Seeing environmental injustices: the mechanics, devices and assumptions of environmental sustainability indices and indicators

At the heart of any colonization project, and therefore any move to de-colonize, are ways of seeing nature and society, that then allow particular ways of governing each. This is plainly visible in a number of tools that exist to measure progress towards (or regress from) environmental sustainability. The tools use indices and indicators constructed mostly by environmental scientists and ecologists. As such, they are not neutral scientific instruments: they reflect the worldviews of their creators. These worldviews depend on three dimensions: the values they prioritize, the explanatory theories they use and the futures they envision. Through these means different tools produce conflicting notions of the sustainability of our economies and societies. In this article, we shed light onto the theoretical and epistemological assumptions that lie behind key international sustainability indices and indicators: the Environmental Performance Index,Domestic Material Consumption, Material Intensity, the Material Footprint, the Carbon Footprint, the Ecological Footprint and CO2 emissions (territorial). The variables included in these indices, the way they are measured, aggregated and weighted all imply a particular way of understanding the relationships between economy, society and environment. This divergence is most clearly visible in the fact that some indices are negatively correlated with each other. Where one index might plot growing environmental sustainability, another shows its decline. Our results highlight that those devices and the theories informing them are particularly interesting for way how colonialism is materialized. Some of these measurements hide the material roots of prosperity and the ecological (and economic) distributional conflicts exported to the poorer countries by the global North, and others show how its production and consumption levels are reliant upon a socio-ecological 'subsidy' imposed on Southern countries. These subsidies represent injustices that present a primafacie case for decolonizing indices and indicators of environmental governance.


Introduction
A considerable number of indicators can be used to track different aspects of environmental change, degradation, and their amelioration across countries and over time. With so many indicators to consider, there is an understandable desire to reduce the number of variables monitored. One aggregate score is easier to compare than many separate numbers, so there have been diverse efforts to turn the multi-dimensional nature of environmental change into single composite indices. These are used in many situations to capture and simplify complex change.
According to Ravallion (2010), two broad types of composite indices of can be identified. In the first, the choices of the component series and the aggregation function are informed and constrained by a body of theory and practice from the literature. This is not the case for the second type, where the analyst identifies a set of indicators that are assumed to reflect various dimensions of a particular concept. An aggregate index is then constructed at the country level, usually after re-scaling or ranking the component series. Therefore, neither the component series "nor the aggregation function is pre-determined from theory and practice but are 'moving parts' of the index -key decision variables that the analyst is free to choose, largely unconstrained by economic or other theories intended to inform measurement practice" (Ravallion 2010: 3).
Ravallion is dismissive of the second type, calling them 'mash-up' indices. However, while their composition and construction may be arbitrary, there are more theoretical assumptions that lie behind them than Ravallion recognizes. Indeed, it is precisely in the composition and construction of mash-up indices that we can read the unstated, unwritten theories and politics that inform them. These indices are not just a means of aggregating facts about the world; they are also lenses into the minds of the aggregators, and the world they want to create. The plethora of indices and indicators of environmental performance and change are particularly interesting for the way in which they encourage us to view the world, and the unwritten theories and politics that inform these views. In other words, they embed in indices different ways in which environmental problems are framed (Lele et al. 2018). These framings depend in turn on three dimensions: the values they prioritize, the explanatory theories they use, and the futures they envision (ibid.). To understand and critically engage with environmental measurements it is necessary to remember the motivations and epistemologies that shape their construction (Thatcher et al. 2016).
Environmental studies are laden with values (Lele and Norgaard 1996). Environmental changes become "problems" only because some group of people in society consider them so. Environmental change is not just a process affecting physical phenomena: it is also human interests, priorities, hopes and desires that attribute negative or positive value to such changes (Lele et al. 2018). In the same way, any aims that are set, such as green growth or sustainable development, as well as the criteria and indicators that will be used to measure progress toward them, are value-laden. Values determine which environmental processes are seen as problems, in what sense and context, and shape the solutions that should be implemented in the future. In these values are found the theoretical influence that Ravallion missed.

The nature and origin of the different measures of environmental sustainability
We have selected seven of the most widely used global sustainability indicators and indices. A brief introduction to them can be found in Table 1. We summarize the essentials below.
Domestic Material Consumption and Material Footprint are the total weight of raw material (biomass, minerals, metals and fossil fuels) used in a specific economy. Domestic Material Consumption does not include the material impact involved in the production and transport of imported goods (Wiedmann et al. 2015). Nor does Domestic Material Consumption take into account the materials needed along the supply chains of traded goods. Material Footprint, in contrast, accounts for the amount of all materials required along the supply chains of all goods and services finally consumed in a country. Therefore, Material Footprint is better at capturing the broader role in the global economy played by any one country since it is more comprehensive than Domestic Material Consumption. Both these indices are presented in per capita terms. A high score for both indicators implies poor environmental performance.
Material Intensity is the ratio of Domestic Material Consumption (in kilograms) to Gross Domestic Product (GDP) at constant prices (e.g. 2010 US$). This indicator is normally used to express the level of efficiency in the use of raw materials. High Material Intensity scores indicate poor environmental performance, because this means that a lot of material has to be used to produce any given unit of GDP, whereas low scores indicate more efficient use.

