How close are European countries to the doughnut-shaped safe and just space? Evidence from 26 EU countries

Ecological Economics 221 (2024) 108189 Available online 23 April 2024 0921-8009/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/bync/4.0/). How close are European countries to the doughnut-shaped safe and just space? Evidence from 26 EU countries M. Rosario G´ omez-Alvarez Díaz a , * , Víctor Ernesto P´ erez Le´ on b , Patricia Fuentes Saguar c a Department of Economics and Economic History, Universidad de Sevilla, Ram´ on y Cajal s/n, 41018 Sevilla, Spain b Department of Applied Economics II, Universidad de Sevilla, Ram´ on y Cajal s/n, 41018 Sevilla, Spain c Department of Economics, Quantitative Methods and Economic History, Universidad Pablo de Olavide, Ctra. Utrera. km 1, 41013 Sevilla, Spain ARTICLE INFO Keywords: Social wellbeing Planetary boundaries Regional sustainability indicators Doughnut economics Sustainable development goals (SDGs) European Union ABSTRACT This study aims to measure how far 26 EU countries are from the safe and just space (SJS) defined in the Doughnut Economics framework by calculating the distance for all social and environmental dimensions and using two synthetic indices for safe space and just space. The distance was measured as a percentage of the threshold to be reached. Therefore, all indicators were interpreted in the same way. Cluster analysis was used to identify countries that displayed similar behaviours. The results showed that none of the countries analysed fall within this space. The environmental and social dimensions that most transgress were identified: CO2 emissions, biochemical flows, and social equality. The groups with the worst results were made up of countries from Southeast and East Europe, and Ireland and Luxembourg, while the best results were from Western Europe, Nordic countries and Slovenia. In conclusion, environmental transgressions and social deprivations exist in all European countries, but prosperity can be achieved through different socioeconomic models, and a high GDP does not correspond to greater sustainability. This information is valuable for the design of public policies aimed at ensuring that European society lives in the Safe and Just Space. 1. Introduction Is there a safe and just space (SJS) in which we could live? Doughnut Economics (DE) (Raworth, 2012, 2017) offers a theoretical model that delimits such a space and proposes an empirical evaluation of SJS worldwide. In this study, we adapt this proposal to evaluate how far 26 EU countries were from the SJS in 2019. Natural resource consumption has accelerated, generating unprecedented biophysical changes, and giving way to the Anthropocene (Crutzen and Stoermer, 2000). The international response has been to sign various treaties since the 1990s, until the most recent Paris Agreement, together with the 2030 Agenda UN Sustainable Development Goals (SDGs) (UNFCC, 2015; UN (United Nations), 2015). SDGs include the term ‘planetary boundaries’ (PBs) (Griggs et al., 2013; UNEP, 2012, 2019), coined by the Stockholm Resilience Centre (Rockstrom et al., 2009) to refer to the nine planetary environmental limits that delimit safe space (SS) for human life (Montoya et al., 2018; Rockstr¨ om et al., 2018; Li et al., 2021a). Steffen et al. (2015) showed that the limits of climate change, biodiversity, nitrogen cycle, and landsystem would have been exceeded; a deterioration that continues (IPCC, 2022). Thus, the transgression of PBs creates a substantial risk of destabilising the Holocene state in which modern societies evolved (Rockstrom et al., 2009). Added to this is a form of societal organisation in which poverty, inequality, discrimination, job instability, or preventable illnesses, among other problems, persist in all countries (Swinburn et al., 2011; Banerjee and Duflo, 2011; Piketty, 2020; Acemoglu and Restrepo, 2021), despite important recent technological developments. In this context, Raworth (Raworth, 2012; Raworth, 2017) proposes that society’s goal should be to guarantee the prosperity of life or a just space (JS), ensuring that everyone has a “social foundation”, below which there is unacceptable human deprivation, within PBs or Safe Space (SS). Raworth (2017) evaluated the situation of the planet with respect to the SJS. Planetary limits were taken directly from the results of Steffen et al. (2015), to which she added the quantification of unmet social needs for the world population. There are a wide variety of Abbreviations: DE, Doughnut Economics; SJS, Safe and just Space; PBs, Planetary Boundaries; SS, Safe space; JS, Just space; EDI, Ecological Distance Index; SDI, Social Distance Index. * Corresponding author. E-mail addresses: [email protected] (M.R. G´ omez-Alvarez Díaz), [email protected] (V.E. P´ erez Le´ on), [email protected] (P. Fuentes Saguar). Contents lists available at ScienceDirect Ecological Economics journal homepage: www.elsevier.com/locate/ecolecon https://doi.org/10.1016/j.ecolecon.2024.108189 Received 11 December 2022; Received in revised form 15 January 2024; Accepted 20 March 2024 Ecological Economics 221 (2024) 108189 2 multidimensional metrics and alternatives to GDP, such as the HDI, Index of Better Life, etc. (Costanza et al., 2016; Hoekstra, 2019). However, DE empirically emphasises that environmental aspects are not just another dimension within the measures of human wellbeing, but rather that we are an integral part of nature, and our life must fall within those natural limits (Georgescu-Roegen, 1971; Meadows and Randers, 1972; Daly, 1991; Washington and Maloney, 2020; Drees et al., 2021). Additionally, prosperity is not a concept linked to growth, but to a life without shortages based on human rights, according to Raworth (2017). Therefore, the focus is not on the average achievements of society in different areas, as reflected in other indicators, but on a decent life for all, as an integral part of the ecosystem. ‘The Doughnut’ (an indicator-led framework that defines a doughnut-shaped Safe and Just Space, measuring performance with respect to social and planetary boundaries) was originally designed at the global level, and translating it to other geographical (e.g., national or subnational) scales poses a significant challenge (Li et al., 2021b; Ferretto et al., 2022). Various studies have attempted to cope with this challenge by applying this adaptation to specific countries (Roy and Pramanick, 2020; Cole et al., 2014; Sayers