Full Terms & Conditions of access and use can be found at
New Political Economy
ISSN: 1356-3467 (Print) 1469-9923 (Online) Journal homepage: https://www.tandfonline.com/loi/cnpe20
Is Green Growth Possible?
Jason Hickel & Giorgos Kallis
To cite this article: Jason Hickel & Giorgos Kallis (2019): Is Green Growth Possible?, New Political
Economy, DOI: 10.1080/13563467.2019.1598964
To link to this article: https://doi.org/10.1080/13563467.2019.1598964
Published online: 17 Apr 2019.
Submit your article to this journal
Article views: 2
View Crossmark data
Is Green Growth Possible?
and Giorgos Kallis
Anthropology, Goldsmiths, University of London, London, UK;
ICREA and ICTA-UAB, Universitat Autonoma de
Barcelona, Barcelona, Spain
The notion of green growth has emerged as a dominant policy response to
climate change and ecological breakdown. Green growth theory asserts
that continued economic expansion is compatible with our planet’s
ecology, as technological change and substitution will allow us to
absolutely decouple GDP growth from resource use and carbon
emissions. This claim is now assumed in national and international
policy, including in the Sustainable Development Goals. But empirical
evidence on resource use and carbon emissions does not support green
growth theory. Examining relevant studies on historical trends and
model-based projections, we ﬁnd that: (1) there is no empirical evidence
that absolute decoupling from resource use can be achieved on a global
scale against a background of continued economic growth, and (2)
absolute decoupling from carbon emissions is highly unlikely to be
achieved at a rate rapid enough to prevent global warming over 1.5°C
or 2°C, even under optimistic policy conditions. We conclude that green
growth is likely to be a misguided objective, and that policymakers need
to look toward alternative strategies.
ecological economics; green
The notion of green growth emerged as a central theme at the Rio+ 20 Conference on Sustainable
Development in 2012, and featured prominently in the outcome document The World We Want (UN
2012), which called simultaneously for a ‘green economy’and ‘sustained economic growth’. Green
growth has since become a dominant response to increasingly serious warnings about climate
change and ecological breakdown (Dale et al. 2016). As a theory, green growth asserts that continued
economic expansion (as measured by Gross Domestic Product, or GDP) is or can be made to be com-
patible with our planet’s ecology. While this idea has been latent in the rhetoric of sustainable devel-
opment since the Brundtland Commission and the ﬁrst Rio Conference, with early formulations taking
shape under names like Ecological Modernization (Ayres and Simonis, 1993, Weizsäcker et al. 1998)or
the Environmental Kuznets curve hypothesis (Dasgupta et al. 2002), green growth theory renders it as
a formal assertion.
Green growth theory is now promoted by leading multilateral organisations and is assumed in
national and international policy. It rests on the assumption that absolute decoupling of GDP
growth from resource use and carbon emissions is feasible (e.g. Solow 1973), and at a rate
suﬃcient to prevent dangerous climate change and other dimensions of ecological breakdown.
This review paper examines this assumption, and tests it against extant empirical evidence. We
ask: how do international organisations deﬁne green growth? Does the theory of green growth
(and speciﬁcally, the assumption that absolute decoupling of GDP growth from material throughput
© 2019 Informa UK Limited, trading as Taylor & Francis Group
CONTACT Jason Hickel email@example.com
NEW POLITICAL ECONOMY
and carbon emissions can be accomplished at a suﬃciently rapid rate) withstand scrutiny in light of
existing data and model-based projections? And if not, what are the implications for policy?
Deﬁning Green Growth
There are three major institutional proponents of green growth theory at the international level: the
OECD, the United Nations Environment Program (UNEP), and the World Bank. Each published ﬂagship
reports on green growth around the time of the Rio+ 20 Conference. In 2011, the OECD launched a
green growth strategy titled Towards Green Growth. That same year, UNEP published a report titled
Toward a Green Economy: Pathways to Sustainable Development and Poverty Eradication. In 2012, the
World Bank published Inclusive Green Growth: The Pathway to Sustainable Development. During the Rio
+ 20 Conference, these institutions joined with the Global Green Growth Institute to create the Green
Growth Knowledge Platform as an instrument for advancing green growth strategy around the world.
Each of the three organisations oﬀers a diﬀerent deﬁnition of green growth. The OECD deﬁnes it as
‘fostering economic growth and development while ensuring that natural assets continue to provide
the resources and environmental services on which our well-being relies’(2011, p. 18). The World
Bank (2012)deﬁnes it as
economic growth that is eﬃcient in its use of natural resources, clean in that it minimizes pollution and environ-
mental impacts, and resilient in that it accounts for natural hazards and the role of environmental management
and natural capital in preventing physical disasters.
UNEP eschews the language of green growth in favour of ‘green economy’, which it deﬁnes as one
that simultaneously grows income and improves human well-being ‘while signiﬁcantly reducing
environmental risks and ecological scarcities’(2011, p. 16).
None of these deﬁnitions are as precise as we might hope (see Jacobs 2013). As Smulders et al.
(2014) points out, the concept of green growth is ‘new and still somewhat amorphous.’The World
Bank’sdeﬁnition is the weakest. The World Bank seeks to ‘minimize’the environmental impact of
growth; but one can minimise environmental impact without reducing impact from its present
levels, and indeed while still nonetheless increasing overall impact. The OECD is slightly stronger
in that it seeks to ‘maintain’resources and environmental services, but here too there is no
demand to reduce impact. The UNEP report oﬀers the strongest deﬁnition in that it calls for reducing
environmental impact and ecological scarcities, and for ’rebuilding natural capital’.
The three institutions agree however on the mechanism for achieving green growth. The promise is
that technological change and substitution will improve the ecological eﬃciency of the economy, and
that governments can speed this process with the right regulations and incentives. But they diﬀer in
the clarity of their claims. The World Bank does not ask whether policy-driven innovations will suﬃce
to reduce environmental impact. The OECD, for its part, clariﬁes that green growth is only possible if
technology becomes eﬃcient enough to achieve ‘decoupling’of growth from environmental impact.
UNEP takes this a step further, and puts decoupling at the centre of the analysis:
A key concept for framing the challenges we face in making the transition to a more resource eﬃcient economy is
decoupling. As global economic growth bumps into planetary boundaries, decoupling the creation of economic
value from natural resource use and environmental impacts becomes more urgent.
UNEP notes that ‘recent trends indicate a moderate tendency of relative decoupling over time’, but
points out that this is not enough: ‘The central challenge …is to decouple growth absolutely from
material and energy intensity’(UNEP 2011, p. 15).
Here again UNEP oﬀers the clearest –and strongest –policy-oriented deﬁnition of green growth,
namely, that green growth requires absolute decoupling of GDP from resource use and environ-
mental impact. This is in keeping with the ecological literature, which insists that in a context of eco-
logical overshoot (Rockström et al.2009, Ceballos et al.,2015,2017, Steﬀen et al.2015), it is not
enough to simply ‘minimize’environmental impact –we must rapidly reduce it down to safe limits.
2J. HICKEL AND G. KALLIS
This leaves us with the question: Is absolute decoupling possible, and, if so, is it possible at a rate
suﬃcient for returning to and staying within planetary boundaries? None of the three reports on
green growth provide any evidence that it is. But since the Rio+ 20 conference, a number of key
studies have emerged to shed new light on this question. We outline the ﬁndings of this empirical
literature in what follows, looking at the two primary dimensions of decoupling –resource use
and carbon emissions –in turn,
before discussing theoretical and policy implications.
Resource Use –Is Absolute Decoupling Possible?
The conventional metric for measuring an economy’s resource use is ‘domestic material consumption’
(DMC), which is the total weight of raw materials (biomass, minerals, metals and fossil fuels) extracted
from the domestic territory, plus all physical imports minus all physical exports. While DMC is not a
direct indicator of ecological pressure, it is a well-established and widely-used proxy in the policy lit-
erature and enjoys robust empirical grounding for this purpose (Krausmann et al.2009, p. 2703). Van
der Voet et al.(2004)ﬁnd that while the mass ﬂows of individual materials are not indicative of their
ecological impacts, and while impacts vary as technologies change, at an aggregate level there is a
high degree of correlation (0.73) between material throughput and ecological impacts.
To assess the relationship between GDP and resource use, many governments have adopted the
practice of dividing GDP by DMC. This gives an indication of the ‘resource eﬃciency’of an economy. If
GDP grows faster than DMC (relative decoupling), the economy is becoming more resource eﬃcient.
