and-
innovation/the-energy-future/scenarios/new-lenses-on-the-future.html.
[38] ExxonMobil, 2017 Outlook for Energy: A View to 2040, ExxonMobil Corporation,
Irving, TX, 2017 (Accessed 20 December 2017), http://cdn.exxonmobil.com/
∼/media/global/files/outlook-for-energy/2017/2017-outlook-for-energy.pdf.
[39] EIA, International Energy Outlook 2016, US Energy Information Administration,
Washington, DC, 2016 (Accessed 6 June 2017), https://www.eia.gov/outlooks/
ieo/pdf/0484(2016).pdf.
[40] P. Donohoo-Vallett, Accounting Methodology for Source Energy of Non-
Combustible Renewable Electricity Generation, US Department of Energy, 2016
(Accessed 23 December 2017), https://energy.gov/sites/prod/files/2016/10/f33/
Source%20Energy%20Report%20-%20Final%20-%2010.21.16.pdf.
[41] IRENA, Chapter 3 of Perspectives for the Energy Transition – Investment Needs for a
Low-Carbon Energy System, International Renewable Energy Agency (IRENA),
2017.
[42] BNEF, New Energy Outlook 2017: Executive Summary. Bloomberg New Energy
Finance, (2017) (Accessed 15 December 2017), https://about.bnef.com/new-
energy-outlook/.
[43] Excel (2010). Microsoft Corporation, https://www.microsoft.com.
[44] REN21, Renewables 2017 Global Status Report, REN21 Secretariat, Paris, 2017
(Accessed 15 August 2017), http://www.ren21.net/wp-content/uploads/2017/06/
17-8399_GSR_2017_Full_Report_0621_Opt.pdf.
[45] Performance Curve Data Base, Fotovoltaica Series, (2017) (Accessed 5 June 2017),
http://pcdb.santafe.edu/graph.php?curve=255.
[46] IEA, Snapshot of Global Photovoltaic Markets 2016. Report IEA PVPS T1-31:2017,
IEA Photovoltaic Power Systems Programme 2017, (2017) (Accessed 9 September
2017), http://www.iea-pvps.org/fileadmin/dam/public/report/statistics/IEA-
112
Energy Research & Social Science 46 (2018) 103–113
PVPS_-_A_Snapshot_of_Global_PV_-_1992-2016__1_.pdf.
[47] Renewables Now, https://renewablesnow.com/news/global-pv-market-grows-to-
almost-100-gw-in-2017-605141/, 2018, (Accessed 24 April 2018).
[48] IEA, World Energy Model Documentation 2017 Version, International Energy
Agency, Paris, 2017.
[49] IEA, Tracking Clean Energy Progress 2017, International Energy Agency, Paris,
2017.
[50] Powerweb, Global Wind & Solar Installations in Gigawatts (GW), (2017) (Accessed
5 June 2017), http://www.fi-powerweb.com/Renewable-Energy.html.
[51] M. Motyka, S. Sanborn, US Solar Power Growth through 2040: Exponential or
Inconsequential? Deloitte Centre for Energy Solutions, Deloitte Development LLC,
2015 (Accessed 6 June 2017), https://www2.deloitte.com/content/dam/Deloitte/
us/Documents/energy-resources/us-er-solar-innovation-growth.pdf.
[52] J.L. Funk, What drives exponential improvements? Calif. Manage. Rev. 55 (3)
(2013) 134–152.
[53] U. Pillai, Drivers of cost reduction in solar photovoltaics, Energy Econ. 50 (2015)
286–293.
[54] G. Nemet, Solar photovoltaics: multiple drivers of technological improvement, in:
A. Grubler, F. Aguayo, K.S. Gallagher, M. Hekkert, K. Jiang, L. Mytelka, L. Neij,
G. Nemet, C. Wilson (Eds.), Case Study 10 in Historical Case Studies of Energy
Technology Innovation. Chapter 24, The Global Energy Assessment, Cambridge
University Press, Cambridge, UK, 2012.
[55] M. Woodhouse, R. Jones-Albertus, D. Feldman, R. Fu, K. Horowitz, D. Chung,
D. Jordan, D. Kurtz, On the Path to SunShot: The Role of Advancements in Solar
Photovoltaic Efficiency, Reliability, and Costs, National Renewable Energy
Laboratory, Golden, Colorado, 2016 NREL/TP-6A20-65872.