Indices and Indicators
Acronym What is measured  The Ecological Footprint is measured using a standardized area unit equivalent to a 'global hectare' (gha) 2 and global hectares per capita (gha/cap). The Ecological Footprint of each country is the sum of six components (cropland, forestland, fishing grounds, grazing land, the built-up area, and carbon footprint). The higher the Ecological Footprint, the worst its environmental performance. Consumption-based accounting of carbon emissions (also called the Carbon Footprint) measures emissions associated with imported and exported goods and materials.
The Carbon Footprint reports the total emissions associated with each country and can be calculated using a global multi-region input-output (MRIO) model which traces global supply chains. A Carbon Footprint is measured using carbon dioxide equivalent (CO2-eq). 3 Emissions physically occurring in a country are its territorial emissions (henceforth, CO2 emissions). This is sometimes called production-based accounting. This is the standard reporting of GHG emissions as reported by CDIAC, IEA, the JRC EDGAR database, UNFCCC, and others. Carbon Footprint and CO2 emissions are presented in per capita terms. A high score for both indicators implies a poor environmental performance.
The Environmental Performance Index is based upon two policy objectives: Environmental Health, which measures threats to human health, and Ecosystem Vitality, which measures natural resources and ecosystem services. A country's EPI score is ranked from 0 to 100. High scores indicate good environmental performance. However, to be consistent with the other indicators' convention of higher values representing poor environmental performance, we have inverted the scores for the purpose of this analysis.
More details about these indices and indicators and their calculation are given in the supplementary document attached to this article. Here it should be highlighted that there are obvious epistemological differences between these indices and indicators. Five of the indicators are quite focused on environmental aspects (Material, Carbon and Ecological Footprints and Domestic Material Consumption and CO2 emissions), whereas the Material Intensity includes a very strong economic element (the GDP denominator) and the Environmental Performance Indicator brings in a strong element of social desirability.

Comparing the different measures of environmental sustainability with each other, and with measures of economic growth
It is reasonable to expect that composite measures of environmental sustainability should generally agree with each other, or, at the very least, not contradict each other. They may have different emphases and constructions, but the thing that they are measuring, sustainability, the impact of people on the environment, and its multiple facets, should be producing a similar sort of signal. This is not the case. Rather we have two groups of measures. There is agreement within each group, and disagreement between them. Large Material or Carbon footprints are positively related to large Ecological footprints and both to Domestic Material Consumption and CO2 Emissions (Table 2, Figure 1e and Figure 1f ). But the measures of material consumption and measures of environmental performance are inversely correlated ( Figure 1a). They oppose each other. Large footprints (either material, carbon or ecological) tend to coincide with good performance as shown by Environmental Performance Index values. This means that, paradoxically, having a high material, carbon or ecological impact is good for the environment according to this Index. On the other hand, having low footprints -either material, carbon or ecological -seems to be bad for the environment according to this Index. This is particularly true for low-income countries. Figure 1 plots indicators and indices tons per hectare) from which it was harvested, or where its waste material was absorbed. The number of hectares that result from this calculation are then converted to global hectares using yield and equivalence factors (see the supplementary document). The sum of the global hectares needed to support a person is that person's total Ecological Footprint. According to the Ecological Footprint Glossary, a global hectare is a biologically productive hectare with world average biological productivity for a given year. Global hectares are needed because different land types have different productivities. A global hectare of, for example, cropland, would occupy a smaller physical area than the much less biologically productive pastureland, as more pasture would be needed to provide the same biocapacity as one hectare of cropland. Because world productivity varies slightly from year to year, the value of a global hectare may change slightly from year to year. 3 CO2-eq is a metric measure used to compare the emissions from various greenhouse gases based on their global-warming potential (GWP), by converting amounts of other gases to the equivalent amount of carbon dioxide with the same global warming potential. Carbon dioxide equivalents are commonly expressed as million metric tons of carbon dioxide equivalents, abbreviated as MMTCDE. Here we will use the unit of metric tons per capita (Mt CO2-eq per capita). against others, grouped by income quartiles. Although low-income countries (orange bubbles) had the lowest per-capita footprints (material, carbon and ecological), none of them had an Environmental Performance Index score below 45. Similarly, Material Intensity is also negatively correlated with footprints, DMC and CO2 emissions. In practice this means that countries with low levels of Material Intensity show also large footprints (see Figure 1b).  We found similar contradictory results when we analyzed the relationships between all sustainability measurements and GDP ( Figure 2 and Table 3). The Environmental Performance Index and Material Intensity indices vary negatively with GDP (Figure 2a and 2b and Table 3). Both measurements indicate that the countries that are performing well in environmental terms are the richest ones, using resources most efficiently to produce high levels of GDP. The other measures of environmental sustainability all get worse as GDP increases. Domestic Material Consumption, CO2 emissions, the Material, Carbon and Ecological Footprints all increase as GDP increases (Figures 2c and 2d   However, there are interesting differences in the precise nature of their relationships with GDP. There are signs that GDP has a non-linear influence on Domestic Material Consumption, CO2 emissions (productionbased) and the Ecological Footprint (see Table 3 and Figures 2 in the Supplementary Document). The best-fit curve for the relationship between Domestic Material Consumption and GDP per capita is positive linearlogarithmic, suggesting diminishing resource use or emissions with higher GDP per capita. The best fit curves for the relationship between CO2 emissions and GDP and also Ecological Footprint and GDP are second-grade Journal of Political Ecology Vol. 28, 2021 10 polynomials and negative square coefficients (drawing an inverted U shape) meaning that various environmental problems will get worse as economies grow, but then decline as economic activity and environmental governance act to clean up and repair current damage. The only exception is the Carbon and Material Footprints, which increase linearly with GDP per capita.

What drives the differences in the relationship between the indices, indicators and economic growth?
The differences between these measures of environmental sustainability are stark. It would be hard to derive a coherent sustainability agenda from them to establish which countries should make what changeseven, distinguishing between the richest or the poorest nations. In order to forge a robust sustainability agenda, we have to understand what drives these differences. It is, we argue, through examining difference that the theories informing different "mash-up" indices and indicators can best be understood. For us to appreciate them properly we must first take a detour, following Joan Martinez-Alier's work, into the different approaches to sustainability that these indices in fact serve.
In particular, understanding the results of the Environmental Performance Index and Material Intensity Index (which improve with GDP) will help us to understand calls for a transition to a ''green growth'' pathway (e.g. UNEP 2011; OECD 2011; World Bank 2012) in which growth continues but is somehow decoupled from key material inputs. The other indices help to explain a rather different set of proposals that push for a focus away from economic growth (