GDP/DMC is used by the European Union to monitor progress toward green growth. It is also the
headline metric of the OECD’s annual Green Growth Indicators report.
By this metric, it appears that many nations have achieved relative decoupling, with GDP growing
at a rate faster than DMC. In the 2017 edition of Green Growth Indicators, the OECD concluded that
‘material productivity has been improving in some OECD countries’(45). The report also indicates
that European OECD nations have achieved absolute decoupling, growing GDP while reducing
DMC. Non-energy material consumption in the OECD declined from 12 tonnes per capita in 2000–
10 tonnes per capita in 2015, with the downward trend beginning after the ﬁnancial crisis in 2008
(it should be noted, however, that the OECD’s version of DMC does not include fossil fuels; this is
not normal practice in the literature on material ﬂows). These data are key to optimistic green
growth narratives, and underpin the popular notion that we have reached ‘peak stuﬀ’(e.g.
Goodall, 2011, Pearce, 2012).
DMC is a problematic indicator, however, as it does not include the material impact involved in the
production and transport of imported goods (Wiedmann et al.2015, Gutowski et al. 2017). In a glo-
balised economy, where rich countries have outsourced much of their production to poorer
countries, this side of material consumption has been shifted oﬀtheir balance sheet. If we bring it
back in, looking at the total resource impact of consumption by any given nation (what Wiedmann
et al refer to as ‘material footprint’, or MF), the picture changes. Wiedmann et al show that while the
USA, UK, Japan, the OECD and EU-27 have achieved relative decoupling of GDP from DMC (including
fossil fuels), material footprint has been rising at a rate equal to or greater than GDP, suggesting no
decoupling at all; indeed, in most cases re-coupling has occurred (see Figure 1). The OECD’sGreen
Growth Indicators partly recognises this problem, stating ‘progress is moderate once indirect ﬂows
associated with trade are considered.’Yet the report does not provide any data on indirect ﬂows;
and the data that is available suggests that progress has been not moderate but negative.
According to Wiedmann et al.(2015), the only signiﬁcant cases of relative decoupling of GDP from
material footprint have been China, India and South Africa. South Africa is the most notable of the
three, with near-zero growth in material footprint since 1990, although no evidence of sustained
On a global scale, resource use has been rising on a steady trajectory. Krausmann et al.(2009) show
that global extraction and consumption of materials (including fossil fuels) increased 8-fold during
the period 1900 to 2005, reaching 59 billion tons per year, growing at annual rates between 1 per
NEW POLITICAL ECONOMY 3
cent and 4 per cent. Giljum et al. (2014)ﬁnd that global consumption grew by 93.4 per cent between
1980 and 2009, at an average rate of 2.4 per cent per year, to reach a total of 67.6 billion tonnes.
Materialﬂows.net (2015), which is run by the Vienna University of Economics and Business, oﬀers
data for the period 1980 to 2013 and shows that global material footprint grew 132 per cent, at
an average rate of 2.5 per cent per year, to reach nearly 85 billion tons (Figure 2(a)).
What is the relationship between global GDP and resource use? Krausmann et al.(2009) show that
during the twentieth century GDP grew at a faster rate (3 per cent per year) than resource use (2 per
cent per year). This represents a relative decoupling or dematerialisation of GDP growth, at a rate of
about 1 per cent per year. But this changed in the twenty-ﬁrst century: the growth rate of global con-
sumption increased between 2000 and 2005, averaging 3.7 per cent per year. As this matched the
growth rate of GDP, no decoupling was achieved. Giljum et al. (2014) also ﬁnd that the growth
rate of global consumption accelerated in the twenty-ﬁrst century, averaging 3.4 per cent per year
between 2000 and 2009; once again, no decoupling was achieved. Wiedmann’s global data shows
a similar trend. Materialﬂows.net (2015) shows a period of modest growth of global material footprint
from 1980 to 2002, at 1.78 per cent per year. As this was slower than the rate of GDP growth, some
relative decoupling was achieved. However, the ﬁnal decade from 2002 to 2013 shows an accelera-
tion of global material use, at 3.85 per cent per year.
Global material use rose more quickly than GDP
during this decade. In other words, the material intensity of the world economy has been increasing
in the twenty-ﬁrst century, not decreasing. The authors state: ‘Currently, the world economy is there-
fore on a path of re-materialization and far away from any –even relative –decoupling.’(Figure 2(b)).
In sum: global historical trends show relative decoupling but no evidence of absolute decoupling,
and twenty-ﬁrst century trends show not greater eﬃciency but rather worse eﬃciency, with re-coup-
ling occurring. Of course, future trajectories could potentially break with these trends if we change
the composition and technology of the global economy (Grossman and Krueger 1995). What does
the data about future prospects show?
One argument is that resource intensity will diminish as economies shift from manufacturing to
services. Historical data do not support this theory, however. As a proportion of world GDP, services
have grown from 63 per cent in 1997 to 69 per cent in 2015, according to World Bank data. Yet during
this same period global material use has accelerated, outstripping global GDP growth. The same is
true of high-income nations. Services represent 74 per cent of GDP in high-income nations (up
from 69 per cent in 1997), but DMC has not diminished and material footprint is outpacing GDP
growth. This may be because services require resource-intensive inputs (in other words, services
embody signiﬁcant amounts of materials), or because the income acquired from selling services is
used to purchase resource-intensive consumer goods (Kallis 2017). Another possibility is that the
resource intensity of primary and secondary sectors has increased to the point of outstripping any
gains made by switching to services. Whatever the cause may be, there is no historical evidence
that switching to services will, in and of itself, reduce the material throughput of the global economy.
Another argument is that technological innovation and government policy might drive decou-
pling in the future. This is the assumption advanced by the World Bank, OECD and UNEP green
Figure 1. Material use trends for EU-27, OECD and USA, 1990–2008. Source: Wiedmann et al.(2015).
4J. HICKEL AND G. KALLIS
growth reports. To our knowledge, there are three major studies that examine this possibility on a
global scale. We discuss their ﬁndings below.
Dittrich et al. (2012) show that a ‘business as usual’scenario will result in material use rising from
68 billion tons in 2008 to 180 billion tons in 2050. This scenario assumes that global South economies
grow to the point where global average per capita consumption in 2030 will equal the OECD’s per
capita consumption in 2008. Dittrich et al conclude that this level of resource use is ‘not an option
for the future’. By contrast, their optimistic scenario assumes (a) medium population growth; (b)
that all countries follow best practice in eﬃcient resource use; and (c) that reduction of consumption
of one material does not require higher consumption of another. In this scenario, resource use
reaches 93 billion tons by 2050. This represents relative decoupling, but no absolute reduction in
In a second study, Schandl et al.(2016) use a model based on 3 per cent average annual global
GDP growth and explore three scenarios between 2010 and 2050. The reference scenario, with no
signiﬁcant change to environmental policies, shows that global resource use grows from 79.4
billion tonnes in 2010 to 183 billion tons in 2050 (similar to the Dittrich et al projection), with
slight relative decoupling. The ‘medium eﬃciency’scenario, with a carbon price of $25 per ton of
(rising by 4 per cent per year), shows that global resource use still grows steadily over the
period, but at about half the rate of global GDP, reaching 130 billion tons by 2050. The ‘high
eﬃciency’scenario, with a carbon price starting at $50 (rising by 4 per cent per year to $236 by
2050) plus a doubling in the material eﬃciency of the economy (from historical average improve-
ments of 1.5 per cent per year up to 4.5 per cent per year), shows that global resource use still
grows steadily, but at about one-fourth the rate of global GDP, reaching 95 billion tons in 2050
(again, similar to Dittrich et al).
It is important to note that the rate of material eﬃciency improvements that Schandl et al assume
(viz., 4.5 per cent per year) has no empirical basis. They provide no evidence that such a rapid rate is
possible to sustain. Yet even with this optimistic assumption, Schandl et al conclude: ‘Our research
shows that while some relative decoupling can be achieved in some scenarios, none would lead
to an absolute reduction in …materials footprint.’
Finally, UNEP has developed a model that explores four diﬀerent future scenarios, which they
discuss in their 2017 report Assessing Global Resource Use (UNEP 2017a, pp. 42–45). Their reference
scenario, extrapolating from existing trends, shows that global resource use rises steadily from 85
Figure 2. (a) Global material footprint, 1970–2013; (b) Change in global material footprint compared to change in global GDP
(constant 2010 USD), 1990–2013. Source: Materialﬂows.net/World Bank.
