Carbon taxes on consumption: Distributional implications for a just transition in the EU

Maier, Sofía; De Poli, Silvia; Amores, Antonio F. Working Paper Carbon taxes on consumption: Distributional implications for a just transition in the EU JRC Working Papers on Taxation and Structural Reforms, No. 9/2024 Provided in Cooperation with: Joint Research Centre (JRC), European Commission Suggested Citation: Maier, Sofía; De Poli, Silvia; Amores, Antonio F. (2024) : Carbon taxes on consumption: Distributional implications for a just transition in the EU, JRC Working Papers on Taxation and Structural Reforms, No. 9/2024, European Commission, Joint Research Centre (JRC), Seville This Version is available at: https://hdl.handle.net/10419/306595 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Maier, S., De Poli, S., Amores, A.F. 2024 JRC Working Papers on Taxation and Structural Reforms No 9/2024 Carbon taxes on consumption: distributional implications for a just transition in the EU JRC138420 Sevilla: European Commission, 2024 © European Union, 2024 The reuse policy of the European Commission documents is implemented by the Commission Decision 2011/833/EU of 12 December 2011 on the reuse of Commission documents (OJ L 330, 14.12.2011, p. 39). Unless otherwise noted, the reuse of this document is authorised under the Creative Commons Attribution 4.0 International (CC BY 4.0) licence (https://creativecommons.org/licenses/by/4.0/). This means that reuse is allowed provided appropriate credit is given and any changes are indicated. For any use or reproduction of photos or other material that is not owned by the European Union permission must be sought directly from the copyright holders. How to cite this report: European Commission, Joint Research Centre, Maier, S., De Poli, S. and Amores, A.F., Carbon taxes on consumption: distributional implications for a just transition in the EU, European Commission, Sevilla, 2024, JRC138420. This document is a publication by the Joint Research Centre (JRC), the European Commission’s science and knowledge service. It aims to provide evidence-based scientific support to the European policymaking process. The contents of this publication do not necessarily reflect the position or opinion of the European Commission. Neither the European Commission nor any person acting on behalf of the Commission is responsible for the use that might be made of this publication. For information on the methodology and quality underlying the data used in this publication for which the source is neither Eurostat nor other Commission services, users should contact the referenced source. The designations employed and the presentation of material on the maps do not imply the expression of any opinion whatsoever on the part of the European Union concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. EU Science Hub https://joint-research-centre.ec.europa.eu Executive summary  Over the past years, a consensus has been growing regarding, first, the need to reduce greenhouse gas (GHG) emissions to limit global warming and second, that carbon pricing is a key tool to achieve this. Yet, carbon taxes often face public resistance, as evidenced by the yellow vest movement, due to their regressivity and perceived unfairness.  This study uses micro-simulations to assess the distributional effects of various hypothetical carbon taxes on household consumption across the 27 EU Member States. We build on the conventional literature by simulating more progressive tax structures, aiming for improved distributional outcomes without relying on compensatory transfers, which are often of limited practical feasibility.  For this purpose, we extend the EUROMOD tax-benefit model with household’s GHG footprints, covering both direct (from fuel use, such as heating or driving) and indirect (from production and trade) emissions from the EXIOBASE multi-regional input-output model. This Green EUROMOD extension allows us to evaluate alternative carbon tax reforms (e.g., with progressive tax designs) within a coherent and standardized framework using EU-HBS and EU-SILC data across all 27 EU Member States.  Our estimates show that household GHG emissions vary significantly across countries and income groups. Notably, only the lowest-income 10% of the EU population exhibits consumption patterns that align with what the literature considers sustainable and consistent with the goals of the Paris Agreement, with per capita emissions averaging approximately 2.5 tonnes of CO₂ equivalent (tCO₂e) per year.  Simulations of a flat carbon tax set at EUR 80 per tCO₂e suggest that this measure would exacerbate inequality across all 27 EU Member States. The tax's regressive pattern arises from the greater tax burden on lower-income households, who allocate a larger proportion of their income to GHG-intensive goods such as food and energy. The extent of this inequality increase varies by country, largely due to the varying tax burdens, which would range from 2% of household disposable income in France and Sweden to up to 9% in Hungary, Poland, and Greece.  This flat carbon tax could generate revenues of up to EUR 208 billion annually. In line with previous studies, redistributing this revenue via lump-sum cash transfers would completely offset the tax's inequality effects.  Notably, we also show that a carbon tax with allowances (i.e., where the first 2.2 tons of tCO₂e are tax exempted), and to a lesser extent, a carbon tax with rate differentiation (across product groups, such as with VAT), perform relatively well in preventing large inequality-increasing effects, without relying on compensatory measures. As such, these alternative tax designs could enhance public acceptability and policy feasibility of taxing household carbon footprints for a just transition. Carbon taxes on consumption: distributional implications for a just transition in the EU Sofía Maier1,3, Silvia De Poli1,2,*, and Antonio F. Amores1 1European Commission, Joint Research Centre, Seville, Spain 2Complutense University of Madrid, Spain 3Antwerp University, Belgium *Corresponding author: sdep[email protected] September 16, 2024 Abstract Carbon taxes on household consumption can simultaneously increase public funding and promote greener consumption habits, an appealing combination for the just transition plans of the European Union (EU). However, concerns about equity and public support pose challenges. This paper assesses the distributional and budgetary effects of various designs for an EU-wide hypothetical carbon tax on households consumption. To this end, we extend the EU tax-benefit microsimulation model, EUROMOD, with greenhouse gas (GHG) emissions data from input-output tables and estimate households’ carbon footprints. We show that a carbon tax on households GHG emissions would be regressive, thereby inequality-increasing. This is primarily due to the low income elasticity of highly GHG-intense necessity goods, such as food and heating, which represent larger shares of income at the bottom of the distribution. Still, we demonstrate that this inequality-increasing impact can be offset with compensatory cash transfers (though these may be challenging to implement), and at least partially reverted with more progressive (and presumably feasible) tax designs, including rate differentiation by products and tax allowances. Keywords: Progressive carbon tax, Just transition, Footprints, Inequality, European Union JEL codes: H23, Q52, D31, C67 Acknowledgements: This paper was developed within the AMEDI projects of the Directorate General for Employment and Social Affairs of the European Commission. We extend our sincere gratitude to Salvador Barrios and Richard Wood for their invaluable support to the Green EUROMOD project. We are particularly indebted to Gerlinde Verbist, whose presentation at the JRC Fiscal Policy Workshop in 2020 inspired us to pursue this research, and for her feedback on the final draft. Special thanks are due to José Antonio Ordoñez, David Klenert and Salvador Barrios for their insightful feedback. We would also like to thank the EUROMOD community for their ongoing development of this powerful micro-simulation tool and its underlying harmonized micro-datasets. We are grateful to Matthias Weitzel for providing information from JRC-GEM-E3 on scenarios for the new 2040 targets. Finally, we would also like to thank the participants of the 2023 EUROMOD Annual Workshop in Seville for their valuable contributions. We also want to thank Ilda Dreoni and Hannes Serruys for their work on statistical matching. 