[56] R. Jones-Albertus, D. Feldman, R. Fu, K. Horowitz, M. Woodhouse, Technology
advances needed for photovoltaics to achieve widespread grid price parity, Prog.
Photovolt. 24 (9) (2016) 1272–1283.
[57] J. Channell, E. Curmi, P. Nguyen, E. Prior, A.R. Syme, H.R. Jansen, E. Rahbari,
E.L. Morse, S.M. Kleinman, T. Kruger, Energy Darwinism II. Why a Low Carbon
Future Doesn’t Have to Cost the Earth, Citi GPS: Global Perspectives & Solutions,
2015 (Accessed 6 June 2017), https://ir.citi.com/E8%2B83ZXr1vd
%2Fqyim0DizLrUxw2FvuAQ2jOlmkGzr4ffw4YJCK8s0q2W58AkV
%2FypGoKD74zHfji8%3D.
[58] M.C. Lewis, Energy Transition and Climate Change: Stranded Assets, Fossilized
Revenues, Kepler Cheuvreux, Paris, 2014 (Accessed 16 August 2017), https://www.
keplercheuvreux.com/pdf/research/eg_eg_253208.pdf.
[59] J. Leaton, The $2 Trillion Stranded Assets Danger Zone: How Fossil Fuel Firms Risk
Destroying Investor Returns, Carbon Tracker, London, 2015 (Accessed 16 August
2017), http://www.carbontracker.org/wp-content/uploads/2015/11/CAR3817_
Synthesis_Report_24.11.15_WEB2.pdf.
[60] D. Tayal, Disruptive forces on the electricity industry: a changing landscape for
utilities, Electr. J. 29 (2016) 13–17.
[61] IIGCC, A Guide on Climate Change for Private Equity Investors. KPMG, Institutional
Investors Group on Climate Change, Principles for Responsible Investment, (2016)
(Accessed 23 June 2017), http://www.iigcc.org/files/publication-files/IIGCC_
Private_Equity_v27.pdf.
[62] N.H. Stern, Stern Review: The Economics of Climate Change, HM Treasury, UK,
2006.
[63] H. Covington, R. Thamotheram, The Case for Forceful Stewardship (Part 1): The
Financial Risk from Global Warming, (2015) (Accessed 6 June 2017), https://
papers.ssrn.com/sol3/papers.cfm?abstract_id=2551478.
[64] Mercer, Investing in a Time of Climate Change, Mercer LLC, International Finance
Corporation, the UK Department for International Development, 2015 (Accessed 6
June 2017), https://www.mercer.com/content/dam/mercer/attachments/global/
investments/mercer-climate-change-report-2015.pdf.
[65] H. Covington, R. Thamotheram, The Case for Forceful Stewardship (Part 2):
Managing Climate Risk, (2015) (Accessed 6 June 2017), https://papers.ssrn.com/
sol3/papers.cfm?abstract_id=2551485.
[66] V. Shah, J. Booream-Phelps, Crossing the Chasm, Solar Grid Parity in a Low Oil
Price Era, Deutsche Bank Securities Inc., New York, 2015 (Accessed 6 June 2017),
https://www.db.com/cr/en/docs/solar_report_full_length.pdf.
[67] Morgan Stanley, Renewable Energy Hits Global Tipping Point, (2017) (Accessed 3
September 2017), https://www.morganstanley.com/ideas/solar-wind-renewable-
energy-utilities.
[68] Generation Foundation, Stranded Carbon Assets. Why and How Carbon Risks
Should Be Incorporated in Investment Analysis, Generation Foundation, London,
2013 (Accessed 6 June 2017), https://www.genfound.org/media/1374/pdf-
generation-foundation-stranded-carbon-assets-v1.pdf.
[69] D. Frankel, K. Ostrowski, D. Pinner, The Disruptive Potential of Solar Power, April
2014 McKinsey Quarterly, McKinsey & Company, 2014 (Accessed 6 June 2017),
http://www.mckinsey.com/business-functions/sustainability-and-resource-
productivity/our-insights/the-disruptive-potential-of-solar-power.