NEW POLITICAL ECONOMY 5
billion tons in 2015 to 186 billion tons by 2050 (similar to Dittrich et al and Schandle et al). Their high
eﬃciency scenario, by contrast, includes strong policy measures: (a) a global carbon price of $5 per
ton of CO
e in 2021, rising by 18.1 per cent per year to $573 in 2050; (b) technological innovation that
improves resource eﬃciency; (c) a resource extraction tax that increases the price of natural resources
relative to other inputs; and (d) progressive changes to government regulations, planning and pro-
curement policies (for full details of the model see UNEP 2017b, p. 287 ﬀ). The high eﬃciency scenario
projects that global resource use rises to 132 billion tons in 2050. While some relative decoupling is
achieved, there is no absolute reduction in resource use.
The UNEP projections are signiﬁcantly worse than either Dittrich et al or Schandl et al predict. The
model’s authors, Ekins and Hughes, say this because they have incorporated the ‘rebound eﬀect’into
their model (UNEP 2017a, 106 ﬀ.). The rebound eﬀect cancels out some gains in resource eﬃciency.
This happens because such gains reduce the cost of a good or service, freeing up income and increas-
ing eﬀective demand (see Herring and Sorrell 2009 for a review of the literature). In light of these
ﬁndings, UNEP acknowledges that improvements in resource eﬃciency will not be enough, in and
of themselves, to achieve sustainability, or green growth. ‘Resource eﬃciency alone is not enough.
Productivity gains in today’s linear production system are likely to lead to increased material
demand through a combination of economic growth and rebound eﬀects’(12). Instead, the report
acknowledges that something else is needed. They suggest further investigation into the principles
of a circular economy: ‘a move from linear to circular material ﬂows through a combination of
extended product life cycles, intelligent product design and standardization, reuse, recycling and
remanufacturing’(12). Improving circularity could reduce the ecological impact of material through-
put, but only a small fraction of total throughput has circular potential. 44 per cent is comprised of
food and energy inputs, which are irreversibly degraded, and 27 per cent is net addition to stocks of
buildings and infrastructure (Haas et al. 2015).
These models suggest that absolute decoupling is not feasible on a global scale in the context of
continued economic growth. These are global studies, however. One might argue that when it comes
to the question of whether green growth is possible, we need to look speciﬁcally at what high-
income nations might be able to achieve, given their greater capacity for technological development.
Hatﬁeld-Dodds et al. (2015) have modelled a number of scenarios for Australia from 2015 to 2050,
with results that have been widely cited in support of green growth theory. Their most optimistic
scenario assumes high levels of policy-driven eﬃciency gains, with an overall 70 per cent drop in
material intensity. They ﬁnd that ‘substantial economic and physical decoupling is possible,’with
GDP increasing at an average rate of 2.41 per cent per year ‘while associated environmental pressures
ease (greenhouse gas emissions, water stress, native habitat loss)’.The model suggests that this can
be accomplished without outsourcing environmental impact to other countries.
Hatﬁeld-Dodds et al have come under criticism for this model, however. First, they provide no evi-
dence for their assumption that a 70 per cent drop in material intensity is possible. Alexander et al.
(2018) have pointed out that this rate of eﬃciency improvement is baseless and unrealistic. Indeed,
the Australian Bureau of Agricultural Economics (ABARE 2008) reports that eﬃciency is likely to
improve by only 0.2 per cent to 0.5 per cent per year into the future –at most one-eighth of the
rate that Hatﬁeld-Dodds assume. Second, even if a 70 per cent drop in material intensity was possible,
it appears that any resulting decrease in resource use may only be achieved over the short term. The
optimistic scenario in the Hatﬁeld-Dodds et al model shows that material use declines from 2015 to
2040, but begins to increase again thereafter.
Ward et al.(2016) have tested the Hatﬁeld-Dodds model over a longer period, to 2100. They
assume a drop in material intensity by 2050 that is 50 per cent more than Hatﬁeld-Dodds et al
propose, for an even more optimistic scenario. They ﬁnd that material extraction declines until
2050 (decoupling at an average rate of about 4 per cent per year) but then ﬂattens oﬀand rises stea-
dily so that by 2100 material use is 20 per cent to 60 per cent higher than its initial value in 2015.
While absolute decoupling from material extraction is achieved in the short term, in the longer
term material extraction rises by 2.16 per cent per year, nearly matching the rate of GDP growth.
6J. HICKEL AND G. KALLIS
Note that the indicator ‘material extraction’is diﬀerent from both DMC and material footprint, in that
it does not include imports; the ﬁgures for DMC and material footprint for Australia would be signiﬁ-
cantly higher (Figure 3).
Ward et al.(2016) argue that this resurgence in material extraction happens because resource
eﬃciency cannot improve forever, as eventually it approaches physical limits. They state:
For non-substitutable resources such as land, water, raw materials and energy, we argue that whilst eﬃciency
gains may be possible, there are minimum requirements for these resources that are ultimately governed by
physical realities: for instance the photosynthetic limit to plant productivity and maximum trophic conversion
eﬃciencies for animal production govern the minimum land required for agricultural output; physiological
limits to crop water use eﬃciency govern minimum agricultural water use, and the upper limits to energy and
material eﬃciencies govern minimum resource throughput required for economic production.
As the physical limits of resource efﬁciency are reached, continued GDP growth drives resource use
back up. Ward et al conclude that ‘decoupling of GDP growth from resource use, whether relative or
absolute, is at best only temporary. Permanent decoupling (absolute or relative) is impossible for
essential, non-substitutable resources because the efﬁciency gains are ultimately governed by phys-
ical limits. Growth in GDP ultimately cannot plausibly be decoupled from growth in material and
energy use, demonstrating categorically that GDP growth cannot be sustained indeﬁnitely. It is there-
fore misleading to develop growth-oriented policy around the expectation that decoupling is
Conclusions and Discussion
The empirical data suggest that absolute decoupling of GDP from resource use (a) may be possible in
the short term in some rich nations with strong abatement policy, but only assuming theoretical
eﬃciency gains that may be impossible to achieve in reality; (b) is not feasible on a global scale,
even under best-case scenario policy conditions; and (c) is physically impossible to maintain in the
longer term. In light of this data, we can conclude that green growth theory –in terms of resource
use –lacks empirical support. We are not aware of any credible empirical models that contradict
this conclusion. There are three counterpoints to consider, however:
First, this conclusion is sensitive to the baseline rate of GDP growth. The studies cited above
project growth at 2–3 per cent per year. As the growth rate approaches zero, absolute decoupling
becomes more feasible, and is likely to last longer. It is reasonable to expect that green growth
could be accomplished at very low GDP growth rates, i.e. less than 1 per cent per year –signiﬁcantly
lower than historical trends and projected pathways.
Second, the studies cited above are based on the existing relationship between GDP and material
throughput. They model the impact of known variables, such as eﬃciency improvements,
Figure 3. Projections for material extraction in Australia under highly optimistic conditions, 2015–2100. Green dots represent the
Hatﬁeld-Dodds projection to 2050. Source: Ward et al.(2016).
NEW POLITICAL ECONOMY 7
technological innovation, taxes, shifts to services, etc. However, one might argue that it is theoreti-
cally possible to break the existing relationship between GDP and material throughput altogether.
We reﬂect on this in the penultimate section of this paper.
Third, one might argue that the aggregate material footprint indicator obscures the possibility of
shifting from high-impact resources to low-impact resources. It is true that diﬀerent materials have
diﬀerent impacts, and that renewable and non-renewable materials have diﬀerent kinds of sustain-
ability thresholds, but the aggregate measure is nonetheless regarded as a useful proxy because all of
the constituent material categories exhibit roughly the same trends as the total (i.e. they all increase
with GDP growth). And because all materials have some impact, indeﬁnite growth of any material cat-
egory is not compatible with ecological principles.
It is important to point out that the standard for green growth we have used above is a conser-
vative one, inasmuch as it regards any reduction of annual resource use, however small, as green. The
academic literature on resource use is signiﬁcantly more stringent than this. An emerging consensus
holds that global material footprint needs to be reduced to 50 billion tons per year in order to be
compatible with the planet’s ecology (Dittrich et al.,2012, Hoekstra and Wiedmann 2014, UNEP
2014, Bringezu 2015). Bringezu (2015) goes further and suggests that this reduction needs to
happen by 2050. Of course, there are reasons to be skeptical of global targets like this, as they
combine renewable and non-renewable materials that should be treated separately, and because
the impacts of material use are locally speciﬁc and thresholds should be tailored to local ecosystems
(except in the case of fossil fuels and land-based biomass extraction, which aﬀect greenhouse gas
emissions). Still, the literature is clear that material footprint needs to be scaled down signiﬁcantly
from present levels. In other words, to be truly green, green growth requires not just any degree
of absolute decoupling, but absolute decoupling that is rapid enough to meet ecological targets.