1 Introduction Over the past years, a consensus has been growing regarding, first, the need to reduce greenhouse gas (GHG) emissions to limit global warming (Stiglitz et al.,2017;Pearce,1991),1and second, that carbon pricing is essential to achieve it (Hepburn et al.,2020). Based on Pigou (1920)’s original idea of incorporating externalities into prices, an entire body of literature has highlighted the theoretical attractiveness of carbon pricing in terms of efficiency (Hepburn et al.,2020;Hassler et al.,2016;Nordhaus,1991), and provided empirical evidence on their positive causal effects on GHG emissions reduction (see, e.g., Dechezleprêtre et al.,2023;Andersson,2019;Martin et al.,2016;Lin and Li,2011). Carbon pricing can be implemented downstream at the production level, or upstream, directly to consumers.2The most common practices follow the first approach, primarily through cap-and-trade schemes known as Emission Trading Systems (ETS) and carbon taxes. According to the World Bank’s Carbon Pricing Dashboard, about 23% of global GHG in 2023 were covered by carbon pricing, 18% by ETS and 5% by carbon taxes (World Bank,2023). ETS set a cap on total GHG emissions and allow firms to trade emission permits, with the price of carbon determined endogenously by the market. In contrast, carbon taxes fix the price and do not guarantee any level of GHG reduction. Unlike ETS or other carbon pricing mechanisms targeting producers, consumption-based carbon taxes do not exert carbon leakages or competitiveness problems (Parry et al.,2022,Nachtigall et al.,2022,Böhringer et al.,2021,Hepburn et al.,2020,Böhringer et al.,2017). More in general, for revenue-raising purposes, environmental taxes are often preferred over other types of taxes, such as those on labour, because they address a market failure and are therefore considered less distortionary (Barrios et al.,2013;Bovenberg,1999). In this paper, we offer fresh insights into the distributional (and budgetary) effects of a hypothetical carbon tax on households’ carbon footprints across the 27 EU Member States. By combining data from household surveys with and input-output tables, we analyze the progressivity and redistributive effects of various tax designs and compensatory measures. In particular, we discuss whether a progressive tax structure could enable governments to achieve a positive redistributive impact without relying on revenuerecycling compensatory measures, which are often challenging to implement on political and practical grounds. Carbon taxes are a particularly appealing tool in the current policy context because they can, in principle, simultaneously boost the de-carbonization efforts the European Union (EU) is pursuing to achieve its objective of becoming the first carbon-neutral economy by 2050,3while improving public finances to implement other transition-related policies. Despite the strong reduction in per capita GHG emissions of the past decades -dropping from 11 to 7 tons of CO2e, according to the European Environmental Agency),4the carbon footprints of EU households are among the top in the world and quite above the 1This has led, among others, to the first-ever international binding agreement in Paris 2015, as a result of the Conference of Parties (COP) 21 (see: UNFCCC,2015). 2In the first case (the so-called “producer responsibility approach”), producers are held responsible only for all the GHG emissions generated directly by them. In the second (or “consumer responsibility approach”), based on the footprint concept, consumers are held responsible for all the GHG emissions generated along the value chain to produce what they consume. Figure A.1 in the Appendix presents an overview of these two strategies. 3To meet its 2050 climate targets, the EU launched the “Green Deal”, which includes carbon-pricing initiatives like the Carbon Border Adjustment Mechanism (CBAM) and the extension of the Emissions Trading System (ETS2). The CBAM, a tax on carbon-intensive imports, will start in 2026, while ETS2, expanding the current ETS to buildings and road transport, begins in 2026/27. Additionally, the Fit-for-55 package, aimed at a 55% GHG emissions reduction by 2030, proposes revising the Energy Taxation Directive (ETD) to update minimum tax rates based on CO2content, a proposal that is still under discussion. 4https://www.eea.europa.eu/data-and-maps/data/data-viewers/greenhouse-gases-viewer 1 “sustainable” levels compatible with international agreements related with global warming (Chancel,2022, Chancel et al.,2022,Gore,2021,Ivanova et al.,2016, and Girod et al.,2014). Furthermore, while the EU has nearly double the global average of 23% of GHG emissions covered by carbon pricing (European Commission,2022), more than half of its GHG emissions remain tax-exempt. Yet, carbon taxes can be quite unpopular, as the yellow vest movement put in clear evidence (Douenne and Fabre,2022,Rubin and Sengupta,2018). In part, this acceptability problem relates to the regressivity and the perceived unfairness of these taxes (Köppl and Schratzenstaller,2023;Andersson,2019;Klenert and Mattauch,2016,Wang et al.,2016), on top of the lack of dissemination or information on the actual effect of these policies (Douenne and Fabre,2022;Klenert et al.,2018;Murray and Rivers,2015). The European Commission has recognized this challenge, as emphasized by its President during the adoption of the European Green Deal Communication in 2019: “This transition will either be working for all and be just, or it will not work at all”.5Still, as documented by the meta-analyses of Wang et al. (2016) and Ohlendorf et al. (2021), evidence on the regressivity of carbon taxes is limited and shows mixed results. In this context, our analysis makes two important contributions to the literature, as outlined below. First, we provide new evidence on the distributional and budgetary effects of a hypothetical EU-wide consumption-based carbon tax, covering all 27 EU Member States. To this end, we extend the EU tax-benefit microsimulation model, EUROMOD,6by integrating greenhouse gas (GHG) emissions data derived from input-output tables. With this “Green EUROMOD” extension we estimate households’ carbon footprints, accounting for emissions linked to both domestic production and imports, including those generated through trade and transport, at the very detailed product level. While the literature on carbon pricing has grown rapidly in recent years (see overviews by Wang et al.,2016,Ohlendorf et al.,2021 and Döbbeling-Hildebrandt et al.,2024), most studies tend to be countryor product-specific and employ a variety of methodologies, making