[70] E. Feller, W. Martindale, R. Sullivan, Fiduciary Duty in the 21st Century – Global
Statement on Investor Obligations and Duties, UNEP FI, PRI, The Generation
Foundation, 2016 (Accessed 6 June 2017), http://www.unepfi.org/fileadmin/
documents/FiduciaryDutyStatement.pdf.
[71] S.D. Campbell, S.A. Sharpe, Anchoring bias in consensus forecasts and its effect on
market prices, J. Financ. Quant. Anal. 44 (2) (2009) 369–390.
[72] IEA, World Energy Outlook 2017 to Include Focus on China’s Energy Outlook and
the Natural Gas Revolution, (2017) (Accessed 9 September 2017), https://www.iea.
org/newsroom/news/2017/march/world-energy-outlook-2017-to-include-focus-
on-chinas-energy-outlook-and-the-natu.html.
[73] IEA, Technology Roadmap: Solar Photovoltaic Energy. Energy Technology
Perspectives, International Energy Agency, Paris, 2014.
G. Carrington, J. Stephenson
[74] IEA, Next Generation Wind and Solar Power; From Cost to Value, International
Energy Agency, Paris, 2016.
[75] K. Mackenzie, Why IEA Scenarios Should Be Treated With Extreme Caution, 25 May
2017. The Financial Times, London, 2017 (Accessed 6 June 2017), https://
ftalphaville.ft.com/2017/05/24/2189189/guest-post-why-iea-scenarios-should-be-
treated-with-extreme-caution.
[76] M. Metayer, C. Breyer, H.-J. Fell, The projections for the future and quality in the
past of the World Energy Outlook for solar PV and other renewable energy tech-
nologies Hamburg, 14–18 September 2015, Proceedings 31st European
Photovoltaic Solar Energy Conference and Exhibition – EU PVSEC (2015)
3220–3246.
[77] T. Osmundsen, How the IEA Exaggerates the Costs and Underestimates the Growth
of Solar Power, 4 March 2014. Energy Post Productions BV, Amsterdam, 2014
(Accessed 6 June 2017), http://energypost.eu/iea-exaggerates-costs-
underestimates-growth-solar-power/.
[78] A. Whitmore, The IEA’s Central Projections for Renewables Continue to Look Way
Too Low, 2 March 2015. The Energy Collective, Energy Post Productions BV,
Amsterdam, 2015 (Accessed 6 June 2017), http://www.theenergycollective.com/
onclimatechangepolicy/2199371/iea-s-central-projections-renewables-continue-
look-way-too-low.
113
Energy Research & Social Science 46 (2018) 103–113
[79] V. Sivaram, S.D. Stranks, H.J. Snaith, Outshining silicon, Sci. Am. 313 (1) (2015)
54–59.
[80] Y. Yang, J. You, Make perovskite solar cells stable, Nature 544 (2017) 155–156.
[81] P. Canepa, S.-H. Bo, G.S. Gautam, B. Key, W.D. Richards, T. Shi, Y. Tian, Y. Wang,
J. Li, G. Ceder, High magnesium mobility in ternary spinel chalcogenides, Nat.
Commun. 8 (2017) 1759.
[82] J. Gaede, J. Meadowcroft, A question of authenticity: status quo bias and the
International Energy Agency’s World Energy Outlook, J. Environ. Policy Plan. 18
(5) (2016) 608–627.
[83] Y. Karneyeva, R. Wüstenhagen, Solar feed-in tariffs in a post-grid parity world: the
role of risk, investor diversity and business models, Energy Policy 106 (2017)
445–456.
[84] P. Huang, S.O. Negro, M.P. Hekkert, K. Bi, How China became a leader in solar PV:
an innovation system analysis, Renew. Sustain. Energy Rev. 64 (2016) 777–789.
[85] M. Deutsch, P. Graichen, What If… There Were a Nationwide Rollout of PV Battery
Systems? Agora Energiewende, 2015 (Accessed 6 June 2017), https://www.agora-
energiewende.de/fileadmin/Projekte/2015/PV-Speicher-Rollout/Agora_
Speicherdurchbruch_2015-10-08_web_EN.pdf.
[86] N.H. Stern, Why Are We Waiting? The Logic, Urgency and Promise of Tackling
Climate Change, MIT Press, Cambridge, MA, 2015.