Carbon Emissions –Is Growth Compatible with the Paris Agreement?
Unlike with resource use, there is a steady long-term trend toward relative decoupling of GDP from
carbon emissions, and we know that absolute reductions in carbon emissions are possible to achieve.
When it comes to climate change, however, the objective is not simply to reduce emissions (a matter
of ﬂows), but to keep total emissions from exceeding speciﬁc carbon budgets (a matter of stocks). For
green growth theory, then, the question is not only whether we can achieve absolute decoupling and
reduce emissions, but whether we can reduce emissions fast enough to stay within the carbon
budgets for 1.5°C or 2°C, as per the Paris Agreement, while still continuing economic growth.
A number of high-income countries have seen declining emissions in the twenty-ﬁrst century,
despite continued economic growth. Figure 4(a) shows declining emissions in the US and EU28, in
both territorial and consumption-based terms, from 2006 to 2016 (i.e. absolute decoupling).
However, emissions from the global South have continued upward, albeit at a slower rate than
GDP (i.e. relative decoupling). China’s emissions declined slightly between 2014 and 2016 (a brief
period of absolute decoupling), before growing again in 2017.
On a global level, CO
emissions have increased steadily, falling only during periods of economic
recession (Figure 4(b)). Global emissions did level oﬀin 2015 and 2016 while GDP continued to rise,
prompting the International Energy Agency, a research arm of the OECD, to announce ‘Decoupling of
global emissions and economic growth conﬁrmed’(IEA 2016), while media outlets celebrated ‘peak
emissions’(Meyer 2016). This news brieﬂy came to constitute a key element of optimistic green
growth narratives, until global emissions began to rise again in 2017 (1.6 per cent) and 2018 (2.7
per cent). Analysts attribute the temporary plateau to a shift in China away from coal and (mostly)
toward oil and gas, and a shift in the US to natural gas.
Once these shifts were complete, continued
economic growth drove emissions up again.
Overall, global carbon productivity has been slowing. World Bank data shows that carbon pro-
per 2010 $US GDP) improved steadily from 1960 to 2000, with decarbonisation hap-
pening at an average rate of 1.28 per cent per year (relative decoupling). However, from 2000 to
8J. HICKEL AND G. KALLIS
2014 there was no improvement in carbon productivity –in other words, not even relative decou-
pling has been achieved in the twenty-ﬁrst century.
High-income nations have done better, at
least in terms of territorial emissions (the World Bank does not track consumption-based emissions),
but even so progress has slowed, from an average rate of 1.91 per cent per year from 1970 to 2000,
down to 1.61 per cent per year from 2000 to 2014.
Existing trends are incompatible with the Paris Agreement targets. Business-as-usual is set to lead
to 4.2°C of warming (2.5°C to 5.5°C) by 2100. Even with the Nationally Determined Contributions and
Intended Nationally Determined Contributions under the Paris Agreement, global warming is still
projected to reach 3.3°C (1.9°C to 4.4°C) –an improvement over the BAU scenario but still far exceed-
ing the 1.5°C and 2°C thresholds.
In order to keep warming below these thresholds, the world will
have to make much more aggressive emissions reductions.
The IPCC’s Fifth Assessment Report (AR5) includes 116 mitigation scenarios that are consistent
with Representative Concentration Pathway 2.6 (RCP2.6), which oﬀers the best chances of staying
below 2°C. All of these scenarios are green growth scenarios in that they stabilise global temperatures
while global GDP continues to rise. Rising GDP is a built-in feature of the Shared Socio-Economic Path-
ways (SSPs), which form the basis for the IPCC mitigation scenarios (Kuhnhenn 2018). AR5 warns,
however, that these scenarios ‘typically involve temporary overshoot of atmospheric concentrations’
and ‘typically rely on the availability and widespread deployment of bioenergy with carbon capture
and storage (BECCS)’(2014, p. 23). Indeed, the vast majority scenarios for 2°C (101 of the 116) rely on
BECCS to the point of achieving negative emissions.
BECCS entails growing large tree plantations to
from the atmosphere, harvesting the biomass, burning it for energy, capturing the CO
emissions at source and storing it underground. Relying on these ‘negative emissions technologies’
allows for a much larger carbon budget (about double the actual size) by assuming that we can suc-
cessfully reduce global atmospheric carbon in the second half of the century.
BECCS is highly controversial among climate scientists. It was ﬁrst proposed by Obersteiner et al.
(2001) and Keith (2001) at the turn of the century. IPCC modelling teams began including it in their
scenarios from 2005, despite having no ﬁrm evidence of its feasibility. With the publication of AR5,
BECCS was enshrined as a dominant assumption. Obersteiner has expressed alarm at the rapid
uptake of his idea; he considers BECCS to be what he calls a ‘risk-management strategy’,ora‘back-
stop technology’in case climate feedback loops turn out to be worse than expected, and says the
IPCC has ‘misused’it by including it in regular scenarios to take pressure oﬀof conventional mitiga-
tion pathways (i.e. emissions reductions) (Hickman 2016). In Keith’s(2001) initial formulation of the
idea, he noted that while ‘measured use’of biomass could help mitigate environmental problems,
‘large scale use of cropped biomass will not.’
Figure 4. (a) Annual territorial and consumption CO
emissions for select regions, 1990–2016; (b) Global CO
Source: Global Carbon Budget (2018).
NEW POLITICAL ECONOMY 9
Anderson and Peters (2016) point out that the ‘allure’of BECCS is due to the fact that it allows
politicians to postpone the need for rapid emissions reductions: ‘BECCS licenses the ongoing com-
bustion of fossil fuels while ostensibly fulﬁlling the Paris Commitments.’There are a number of con-
cerns. First, the viability of power generation with CCS has never been proven to be economically
viable or scalable; it would require the construction of 15,000 facilities (Peters 2017). Second, the
scale of biomass assumed in the AR5 scenarios would require plantations covering land two to
three times the size of India, which raises questions about land availability, competition with food
production, carbon neutrality, and biodiversity loss (Smith et al.2016; Heck et al. 2018). Third, the
necessary storage capacity may not exist (De Coninck and Benson 2014, Global CCS Institute
2015). Anderson and Peters conclude that ‘BECCS thus remains a highly speculative technology’
and that relying on it is therefore ‘an unjust and high stakes gamble’: if it is unsuccessful, ‘society
will be locked into a high-temperature pathway.’This conclusion is shared by a growing number
of scientists (e.g. Fuss et al. 2014, Vaughan and Gough, 2016, Larkin et al.2017, Van Vuuren et al.
2017), and by the European Academies’Science Advisory Council (2018).
It is not clear that we can justiﬁably rely on BECCS, an unproven technology, to underwrite green
growth theory. If we accept this point, then we must return to asking whether it is possible to main-
tain growth without relying on BECCS to stay within the carbon budgets consistent with the Paris
Agreement. Without BECCS, global emissions need to fall to net zero by 2050 for 1.5°C, or by 2075
This entails reductions of 6.8 per cent per year and 4 per cent per year, respectively
(Figure 5). Theoretically, this can be accomplished with (a) a rapid shift to 100 per cent renewable
energy to eliminate emissions from fossil fuel combustion (Jacobson and Delucchi 2011); plus (b)
aﬀorestation and soil regeneration to eliminate emissions from land use change; plus (c) a shift to
alternative industrial processes to eliminate emissions from the production of cement, steel, and
plastic. The question is, can all of this be accomplished quickly enough?
Only 6 of the 116 scenarios for 2°C in AR5 exclude BECCS. These work by assuming ‘optimal full
technology’in all other areas, plus mass aﬀorestation, and with high mitigation costs. These represent
theoretically possible pathways, but without any empirical evidence as to their feasibility.