cross-country comparisons difficult. Our study leverages detailed, harmonized data from multiple sources, addressing gaps in previous research. Second, while much of the existing literature focuses on compensatory measures like cash transfers or labor tax reductions for households affected by carbon taxes, our study explores alternative tax designs that aim for more progressive outcomes. This approach avoids over-reliance on compensatory measures, which can be costly to administer and may face low uptake, potentially undermining public support for the initiative. The most related study to ours is Feindt et al. (2021), which reports progressive patterns (at the countrylevel) from a simulated carbon tax in 23 EU countries. We extend and depart from this paper in several ways, by expanding country coverage, using more updated data, considering all GHG emissions for the tax base and not only CO2; and more fundamentally, varying the type of simulated tax and using incomes from SILC instead of expenditures to assess the tax burden, regressivity and distributional effects. This directly influences the less regressive patterns they identify, as discussed also by Maier and Ricci (2024) and Linden et al. (2024). We begin by simulating a flat carbon tax of EUR 80 per tonne of CO2e on household consumption. This tax covers direct GHG emissions from household fuel use as well as emissions embedded throughout the supply chain of goods and services, including those from abroad. The chosen tax rate of 80 EUR per tCO2e is consistent with the prevailing price under EU ETS and aligns with the emissions reduction 5See, e.g., speech at: https://ec.europa.eu/commission/presscorner/detail/fr/speech_19_6749. 6EUROMOD is a publicly available model developed by an international community of scientists and policymakers since the 1990s. The recent extensions to cover consumption taxes and GHG emissions are not yet public, but shall be in the future. More information: https://euromod-web.jrc.ec.europa.eu/. 2 targets stipulated by the Paris Agreement, as suggested by Stiglitz et al. (2017). We then discuss to what extent revenue-neutral compensatory measures could improve the distributional outcomes. Finally, we simulate two alternative carbon tax regimes, where we try a less regressive design. With this, we avoid relying on a compensatory measure. Specifically, we simulate a scenario with what we call “green allowances” (tax exemptions below a threshold, similar to the global progressive tax proposed by Chancel and Piketty,2015) and another consisting of different carbon tax rates across product categories, like the different rates of VAT. To simulate these carbon tax scenarios, we extend the EUROMOD model with carbon footprints, following two main steps. First, we statistically match two EU-level harmonized household surveys, EU-SILC (EU Statistics on Living Conditions) and HBS (Household Budget Surveys), which provide detailed data on household income and consumption, respectively. Second, we impute GHG emissions per euro spent from EXIOBASE at a highly disaggregated product level into our HBS-SILC matched files.7This Green EUROMOD extension allows us to simulate various carbon taxes, considering each country’s fiscal policies and evaluating their distributional effects based on household disposable income. By using SILC-based incomes, our results can be directly compared to the EU’s official figures on income inequality and poverty published by Eurostat. While our analysis does not account for behavioral responses and should be interpreted as “overnight” effects, the literature suggests that demand responses to price changes in necessity goods (e.g., food and energy) are typically small in the short run (Linden et al.,2024;Labandeira et al., 2017). Therefore, we offer a reliable insight into the expected short-term impacts of these tax reforms. Our results suggest that this simulated carbon tax would represent, on average, 5% of household disposable income, a tax burden that ranges from about 2% in Sweden to almost 9% in Greece. Generally, the burden is more pronounced in lower-income countries, largely due to the higher proportion of income these households allocate to expenditures compared to their wealthier counterparts. This regressive pattern is also evident within countries, where the tax burden disproportionately affects lower-income households. On average, the tax would represent 7.7% of disposable income in the poorest income quintile, more than double the 3.6% tax burden faced by the richest quintile. This pattern leads to a negative distributional effect within all 27 EU countries, with no exceptions. This is particularly pronounced under the flat tax scenario (with no allowances, rate differentiation nor compensatory measures), with the Gini coefficient increasing by up to 0.02 points in some countries.8 This flat carbon tax of EUR 80 per tCO2e could generate up to EUR 208 billion annually for the EU, more than double the projected budget for the Social Climate Fund. Although the increase in inequality driven by the simulated tax can be fully reversed through revenue-neutral cash transfers —whether designed at the national or EU level— the practical feasibility of such transfers remains a significant challenge. Importantly, our findings also show that, even without revenue-recycling cash transfers, the inequality effects can be partially mitigated through alternative, more progressive tax designs, such as implementing tax allowances or varying tax rates across different products. The rest of the paper is organized as follows. In the next section, we briefly present the EU-level GHG emissions reduction targets and describe the current carbon pricing initiatives to understand where we stand in the baseline. In Section 3, we present the methodology and data, while in Section 4, we present the results from our analysis. In Section 5, we discuss the scope of our results in light of the assumptions 7EXIOBASE data is publicly available at their website (https://www.exiobase.eu/), and documented by Stadler et al. (2018), Tukker et al. (2013), Wood et al. (2014). 8For context, this figure can be compared to the average redistributive effect of consumption taxes in the EU, and it is four times greater than the redistributive effect of current energy taxes (Amores et al.,2023c). 3 we make, as well as the relative advantages and disadvantages of each of these simulated scenarios. At the end, Section 6concludes. 