Productive Confidence
The position of Productive Confidence is constituted by an assemblage of positions, as Martinez-Alier (1995a) has shown. These posit the environmental benefits of continued economic growth combined with changes to the way that nature and the environment are perceived and valued, which will result in stronger protection for the natural world.
Two theories underpin this reasoning. First, the Environmental Kuznets curve -based upon "The Kuznets curve" (Kuznets 1950) and used for the first time by Grossman and Krueger 1991 as an Environmental Kuznets Curve). This suggests that various environmental problems (e.g. air / river pollution) will get worse as economies grow, but then decline as economic activity and environmental governance act to clean up and repair current damage. Cleaning up thus becomes a profitable industry, damage is costly, and technology improves to become less damaging (and less expensive). Economic growth is thus ultimately good for the environment. Second, the idea of "weak sustainability" (based upon the work of Solow [1957] and Pearce [Pearce and Atkinson 1993]) assumes that sustainability depends on the maintenance of the capital stock, with substitution allowed between different capital forms (Pearce and Atkinson 1993: 104). Weak sustainability thus assumes almost complete substitutability between 'natural' and 'manufactured' capital. For example, a tree plantation that is being cut down to make paper would be sustainable as long as the paper sales generated enough revenue to re-plant and sustain the new plantation. This ensures that there is a constant flow of capital, and none is lost. In fact, capital is gained in the form of the paper product being sold (Simko et al. 2014). According to such perspectives, economic growth is sustainable if it saves more than the combined depreciation of the two forms of capital. Hence, manufactured capital can growth indefinitely and substitute natural capital -although it is a moot point what sort of 'environment' is created and conserved.
The implication of the Environmental Kuznets curve and the idea of weak sustainability is that the material economy can grow without plundering natural resources, because technical improvements and substitutions allow decoupling to take place. As Jackson puts it "Economic output becomes progressively less dependent on material throughput. In this way, it is hoped, the economy can continue to grow without breaching ecological limits" (Jackson 2009: 67, see also figure 3).
Progressive decline of 'material throughput' is also known as decoupling -the attempt to allow economic growth to occur without attendant growth in material inputs. But it is necessary to distinguish between relative and absolute decoupling. Relative decoupling refers to a decline in the ecological intensity per unit of economic output. In this situation, resource uses decline relative to the GDP. But they do not necessarily decline in absolute terms. Impacts may still increase, but at a slower pace than growth in GDP. The situation in which resource impacts decline in absolute terms is called 'absolute decoupling.' Those who advocate for Productive Confidence would argue that some countries, especially the most mature economies, are experiencing real (absolute) decoupling. This decoupling between economic growth and resource consumption has intensified because of technological improvements, and the shift of economic activity towards services and an information and knowledge economy (Garcia 2006). This perspective is sometimes called 'ecological modernization', because it envisages that economic growth will transform what the economy does to the environment (for the better) (Martinez-Alier 2002).
Productive Confidence is also imbued by three other ideas that hold that nature will be valued better and more effectively -society as a whole will evolve in such a way that continued destruction of valued nature becomes uneconomic and unpopular. These include, first, Inglehart's post-materialist thesis (Inglehart 1977(Inglehart , 1995, which posits that as societies become richer so they seek more amenity value in nature and will come to value the natural world more highly, directing political pressure and entrepreneurial activity towards strong environmental protection. Second, Krutilla's criterion for the valuation of environmental amenities (Krutilla 1967), which proposes to modify cost-benefit analyses to allow environmental amenities to retain, or even increase their value over time. Krutilla observed that economic discounting practices tended to devalue the long-term contribution of environmental amenities, and he sought to counteract that tendency. Third and finally, Hirsch's notions of "positional goods" (Hirsch 1976) which distinguishes between two categories of goods. There are material goods whose consumption generates utility because of their intrinsic characteristics. This kind of good can grow indefinitely by technological advances. There are also positional goods, whose intrinsic qualities are scarce, and by virtue of that scarcity provide social status. Where particular environments or places acquire value because of their beauty (and especially if that beauty is spoilt elsewhere) then society will move to protect them because that beauty is now valuable. This can either be through things like property prices, or through planning legislation which moves dirty industries to poor neighborhoods (for which see the huge literature on 'environmental justice' for instance : Bullard, 1996;Martinez-Alier, 1997;Apostolopoulou and Cortes-Vazquez, 2018).
These three ideas, post-materialism, Krutilla's criterion and positional goods combine to promote a particular form of environmentalism that is characteristic of wealthy societies. It is both produced by, and justifying of, the economic growth that creates it. It assumes an elastic supply of material goods and that technological change and economic growth can reduce unwelcome environmental change. These ideas theoretically inform some of the indices and indicators analyzed here.