Results of empirical studies are not promising. Schandl et al.(2016) model what might be achieved
with aggressive mitigation policies, without relying on BECCS. Their high-eﬃciency scenario has a
carbon price starting at $50 per ton (rising by 4 per cent per year to $236 by 2050) plus a doubling
in the material eﬃciency of the economy due to technological innovations (improving from a histori-
cal average rate of 1.5 per cent per year up to 4.5 per cent). Schandl et al provide no evidence for the
feasibility of the eﬃciency improvements that they assume. Even so, the result shows that with global
growth of 3 per cent per year, annual emissions plateau to 2050 but do not decline. In this scenario,
Figure 5. CO
mitigations curves for 1.5°C and 2°C. Source: Global Carbon Budget (2018).
10 J. HICKEL AND G. KALLIS
growth in energy demand outstrips the rate of decarbonisation, violating the carbon budgets for
1.5°C and 2°C.
The International Renewable Energy Association (IRENA 2018) have modelled a scenario for con-
tinued GDP growth compatible with 2°C by relying on a rapid shift to renewable energy (consistent
with Jacobson and Delucchi 2011). The scenario requires adding 12,200 GW of solar and wind
capacity by 2050, with a dramatic increase in installation rates (2.3 to 4.6 times faster than the
The scenario also requires that the energy intensity of the global economy falls by two-
thirds (by 2.8 per cent per year, double the historical rate), lowering energy demand in 2050 to
slightly less than 2015 levels.
This is feasible inasmuch as the transition to wind and solar itself
improves energy eﬃciency (Jacobson and Delucchi 2011).
Still, even this optimistic scenario accom-
plishes only 90 per cent of the necessary emissions reductions for 2°C (likely because it pays no atten-
tion to emissions from land use change and cement production). The model relies on negative
emissions technology to cover most of the remainder.
Van Vuuren et al.(2018) consider ‘alternative pathways’for meeting the Paris Agreement targets
without relying on widespread use of negative emissions technologies. They model rising GDP in
accordance with SSP2. In addition to a carbon tax and other aggressive mitigation strategies, their
optimistic scenario includes the following settings: global population peaks at 8.4 billion in 2050
and declines to 6.9 billion by 2100; meat consumption declines 80 per cent by 2050; all new cars
and airplanes are eﬃcient from 2025; the world shifts to the most eﬃcient technologies for steel
and cement production, etc. Even with these highly optimistic assumptions in place, they ﬁnd that
the pressures of continued growth drive emissions to exceed the carbon budgets for 1.5°C and
2°C, without negative emissions technologies.
Another way to approach this question is by looking at projected rates of decoupling. If we assume
global GDP continues to grow at 3 per cent per year (the average from 2010 to 2014), then decou-
pling must occur at a rate of 10.5 per cent per year for 1.5°C, or 7.3 per cent per year for 2°C. If global
GDP grows at 2.1 per cent per year (as PWC predicts), then decoupling must occur at 9.6 per cent per
year for 1.5°C, or 6.4 per cent per year for 2°C. All of these targets are beyond what existing empirical
models indicate is feasible. The Schandl et al model indicates that decoupling can happen by at most
3 per cent per year under optimistic conditions. Other models arrive at similar conclusions. Before
adopting BECCS assumptions, the IPCC (2000) projected decoupling of 3.3 per cent per year in a
global best-case scenario. The C-ROADS tool (developed by Climate Interactive and MIT Sloan) pro-
jects decoupling of at most 4 per cent per year under the most aggressive possible abatement pol-
icies: high subsidies for renewables and nuclear power, plus high taxes on oil, gas and coal. All of
these results fall short of the decoupling rate that must be achieved if the global economy continues
to grow at expected rates. Holz et al. (2018)ﬁnd that if we rule out widespread use of negative emis-
sions technologies, the required rate of decarbonisation for meeting the Paris Agreement is ‘well
outside what is currently deemed achievable, based on historical evidence and standard modelling.’
The challenge is even more diﬃcult for rich nations. Anderson and Bows (2011) have modelled the
emissions reductions necessary for achieving a 50 per cent chance of staying under 2°C (more relaxed
than the two-thirds chance that the UNFCC calls for), without BECCS. They proceed from the principle
of ‘common but diﬀerentiated responsibility’, whereby rich nations (Annex-1 nations) make more
aggressive emissions reductions than poor nations, owing to their greater historical responsibility
for emissions and their greater capacity for managing the costs of transition. They assume that
Non-Annex 1 nations defer peak emissions until 2025, and thereafter reduce emissions by 7 per
cent per year. They acknowledge that these are extremely ambitious assumptions but consider
them to be the most feasible compromise between practicality and equity. To stay within the remain-
ing carbon budget, Annex 1 nations need to reduce emissions by 8–10 per cent per year, beginning in
2015. This model was developed with data up to 2010; as the remaining carbon budget is now
smaller, Anderson estimates that Annex 1 nations need to reduce emissions by 12 per cent per year.
If we accept that Annex 1 nations need to achieve emissions reductions of 12 per cent per year,
and if we assume that GDP growth in Annex 1 nations continues at 1.86 per cent per year (the
NEW POLITICAL ECONOMY 11
average from 2010 to 2014), then decoupling must occur at a rate of 15.8 per cent per year.
perspective, this is eight times faster than the historic rate of decoupling in Annex 1 nations (viz.,
1.9 per cent per year from 1970 to 2013), and it is important to bear in mind that the rate of decou-
pling has generally slowed over this period.
It also exceeds the decoupling rate implied by the
average G20 Nationally Determined Contributions under the Paris Agreement (viz., 3 per cent per
year) by a factor of ﬁve.
There is one empirical model that feasibly accomplishes emissions reductions consistent with the
Paris Agreement, without relying on negative emissions technologies. Published by Grubler et al.
(2018), it was included in the IPCC Special Report on 1.5°C (2018) in response to growing critiques
of the IPCC’s reliance on BECCS. The scenario, known as ‘Low Energy Demand’(LED), accomplishes
emissions reductions compatible with 1.5°C by reducing global energy demand by 40 per cent by
2050. In addition to decarbonisation and aﬀorestation, the key feature of this scenario is that
global material production and consumption declines signiﬁcantly: ‘The aggregate total material
output decreases by close to 20 per cent from today, one-third due to dematerialization, and two-
thirds due to improvements in material eﬃciency.’Dematerialisation is accomplished by shifting
away from private ownership of key commodities (like cars) towards sharing-based models. LED
diﬀerentiates between the global North and South. Industrial activity declines by 42 per cent in
the North and 12 per cent in the South. With eﬃciency improvements, this translates into industrial
energy demand declining by 57 per cent in the North and 23 per cent in the South.
The LED scenario projects continued GDP growth at just over 2 per cent per year, which would
make it consistent with green growth theory. However, the empirical basis for this GDP trend is
not robust. It is derived from the MESSAGE-Globium model, which calculates GDP from only two
inputs: labour supply (population size and productivity) and energy. The low energy demand in
the LED scenario does not aﬀect growth because it is oﬀset by eﬃciency improvements. As the
model is insensitive to changes in material throughput, reductions in production and consumption
do not aﬀect output. The paper oﬀers no evidence that GDP will continue to grow despite such
reductions. Charlie Wilson, one of the paper’s authors, acknowledged that ‘we did not consider
broader questions of GDP growth or degrowth, and we did not explicitly report relationships
between our scenario and GDP outcomes for this reason.’
Conclusions and discussion
The empirical data demonstrate that while absolute decoupling of GDP from emissions is possible
and is already happening in some regions, it is unlikely to happen fast enough to respect the
carbon budgets for 1.5°C and 2°C against a background of continued economic growth. Growth
increases energy demand, making the transition to renewable energy more diﬃcult, and increases
emissions from land use change and industrial processes. Models that do project green growth
within the constraints of the Paris Agreement rely heavily on negative emissions technologies that
are either unproven or dangerous at scale. Without these technologies, the rates of decarbonisation
required for 1.5°C or 2°C are signiﬁcantly steeper than extant models suggest is feasible even with
aggressive mitigation policies.
This conclusion changes somewhat if we adjust the baseline growth rate. All of the studies cited
above project global GDP growth at 2–3 per cent per year. A lower rate of growth requires a lower
rate of decarbonisation. A growth rate of 0 per cent requires decarbonisation of 6.8 per cent per year
(for 1.5°C) and 4 per cent per year (for 2°C). There is no empirical evidence that 6.8 per cent can be
achieved on a global scale, but 4 per cent is nearly within reach. In other words, it is empirically feas-
ible to achieve green growth within a carbon budget for 2°C with the most aggressive possible miti-
gation policies if the growth rate is very close to zero and if mitigation starts immediately. This
conclusion is in line with research by Schroder and Storm (2018), which ﬁnds that reducing emissions
in line with the 2°C target is feasible (under optimistic assumptions) only if global economic growth is
12 J. HICKEL AND G. KALLIS
less than 0.45 per cent per year. This conclusion does not hold for 1.5°C, however; emissions
reductions in line with 1.5°C are not empirically feasible except in a de-growth scenario.