2 GHG targets and carbon prices in the EU The policy relevance of the estimated budgetary and distributional consequences of carbon taxes depends on the feasibility of our simulated scenarios. The parameters that we use (e.g., the carbon price that we impose on each tCO2e, as well as the thresholds that we use for tax allowances) have to be consistent with the current context and, in particular, with the carbon pricing practices and EU’s climate policy agenda. Therefore, before introducing the data and methodological strategy in Section 3, we provide here a brief overview of the main international and EU-level climate targets related to GHG emissions reduction, as well as the shares of GHG that are currently covered by some carbon pricing. GHG targets The “Paris Agreement”, signed in the 2015 Paris Conference of the Parties (COP) 21 of the UNFCCC, was the first-ever internationally binding agreement. It established a long-term goal to limit the increase in global average temperature to well below 2°C above pre-industrial levels by the end of this century and to pursue efforts to limit this increase to 1.5°C (UNFCCC,2015). Since then, climate policy agendas have considered and pushed forward several intermediate targets related to GHG emissions reduction. According to the United Nations’ Intergovernmental Panel on Climate Change (IPCC), to reach the 1.5°C goal, GHG emissions must peak before 2025 at the latest and decline by 43% by 2030 (with respect to the 1990s). Other studies have estimated the “sustainable levels” of consumption compatible with these targets. For example, Gore (2021) and Ivanova et al. (2016) suggest that per capita GHG emissions should be around 2.2 tCO2e to be consistent with this 1.5°C goal, similar to the 1.9 tCO2e estimate of Chancel (2022). In this context, the EU has committed to becoming the first carbon-neutral economy (i.e., zero net GHG emissions) by 2050. To achieve this, it has set an ambitious climate agenda under the road-map of the Green Deal, launched in 2021, with specific targets towards 2030 (55% net GHG emissions reduction with respect to 1990) and -more recentlyalso so for 2040 (90% reduction).9To be consistent with carbonneutrality and the new 2040 targets, in-house estimates from the European Commission suggest that per capita GHG emissions have to be reduced to about 1.7 in 2040, starting from 4.9 in 2025.10 GHG emissions covered by carbon pricing The World Bank provides a publicly accessible dashboard documenting the type of carbon price initiatives implemented worldwide and the share of GHG emissions they cover. According to World Bank (2023), referring to data from 2023, about 23% of global GHG are covered by some form of carbon pricing: 18% by ETS, and 5% by carbon taxes. The carbon price per tCO2differs widely between countries and carbon price instruments, from US$0.46 to 167. In this context, the EU has one of the largest GHG coverage (about 45%, mostly covered by the EU ETS). Moreover, a few European countries (especially some Scandinavian countries) have among the highest carbon prices in the form of specific carbon taxes. 9More information: https://climate.ec.europa.eu/eu-action/climate-strategies-targets/ 2050-long-term-strategy_en. 10These values were obtained from the JRC-GEM-E3 modelling that informed the impact assessment underlying the European Commission’s proposal for a 2040 climate target (European Commission,2024). 4 ∆Zg=Zag −Zbg (6) To estimate the expenditures after the tax under compensating variation (Zag, equation 7), we subtract from baseline expenditures (Zbg) total simulated carbon taxes. These taxes are obtained by multiplying the tax rate (τ, defined in EUR per ton) by the household carbon footprint (F PGHGg). In that stage is where we assume a full pass-through (inelastic demand), as the quantities consumed (that enter the tax base, through the carbon footprints) are fixed in the baseline. Zag =Zbg +τFPGHGg(7) At the end, AY d gis obtained by subtracting these simulated tax liabilities from from baseline equivalised household disposable incomes. AY d g=Yd g−∆Zg=Yd g−FPGHGgτ(8) To assess the distributional effects under this compensating variation welfare approach we then compare this after-tax income and the distribution of the tax with baseline disposable income with two standard indicators: the Gini coefficient -based on the concentration indexand the S8020 ratio -the difference between the income shares of the top and bottom 20%. Although this serves to provide a welfare magnitude of the effect in the short run, the main motivation for the implementation of carbon taxes is to reduce demand, especially of carbon-intensive goods and services. Since we do not model it in this paper (see Section 5), our ‘overnight’ results should be interpreted as short-run effects and probably upper bounds of the overall effect (as the negative demand response should actually reduce both the tax burden on households as well as the revenues from tax collection). Nevertheless, the literature suggests that demand responses to changes in prices of necessity goods (i.e., food and energy, which are precisely among the most carbon intensive) tend to be small, especially in the short-run (Linden et al.,2024;Labandeira et al.,2017). 4 Results In this section, we present the main results of our study. We start, in sub-section 4.1, with a brief descriptive analysis of the distribution of carbon footprints across households and countries. Then, in sub-section 4.2, we present the results from the simulation of alternative carbon tax designs. 4.1 Carbon footprints - EU households Figure 1shows the distribution of per capita GHG footprints (tCO2e per year) across income and expenditure quintiles. These quintiles are based on each country’s distribution, with the EU average calculated as a weighted population average. As illustrated in Figure 1, individuals in high-income households (within countries) have much larger carbon footprints than those in low-income households. On average, our estimates for 2019 suggest that the average GHG embedded in the consumption of the top 20% (fifth quintile, Q5) is slightly above 8 tCO2e, doubling that of the bottom 20% (first quintile, Q1). The Q5/Q1 gap is even larger -almost five 11 timeswhen individuals are ranked by their equivalised expenditures (expenditures quintiles, right-hand side). Three consumption categories -residential energy, transport, and foodare responsible for most of the GHG emissions. The predominance of these categories is well-documented in the literature. Not coincidentally, the ETS2 covers the first two groups. This can be explained by the high level of expenditure on these consumption categories and their high GHG intensity. Figures B.1 and B.2 highlight the large heterogeneity across countries in expenditures and GHG emission intensities by type of consumption category. While this pattern holds true across all income and consumption quintiles, emissions from food and energy remain relatively constant, whereas GHG emissions from transport substantially increase among high-income and high-consumption households. Figure 1: Per capita GHG emissions across income and expenditure quintiles (EU average, 2019) Note: GHG footprints (tCO2e) based on environmental extensions from EXIOBASE (2019), EU HBS 2015 and SILC 2015 and EUROMOD tax-benefit rules of 2019. EU average: population weighted. Consumption categories at the most aggregate level (COICOP two digits) - for illustrative purposes, we group health with education (COICOP 06 and 10), communication with recreation and culture (COICOP 08 and 09), housing with furnishings, household equipment and routine maintenance of the house (COICOP 04 and 05), excluding energy. To better visualize the cross-country dispersion of these per capita emissions across the income distribution, Figure 2presents a map with per capita GHG emissions at the bottom (Q1) and at the top of the income distribution (Q5). The same information per country and type of good is provided in Figure B.3 in the Appendix B. At the bottom of the distribution (left panel of Figure 2), some Eastern Countries and Spain exhibit very low emissions (less than 3 tonnes per capita). In contrast, some Western central 12 countries (Germany, the Netherlands, Luxembourg, Austria) and Ireland have more than 8 tones per capita CO2e of GHG emissions. As expected, in quintile 5 (right panel of Figure 2), per capita GHG emissions are much higher, with about half of the countries showing more than 10 tonnes per capita emissions. Here we can also appreciate that the cross-country dispersion of the carbon footprints of the top 20% income groups is higher than the one of the bottom 20%. Figure 2: Per capita