Ecological Prevention
The views described as Ecological Prevention in Figure 4 include the idea that resource use has already exceeded wise limits, and that dematerialization of economic growth is challenging. Some also insist that levels of resource use must shrink, after a reduction in economic growth and resource use (Garcia 2006). There is much debate between those who think that we need to reduce economic growth and resource use, i.e. degrowth, and those who advocate for stabilizing these variables instead; i.e. Steady State economics.
Economic degrowth ideas first appeared during the mid-to late-1970s. The main idea behind the concept challenges the "growthmania" (Georgescu-Roegen 1977: 270) of orthodox economists and has latterly been defined as an equitable downscaling of production and consumption that increases human well-being and enhances ecological conditions at the local and global level, in the short and long-term (Schneider et al., 2010). Some degrowth proponents would add that such downscaling has to assure that society's throughput -resource use and waste -stays within safe planetary boundaries 4 (Kallis 2011;Schneider et al. 2010).
Contrary to Georgescu-Roegen, Daly (1973) argues that attaining a sustainable state of the human economy is possible through a transition to a stationary or steady state. Swayed by John Stuart Mill's description of a stationary state and based on the thermodynamic world view of Georgescu-Roegen, Daly developed the concept of a desirable zero-growth economy (Kerschner 2010). However, advocates of a Steady State economy increasingly call for a period of degrowth in the transition to a steady state (Farley, 2014). A growing number of studies suggest that the global economy has exceeded critical planetary boundaries (O'Neil et al. 2019;Rockström et al. 2009). According to Farley (2014), for those who promote a Steady State economy, the question is no longer when to stop economic growth, but rather how much degrowth is necessary before we transition to a steady state.
In contrast to weak sustainability, which assumes complete substitutability between different capital stocks, these theories do not presume such substitutability is possible, meaning that all-natural capital must be conserved. For proponents of strong sustainability, the substitutability of manufactured for natural capital is seriously limited by specific characteristics of the latter, such as irreversibility, uncertainty, and the existence of critical components of natural capital which make a unique contribution to welfare (Costanza and Daly, 1992).
Behind notions of Ecological Prevention is the assumption that the growth of the material economy is linked to increased flows of energy and materials (Steinberger et al. 2013) and that these imply a heavier, negative environmental impact on ecosystems. High use of energy and material mostly derives from the wealthy countries. Wiedmann et al. (2015) has pointed out that "a number of studies have identified affluence, along with other factors, as a key driver for consumption based indicators, such as land (Weinzettel et al. 2013), carbon (Hertwich and Peters 2009), energy (Lenzen et al. 2006), ecological footprints (Wiedmann et al. 2006), and water footprints (Lenzen and Foran, 2001), as well as resource use (Schand and Turner, 2009)" (Wiedmann et al. 2015: 6273). More recently, Wiedmann et al. (2020) posited that the wealthiest 0.54 percent, about 40 million people, are responsible for 14 percent of lifestyle-related greenhouse gas emissions, while the bottom 50 percent of income earners, almost 4 billion people, only emit around 10 percent. For these reasons economic growth is ultimately bad for the environment. 4 The planetary boundaries jointly define a 'safe operating space', within which it is argued the relatively stable conditions of the Holocene may be maintained (Rockström et al. 2009). Some degrowth proponents have criticized the concept of planetary boundaries. According to them, this concept is underpinned by a faith in factual information presented by scientists as unquestionable truths, without questioning the political system into which that information enters (Smith-Spash, 2014). They argue that all data suffer from uncertainty and some more than others, some of the indicators are weak, the proposed boundaries are arbitrary and ignore feedback loops, and the focus is global but several of the systems covered are more interesting to look at from a local or regional level (e.g. freshwater use). Looking at sources rather than sinks may be more useful to deal with issues of resource scarcity (ibid.). Some have claimed the concept is a social construction, meaning the limits are not only biophysical, but also political as they depend on "perceptions of risk, on public debate and powerful lobby groups, and on international political power" (Raworth 2012: 12). Alternatives have been suggested such as building stable institutions that guarantee a safe transition and a co-evolution of natural and social systems at planetary scale (Biermann 2012, p.7), rather than planetary boundaries (Biermann 2012) or moving beyond a focus on biophysical limits and global scale analyses, towards solution-oriented research from household to global scales (DeFries et al. 2012).

Journal of Political Ecology
Vol. 28, 2021 14 The idea that economic growth must result in an increase in resource use, making decoupling difficult, is captured in Jevons' Paradox (1865) or the 'rebound effect.' Jevons' Paradox states that an increase in resource efficiency leads to an absolute increase in the use of that resource, as a falling price brings formerly unprofitable resource-consuming processes into production, and rising income drives consumption (Kemps-Benedict 2018). For example, efficiency gains in internal combustion engines have help made private car transportation effective and affordable. This ubiquity of private car transport has in turn driven up car ownership, and the extensive reconfiguration of cities that now rely on the use of private cars. This scale modification of the system of needs now results in dramatically higher energy consumption from the transport sector (Parrique et al. 2019). So the benefits of increased eco-efficiency can easily be nullified or even can increase the total use of natural resources consumed (Martinez-Alier 2014).
To sum up, in the Ecological Prevention scenario, relative decoupling could be a signal of improvement in the efficiency of the economy, but it does not necessarily mean we are using fewer materials (or emitting fewer pollutants) overall. From this perspective, what is needed is an absolute decrease, where resource use (or emissions) declines in absolute terms and on a global scale, even as economic output continues to rise (Jackson 2017).

What types of theoretical assumptions are informing the index and indicators?
As we have mentioned, the Environmental Performance Index and Material Intensity scores improve as GDP rises. But why do wealthy countries appear to perform well with respect to Environmental Performance Index and Material Intensity but are the worst countries in terms of their footprints? In this section we explore how the theories we have described above inform different indices. In the account below 'theories of productive confidence' refers to all those theories which show how environmental sustainability can improve with economic growth, whereas 'theories of Ecological Prevention' hold that economic growth can harm the environment.

The Environmental Performance Index
The negative relationship between GDP per capita and the Environmental Performance Index score shows that economic growth is "good for the environment because it will provide the means for cleaning up the environment" (Martinez-Alier 1995a: 1). Note that we have inverted the scores for the purpose of this analysis so that higher values indicate poor performance. Improved quality of air, water and sanitation are all captured in this index. As predicted in the World Bank's World Development Report 1992 (IBRD 1992), it seems that "as incomes rise, the demand for improvements in environmental quality will increase, as will the resources available for investment" (p. 39). Similarly, the main message of the Brundtland report was that poverty is the main cause of environmental degradation. Hence, economic growth, renamed as sustainable development, is a remedy for both poverty and environmental degradation (Guha and Martinez-Alier 1997). Countries with large footprints and high economic capacity can therefore invest in reducing pollution stresses, improving resource efficiency and addressing some of the consequences of their high natural resource consumption rates. The everyday environment needed for subsistence presents no problems for high-income countries and environmental amenity becomes a luxury.
The 2018 Environmental Performance Index ranks 180 countries on 24 performance indicators across ten issue categories covering environmental health and ecosystem vitality. According to the Environmental Performance Index's developers, the Index bridges a tension between two fundamental dimensions of sustainable development: (1) environmental health, which rises with economic growth and prosperity, and (2) ecosystem vitality, which comes under strain from industrialization and urbanization. However, as we can see in Figure 4, western countries do not experience the aforementioned tension and show a good performance in both Ecosystem Vitality and Environmental Health (higher values for both categories indicate poor performance). Such a relation could be accounted for by Environmental Kuznets curves. Western countries are at the bottom-right corner (Figures 4) not only because they are rich, but also because they are post-industrial economies that passed through a process of modernization that has entailed the relocation and cleaning of the most polluting industries. Thus, according to the Environmental Performance Index, economic growth is good for the environment and the future that the Index envisions is close to the Green Growth or Productive confidence scenario.