and Trebeck, 2015; Swaffield and Egan, 2020; Saunders and Luukkanen, 2022), regions (Dearing et al., 2014), as well as by comparative analyses of different cities (Hoornweg et al., 2016) and countries (O’Neill et al., 2018; Hickel, 2019; Fanning et al., 2022). The results of all these studies show that no territory is located within the SJS. Similarly, this proposal has attracted the attention of various organisations, such as the UN General Assembly (UN (United Nations), 2012). Evaluations are currently being carried out through participatory processes to design local policies, with Amsterdam being the pioneer (DEAL, 2020), followed by projects undertaken in Melbourne (Regen Melbourne, 2021) and London (Eboli et al., 2022), among other places. To our knowledge, O’Neill et al. (2018); Hickel (2019) and Fanning et al. (2022), are the only studies that compare >140 countries, including European countries. They used some of the dimensions of Raworth (2017), downscaling six PBs to the national level and applying eleven social thresholds. This article includes all the dimensions from Raworth as well as two new dimensions proposed for Amsterdam (DEAL, 2020), using variables at the national scale (H¨ ayh¨ a et al., 2018) for each European country, except Malta due to the lack of data, from official EU statistics for 2019 in most dimensions. Furthermore, the use of environmental resources is evaluated using a production-based approach, except for emissions of CO2. The indicator of GHG emissions per capita has been calculated from the consumption perspective, so that the true responsibility of each Member State with regard to “Climate change” is more faithfully reflected (Lucas et al., 2020). The distance that separates every country from the SJS for each of the social and environmental dimensions is calculated as well as the average number of transgressions. In addition, the average deficit for the European Union (EU-27), of each dimension is analysed, even if we do not strictly have data for Malta. Two synthetic indicators were calculated to communicate in a summarised manner concerning the degree to which every country is distant from the safe space (SS), through the ecological distance index (EDI), and from the just space (JS) via the social distance index (SDI). Previously, O’Neill et al. (2018) calculated the number of environmental transgressions and social achievements of each country. Fanning et al. (2022) only quantified the deficits, like our work, whereas Hickel (2019) also measured “undershoot”. This is the first study to evaluate all dimensions applying the same normalisation methodology, by adapting the multidimensional poverty index used by the UN (Alkire and Foster, 2011). In this way, all the dimensions and the two indices measure the deficit as a percentage of the threshold to be reached; therefore, all are interpreted in the same way. Our proposal is the same as that of Fanning et al. (2022), but only for the quantification of environment pressures. These authors evaluated social deficit as a percentage of the interval between the target and the minimum value among all the countries studied, so the deficit of a country is conditioned by the performance of another country, and not only by its situation with respect to the set goal. Synthetic indicators are used because their ability to summarise the complexity of many variables into manageable amounts of meaningful information (Nardo et al., 2005; Greco et al., 2019). The application to sustainability studies is extensive, and the methodologies are very diverse, where the choice of one or the other depends on the objective of the research (Singh et al., 2012; Costanza et al., 2016; P´ erez Le´ on et al., 2022). Its use in DE studies identifies countries that could serve as reference models and determines a country’s closeness or otherwise to the SJS over time in a simple way. This information is valuable to show the public and public managers an easy-to-understand alternative metric to GDP. Composite indicators implementation to evaluate the doughnutshaped Safe and Just Space and its interpretation has several caveats. First, it is a strong sustainability model (Dietz and Neumayer, 2007), therefore we constructed two indicators, distance to SS and distance to JS. Second, the form of aggregation is the arithmetic mean that compensates for the deficit of one dimension with a smaller value of another. However, it is only partial, because the value of a dimension with a surplus takes the value zero (Alkire and Foster, 2011; Fanning et al., 2022). Third, the DE framework considers all the dimensions as equally important, which is why they are all equally weighted. Finally, no dimension should disappear from the evaluation, following the proposal by Constanza et al. (Costanza et al., 2016, 350) “recognizing that the dashboard and an aggregated indicator of overall progress toward our shared goal are both necessary if we hope to achieve our goal.” Cluster analysis was applied to group countries according to the values obtained in the two indices (EDI and SDI) to identify common patterns and improve the analysis. Section two presents the theoretical framework and a literature review on safe space and just space. Section three presents the method, indicators, and sources used to analyse to what extent the EU is situated below the social foundation and/or has transgressed the ecological ceiling. Section four presents an empirical evaluation. Finally, we discuss the applicability of the tool in the EU, local-regional-global links and the SJS, as well as data limitations, scientific challenges, and further research needs. 