As John O’Neill (2017) writes, whereas
it is logically possible to have increasing GDP and a decreasing physical and energy throughput in an economy …
it is a fallacy to move from claims about what is logically possible to claims about what is physically possible and
another from what is physically possible to what is empirically actual.
Green growth, we have shown, is not empirically actual –but is it possible in theory?
This question is often approached in terms of the IPAT equation (Environmental Impact = Popu-
lation * Aﬄuence * Technology), which says that the impact of an economy (e.g. tons of C per
capita) is equal to the scale of the economy (GDP per capita) times its eﬃciency (e.g. GDP per tons
of carbon). Eﬃciency is in principle determined by technology and policy and there is no a priori
reason why it cannot increase faster than scale, or even as fast as necessary to reduce impact to a
sustainable level. Furthermore, insofar as GDP measures what people are willing to pay for things,
as opposed to the amount of energy and resources people consume, there is no reason why the
economy cannot in theory grow using progressively less energy and resources: peoples’preferences
may shift to goods and services with ever-lower energy and material requirements. One may con-
clude then that absolute decoupling should theoretically be possible –and in fact this is precisely
the reason that advocates of green growth are not deterred by claims that it has not happened
yet and does not seem likely to happen in the future. They attribute this to lack of eﬀort.
Ward et al.’s(2016) study provides perhaps the most compelling counter-argument to this claim.
As there is a thermodynamically deﬁned maximum of eﬃciency, indeﬁnite growth will sooner or later
lead to increase in resource and energy use. Any absolute reductions due to substitution or eﬃciency
will at best be temporary. Imagine a hypothetical economy powered by the sun, with a steady supply
of food and necessities from renewable sources where goods are reused and materials recycled. In
the transition to such an economy, resource use will decline. But even such an economy will still
have some minimal requirement for material inputs, land, etc –so after the transition takes place
then any further growth in this economy will lead to a growth in resource use. Given that compound
growth quickly turns to inﬁnity, so too will resource use and impact.
One may respond by arguing that we are still far from reaching limits in eﬃciency and substitution.
We cannot rule out substitutions or technological breakthroughs that will push such limits so far into
the future as to render them irrelevant (e.g. nuclear fusion, 100 per cent recycling of materials fuelled
by fusion or solar power, etc). Plus, the economy still has signiﬁcant room for structural change
towards less resource intensive services. In other words –the argument might go –maybe green
growth is not sustainable indeﬁnitely, but it can nonetheless happen now and can be sustained
for a time horizon relevant for our civilisation (although note that Ward et al indicate that the
limits of resource eﬃciency may be reached by 2050).
So let us assume that green growth is theoretically possible in the short to medium term. Still, we
must ask if there is a fundamental, as opposed to historically contingent reason why it has not hap-
pened yet. Is there some underlying reason why throughput and output are so tightly coupled in the
It is worth noting that the IPAT model gives the impression that A and T, or scale and eﬃciency, are
independent factors, when in fact they aﬀect one another (Ekins 2012). But note that IPAT is a tau-
tology, true by deﬁnition of the quantities involved, and should not be confused with a causal
model. Furthermore, P, A and T are not independent from one another. We know for example
from basic growth economics that technological development (T) causes economic growth and
growth in consumption (A). Ecological economists have also shown that the more eﬃciently an
economy uses resources, the more it grows, and the more resources it ends up consuming –the
NEW POLITICAL ECONOMY 13
so-called Jevons’paradox (Polimeni et al.2008). This is not just a matter of rebounds eating eﬃciency
gains at the micro-level –it refers to a more fundamental macro-mechanism through which industrial
economies grow by using resources more productively. For example, when technology improves
labour productivity, we expect that this will lead to more growth and more jobs as the relative
cost of labour declines –why some expect this to work diﬀerently in relation to resources is not
clear (Kallis 2018).
Another fundamental reason why eﬃciency might be coupled with scale is that as we know from
biology and ecology, the metabolism of a larger organism, say an elephant, is more eﬃcient than that
of a smaller one, say a mouse, but this is because the elephant is bigger (Polimeni et al.2008). It is true
that relative resource or energy decoupling often accompanies the growth of an economy –but this
might simply be an artifact of scale. And it does not follow that more and more relative decoupling
will amount to absolute decoupling. The U.S. economy, like an elephant, could not be so much bigger
than others were it not also more eﬃcient, and it is big because it is eﬃcient –but this doesn’t mean
that by getting bigger and bigger it will burn less energy, just as an elephant does not burn fewer
calories than a mouse. All this does not amount to a theoretical refutation of absolute decoupling,
but it shows that there might exist a more fundamental mechanism that links the scale of an
economy to its throughput that is worth exploring.
That said, one might argue that unlike the scale of an animal, the scale of the economy (i.e. GDP) is
a measure of value, not of physical size, and it can therefore grow without limit even while resource
and energy throughput diminishes. GDP, one might argue, merely measures what people are willing
to pay for, which is not necessarily connected to the use of resources and energy.
Can value grow independently of throughput? This begs for a clear theory of value. Unfortu-
nately, the green growth literature provides no such theory. There are two general possibilities
that we might consider. (1) The neoclassical theory of value, whereby value represents utility
(how useful we ﬁnd goods), which is revealed in prices (how much we are willing to pay for
them). In this schema, GDP is the amount of valuable goods and services bought and sold, multi-
plied by their value. To the extent that the green growth literature considers GDP to be a proxy for
total value, we can assume that it accepts this neoclassical theory of value. (2) The labour or energy
theories of value, which claim that value is ultimately determined by the work or energy that goes
into production, hinting at a more fundamental coupling between value and throughput (Kallis
2018). From this perspective, value cannot grow without more human labour or energy put into
Neither the neo-classical nor the labour or energy theories of value have been empirically proven;
in other words, they cannot accurately predict the price at which goods trade. It is impossible to cal-
culate the total labour or energy that has gone into the production of a good, or the utility it provides.
Indeed, no one has ever independently measured utility to test whether it correlates with prices or
willingness to pay (Sagoﬀ2008). We therefore do not have a theory of value that allows us to deter-
mine whether value can be absolutely decoupled from throughput. Of course, one might say there is
a third way: we can think of value as the sum of all the ‘values’people hold. There is of course no
reason why the things a society values cannot increase while throughput decreases. There are two
problems with this approach, however. First, if values are incommensurable, it is impossible to aggre-
gate them and determine whether total value is growing or not. Second, one can imagine a society
that values the quality of the natural environment above all else; such a value could of course grow
while throughput decreases, but to call such a scenario ‘green growth’is to stretch the meaning of
the term beyond relevance.
In sum, it cannot be proven that green growth of value is theoretically possible, unless we accept a
framework that makes it by deﬁnition possible –a framework that assumes that value and output are
determined by some undeﬁned, limitless quality called utility that is uncoupled from the physical
world. Conversely, though, and by the same token, it cannot be proven either that green growth
is theoretically impossible, at least not as long as ultimate limits in eﬃciency and substitution have
not been reached. As a result, our only reliable guide to the green growth/decoupling question
14 J. HICKEL AND G. KALLIS
must be empirical. And, as we have demonstrated, existing empirical studies demonstrate that green
growth is at best highly unlikely. One may insist that green growth hasn’t occurred because it has not
been tried, the fact that it hasn’t been empirically observed till now then becoming irrelevant. We
follow instead a more precautionary approach and argue that policy should be made on the basis
of robust empirical evidence, rather than on the basis of speculative theoretical possibilities, particu-
larly given the severity of the crisis that is at stake.