Domestic Material Consumption, Material Footprint and Material Intensity
Indicators of material consumption measure extraction and use of biomass, minerals, fossil fuels and construction materials. Such measurements are key indicators in that they pertain to a broad range of ecological concerns, including deforestation, meat consumption, overfishing, greenhouse gas emissions and environmental damage due to mining (Hickel 2018). Nevertheless, behind these three indicators we found very different types of environmentalist thinking. We summarize in Figure 6 the theoretical assumptions that underline these measurements. The Material Footprint per capita shows a strong relationship between the use of materials and wealth -as the proponents of the 'environmentalism of the poor' and in general the Ecological Prevention scenario (Figure 3) are claiming -while the Domestic Material Consumption per capita shows that as the economy grows, initially the environment suffers but eventually the relationship between the environment and the economy improves -as the Environmental Kuznets Curve hypothesis proposes. In fact, Domestic Material Consumption indicates that high-income countries are facing a relative and, in some cases, an absolute decoupling, which is consistent with green growth (Steinberger et al. 2013, Giljum et al. 2014, Wieldmann et al. 2015, Hickel and Kallis 2019. However, both Material Footprint and Domestic Material Consumption, when measured against a specific planetary boundary 5 , indicate that the majority of the countries analyzed have transgressed such boundaries. Not even 40% of the analyzed countries are within the per capita boundary for these indicators. Hence, these indices support those in the Ecological Prevention school (Figure 4) who argue that planetary boundaries have been overstepped and so many countries, especially the wealthy ones (see     Figure 5). Material intensity declines as GDP increases (Figure 2), demonstrating that economic growth is good for the environment. As Green growth theory asserts continued economic expansion is compatible with our planet's ecology, as technological change and substitution will allow us absolutely to decouple GDP growth from resource use (Hickel and Kallis, 2019). Hence, according to Material Intensity, the material economy can grow without plundering natural resources. Indeed, this indicator shows that wealthy countries are decoupling, meaning that there is a decline in the material intensity (or the emission intensity) of economic output, at least in relative terms.

CO2 Emissions and Carbon Footprint
CO2 emissions are a common indicator for strong sustainability. When comparing production-based CO2 emissions and Carbon Footprint (CO2 emissions at consumption-based) against GDP per capita, our results show also different theoretical assumptions at work. First, relationships between CO2 emissions (productionbased) and GDP would partially support the Environmental Kuznets Curve hypothesis. In Figure 2 and Figure  6, we observe that GDP has a lesser influence on CO2 emissions (production-based) than on the Carbon Footprint. Just as the proponents of Environmental Kuznets Curve have theorized, CO2 emissions first rise, and then fall or stabilize with increasing income per capita (Table 3). However, when switching from CO2 emissions to Carbon Footprint, emissions increase linearly with GDP. This second situation would be consistent with the Ecological Prevention scenario which claims that the growth of material economy is linked to increase in the flows of energy and materials (Steinberger et al., 2013) and it implies a heavier negative environmental impact on ecosystems. This difference reflects the fact that the Carbon footprint captures the whole value chain beyond borders, whereas CO2 emissions are restricted to national boundaries. So, a country can appear to reduce its planetary impact using the CO2 emissions measure when it is in fact merely exporting them.
Finally, as can be observed in Figure 2, few of the analyzed countries are within the per capita boundary for CO2 emissions and Carbon Footprint 6 (24% of the countries). Hence, as Kallis (2017) emphasized, the emission targets agreed in Paris require a radical reduction of emissions and material extraction, use and disposal. Under this perspective, both CO2 emissions indicators (production-based and consumption-based)

The Ecological Footprint
The Ecological Footprint measures the ecological assets that a given population requires to produce the natural resources it consumes (including plant-based food and fiber products, livestock and fish products, timber and other forest products, space for urban infrastructure) and to absorb its waste, especially carbon emissions. The relationship between GDP per capita and Ecological Footprint per capita is positive (that is, higher GDP per capita is associated with higher resource use). However, the best-fit curve is quadratic. This shape suggests diminishing resource use with higher GDP per capita. Such result would be consistent with the Productive Confidence scenario.
Nevertheless, grouping Ecological Footprint per capita by national income level reveals the inequality in per country demand for renewable resources and ecological services (see Figure 7). The Ecological Footprint indicator has also calculated the global biocapacity per capita (1.6 g/ha in 2017). Most of the analyzed countries -74% -, in particular the wealthy ones, are using resources at levels above per capita biophysical boundaries (see Figure 7). Thus, the Ecological Footprint supports the degrowth objective which is to scale down the material and energy throughput of the global economy, focusing on high-income nations with high levels of per capita consumption (Hickel 2019). This indicator is a well-known gauge of strong sustainability, that fits with the assumptions of the Ecological Prevention scenario (Figure 3). As a matter of fact, it is the only one of the analyzed indicators and indices that takes into account the biocapacity of the Earth, i.e. it is the only metric that measures what resources we have and how much we use. Table 4 extends Table 3 to summarize the theoretical implications of each of the selected indicators.

The elephant in Green Growth: devices and tricks that make rich countries look green
We have explored how particular forms of environmentalism and environmentalist action are represented through assemblages of ideas. We have also analyzed how these ideas are served by indices and indicators that embody their theories and values. However, comparing indices and indicators produces contradictory results. We still do not have definitive answers to the crucial questions that have brought these indices into existence in the first place: What changes are required to save the planet? Should rich countries We believe that empirical support for Green Growth is, in fact. weak. This is plain when examining the tricks and devices present in some indicators which make rich countries look green. These show flaws in the empirical support for green growth based on business-as-usual scenarios and Productive Confidence scenarios.