2. Theoretical framework and literature review 2.1. A safe space The first step in assessing the transgression of global boundaries is to translate them into the corresponding territorial scale. Some scholars employ bottom-up approaches that do not relate directly to global limits on the premise of respecting local boundaries, such as observed local environmental boundaries (Cole et al., 2014; Dearing et al., 2014; McLaughlin, 2018). However, from an up-bottom methodology, a distribution criterion that considers environmental, social and ethical aspects must be applied (H¨ ayh¨ a et al., 2018). Firstly, there are three limits that are global public goods (climate change, ozone depletion, ocean acidification) where there is no direct relationship between transgression, which is the sum of all planetary activity, and local impact. However, there are many complications for global environmental governance. Consequently, downscaling limits raises awareness and holds local governments accountable (McLaughlin, 2018). In DE studies, climate change is determined by carbon dioxide (CO 2 ) concentration in the air; but territorial evaluation is carried out based on emissions (O’Neill et al., 2018; Cole et al., 2014; Saunders and Luukkanen, 2022). Similarly, the reduction of the ozone layer is evaluated through the emission of the HCFCs (Cole et al., 2014; Swaffield and Egan, 2020). The driver of ocean acidification is the rising level of CO 2 , which is dealt with in the climate change domain. For this reason, in empirical studies either it is not evaluated or variables on the health of M.R. G´ omez-Alvarez Díaz et al. Ecological Economics 221 (2024) 108189 3 oceans and rivers or water quality are used (Cole et al., 2014; DEAL, 2020). Secondly, the six global boundaries involve processes that aggregate from regional to planetary scales; as a result, remaining within territorial boundaries is a prerequisite for respecting planetary limits (Nykvist et al., 2013; H¨ ayh¨ a et al., 2018). The transgression of individual regional thresholds can provide feedback to processes that do have large-scale thresholds, affecting the Earth system at the global level (Nykvist et al., 2013; Dao et al., 2018). For that reason, Steffen et al. (2015) propose regional thresholds for land-system change, biodiversity loss, freshwater use, biochemical flows (phosphorus and nitrogen flows), expressed as a percentage for the former three and in an absolute amount for the last, from which a per capita distribution is established. However, the lack of data has led other empirical studies to apply per capita distribution (O’Neill et al., 2018; Roy and Pramanick, 2020; Nykvist et al., 2013). Thirdly, atmospheric aerosol loading, and chemical pollution or novel entities have global impacts which are not well understood, and no planetary boundaries have yet been established. Aerosol loading has been replaced by atmospheric pollution or air quality and is quantified using PM10s and PM5s, a wide range of particulate pollutants associated with aerosol loading, as a result of their proven health impacts (Cole et al., 2014; Sayers and Trebeck, 2015). The evaluation of novel entities was conducted based on the chemical status of the water bodies (Swaffield and Egan, 2020). Finally, some studies have incorporated new dimensions such as the ecological footprint, deforestation, waste, etc. (Dearing et al., 2014; Roy and Pramanick, 2020; O’Neill et al., 2018; Saunders and Luukkanen, 2022; Hoornweg et al., 2016). In most instances, thresholds are determined according to two criteria: distribution per capita or the limits of the territory itself (Dearing et al., 2014; Dao et al., 2018; O’Neill et al., 2018; DEAL, 2020): But there are different criteria, such as historical responsibility (Fanning et al., 2022), capability, right to develop or grandfathering, which imply different outcomes (Ryberg et al., 2020; Lucas et al., 2020). The second step is to determine the use of resources within a territory. A large body of the literature uses two major approaches. First is LCA methodology, that measures the environmental impacts of products all through their life cycles, which demands integrated micro-data (Bjørn et al., 2020; Sala et al., 2020). The second approach, based on MRIO, can be used to link planetary biophysical resource flows and environmental impacts with the monetary exchange of goods and services through global supply chains (Wiedmann and Lenzen, 2018). The environmental footprint obtained can be used to directly compare with downscaled PBs (Fang et al., 2015; O’Neill et al., 2018; Li et al., 2020; Suarez-Eiroa et al., 2022). Another aspect to assess is whether the analysis is carried out from a production-based accounting (Franzen and Mader, 2018), consumption-based accounting (Heinonen et al., 2020), or even different shared producer-consumer allocations as a middle road (Jakob et al., 2021). European countries are among the most-studied from a consumption-based perspective because of their inclusion in global MRIO databases, especially Exiobase (Lenzen et al., 2013; Beylot et al., 2019; Stadler et al., 2018; Tukker et al., 2016; Ivanova and Wood, 2020). 1 To the best of our knowledge, studies on PBs have been realised for Switzerland (Dao et al., 2018), Sweden (Nykvist et al., 2013), the Netherlands (Lucas and Wilting, 2019) Canada and Spain (Fanning and O’Neill, 2016) and Spain (Suarez-Eiroa et al., 2022). Regarding the EU as a territorial entity, PBs studies (Hoff et al., 2014; H¨ ayh¨ a et al., 2018; EEA-FOEN, 2020), have concluded that Europe has overshot its nitrogen, phosphorus, climate change and land system boundaries, but not its freshwater boundary. Focusing on comparative studies of the Doughnut for European countries, O’Neill et al. (2018) evaluated climate change, biogeochemical flows, land-system change, blue water, HANPP and added a material and ecological footprint, based on the per capita distribution of global limits. Fanning et al. (2022) extended the period of study, blue water could not be included, applying historical responsibility for the distribution of CO2 emissions, and the HANPP indicator. Both articles used the EORA MRIO database from a consumption-based approach. The main results were that all European countries crossed five of the six proposed limits. 2.2. A just space In Raworth’s SJS framework, “just” describes the avoidance of unacceptable human deprivation and extreme inequality. As a result, prosperity involves ensuring that no one suffers deprivation within environmental limits (Raworth, 2012; Raworth, 2017). From a theoretical perspective, there are different proposals based on the multidimensionality of needs (Max-Neef, 2017; Doyal and Gough, 1991) and capabilities (Sen, 1999; Nussbaum, 2001). All these theories argue that there are a finite number of basic human needs and capabilities that are universal, satiable, and non-substitutable, such that they are multidimensional (Nussbaum, 2001). However, need satisfiers are different between individuals and societies, but arguably have certain universal characteristics that may be measured empirically. That said, theories of human needs are commonly criticised on two grounds. First, they face claims of paternalism, being derived from experts and academics, with little apparent