GHG emissions in quintile 1 (Q1) and quintile 5 (Q5), 2019 Q1 Q5 (13,17] (10,13] (8,10] (5,8] (3,5] [0,3] Note: GHG footprints (tCO2e) based on environmental extensions from EXIOBASE (2019), EU HBS 2015 and SILC 2015 and EUROMOD tax-benefit rules of 2019. Overall, we can see that most of the first quintiles across the EU countries have, on average, carbon footprints that are above “sustainable levels” (of 2-2.5 tons a year). In fact, only if we rank households from the poorest to the richest (in terms of incomes) in the EU and look at the first decile (the 10% with the lowest incomes, regardless of the country of residence) We observe a carbon footprint that is aligned with this threshold. Consumption taxes are generally regressive (Maier and Ricci,2024,Decoster et al.,2010). Their overall redistributive effect depends on two main features: the average tax rate or tax burden, and how these tax liabilities are distributed across income groups, see, e.g., Kakwani (1977) and Reynolds and Smolensky (1977). We now explore the second dimension (distribution of GHG emissions across income) using concentration indices, as this can help us interpret the results from the simulation and provide inputs for the design of a progressive carbon tax (scenario 3). Figure 3presents the distribution of concentration indices -see equation 3across countries and aggregate consumption categories. These are ranked from left to right by the median concentration index of the 13 27 EU countries. GHG emissions from food, energy, alcohol and tobacco show the lowest concentration indices, suggesting that they are more equally distributed across income groups than the rest of the consumption categories. Conversely, transport, other, clothing and footwear, communication, recreation and culture, restaurants, and hotels are mostly concentrated at the top of the distribution. These results feed the design of our third tax scenario, where we look at improving the progressivity of the carbon tax by imposing different rates across consumption categories. To make carbon taxation more progressive with this strategy, tax rates should be higher among those products with higher concentration indices (GHG emissions vs incomes). Still, the improvement of the overall distributional effect (i.e., to minimize the inequality-increasing effects) depends also on whether those products with higher tax rates represent large shares of the total carbon footprints. For instance, in most of the countries, we could expect an improvement in the distributional results driven by transport, which has both a high concentration index (and as such, it will be taxed at higher rates in our simulations) while it also represents a large share of the tax burden. Figure 3: Concentration index: GHG emissions by consumption category (EU 27 countries, 2019) Note: GHG footprints (eq-CO2) based on data from EXIOBASE (2019), HBS 2015 and EU-SILC 2015, with values uprated and tax-benefit rules of 2019 applied with EUROMOD. Concentration index of equivalised GHG emissions with respect to equivalised household disposable income. The line within the boxes indicates the median across the 27 EU Countries. The boxes represent the range of the percentiles 25 and 75 (inter-quartile range). 14 4.2 Simulation results We now present the results from our simulations. We start with the results from scenario 1, where we impose a flat CO2tax on consumption, and discuss the overnight budgetary and redistributive effects both without and with revenue-recycling compensatory measures. Next, we move to scenarios 2 (green allowances) and 3 (different tax rates by products), where we try to make the carbon tax less regressive by design (without compensatory measures). Before presenting the distributional effects of our simulated scenarios, Table 3provides an updated version of Table 2, with an overview of the estimated budgetary effects (aggregate tax revenues). Additionally, it provides the implicit carbon tax for those scenarios where this variable is endogenously determined (scenarios 2.2 and 3.1). Table 3: Simulated scenarios: price and budget Scenario Name Description Carbon price Budget* (EUR/tCO2e) (EUR bn) S1 Flat Tax on GHG emissions S1.1 Flat80 No compensation 80 208 S1.2 . . . + cnt compensation Country specific compensation 80 0 S1.3 . . . + EU compensation EU compensation 80 0 S2 Green allowances Tax on GHG emissions with allowances S2.1 GA 2.2 Exempted 2.2 tCO2e 80 135 S2.2 GA 2.2 - FB Exempted 2.2 tCO2e (FB) 145** 208 S3 Different GHG tax rates by product S3.1 Prog40 Three rates (40, 80, tph) 40-144** 208 Note: FB = fixed budget - equivalent to scenario 1.1. *Budget effect for the whole EU27. Detailed results by country are reported in Table 4for scenario 1, and in the Appendix Bfor the rest. EU simple average of the carbon tax (country-specific values in Table B.5). 4.2.1 Results from scenario 1: hypothetical 80 EUR CO2flat tax Let us start with the simulated flat carbon tax of EUR 80 (per tCO2e) for the 27 EU countries. In the horizontal axis of Figure 4, we plot the average tax rate, or “tax burden” (i.e., total tax liabilities expressed as a share of disposable income). In the vertical axis we plot the redistributive effect, measured as the change in the Gini coefficient before and after the simulated carbon tax. The tax burden ranges from about 2% of household disposable income in Sweden (SE), France (FR) and Denmark (DK) to 9% in Hungary (HU), Greece (EL) and Poland (PL). In general, we observe a larger tax burden in lower-income countries compared to higher-income ones. While, in fact, higher income countries consume more and have larger per capita carbon footprints, (and as such, they end-up paying more in absolute terms), the relative weight of the carbon tax on incomes is smaller. This is primarily because of their lower income shares of consumption expenditures. Additionally, differences in consumption patterns contribute to this variation: lower-income countries tend to have a higher proportion of expenditures allocated to GHG-intensive categories such as food, heating, and transport. These categories also account for a greater share of total GHG emissions in these countries. As illustrated in Figures B.1 and B.2 (Appendix B), food, heating and transport ranges from about one-third of household consumption 15 expenditures in Luxembourg, Netherlands and Finland to almost two-thirds in Estonia, Croatia and Romania. In the same figure we can observe that the tax leads to negative redistributive effects in all EU 27 countries. We measure it as the change in the Gini coefficient of equivalised disposable income before and after the tax (by “after” we mean incomes that would be needed to purchase the baseline consumption basket with the new prices/taxes, following the compensating variation welfare approach discussed at the end of section 3). A positive value suggests that the Gini coefficient has increased after the tax, which means that the after-tax income distribution is more concentrated than in the baseline (i.e., “inequality increases”). Although this inequality-increasing effect is observed in all EU countries, its magnitude widely varies from around 0.0025 Gini points in Sweden (in general in all Scandinavian countries, as well as in France and Belgium) to 0.02 (in Greece). These results are robust to the use of an alternative indicator to measure income concentration, such as the S8020 ratio, as we show later. From the well-known Kakwani decomposition (Reynolds and Smolensky,1977,Kakwani,1977), that is the difference in Gini coefficient before and after a tax can