Relationships between the indicator and GDP
Theoretical scenario

Material Footprint
Material Footprint increase linearly with GDP Ecological Prevention. Degrowth critique

Domestic Material Consumption
Diminishing Domestic Material Consumption with higher GPD per capita Productive Confidence

Material Intensity
As GDP rises Material Intensity decreases Productive Confidence Green Growth

Environmental Performance Index
As GDP rises the Environmental Performance Index decreases Productive Confidence Green Growth

Ecological Footprint
Ecological Footprint increases in the initial levels of GDP to decrease later on, after a certain point of return.

Carbon Footprint
Carbon Footprint increase linearly with GDP Ecological Prevention. Degrowth critique

CO2 Emission (productionbased)
CO2 Emission (production-based) increases in the initial levels of GDP to decrease later on, after a certain point of return.  (Athanasoglou et al. 2014;Saisana and Saltelli 2010). According to them, although several and different indicators compound the EPI, few are strong determinants of good environmental performance. They have questioned the differences between the weights that have been given to some indicators (that are compounding this index) and their real impact on it (ibid.). According to Saltelli et al. (2011) the weights used in Environmental Performance Index were not measures of importance, but rather measures of trade-offs among variables (see the Supplementary Document, Figure 1). This is also true for the most recent 2018 Environmental Performance Index. Although it includes 24 different indicators, just five contribute the most to the ranking, namely: Exposure to indoor air pollution from Household Solid Fuels; Unsafe drinking water; Unsafe Sanitation; Lead exposure, and Wastewater treatment (for more information see the Supplementary Document). These five indicators are mainly about quality of life. They are measuring the quality of the air that we breathe, the quality of water that we drink or the access to or use of improved sanitation facilities. But none relate to increased resource depletion, or the use of water, materials, land, forests, energy and so on. Neither are they related to the generation of waste. As Martinez-Alier states, attention is drawn "away from material consumption and away from distributional conflicts" (Martinez-Alier 1995a: 2). Although not "arbitrary" as Ravallion (2010) might argue, the choice of indicators is fundamental to the scores that the Index provides.

Productive Confidence Environmental Kuznets Curve
In common with the post-materialist thesis (Inglehart 1977), the Environmental Performance Index describes environmentalism not in terms of concern for the increasing depletion of resources, but on the contrary, in terms of a change in cultural values towards "quality of life" issues. This makes some countries -specifically richer ones -look good (or clean) in terms of environmental sustainability (Morse and Fraser 2005;Jesinghaus 2000). Further note that the indices produce this result by virtue of their parochialism and of the institutions and actors who decide to use them. They are only likely to be able to achieve this high level of environmental performance domestically -the international consequences of the environmental problems they export by virtue of their consumption demands are not within their purview. This is because the use of natural resources and emissions are measured on a territorial basis (IPCC, 2006). Thus, these indicators have not taken into account the material effects of relocation of resource intensive production.
As Jackson (2017) has argued, a significant proportion of the carbon emissions for which rich consumers are ultimately responsible (i.e., the carbon emitted from the factories in China where their consumption goods are made) end up being attributed to citizens in poorer countries. The material effects of relocation of resource intensive production are not being accounted for. Moreover, the indicators that are included in the Climate and Energy and Air Pollution issue categories are measured by unit of GDP. As was explained in the 2018 Environmental Performance Index Report (Wendling et al. 2018), Greenhouse gas (GHG) emissions must be divided by the size of each country's economy, as measured by GDP, to calculate carbon intensity. 7 However, consuming less energy per unit of GDP does not mean that the total amount of consumed energy has decreased. In fact, countries that derive 'good' scores in terms of material intensity (the wealthy nations) are the ones who consume the most energy and materials per capita. We will return to this issue of dividing indicators by GDP when discussing Material intensity.
For instance, according to the CO2 Emission intensity indicator -included in the 2018 Environmental Performance Index-Seychelles was ranked well, and Burundi as one of the worst countries. However, if we do not control for GDP, but instead examine how emissions vary per capita, then the production-based CO2 emissions per capita in Seychelles in 2016 were 5.75 and in Burundi 0.54. What is worse, consumption-based CO2 emissions per capita in Seychelles were 12.45 metric tons, against 0.54 per capita metric tons in Burundi. High levels of CO2 emissions can be concealed.

Journal of Political Ecology
Vol. 28, 2021 20 Finally, none of the 24 indicators that compound the Environmental Performance Index have been measured against planetary boundaries. This representation creates the illusion that the economy can continue to grow, without knowing if ecological limits have been breached.

Domestic Material Consumption and CO2 emissions (production-based)
Domestic Material Consumption per capita has a weak but still significant correlation with GDP. We found some high-income countries present a low level of Domestic Material Consumption in tons (see Figures  2). However, the scope of Domestic Material Consumption is limited to the amount of materials directly used by any national economy. It does not include the upstream raw materials related to imports and exports originating from outside the national economy. Hence if a car is imported, only its weight is counted, and not all the energy, water and other materials required to produce the metal, rubber and plastic it contains. In a globalized economy, where rich countries have outsourced much of their production to poorer countries, this side of material consumption has been shifted off their balance sheet (Hickel and Kallis 2019).
In 2017, high-income countries had the highest material footprint per capita (approximately 28 tons per person, Figure 6). The material footprint of wealthy nations is greater than their domestic material consumption, indicating that consumption in those countries relies on materials from other countries through international supply chains. Taking into account Material Footprint per capita, high-income countries rely on approximately 11.5 tons 8 per capita of primary materials extracted elsewhere in the world, over and above their Domestic Material Consumption. Hence, the Material Footprint indicator shows that the lifestyles of people in the richest countries are heavily dependent on resources extracted from poorer ones.
As we saw with Domestic Material Consumption, the scope of the CO2 emissions indicator is limited to the volume of emissions directly used by an economy. As we can observe in Figure 2 in the Supplementary Document when switching from CO2 emissions (production-based) to the Carbon Footprint measure, emissions increase linearly with GDP. This difference between production-based and consumption-based emissions have been labelled as "accumulation by displaced emissions" (Isenhour and Feng 2014), given that they allow for the reproduction of affluence in one context while attributing the mitigation responsibility and associated investment to another (Isenhour 2016). This avoids the root causes of the problem -growth in resource use and materials throughput-and it delays more substantive climate mitigation policy and action. Hornborg (2009) states that this situation outlines a zero-scenario, in which gains in one country are linked to increased costs in another. One of the tricks in the Environmental Performance Index is exposed: not taking the material effects of relocation of resource intensive production into account.