recourse to individual desires and sovereignty over life choices. Second, many see a project to define universal human attributes as impossible due to the relative, historical, and socially constructed nature of individual needs (Gasper, 2002; Gough, 2004). In responding to these issues, much emphasis has been placed on the separation of needs and satisfiers or the separation of functionings and capabilities (Davis and Wells, 2016). Since normative political goals are focused on extending the ability to fulfil needs, not the fulfilment itself, individual choice and cultural diversity can be preserved. Furthermore, the approach is objective in the sense that collective reasoning is applied to understand needs and well-being (Gough, 2020). Theoretical argumentation and deliberative consensus are two construction tools for the dimensions that make up basic human needs (Doyal and Gough, 1991; Alkire, 2002; Sen, 1999; Nussbaum, 2001; Gough, 2020). In the case of the DE, the proposed social dimensions were the result of a review of documents presented at the Rio +20 conference. Subsequently, after various participatory processes involving an actual application of DE (Sayers and Trebeck, 2015; Sayers, 2015; Swaffield and Egan, 2020). Raworth (2017) proposed 12 dimensions: food, health, education, housing, water and sanitation, energy, networks (communication and personal support), along with income and employment, all within a framework of gender equality, peace and justice, social equity, and citizen participation (voice) in the political sphere. In comparison, the SDGs identified 17 goals, of which 12 could be categorised as social objectives, four were environmental and one concerned the process of implementation. At a high level, these goals aligned well with the social foundation of the SJS framework, although there were some differences in the proposed specifics. The largest difference was the inclusion of goals related to sustainable cities and industry/innovation in the SDGs and the exclusion of social support. In addition, DE proposes that the evaluation of these dimensions should not be carried out in terms of average achievements, but rather to ensure that no one suffers a lack in these dimensions. The justification for this approach can be found on the one hand, in the defence of human dignity, based on human rights (Raworth, 2017). On the other hand, social opportunities must aim to protect the less privileged by following Rawls’ principle of justice (Rawls, 1971). Likewise, human development implies the expansion of the capacities that an individual chooses; yet at the very least, we must be free from key deficiencies (Sen, 1999). 1 See more European studies: https://www.exiobase.eu/, and surveys: Chen et al., 2022; Downing et al., 2019, EEA-FOEN, 2020, H¨ ayh¨ a et al., 2018; Li et al., 2021b; Vea et al., 2020. M.R. G´ omez-Alvarez Díaz et al. Ecological Economics 221 (2024) 108189 4 Raworth’s proposal (Raworth, 2017) of the dimensions and metrics of safe space at a planetary level was modified in applied studies owing to a lack of data or differences between the global scale and other territories. The most similar work is the study conducted in India (Roy and Pramanick, 2020), with the only difference being that social support is not evaluated, an aspect excluded from the SDGs; and inequality is evaluated using the Gini index rather than the Palm index. In addition, the thresholds, of which only five were reached, were established by the authors between 90% and 95% of the population, considering that 100% was impossible to get. The study carried out in South Africa omitted gender equality and equity because they are considered as cross-cutting other issues, and political voice was not evaluated due to lack of data (Cole et al., 2014). The variables employed were from the country’s statistics, and the threshold was set at 0% deprivation. None of these has been achieved, with those furthest away being safety, income, and jobs. The problem of lack of data becomes more evident in a study of Chinese regions (Dearing et al., 2014) where seven dimensions were evaluated, with none reaching the threshold established at 0%. In a comparative study of large cities by Hoornweg et al. (2016), neither food nor education was evaluated, as they are not municipal powers, and political participation was replaced by institutions, in addition to making a subsequent grouping of the dimensions, using an adaptation of the SDGs. Only Toronto was above the thresholds, as determined by the world average. The Cuban study evaluated only six dimensions, with the threshold being reached in four of them (Saunders and Luukkanen, 2022). In a comparative study of countries by O’Neill et al. (2018) and Fanning et al. (2022), life satisfaction was included in addition to health to provide both subjective and objective measures of well-being, while the remaining nine social indicators were classified as need satisfiers. Gender equality, peace, and justice were not evaluated, and the variables used differed from those used by Raworth, although they are global limits. The authors chose a threshold of 95% for recognizing the difficulty associated with extending universal access to the last 5% of the population. The results were rather mixed. Only three countries (Austria, Germany, and the Netherlands) achieved all 11 social thresholds, although an additional 7 (mostly European) countries achieved 10 of them, and 35 countries failed to achieve more than a single social threshold. In Doughnut studies based on participatory processes, new dimensions appear, but above all different variables. This is because DE focus on global objectives, and participatory processes were adapted to each area’s reality. Therefore, in the studies of Wales (Swaffield and Egan, 2020) and the UK (Sayers and Trebeck, 2015), both social and gender inequalities were not included, since they are considered to be present in all dimensions, rather than only being evaluated through a single variable. The dimension “access to water” was not included, because a lack did not exist in these areas. However, the authors do evaluate specific variables, such as the percentage of people without access to public transport, those who have problems paying for their house, or those who experience anxiety. Currently, the methodological reference study for cities is that of Amsterdam (DEAL, 2020), which results in an expansion to 16 dimensions. Despite being applied to developed countries, all these studies evidence the existence of disadvantaged groups. 