be decomposed into a progressivity effect (measured with the Kakwani index, capturing how the tax is distributed along pre-tax income), a size effect (the average rate or tax burden) and a re-ranking effect. In Figure 4, we observe a clear positive cross-country correlation between the tax burden and the redistributive effect. This suggests that most of the cross-country disparities are explained by the relative size of the carbon tax with respect to household incomes, rather than by differences in progressivity/regressivity. Nevertheless, the relationship is not perfectly linear. We can see, for example, that there are countries with very similar average tax burdens (e.g., Estonia and Romania, where the carbon tax would represent on average 7% of household disposable income) with quite different redistributive effects (0.017 vs 0.010, for Estonia and Romania, respectively). In these cases, we do observe substantial differences in regressivity - the simulated carbon tax is more disproportionately affecting lower-income households in Estonia than in Romania. As it can be appreciated in Figure B.3 (Appendix B), per capita GHG emissions in quintile 1 are much higher in Estonia than in Romania. 16 Figure 4: Distributional effect and average tax rate of a hypothetical 80 EUR flat CO2tax (S1.1) Note: Simulated results based on Green EUROMOD. Average tax rate (or tax burden) expressed as % of equivalised household disposable income. GHG footprints (tCO2e) based on environmental extensions from EXIOBASE (2019), HBS 2015, EU-SILC 2015 and EUROMOD tax-benefit rules of 2019. An advantage of the Green EUROMOD model is that we can simulate not only household disposable income (after direct taxes and cash benefits) but also consumption tax liabilities (VAT and excises), with a high level of precision and cross-country harmonization. This allows us to compare the tax burden and redistributive effect of our simulated carbon tax with the one from consumption taxes.20 Our estimates suggest that consumption taxes represent approximately 13.5% of household disposable income (in 2019, on the EU-27 average). This means that our hypothetical carbon tax would increase the overall consumption tax burden by about one-third (35%). Unsurprisingly, this EU average also masks substantial cross-country heterogeneity, as this share (carbon tax to total consumption taxes) ranges from 16-19% in Sweden and Denmark to about 70% in Poland (see Figure B.5, Appendix B). Although it is well true that some countries have higher consumption tax rates in the baseline -what ceteris paribus makes the relative increase from carbon taxation smallerdifferences here are again mainly driven by the heterogeneity in the tax burden of the simulated carbon tax. In terms of the distributional outcomes, the inequality-increasing effect of the simulated carbon tax is about one-third of the inequality-increasing effect of 2019 consumption taxes in the EU. Again, this varies across countries, from about 13% in Sweden and Denmark to more than 100% in Czechia and Poland (see Figure B.5, in Appendix B). We now explore to what extent a compensatory measure such as a cash transfer to households (recycling the tax revenues from this simulated tax) can offset these inequality-increasing effects. To do so we need first to estimate the total tax revenues. According to our micro-simulations, this EUR 80 flat carbon tax would increase government revenues by EUR 208 billion. Table 3reports the results for the EU 20For more information on simulated carbon taxes in EUROMOD, see, e.g., Maier and Ricci (2024). 17 aggregates and Table 4by country.21 For context, the EUR 208 bn represents more than twice the Social Climate Fund for 2026-2032 (EUR 87 bn). Table 4: Results from a hypothetical EUR 80 carbon tax by country (scenario 1) Average tax burden Budgetary effect Redistributive effect (change in Gini) cnt % (tax/yd) EUR bn %GDP S1.1 (no comp) S1.2 (cnt comp) S1.3 (EU comp) AT 3.79 6.29 1.6% 0.004 -0.006 -0.003 BE 3.56 6.63 1.4% 0.003 -0.005 -0.003 BG 5.71 1.48 2.4% 0.004 -0.017 -0.042 CY 3.70 0.38 1.6% 0.004 -0.008 -0.009 CZ 7.60 6.09 2.7% 0.013 -0.007 -0.004 DE 3.62 56.94 1.6% 0.007 -0.003 0.000 DK 2.40 3.19 1.0% 0.003 -0.003 -0.002 EE 6.95 0.85 3.0% 0.016 -0.006 0.000 EL 8.51 5.58 3.0% 0.020 -0.010 -0.007 ES 2.95 14.34 1.2% 0.005 -0.006 -0.011 FI 3.08 3.23 1.3% 0.003 -0.005 -0.003 FR 1.96 22.42 0.9% 0.002 -0.003 -0.005 HR 5.02 0.97 1.7% 0.006 -0.009 -0.024 HU 8.73 3.86 2.6% 0.010 -0.018 -0.023 IE 4.37 3.38 0.9% 0.009 -0.004 0.000 IT 3.82 29.64 1.6% 0.006 -0.006 -0.006 LT 5.28 0.97 2.0% 0.012 -0.007 -0.014 LU 3.52 0.43 0.7% 0.004 -0.005 -0.001 LV 5.58 0.92 3.0% 0.003 -0.016 -0.016 MT 4.86 0.11 0.8% 0.003 -0.003 -0.008 NL 3.36 10.55 1.3% 0.005 -0.004 -0.001 PL 8.44 16.76 3.1% 0.016 -0.010 -0.012 PT 4.09 3.36 1.6% 0.009 -0.006 -0.012 RO 6.93 3.97 1.8% 0.010 -0.017 -0.049 SE 1.74 3.35 0.7% 0.002 -0.002 -0.004 SI 4.23 0.88 1.8% 0.007 -0.004 -0.005 SK 4.42 1.31 1.4% 0.006 -0.005 -0.014 Note: Simulated results based on Green EUROMOD. GHG footprints (tCO2e) based on environmental extensions from EXIOBASE (2019), HBS 2015, EU-SILC 2015 and EUROMOD tax-benefit rules of 2019. “bn”: billions, “yd”: household disposable income. Change in Gini is calculated as the difference between the Gini coefficient of equivalised household disposable income after the carbon tax (under compensating variation) and the Gini coefficient of baseline equivalised household disposable income. The absolute size of the budgetary effect depends on each country’s per capita GHG emissions and population size and varies from below EUR 0.5 bn in small countries (Malta, Cyprus, Luxembourg) to EUR 57 bn in Germany. On average, it represents about 1.74% of GDP, a share that varies from 0.7% (Sweden or Luxembourg) to around 3% (Latvia, Poland, Estonia or Greece). When the compensatory measure is given in a revenue-neutral way for all EU inhabitants in the same magnitude (scenario 1.3), the cash transfer amounts to EUR 480 per year. In contrast, when the cash transfer is given at the Member State level (scenario 1.2), it varies from EUR 200-250 in Bulgaria, Romania, Croatia, Malta and Slovenia to close to EUR 700-800 in Austria, Germany, Ireland and Luxembourg. 21These budgetary estimates are, as well as our distributional estimates, ‘overnight’ effects (i.e., do not consider potential behavioural responses, nor at the level of consumer demand nor general equilibrium effects related with competitiveness and changes in functional income). We further discuss the implications of this assumption in the Discussion section 5. 