Material Intensity
Large footprints -either material, carbon or ecological-are negatively related with Material Intensity (Figure 1b). What is driving such contradictions? An improvement in the material intensity (i.e. the efficiency of material use) of economic output does not necessarily mean that material use itself is falling (Jackson and Victor 2019). Resource efficiency alone is not enough because "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' (UNEP 2017: 12  In fact, if an international sustainable policy is guided by Material Intensity, it suggests that poor countries are the ones who should reduce the use of natural resources, or at least improve the efficiency with which they use them. However, little attention is paid to the fact that in absolute amounts the richest countries are the ones who emit and consume the most. Here we find another device that has also been reported in the Environmental Performance Index: monitoring material or energy use per unit of GDP does not consider change in absolute or per capita terms.
The index of Material Intensity contains two further devices. Material consumption is measured as Domestic Material Consumption, which does not count the true costs of imported consumption. And, in addition to that, Domestic Material Consumption is divided by GDP. This means larger economies, particularly those inflated by large financial sectors, will receive smaller scores. That is why large footprints -either Material, Ecological or Carbon-correspond to lower Material Intensity.
Finally, Material Intensity has not been measured against ecological limits. This gives the illusion that rich countries can grow indefinitely because natural resources are unlimited and/or substitutable with manufactured capital. However, as Hickel (2018) has argued, materials are non-substitutable resources, and the efficiency gains are ultimately governed by physical limits. As efficiency improvements reach physical limits, the scale effect of growth drives total resource use up (Ward et al. 2016). Hence, growth in GDP ultimately cannot plausibly be decoupled from growth in material and energy use, demonstrating categorically that GDP growth cannot be sustained indefinitely (Hickel 2018).

Ecological Footprint
Although the Ecological Footprint has been identified as a strong sustainability indicator, we found a ruse in the way that this indicator calculates the Carbon Component. The Carbon Component estimates the area of forestland that is required to absorb all the carbon emissions from human activity in excess of what the oceans already absorb (for more details see Mancini et al., 2015 and the Supplementary Document). Such an approach can be associated with the theory of weak sustainability: the negative environmental consequences of producing manufactured capital (CO2 emissions) can be easily replaced by natural capital (forestation). This fact involves the assumption of a so-called sustainable energy scenario under which CO2 is being captured by planting trees or forestation (Van den Bergh and Grazi 2013). Little attention is paid to the fact that carbon sequestration strategies need to consider their full environmental and social consequences. 9 This may create a bias because at some point wealthy countries, which are also the counties who emit the most CO2, can offset their emission by planting trees. Hence, this represents another trick that could make rich countries look green and absolve them for taking drastic actions against climate change: assuming that carbon can be sequestered irrespective of the social and ecological contexts where that occurs. This in turn is part of a broader mechanism that seeks to use commodity trading (which tends to strip goods of their social and ecological content, see Kosoy and Corbera 2010) to attempt to solve environmental problems.

Discussion: environmental sustainability measurements, tricks, and colonization
Indicators that absolve wealthy countries of their culpability for causing pollution and environmental degradation perpetuate unequal exchange. They permit colonial forms of environmental governance, as political ecologists have repeatedly noted.
The forms, variety and intensity of colonization varies with each sustainability index or indicator and their theoretical and epistemological assumptions. The devices that we have identified (Table 5) make certain forms of environmental change, and environmental degradation, visible. Or, to paraphrase Robertson (2006) they create particular sorts of environments, and environmental change, that can be 'seen' by governments and international institutions.
The environment that wealthy countries can see most clearly is that captured by the Environmental Performance Index which is concerned with quality-of-life issues such as "air quality, access to clean water,

Journal of Political Ecology
Vol. 28, 2021 22 adequate sanitation or waste disposal" (Inglehart 1995, 57) or conserving pristine nature away from material consumption and away from economic distributional conflicts. By providing the results of a few indicators (such as water quality or the quality of the air) it can be argued not only that wealth increases appreciation of environmental values, but also that wealth is good for the environment. However, as we have stressed, the Environmental Performance Index ignores planetary boundaries. This truncation might mislead assessments of national natural resource use and could hide the environmental implications of an infinite economic growth. If international institutions, such as UN, want to pursue a reduction of CO2 emissions rapid enough to stay with the carbon budget of 2ºC targets, it should start to with a firm commitment to respecting planetary boundaries and seek to achieve human development within those limits (Hajer 2015). Such planetary boundaries, as we have shown, indicate that rich countries need to adopt degrowth strategies (Hickel 2018).
The other kind of environmentalism that wealthy countries are prepared to see is that of Ecological Modernization: the idea that continued economic expansion is compatible with our planet's ecology, because technological change and substitution will allow us absolutely to decouple GDP growth from resource use and carbon emissions (Hickel and Kallis 2019). However, such ideas conceal environmental problems hidden by capital.
First, decoupling is measured by indicators rendered in domestic production-based terms. Consequently, the richest countries can appear to reduce their national material consumption and emissions by moving material-intensive industries to other countries, generally to the Global South. Established privilege is blind to ecologically unequal exchange. Sharing responsibility for emissions and material use among producers and consumers could facilitate international agreement on global environmental policies that is now hindered by concerns over the regional and historical inequity of emission and material use. They export not just the causes but also the worst consequences of climate change. Although high-income countries are those emitting more CO2 emissions per capita, scientists at Notre Dame Global Adaptation Index calculated that people living in poor countries have 10 times more chance of being affected by a climate disaster than those in wealthy countries each year (Coffe 2016).
Second, indicators that are measuring decoupling are normally divided by GDP. As we have seen with the index of material intensity (DMC/GDP) and carbon intensity indicators, this mathematical device suggests poor countries should improve their environmental performance, belying the fact that the richest countries are those who consume the most natural resources per capita. As argued by Fletcher and Rammel (2017), decoupling acts as a distracting fantasy that warrants a (continuously more) destructive path with a promise of success. Demonstration of its impossibility is deferred into the future. But as absolute decoupling fails to materialize, natural resources continue to be depleted. In that sense, decoupling is more a threat than an opportunity. Ultimately, at the global level, what counts is the total amount of resources that are extracted from the ground, emitted into the atmosphere or consumed, and the possibility of regenerating these resources (Jackson, 2017). The final arbiter on material decoupling and the possibilities for escaping the dilemma of growth are worldwide trends on primary resource extraction and consumption (ibid.). In this vein, the amount of carbon dioxide going into the atmosphere today is more than 60% greater than in 1990 (Jackson and Victor, 2019, see also the supplementary document). Human consumption of material resources has reached 91 billion tons per year, overshooting a sustainable level by 82% (Wiedmann et al. 2015;Bringezu et al. 2015). Needless to say, on per capita basis, the majority of this overshoot is due to consumption in high-income nations (Hickel 2019). Ultimately, "trying to reduce impacts while growing makes as little sense as trying to brake while accelerating in front of an obstacle" (Parrique et al. 2019: 59).
It is only by explaining economic growth without empirical reference to energy and material inputs that we are able to forget the environmental conditions and consequences of affluence (Martinez-Alier 1995b). Measurements indicating that economic growth is good for the environment downplay or ignore transboundary or spillover effects of the unsustainable consumption of nations in the Global North (Morse and Fraser 2005). Not making the mess in the first place should be preferable to cleaning it up later, but this is overlooked by such approaches. If we are going to label countries "good" or "bad" in environmental terms, we must get our measurements right. These measurements are "based on misleading data, which fail to take into account the Journal of Political Ecology Vol. 28, 2021 23 true environmental costs that rich countries impose on the world, are designed to make dirty nations look clean" (The Ecologist 2001: 47).