3. Materials and methods 3.1. Ecological distance index (EDI) and social distance index (SDI) Evaluation of a society’s situation vis-` a-vis the Dougnhut is determined by the ecological distance index (EDI) and the social distance index (SDI). These indices are an adaptation of Alkire and Foster (2011) methodology, based on three steps. First, once the dimensions have been identified, they are quantified using a series of variables xi,(i=1,…,n), and a limit or boundary is established for each of them zi,(i=1,…,n). Next, the normalised gap of the variable is calculated di,(i=1,…,n), where di= |zi−xi|/zi. The normalised gap is truncated; that is, once the achievement has been accomplished or the limit has been reached di=0. The average of the indicators obtained for each dimension is calculated for the dimensions in which more than one variable is used. The indicator di evaluates a deficiency or shortfall in a certain dimension and is expressed as a percentage of the boundary required to fall within the JS or the SS. The di value can take a value between zero (a situation in which there is no deficiency) and 100% (maximum deficiency), except for a series of environmental dimensions where the real transgression of some limits exceeds 100% (Rockstrom et al., 2009). The second step is to determine the weight of each dimension (Nardo et al., 2005). In this study, we propose using the same weight for each dimension. The theoretical justification is that each of these dimensions corresponds to facets that allow a person to develop. Therefore, we could not determine which of these factors is more important (Nussbaum, 2001; Max-Neef, 2017; Raworth, 2017). Similarly, in the case of exceeding each of the environmental limits, the stability of the ecosystem, as we know it, is in danger (Rockstrom et al., 2009). The third step is to aggregate the dimensions, using the arithmetic mean. A zero value indicates that the territory is operating in the safe space and/or in the just space, and positive values indicate that the distance, expressed as a percentage, is within each space, for both spaces. In the case of the EDI, the value can exceed 100%, because the limits can be violated by >100%. In the case of the SDI, the index is normalised between 0% and 100%, where 100% indicates a society with maximum deficiencies in all dimensions. In the case of a variable taking the value 0%, it is not removed from the calculation, as it is counted as an assessed dimension, and is part of the denominator and the mean value. In addition, hierarchical cluster analysis was performed to identify possible clusters according to the values of the two indices: EDI and SDI. Ward’s method was used for the extraction procedure, with the square of the Euclidean distance as a measurement. This guarantees minimisation of the differences within clusters. 3.2. Dimensions, variables, boundaries, and indicators of the safe space Adaptation of the nine global environmental limits proposed by Steffen et al. (2015) to the EU countries for 2019, for most dimensions, is carried out in different ways, using a production-based approach, except for GHG emissions(see Table 1). The first is to apply the planetary limit directly, which is only possible for the land-system change dimension, where 50% of the land must be forest. Second, the emission of ozone-destroying gases from the EU was 0%, therefore adaptation was not necessary. Third, the per capita distribution between 7673 million people (World Bank, 2022) was applied to global limits of phosphate (6.2Tg yr- 1) and nitrogen (62 Trg yr-1) (Steffen et al., 2015; FAOSTAT, 2019). Fourth, there are three dimensions: freshwater, biodiversity, and the health of oceans and rivers, where the global boundary raises many concerns owing to the importance of context-specific criteria in defining sustainable levels (H¨ ayh¨ a et al., 2018). Specifically, the loss of diversity was determined by the percentage of species and habitats whose conservation status “is not good (European Environmental Agency, 2021). As regards the level of water consumption, the Water Exploitation Index plus (WEI+) was used as a measure of total freshwater use as a percentage of renewable freshwater resources (groundwater and surface water) at a given time and place. It quantifies how much water is extracted and how much water is returned after use to the environment. In the absence of formal goals agreed upon across Europe, values >20% are generally considered an indication of water scarcity (EEA (European Environment Agency), 2016). Ocean acidification is tightly coupled with climate change and shares the same primary driver, anthropogenic CO 2 emissions (IPCC, 2022). For this reason, we evaluated the water M.R. G´ omez-Alvarez Díaz et al. Ecological Economics 221 (2024) 108189 5 quality or health of oceans and rivers following other works (Cole et al., 2014; Fanning et al., 2022). Fifth, although the planetary limits for atmospheric and chemical pollution have not been quantified, data are available to assess their existence in the European context. The former, through the population weighted annual mean concentration of particulate PM10 and PM5 matter at urban background stations in agglomerations (ug/m 3 ), can be carried deep into the lungs, causing inflammation, and exacerbating the condition of people suffering from heart and lung diseases. Here, the limit established by the WHO (World Health Organization) (2021) was used. Chemical pollution takes various forms, but the EU only offers data on the generation of waste, excluding major mineral waste by hazardousness at a national level. Finally, climate change, ocean acidification, and ozone depletion can be seen as global public goods, so the metrics used to assign responsibility must likewise have a global scope to be meaningful. We focus on (regional) healthy oceans and that consumption-based data on ozone-depleting substances is arguably less reliable. However, the indicator of GHG emissions per capita has been calculated from the consumption perspective, so that the true responsibility of each Member State with regard to “Climate change” is more faithfully reflected (Lucas et al., 2020). CO2 emission limits depend on different scenarios, targets and period analysed. In this paper, we opt for the proposal of Lucas et al. (2020). Limiting the global temperature increase to 1.5 ◦C corresponding to