18 To what extent do these compensatory measures offset the negative redistributive consequences of the carbon flat tax? The different redistributive outcomes across countries and scenarios are plotted in Figure 5, where countries are ranked from left to right according to the increase in the Gini coefficient in scenario 1.1 (no compensation). There, we can see that the effect is reverted in all countries: while the tax is inequality-increasing in all EU countries, inequality is reduced in all EU countries with compensatory measures.22 These values, together with the average tax burden and budgetary effects per country are also reported in Table 4. Figure 5: Distributional effect: without and with compensatory measures (scenario 1) Note: Simulated results based on Green EUROMOD. GHG footprints (tCO2e) based on environmental extensions from EXIOBASE (2019), HBS 2015, EUSILC 2015 and EUROMOD tax-benefit rules of 2019. Generally, in the budget-neutral scenarios with compensatory measures (1.2 and 1.3), countries in Central and Eastern Europe experience the strongest reductions in inequality. These countries tend to have higher levels of inequality than Western, Northern and Southern European countries. Therefore, scenario 1.1 not only increases inequality everywhere but also broadens divergences across countries, whereas scenarios 1.2 and 1.3 would not only lead to a reduction in within-countries inequality but also to smaller cross-country divergences in this dimension. In countries where the EU-level transfer is much larger than the national-level one (i.e., in Bulgaria, Romania, Slovakia and Croatia), the positive redistributive effects are stronger in scenario 1.3 than in 1.2. Extreme cases are Bulgaria and Romania, where the redistributive effect in scenario 1.3 (of about 0.04 Gini points) more than doubles that in scenario 1.2 (of slightly less than 0.02). In contrast, in most Western/Northern high-income countries, the inequality-reducing effect after the EU-level transfer (scenario 1.3) is weaker than with the national-level transfer (1.2). This is clearly the case in Germany, Ireland, Austria and Luxembourg. 22The only exceptions are Ireland, Germany and Estonia, that in scenario 1.3 (EU-level compensatory transfer) experience an inequality-neutral effect of the combined reform (tax + compensatory measure). 19 Overall, results from scenario 1 suggest that an EU flat carbon tax can lead to positive redistributive effects only if accompanied by revenue-recycling compensatory measures. We illustrate this with a lumpsum transfer, which turns out to be enough to compensate the increase in the post-tax concentration of income. While compensatory measures -typically in the shape of revenue-recycling cash transfers to households, such as those we have simulatedare widely discussed in the literature (e.g., Feindt et al., 2021;Vandyck et al.,2021), they are not automatic and can be challenging to implement from the point of view of their public and political support. Among other limitations, it is difficult to define how long these transfers would endure, and how to phase them out, and even more to identify eligibility based on potential winners/losers. Moreover, they may also bring undesired rebound effects (e.g., see Murray, 2013). 4.2.2 Results from scenarios 2: green allowances In what comes next, we turn our attention to evaluating the budgetary and distributional impacts of alternative tax designs, specifically scenarios 2 and 3 outlined in Table 2. We simulate these other scenarios as alternative approaches to enhancing distributional outcomes from this simulated carbon pricing initiative without the need for additional compensatory cash transfers. This is achieved by making the carbon tax more progressive. Moreover, a progressive tax could, in fact, increase public support for this policy, as suggested by Klenert et al. (2018). We start by comparing scenarios with “green allowances” (S2.1 and S2.2) with the flat tax scenarios, without and with compensation (1), see Section 3.2. If we keep the original carbon price of EUR 80, under the green allowances (S2.1), total tax revenues would be EUR 135 bn (see Table 3) instead of EUR 208 (S1). Alternatively, we simulate a budget-equivalent variant (S2.2) where we endogenously calculate the new carbon tax rate that would be needed to collect the same tax revenues per country as in the flat tax without compensation (S1). This, of course, leads to a larger carbon price on all those GHG emissions that are generated above the 2.2 tCO2e per year threshold (a price of EUR 145 per tCO2e, on EU average), with wide cross-country variability, as reported in Table B.5. In Figure 6, we can see that, in general, across scenarios, the bottom 20% (first quintile, based on equivalised disposable income) in each of the 27 EU countries would face a much larger tax burden than their richest counterparts (top 20%, quintile 5), suggesting that these simulated tax reforms are generally regressive. The largest gap between quintiles is observed in the flat tax (scenario 1.1), where, on average, the tax burden of the first quintile is 7.7%, more than doubling that of the fifth (richest) quintile (3.6%). We also identify substantial cross-country dispersion of the average tax burden in the flat tax scenario, which is particularly pronounced at the bottom of the income distribution. Table B.6 reports the countryspecific estimates (tax burdens for quintiles 1 and 5 across all simulated scenarios). The simulated tax burden of the bottom quintile in scenario 1.1 ranges from about 2.5% in France and Sweden to almost 20% in Greece. By contrast, the tax burden in the fifth quintile ranges from 1.5% to 6%. 20 this line, Böhringer et al. (2021) argue that adding a consumption tax on all use of goods that are covered by the free allowances of the ETS is optimal from both a regional and global welfare perspective, in line with previous studies (Böhringer et al.,2017;Holland,2012;Eichner and Pethig,2015). In our study, the simulated carbon taxes come on top of the already existing carbon pricing initiatives embedded in our baseline prices. Note that the forthcoming ETS2 on households does not foresee free allowances, so there might be some overlapping. In any case, the tax designs assessed in this paper could be fine tuned to exclude products under ETS2 or any other product. In this sense, an interesting complement to our study would be to further refine the tax design scenarios to make them more compatible (complementary) to other carbon pricing initiatives by, for example, considering only those GHG emissions exempted from ETS and ETS2. Another important consideration regarding the feasibility of our scenarios is that a flat carbon tax implemented at the EU level, without adjustments for differences in purchasing power or income, would, ceteris paribus, disproportionately affect low-income households within countries and even more so those in low-income countries. Such hypothetical tax could be accompanied by compensatory measures or adjustments based on each country’s purchasing power. Note that energy products are among the most carbon-intensive, and while all countries face the same international energy prices, their purchasing power varies. This disparity leads to significant differences in the energy bill burden across countries, even at similar efficiency levels. Implementing a flat carbon tax across countries would further widen this gap. In this sense, we have shown that low-income households in low-income countries would be those benefiting the most from the EU-level revenue-neutral compensatory cash transfer (S1.3). And more in general, both compensatory measures (S1.2 and S1.3) would revert the inequality-increasing effect within countries. Improving the distributional outcomes of the tax are fundamental to justify its just transition objectives, as using carbon taxation with the only purpose of raising funds suffer from little support (Valencia et al., 2023).27 In this sense, a flat carbon tax with revenue-neutral compensatory cash transfers to households (S1.2 and S1.3) could be promising, as it is not revenue-increasing and leads to the best distributional outcomes. However, these have at least three important limitations: i) they depend on two different policies (the tax and the transfer), which could challenge their accountability and with it their public support, ii) they are difficult to implement (phase-out strategy, eligibility of beneficiaries, etc.), iii) they do not take into account changes in consumption patterns nor rebound effects, which together