This review ﬁnds that extant empirical evidence does not support the theory of green growth. This
is clear in two key registers. (1) Green growth requires that we achieve permanent absolute decou-
pling of resource use from GDP. Empirical projections show no absolute decoupling at a global
scale, even under highly optimistic conditions. While some models show that absolute decoupling
may be achieved in high-income nations under highly optimistic conditions, they indicate that it is
not possible to sustain this trajectory in the long term. (2) Green growth also requires that we
achieve permanent absolute decoupling of carbon emissions from GDP, and at a rate rapid
enough to prevent us from exceeding the carbon budget for 1.5°C or 2°C. While absolute decou-
pling is possible at both national and global scales (and indeed has already been achieved in
some regions), and while it is technically possible to decouple in line with the carbon budget for
1.5°C or 2°C, empirical projections show that this is unlikely to be achieved, even under highly opti-
The empirical evidence opens up questions about the legitimacy of World Bank and OECD eﬀorts
to promote green growth as a route out of ecological emergency, and suggests that any policy pro-
grammes that rely on green growth assumptions –such as the Sustainable Development Goals –
need urgently to be revisited. That green growth remains a theoretical possibility is no reason to
design policy around it when the facts are pointing in the opposite direction.
Of course, we need all of the technological innovations we can get, and we need to gear govern-
ment policy toward driving these innovations, but this will not be enough in and of itself. The evi-
dence presented above indicates that in order for eﬃciency gains to be eﬀective, we will need to
scale down aggregate economic activity too. It is more plausible that we will be able to achieve
the necessary reductions in resource use and emissions without growth than with growth. Indeed,
there are no scientiﬁc grounds upon which we should not question growth, if our goal is to avoid
dangerous climate change and ecological breakdown. Staying within planetary boundaries may
require a de-growth of production and consumption in high-consuming nations (Victor 2008, Alier
2009, Jackson 2009, Kallis 2011, Kallis et al. 2012), and a shift away from the narrow growth-
focused development agenda in the global South. As Gough (2017) notes, combatting climate
change might require not only new clean and eﬃcient energy technologies, but also a reduction
and re-composition of consumption, with a shift from carbon-intensive to low or zero carbon
sectors. Legislative limits, green taxes, shifts in public investment and working hour-reductions or
new social security institutions such as a basic income all have a role to play in such a transition
(Gough 2017, Kallis 2018). The objective could be to ﬁnd ways to decouple prosperity and develop-
ment from growth (e.g. Jackson, 2009,O’Neill et al.2018) rather than to continue to chase the
phantom of green growth.
It seems likely that the insistence on green growth is politically motivated. The assumption is that
it is not politically acceptable to question economic growth and that no nation would voluntary limit
growth in the name of the climate or environment; therefore green growth must be true, since the
alternative is disaster. But it might well be the case that, as Wackernagel and Rees (1998) put it, ‘the
politically acceptable is ecologically disastrous while the ecologically necessary is politically imposs-
ible’. As scientists we should not let political expediency shape our view of facts. We should assess the
facts and then draw conclusions, rather than start with palatable conclusions and ignore inconveni-
NEW POLITICAL ECONOMY 15
1. Steﬀen et al.(2015) have identiﬁed biosphere integrity and climate change as the core planetary boundaries mer-
iting most concern.
2. Wiedmann et al.(2015) come up with a similar ﬁgure, 70 billion metric tons in 2008.
3. This trend was driven primarily by growth in industrial and construction materials, primarily in Asia. It is not clear,
however, how much of this material use has been consumed domestically and how much has been exported for
4. The UNEP model suggests that decoupling can be achieved at a max rate of 1 per cent per year. Therefore GDP
growth would have to be less than 1 per cent per year in order for resource use to be reduced.
5. Even while CO
emissions had plateaued, methane emissions were growing, by more than 30 per cent between
2002 and 2014 (Turner et al.2016).
6. The trend looks somewhat more promising if we use PPP dollars instead of constant USD, but PPP calculations are
unreliable and tend to overstate the purchasing power of poor countries.
7. ‘Climate Scoreboard’, Climate Interactive.
8. Another 9 scenarios include some BECCS, but not to the point of achieving negative emissions.
9. PWC Low Carbon Economy Index 2017.
10. 150 GW were installed in 2017; the IRENA scenario requires that 350 GW be installed per year on average to 2050.
This is feasible with existing growth rates (from 2016 to 2018 solar and wind capacity grew by 8 per cent per year),
but IRENA do not specify the trajectory necessary for 2°C. Jacobson and Delucchi (2011) indicate that 700 GW
need to be added per year to 2030–4.6 times the existing rate. This requires a growth rate of 25 per cent per
year on existing rates.
11. Global energy intensity improved by 1.3 per cent per year from 2000 to 2010, and 1.8 per cent per year from 2010
12. Jacobson and Delucchi (2011) claim that global energy demand will decline by 36 per cent (relative to business as
usual by 2050) as fossil fuels are replaced by wind and solar, which means that demand in 2050 will be less than
demand in 2012.
13. This is the ﬁgure that Anderson used in various public talks in 2018. In 2019 he conﬁrmed a range of 10–15 per
cent per year, inpersonal correspondence.
14. Using the equation: Rate of necessary decoupling = GDP growth rate/(1 –Rate of necessary emissions reductions).
15. Decoupling slowed from an average of 2.3 per cent per year in the ﬁrst half of the period to an average of 1.6 per
cent in the second half, according to the World Bank, Databank, CO
emissions (kg per 2010 US$ GDP).
16. Personal correspondence, 2018. Also, it is worth noting that Grubler et al state that LED does not incorporate
rebound eﬀects; they acknowledge that this is a relevant shortcoming of the work.
17. For a detailed discussion of this question, see Ekins (2012).
Kallis’s research beneﬁted from support from the Spanish Ministry of Economy and Competitiveness (MINECO) under the
“María de Maeztu”Unit of Excellence (MDM-2015-0552) and the COSMOS (CSO2017-88212-R) grant.
No potential conﬂict of interest was reported by the authors.
Notes on Contributors
Jason Hickel is an anthropologist at Goldsmiths, University of London, and a Fellow of the Royal Society of Arts. He writes
on global inequality, political economy and ecology.
Giorgos Kallis is an ICREA professor at the Institute of Environmental Sciences and Technology at the Autononomous
University of Barcelona, an ecological economist and political ecologist writing on limits to growth.
Alexander, S., Rutherford, J., and Floyd, J., 2018. A critique of the Australian national outlook decoupling strategy: a ‘limits
to growth’perspective. Ecological economics, 145, 10–17.
Alier, J.M., 2009. Socially sustainable economic de-growth. Development and change, 40 (6), 1099–1119.
Anderson, K., and Bows, A., 2011. Beyond ‘dangerous’climate change: emission scenarios for a new world. Philosophical
transactions of the royal society of London a: mathematical, physical and engineering sciences, 369 (1934), 20–44.
16 J. HICKEL AND G. KALLIS
Anderson, K., and Peters, G., 2016. The trouble with negative emissions. Science, 354 (6309), 182–183.
Australian Bureau of Agricultural and Resource Economics (ABARE), 2008.Energy in Australia. Canberra: ABARE.
Ayres, R.U., and Simonis, U.E., 1993.Industrial metabolism: restructuring for sustainable development. Tokyo, New York: UN
Bringezu, S., 2015. Possible target corridor for sustainable use of global material resources. Resources,4,25–54.
Ceballos, G., et al., 2015. Accelerated modern human–induced species losses: entering the sixth mass extinction. Science
advances, 1 (5), e1400253.
Ceballos, G., Ehrlich, P.R., and Dirzo, R., 2017. Biological annihilation via the ongoing sixth mass extinction signaled by
vertebrate population losses and declines. Proceedings of the national academy of sciences, 114 (30), E6089–E6096.
Dale, G., Mathai, M.V., and de Oliveira, J.A.P., eds., 2016.Green growth: ideology, political economy and the alternatives.
London: Zed Books Ltd.
Dasgupta, S., et al., 2002. Confronting the environmental Kuznets curve. Journal of economic perspectives, 16 (1), 147–168.
De Coninck, H., and Benson, S.M., 2014. Carbon dioxide capture and storage: issues and prospects. Annual review of
environment and resources, 39, 243–270.
Dittrich, M., et al.,2012.Green economies around the world: implications of resource use for development and the environ-
ment. Vienna: SERI.
Ekins, P., 2012. Sustainable growth revisited: technology, economics and policy. Mineral economics,24(2–3), 59–77.
European Academies Science Advisory Council. 2018. Negative emission technologies: what role in meeting Paris agree-
ment targets? EASAC Policy Report 35.
Fuss, S., et al., 2014. Betting on negative emissions. Nature climate change, 4 (10), 850–853.
Giljum, S., et al., 2014. Global patterns of material ﬂows and their socio-economic and environmental implications: a MFA
study on all countries world-wide from 1980 to 2009. Resources, 3 (1), 319–339.