a global budget of 570 GtCO2 from 2018 to 2100 onwards was used. The global CO2 budget was equally distributed between 2018 and 2100, resulting in an average budget of 7.0 GtCO2 yr −1 for the remainder of this century, and 0.9tCO2 yr −1 per capita. The distribution period chosen is consistent with the time horizon of most mitigation scenarios (IPCC, 2018). This variable is not directly available in official statistics, unlike the others. Emissions contain, in terms of CO2 eq., emissions of Carbon dioxide (CO2), Methane (CH4), Nitrous oxide (N2O), Hydrofluorocarbons (HFCs), Perfluorocarbons (PFCs), Sulfur hexafluoride (SF6), Nitrogen trifluoride (NF3). Also included here are the values of the so-called precursors: Carbon monoxide (CO) and NMVOC (Non-methane volatile organic compounds) and exclude LULUCF. To do this, the MRIO (multiregional Input-output database) and EXIOBAS (EXIOBASE, 2019) see Stadler et al., 2018) has been taken, which contains da ta on emissions by industry and by product, allowing the calculation of emissions embodied in the consumption of each country. 3.3. Dimensions, variables, boundaries, and indicators of the just space From a procedural perspective, the selection of the dimensions and variables used to evaluate the just space carried out is based on Raworth (2017) proposal, the SDG, and on the participatory process conducted in Amsterdam (Alkire, 2002; Gough, 2020; Cole et al., 2014; DEAL, 2020). In the latter, transport and culture were incorporated into new dimensions. Income and employment were separated because, although their relationship is evident, achieving standards in one does not directly imply an improvement in the other. Work is a means to income for most of the population, but it is also an end in itself, as long as it is work that is dignified and of intrinsic value to the individual. Additionally, we felt it pertinent to separate the dimension of Internet connectivity and personal support, because they respond to different human needs (Max-Neef, 2017). Finally, the Amsterdam proposal added equality to diversity, embracing gender inequality and social discrimination. In this study, we chose to include discrimination as a sub-dimension within social inequality, which only included economic inequality, and to assess gender inequality independently, given its relevance to the SDG and in Raworth’s proposal (Raworth, 2017). Accordingly, we evaluated 16 dimensions, splitting the compound dimensions of Raworth (income and jobs; connectivity and personal support) and two new dimensions (culture and transport) (Table 2). On the other hand, Raworth (2017) proposes variables to assess the situation of the world population. However, we use variables that aim to quantify the position of the worst-off population in the European context, always drawing on official statistics published by the EU. For example, the food dimension is evaluated worldwide through caloric intake, although the food problem in Europe is that some of the population cannot afford adequate food. The selection of variables followed four criteria. First, these variables were used to quantify the SDGs. Eleven were used in the European SDGs (EUROSTAT, 2019). Likewise, political voice and social equality are evaluated in the European SDGs, although we chose to use Eurobarometer (EC (European Comission), 2018, 2019; EP (European Parlement), 2019). This was because the answers in the Eurobarometer were presented in terms of the degree of satisfaction of those evaluated, which allowed us to determine the percentage of the population who were dissatisfied in these dimensions, without the need for us to determine the appropriate threshold. Finally, the dimensions of social support and culture were not included in the SDGs, and the transport variable in the SDGs did not allow the evaluation of the deprived population. In these cases, we used other official European statistics (EUROSTAT, 2019). Table 1 Biophysical boundaries. Dimension Variable (xi)Boundary (zi) Unit Proposal for/ data source/year Atmospheric Pollution (SDG 13) The population weighted annual mean concentration of particulate matter at urban background stations in agglomerations 15 PM10 (ug/m3) WHO (World Health Organization) (2021)/Eurostat (2019) 5 PM2.5 (ug/m3) Biodiversity loss (SDG 15) Habitats with good conservation status 100 Habitats (%) European Commission (2010)/ EEA/2018 Species with good conservation status 100 Species (%) Biochemical flows (SDG 12) Nutrient phosphate use 0.8 Kgyr −1 P Steffen et al. (2015)/FAOSTAT/ 2019 Nutrient nitrogen use 8.08 Kgyr −1 P Chemical pollution (SDG 12) Waste without hazardous chemicals 100 Waste (%) European Commission (2014)/ Eurostat (2018) Climate change (SDG 13) CO 2 emissions 0.9 Tyr −1 P IPCC (2018), Lucas et al. (2020)/ EXIOBASE (2019) Freshwater use (SDG 6) Water Exploitation Index plus (WEI+) 20 Freshwater used as a % of renewable freshwater resources EEA (European Environment Agency) (2016)/Eurostat (2017) Health of oceans and rivers (SDG 15) Inland water excellent 100 Km (%) EEA (European Environment Agency) (2016)/Eurostat (2019) Coastal water excellent 100 Km (%) Land-change system (SDG 15) Temperate forest 50 Total land area (%) Steffen et al. (2015)/Eurostat (2018) Stratospheric (SDG 15) ozone depletion Free of consumption of ozone-depleting substances (ODS) 100 % (Achieved) UNEP (1987)/EEA/2019 M.R. G´ omez-Alvarez Díaz et al. Ecological Economics 221 (2024) 108189 6 The second criterion was to select the variable that evaluated the performance of those that are least favourable, following Raworth (2017) proposal of creating a just space for all and in the quest for a society which ensures that no one suffers from deficiencies. In this study, we chose to use poverty at work as a measure of the quality of employment and long-term unemployment to identify the expulsion of the population from access to employment. The third criterion was the priority of the variable that refers to the population, as this facilitates an understanding of the indicator and the scope of the deficiency, except for gender equality and work. The fourth criterion was that one variable per dimension should be used, when possible, although all issues are complex and could have been evaluated from different perspectives. 