Tricks Indicators Actors
Focusing on "environmental issues" away from material consumption and away from distributional conflicts"

Conclusion
As we stated at the beginning of this article, central to any colonial project is a particular way of seeing nature and society. This vision -of the present and for the future -makes it possible to identify groups of people who need to be governed in particular ways, or nature (land, resources) that needs to be used in particular ways. Colonization requires mapping, categorizing, surveillance and control.
Decolonization therefore requires new ways of seeing nature and society, and some understanding of how the colonized visions operated -how colonizers constructed their image of what the world is like, and how it should be transformed. With respect to environmental and conservation policies we have to understand how different inequalities and injustices become engrained and sedimented into the world views we are challenging. Critics have long challenged the problems associated with particular categories of conservation interest (such as 'wilderness' and even biodiversity) that can structure conservation thought (Cronon 1995, Guyer andRichards 1996). But it is plain from our arguments above that the decolonization project must also extend to include the way in which we measure environmental change, degradation, amelioration and sustainability. Environmental sustainability indeces and indicators should consider the process of colonization that reproduces injustices within and across countries and enacts violence on racialized bodies, exacerbates class divisions, and damages our relationship with nature (Ricaurte, 2019).
The SDGs have stemmed from a global call to improve prosperity and environmental well-being. Through them international institutions are calling for action to protect the planet from degradation, through a wide range of changes to lifestyles, livelihoods, economies and societies. But this agenda begs vital questions: who should make which changes? What should be rich countries try and do, what should be poorer countries aim for?
The answers to these questions are plainly relevant to debates about decolonizing conservation and environmental governance. Simply put, current forms of environmental degradation are both a legacy of past trade, practices and resources use which have depended on unequal exploitation and consumption of resources in poor countries, and a consequence of ongoing international value chains in which production, and pollution, occurs in one country, but relatively benign consumption in others.
Indices and indicators that make rich countries look clean, are not just showing the environment that established wealth wants to see, it also hides the injustices and ecological-distributional conflicts that capital and society needs to see. Indices and indicators that try to hide the ecological debt are a product of what Quijano (2007Quijano ( : 1968) calls "Eurocentered colonial domination", mistakenly "… assumed to be 'objective', 'scientific', categories" and "…as natural phenomena, not referring to the history of power. This power structure was, and still is, the framework within which operate the other social relations of classes or estates." By analyzing indicators and indices that are relatively free from tricks it is very clear that the ones that are being dispossessed of natural resources "…are precisely the members of the 'races', 'ethnies', or 'nations' into which the colonized populations, were categorized in the formative process of the world power, from the conquest of America and onward" (Quijano, 2007: 169).
Sustainability indicators that are relatively free from tricks (i.e. the Carbon Footprint and Material Footprint) are easier to relate to environmental conflicts. For instance, in a country where exports of materials are much greater than imports of materials in tons (as is usually the case in Latin American and African countries) there are local environmental conflicts where minerals are mined; and within society at large, the perception of the economic and ecological inequality of foreign trade should grow (Muradian & Martinez-Alier 2001;Giljum & Eisenmenger 2004). There are historical and current conflicts over coal extraction (regarding the health of miners, land subsidence, production of Sulphur Dioxide) and oil extraction (over gas flaring, biodiversity loss), but also about the export of waste from rich to poor countries. At the global level, more fossil fuel extraction means more carbon dioxide production, which creates a conflict over the distribution of responsibilities and damages from climate change. These damages have been catalogued as an ''ecological debt'' from North to South, which includes the ''carbon debt'' i.e., damages from rich countries caused by excessive per capita emissions of carbon dioxide (the main effluent of affluence) and claims resulting from biopiracy, ecologically unequal exchange, and environmental liabilities by private corporations or northern governments (Martinez-Alier 2009b). This is how the indicators of ecological unsustainability of the economy are linked to social and environmental conflicts (Martinez-Alier 2004).
At a global scale, Footprints show that Northern historical development and present production and consumption levels are reliant upon a socio-ecological 'subsidy' imposed on Southern countries (Rice 2009). The socioecological subsidy refers to the underpayment and, at times, explicit plundering, expropriation and looting of their natural resource assets and labor power. Such a situation, which began in the colonial era and continues unabated even today, not only enriches the North but also "impoverishes and degrades the land,