Global Carbon Budget. 2018.Global carbon project. Available from: http://www.globalcarbonproject.org/carbonbudget/
[Accessed 15 Dec 2018].
Global CCS Institute, 2015.Global status of CCS 2015: summary report. Melbourne.
Goodall, C. 2011.‘Peak Stuﬀ’: Did the UK reach a maximum use of material resources in the early part of the last decade. A
research paper for Carbon Commentary, 13.
Gough, I., 2017.Heat, greed and human need: climate change, capitalism and sustainable wellbeing. Cheltenham: Edward
Grossman, G.M. and Krueger, A.B., 1995. Economic growth and the environment. The quarterly journal of economics, 110
Grubler, A., et al., 2018. A low energy demand scenario for meeting the 1.5C target and sustainable development goals
without negative emissions technologies. Nature energy, 3, 515–527.
Gutowski, T., Cooper, D., and Sahni, S., 2017. Why we use more materials. Philosophical transactions of the royal society a:
mathematical, physical and engineering sciences, 375, 20160368.
Haas, W., et al., 2015. How circular is the global economy? An assessment of material ﬂows, waste production, and recy-
cling in the European Union and the world in 2005. Journal of industrial ecology, 19 (5), 765–777.
Hatﬁeld-Dodds, S., et al., 2015. Australia is ‘free to choose’economic growth and falling environmental pressures. Nature,
527 (7576), 49–53.
Heck, V., et al., 2018. Biomass-based negative emissions diﬃcult to reconcile with planetary boundaries. Nature climate
change, 8, 151–155.
Herring, H. and Sorrell, S., 2009.Energy eﬃciency and sustainable consumption. Hampshire: The Rebound Eﬀect.
Hickman, L. 2016. The history of BECCS. Carbon Brief.https://www.carbonbrief.org/beccs-the-story-of-climate-changes-
Hoekstra, A.Y., and Wiedmann, T.O., 2014. Humanity’s unsustainable environmental footprint. Science, 344, 1114–1117.
Holz, C., et al., 2018. Ratcheting ambition to limit warming to 1.5 C–trade-oﬀs between emission reductions and carbon
dioxide removal. Environmental research letters, 13 (6), 064028.
International Energy Agency, October 2016. Decoupling of global emissions and economic growth conﬁrmed. https://
IPCC, 2000. Special report on emissions scenarios.
IPCC, 2014. Climate change 2014 synthesis report –summary for policymakers.
IPCC, 2018.Global warming of 1.5C –summary for policymakers. Geneva: IPCC.
IRENA, 2018.Global energy transformation: a roadmap to 2050. Abu Dhabi: International Renewable Energy Agency.
Jackson, T. 2009. Prosperity without growth: the transition to a sustainable economy.
Jacobs, M., 2013. Green growth. In: Robert Falkner, ed. The Handbook of global climate and environment policy. Chichester,
UK: Wiley-Blackwell, 197–214.
Jacobson, M.Z., and Delucchi, M., 2011. Providing all global energy with wind, water, and solar power, part i: technologies,
energy resources, quantities and areas of infrastructure, and materials’.Energy policy, 39 (3), 1154–1169.
Kallis, G., 2011. In defence of degrowth. Ecological economics, 70 (5), 873–880.
Kallis, G., 2017. Radical dematerialization and degrowth. Philosophical transactions of the royal society a: mathematical,
physical and engineering sciences, 375 (2095), 20160383.
NEW POLITICAL ECONOMY 17
Kallis, G., 2018.Degrowth. Newcastle-upon-Tyne: Agenda Publishing.
Kallis, G., Kerschner, C., and Martinez-Alier, J., 2012. The economics of degrowth. Ecological economics, 84, 172–180.
Keith, D.W., 2001. Sinks, energy crops and land use: coherent climate policy demands an integrated analysis of biomass.
Climatic change, 49 (1), 1–10.
Krausmann, F., et al., 2009. Growth in global materials use, GDP and population during the 20th century. Ecological econ-
omics, 68 (10), 2696–2705.
Kuhnhenn, K., 2018.Economic growth in mitigation scenarios: a blind spot in climate science. Berlin: Heinrich Boll
Larkin, A., et al., 2017. What if negative emissions technologies fail at scale? Climate policy, 18, 690–714.
Meyer, R. April 2016. Not doomed yet: the biggest political-economy news this millennium. The Atlantic. Available from:
Obersteiner, M., et al.,2001.Managing climate risk. Laxenburg: International Institute for Applied Systems Analysis.
O’Neill, J., 2017. Happiness, austerity and inequality. In: H. Rosa, ed. Good life beyond growth: critical perspectives.
Abingdon: Routledge, 141–152.
O’Neill, D.W., et al., 2018. A good life for all within planetary boundaries. Nature sustainability, 1 (2), 88–95.
Organisation for Economic Cooperation and Development (OECD), 2011.Towards green growth. Paris: OECD.
Organisation for Economic Cooperation and Development (OECD), 2017. Green Growth Indicators 2017.
Pearce, F., 2012. Peak planet: are we starting to consume less? New scientist, 214, 38–43.
Peters, G., 2017. Does the carbon budget mean the end of fossil fuels? Climate News. Available from: https://www.cicero.
Polimeni, J.M., et al., 2008.The Jevons paradox and the myth of resource eﬃciency improvements. London: Earthscan.
Rockström, J., et al., 2009. Planetary boundaries: exploring the safe operating space for humanity. Ecology and society,14
Sagoﬀ, M., 2008. On the economic value of Ecosystem services. Environmental values, 17 (2), 239–257.
Schandl, H., et al., 2016. Decoupling global environmental pressure and economic growth: scenarios for energy use,
materials use and carbon emissions. Journal of cleaner production, 132 (2016), 45–56.
Schroder, E., and Storm, S. 2018.Economic growth and carbon emissions: the road to ‘hothouse earth’is paved with good
intentions. Institute for New Economic Thinking, Working Paper 84.
Smith, P., et al., 2016. Biophysical and economic limits to negative CO
emissions. Nature climate change, 6 (1), 42–50.
Smulders, S., Toman, M., and Withagen, C. 2014. Growth theory and ‘green growth’.Oxford review of economic policy,30
Solow, R.M., 1973. Is the end of the world at hand? Challenge, 16, 39–50.
Steﬀen, W., et al., 2015. Planetary boundaries: Guiding human development on a changing planet. Science, 347 (6223),
Turner, A.J., et al., 2016. A large increase in U.S. methane emissions over the past decade inferred from satellite data and
surface observations. Geophysical research letters, 43 (5), 2218–2224.
UNEP, 2014.Managing and conserving the natural resource base for sustained economic and social development. Nairobi:
United Nations Environment Programme.
United Nations, 2012. The world we want.
United Nations Environment Programme (UNEP), 2011.Towards a green economy: pathways to sustainable development
and poverty eradication –a synthesis for policy makers. Nairobi: UNEP.
United Nations Environment Programme (UNEP), 2017a.Assessing global resource use. Nairobi: UNEP.
United Nations Environment Programme (UNEP), 2017b.Resource eﬃciency: potential and economic implications. A report
from the International Resource Panel.
Van Vuuren, D.P., et al., 2017. Open discussion of negative emissions is urgently needed. Nature energy, 2, 902–904.
Van Vuuren, D.P., et al., 2018. Alternative pathways to the 1.5C target reduce the need for negative emission technologies.
Nature climate change, 8, 391–397.
Vaughan, N.E., and Gough, C., 2016. Expert assessment concludes negative emissions scenarios may not deliver.
Environmental research letters, 11, 095003.
Victor, P., 2008.Managing without growth: slower by design, not disaster. Cheltenham: Edward Elgar Publishing.
Voet, E., Oers, L., and Nikolic, I., 2004. Dematerialization: not just a matter of weight. Journal of industrial ecology, 8 (4),
Wackernagel, M. and Rees, W., 1998.Our ecological footprint: reducing human impact on the earth (No. 9). Gabriola Island:
New Society Publishers.
Ward, J.D., et al., 2016. Is decoupling GDP growth from environmental impact possible? Plos one, 11 (10), e0164733.
Weizsäcker, E.U., Lovins, A.B., and Lovins, L.H., 1998.Factor four: doubling wealth, halving resource use. Club of Rome.
Wiedmann, T.O., et al., 2015. The material footprint of nations. Proceedings of the national academy of sciences, 112 (20),
World Bank, 2012.Inclusive green growth: the Pathway to sustainable development. Washington, DC: World Bank.
18 J. HICKEL AND G. KALLIS