4. Results The results showed that all the countries were within the stratospheric ozone depletion limit. Three countries have 8 transgressions (Cyprus, Greece, and Spain), 17 are within the water use limit (7 deficits), 5 countries (Croatia, Finland, Latvia, Slovenia, and Sweden) do not exceed the land-system-change limit (6 deficits), and Estonia has no atmospheric pollution (5 deficits) (see Table A.1). However, the amount of transgression, both minimum and maximum, was not concentrated within a specific country profile (Fig. 1). If we consider the mean value of the transgressions, this is higher for climate change (1246.6%) and biochemical flows (512%) and lower for chemical pollution (15%) and freshwater use (14.1%). In the case of just space, the initial results show that 100% of the population have adequate conditions of water and sanitation only in Sweden, the Netherlands and Germany, whereas the remaining countries are outside of JS in all the dimensions (see Table A.2). Second, to carry out a more synthetic analysis, we grouped the 16 dimensions according to their descriptions, adapting the Amsterdam proposal (DEAL, 2020) in healthy, connected, trained, and empowered societies. According to the data (Fig. 2), European countries are closer to being healthy societies than in the other three spheres; on average 2.3% of the population did not have access to water and sanitation, 2.6% to health and 5.5% to food, although 8.2% cannot have a house at the right temperature and there is an 18.6% overcrowding rate in housing. These countries, as connected societies, show that few people lack social support (5.3% of the population), but there are more people with problems accessing cultural activities (10%), public transport (19.4%) and high-speed Internet (45.2%). Regarding trained societies in which people can develop individual life projects, the data indicate an average poverty rate of 16.3%, while 23.5% of 15-year-old students lack basic educational skills. In the workplace, the shortage is 5.1%, which is calculated as the average of the long-term unemployment and the inwork poverty rates. Finally, European societies show significant deficiencies in empowerment. On average, 10.3% of the population are dissatisfied with the functioning of democracy, 21.2% are not satisfied with the level of peace and justice, and there is also a perception of high social inequality, which is greater in economic inequality (83.3%) than in social discrimination (15.9%). The empowerment of women stands at 26.4%. This is evaluated as the average salary gap, women earn 39.8% less than men, and lack of political representation (12.9%). A descriptive analysis of the data revealed that the most homogeneous behaviours for the 26 countries corresponded to the biochemical flows, income, and social equality, with more representative means. Except for biochemical flows, gender equality, and social equality, most countries had higher values than the mean in almost all the dimensions. However, 15 of the 26 dimensions had values that were highly centred with respect to the mean. In addition, the environmental dimensions have a higher variation than the social dimensions, being the income, social support and work, the most stable dimensions of the set. Calculating the indices allowed us to summarise the detailed information obtained for each dimension (Fig. 3). The data indicate that no country is located within the limits of the Doughnut. Nor is there a country that is closest to the two limits of the Doughnut simultaneously. Portugal is the country closest to safe space (129.6%) and Ireland is the furthest (479.6%). In relation to just space, Denmark had the lowest EDI (9.1%) whereas Bulgaria had the highest EDI value (26.8%). Therefore, Table 2 Social boundaries. Dimension Variable(xi)and Boundary (zi) = 100% population aged 16 years over (unless otherwise stated) Data Source/year Connectivity (SDG 11) Households with high-speed internet coverage Eurostat (2019) Culture (SDG, 11) Population reports that non-participation in cultural or sport activities is not for financial reasons or because of there being none in the neighbourhood Eurostat (2015) Education (SDG 4) 15-year-old students achieve level 2 (‘basic skills level’) on the PISA scale for the subject of reading (zi=15-year-old students) Eurostat (2018) Energy (SDG 7) Households are able to keep their home adequately warm Eurostat (2019) Food (SDG 2) Population with appropriate meals: the ability to eat a meal with meat, chicken, fish (or vegetarian equivalent) every second day Eurostat (2019) Gender equality (SDG 5) Proportion of women in national parliaments and national government (zi=50% of seats) Eurostat (2019) Gross hourly earnings of females (zi=earnings of males) Eurostat (2019) Health (SDG 3) Population covers needs for medical care without problems due to ‘Financial reasons’, ‘Waiting list’ or ‘Too far to travel’ Eurostat (2019) Housing (SDG 11) Population without overcrowding Eurostat (2019) Income (SDG 1) Population without risk of poverty Eurostat (2019) Peace and Justice (SDG 16) Population reports that they do not face the problem of crime, violence, or vandalism in their local area Eurostat (2019) Population does not perceive the independence of the courts and judges in a country as “being very bad” or “fairly bad” Eurostat (2019) Political voice (SDG 16.7) Population does not answer “not at all satisfied” with the way democracy works in the country Eurobarometer 92.2, 2019 Social equality (SDG 10) Population thinks differences in people’s incomes in the country are not today too high Eurobarometer 471, 2017 Population affirms that they do not feel discriminated against Eurobarometer 493; 2019 Social support Population has someone to ask for help Eurostat (2015) Transport (SDG 11) Population without great or very great difficulty in accessing public transport Eurostat (2012) Water and sanitation (SDG 6) Population with bath or shower and indoor flushing toilet in their household Eurostat (2019) Work Total active population without long-term unemployment (zi=Total active population) Eurostat (2019) Employees not at-risk-of-poverty (zi=Total employees) Eurostat (2019) M.R. G´ omez-Alvarez Díaz et al. Ecological Economics 221 (2024) 108189 7 the dispersion in the absolute distance to the SS was much greater than that in the case of the JS, and in relative terms, the relationship between the maximum and minimum values is 3.7 and 3, respectively. Likewise, the average value for the EU is 221.6% (EDI) and 16.8% (SDI). A hierarchical cluster analysis was performed to identify possible clusters according to the values of the two indices: EDI and SDI. Ward’s method was used for the extraction procedure with the square of the Euclidean distance as the measurement. This guarantees the minimisation of differences within clusters. The dendrogram (Fig. 1, Appendix B) reveals the presence of six well-defined groups. An analysis of the error bar graphs shows, with a 95% confidence level, that there are significant differences between the means of the groups in both indices (Fig. 4), and the Kruskal-Wallis test demonstrates that the indices created provide significant differences among the groups (Table 3). a. Kruskal Wallis Test . b. Grouping Variable: Cluster. Table 4 below lists the countries grouped in each cluster and their average values in each of the indices analysed. The six clusters (Fig. 5) have been ranked e for each index, based on the mean values obtained Fig. 1. Distance to ecological dimensions (%). Average distance for UE-27, countries with maximum and minimum values. Fig. 2. Distance to social dimensions (%). Average distance for UE-27, Countries with maximum and minimum values. M.R. G´ omez-Alvarez Díaz et al. Ecological Economics 221 (2024) 108189 8 (Table 4). The first category corresponds to the best result (Best) or minimum value, and the following are: fairly good, good, bad, fairly bad, and worst result. From the data collected in Table 4 and Fig. 5, the following results are derived. Firstly, no cluster can be deemed the best in comparison to others because its social deficit and environmental deficit are the lowest. Secondly, we identify two clusters (A and B) characterised by values lower than the European average for both spaces. Two clusters (C and E) have average values higher than the European average, and Group D exhibits a significant social deficit but is closer to SS than the European average, with Group E exhibiting the opposite behaviour. Concerning groups with deficits below the EU average, Group A has the smallest environmental distance (166.67%) of the six, although its social distance widens by 3.25 p.p. compared to Group B. Conversely, Group B performs better in the social context (11.54%), but the average EDI value exceeds that of the group by 45.62 p.p. Both groups share, from a geographical perspective, the presence of Nordic, Central European, and South-Western countries. Additionally, Group A includes Fig. 3. EDI and SDI for European countries and Average Deficit UE-27. Fig. 4. Error bars of six socio-environmental groups for EDI and SDI. Table 3 Kruskal-Wallis test. SDI EDI_Consumption Kruskal-Wallis H 21.834 20.325 df 5 5 Asymp. Sig. <0.001 0.001 M.R. G´ omez-Alvarez Díaz et al. Ecological Economics 221 (2024) 108189 9 Slovenia, belonging to the historical trajectory of Eastern European countries. Both include countries of very different sizes, political tradition, economic specialisation, geographical situation and per capita income. However, most of them belong to Western Europe. On the other hand, Groups C and E have an average deficit above the European value for both areas. Group C, comprising Bulgaria and Greece, is socially the worst in the EU, with a value of 25.36%. Group E, consisting of three Eastern European countries (Czechia, Hungary, Poland), three former Russian republics (Estonia, Latvia, and Lithuania), and Cyprus, has a lower social deficit (18.02%), but its environmental deterioration is 242.82%, compared to 228.38% for Group C. Finally, Group D has an EDI only 0.9 p.p. higher than the smallest average distance of Group 1, but its social deficit is one of the most significant, with a value of 21.97%, 3.3 p.p. lower than that of Group C, which is the worst. This group consists of Croatia, Romania, Slovakia, and Italy. Conversely, Group F (Ireland, Luxembourg) has an SDI very close to the minimum value of Group 2, separated by 1.7 p.p., but is by far the group furthest away from the safe space. 5. Discussion and conclusions The EU is considered a reference for the standard of living and protection of social rights. However, despite this “privileged” situation in terms of average world income and the economic growth recorded throughout the 20th century until the present day, notwithstanding the COVID crisis, this has not translated into a society that is free from deficiencies nor one that guarantees the well-being of future generations or of nature. In fact, no countries fall within the doughnut, although the results are very diverse. Regarding safe space, all countries crossed the same four thresholds, and the majority (17 countries) crossed seven of them. Our results indicate that, from an environmental perspective, European policies must impact all the dimensions, especially CO 2 emissions and biochemical flows. The average deficits for the EU in these dimensions are particularly of concern, reaching 1245% for the former and 512% for the latter. In addition, national policies are required for specific dimensions, such as water scarcity in Spain. Concerning the just space, all countries had deficits in all dimensions, except for the three countries that had access to water and sanitation for 100% of the population. There are significant deficiencies in housing, treated as a commodity despite being a right. The same can be said for energy, Internet access, and public transport. Likewise, poverty is high, and working does not guarantee emancipation. Additionally, there are educational deficiencies among society’s younger members. Results concerning satisfaction with peace and justice, economic inequality, and gender are worrying. The use of synthetic indices allows us to affirm that Portugal is the country closest to the safe space (129.6%) and Denmark to the just space (9.1%). Ireland is the furthest from to the safe space (479.6%) while Bulgaria is furthest from the just space (26.8%). No country falls within the SJS, and no country has the lowest EDI and SDI values. A cluster analysis was conducted to find patterns. The results identify Table 4 Clusters: Countries, average values of EDI and SDI. A B C D E F UE Austria Belgium Bulgaria Croatia Cyprus Ireland France Denmark Greece Italy Czechia Luxembourg Germany Finland Romania Estonia Portugal Netherlands Slovakia Hungary Slovenia Spain Latvia Sweden Lithuania Poland AVGEDI 166.67 212.30 228.40 167.58 242.83 436.3 221.6 AVG SDI 14.78 11.54 25.35 21.98 18.03 13.25 16.8 Fig. 5. Clusters labelled according to their position relative to the best-positioned cluster (minimum value EDI, minimum value SDI). M.R. G´ omez-Alvarez Díaz et al.