with potential strategic behaviours by households can limit their efficacy.28 As revenue-neutral compensatory measures are not very feasible in the current policy context, we may want to explore further the advantages and disadvantages of the other simulated scenarios, where we try to make the carbon tax more progressive (or less regressive) by design. The different carbon tax rates across aggregate consumption categories (scenario 3) are, in principle, also relatively easy to implement, as they mimic what consumption taxes do (e.g., by applying reduced VAT rates to first-necessity goods 27In their literature review covering 35 studies and 70 surveys suggests that greener spending has higher public support than raising taxes for the only purpose of raising public funds, while evidence is mixed on the support for cash transfers, with wide cross-country variation. Along this line, a recent study for Germany (Sommer et al., 2022) suggests that green spending also increases public support for carbon pricing, but for the case of cash transfers, untargeted measures are more preferred than targeted ones. An interesting discussion on targeted vs untargeted price-related measures as well as income support measure related to the recent hike in energy prices, including the case of Germany, can be found in Amores et al. (2023b) and Amores et al. (2023a). 28A way to deal with the latter, and encourage greener demand habits for the additional incomes received as compensation would be to restrict its use to the purchasing of greener products and/or to invest in greening the capital stock, such as in house insulation, etc. 27 or higher rates to health-detrimental consumption categories). However, we have shown that the gain in progressivity obtained from applying different carbon taxes to consumption categories (e.g., by taxing transport at a higher level than food or heating) is limited and involves the establishment of arbitrary thresholds.29 In this sense, the carbon tax with green allowances (S2.1 and S2.2), where the first 2.2 tCO2e are tax exempted) seem more attractive. Overall, although these tax exemptions are not progressive enough to prevent an increase in inequality, they offset a large part of the negative redistributive consequences of the flat carbon tax. The main drawback of these tax-exempted instruments is their challenging practical implementation, requiring CO2footprint labeling for all consumption items and a voucher system for households to purchase products accounting for their first 2.2 tCO2e without the carbon tax, similar to free allowances in the EU ETS. This system could be operationalized through a centralized process where the retailer registers purchases against consumers’ ID, mimicking the VAT collection system across the supply value chain. However, beyond the technical challenge that could be solved with blockchain technology, this poses privacy concerns as the central system would track purchases and identify frequented retailers. An alternative based on an anonymous voucher also has disadvantages, as it might be prone to the loss of the vouchers or even a black market of free carbon allowances. Therefore, a technological solution that combines privacy with nontransferable allowances could be envisaged, at least theoretically. Finally, there could also be potential concerns related to cross-border fraud (in case of a flat tax adjusted by the purchasing power of countries to avoid distortions of the single market). 6 Concluding remarks In this study, we assess the distributional consequences of various hypothetical carbon taxes on household consumption across the 27 EU countries. This is especially relevant in the current economic and political landscape in Europe, where fostering "greener" consumer habits and ensuring the public acceptability of such measures are crucial for a just transition. To simulate these alternative carbon tax designs, we extend the EU tax-benefit micro-simulation model EUROMOD, with information on GHG emission intensities at a very detailed product level from an environmentally extended input-output model (EXIOBASE). With this “Green EUROMOD” extension, we estimate household-level carbon footprints, including emissions from both domestic production and imports, as well as trade and transport services. This is what allows us to simulate carbon taxes for the EU27 countries within a consistent and harmonized empirical framework, based on EU-SILC incomes. As such, our distributional analysis uses the same income concept as Eurostat for the EU-level official measurement of income poverty and inequality. Our estimates suggest that per capita GHG emissions in the EU are still well above the “sustainable levels” of about 2/2.5 tons of CO2e a year) consistent with the international (Paris) agreements (see, e.g., Chancel et al.,2022;Gore,2021; and Ivanova et al.,2016). These emissions vary widely across countries and income groups. Per capita GHG emissions of the lowest income quintiles of the 27 EU countries (bottom 20%) average slightly above 4 tCO2e, while the richest quintile averages over 8 tCO2e. In fact, only the poorest (in terms of incomes) 10% of all individuals in the EU (pooled together, regardless of 29The design with different tax rates across products (S3.1) could be enhanced through a better identification of the goods to tax at a lower or higher rate, depending on their share of emissions and the impact on inequality. This can be observed in figures B.6,B.8 and B.7, where we show (for the subcategories of food, transport and energy, for each country) how the impact of a EUR 80 flat tax on every single good would affect inequality and the related share of emissions. 28 their country of residence) have consumption patterns that are compatible with the 2/2.5-sustainable threshold. The results from our microsimulations suggest that a flat carbon tax on household consumption (EUR 80 per tCO2e) would be inequality-increasing in all 27 EU Member States, without exception. This negative redistributive effect comes from the regressivity of the tax when it is assessed against the distribution of household disposable incomes. This regressivity arises because lower-income households, which allocate a larger portion of their income to consumption rather than savings, bear a heavier tax burden. This aligns with the findings of Linden et al. (2024) and mirrors the regressive nature of consumption taxes relative to disposable income (Maier and Ricci,2024;Decoster et al.,2010). The tax’s regressivity is further exacerbated by differences in consumption patterns, with low-income households spending more (higher income shares) on GHG-intensive categories like food and residential energy. The distributional effect largely varies across EU Member States. Countries that would experience the strongest inequality-increasing effects are mostly Central and Eastern European countries. This crosscountry dispersion is primarily explained by the differences in the tax burden. While the simulated consumption-based carbon tax would represent almost 9% of disposable income in Hungary, Poland and Greece, it is only 2% in Sweden and France. Differences in the tax burden are mainly driven by the decreasing gradient of income shares of household consumption expenditures across countries’ per capita income, the same that drives the regressivity within countries. Again here, there are differences in consumption profiles reinforcing this pattern: the expenditure shares of highly polluting goods (such as heating, food and transport) tend also to be larger in lower-income countries (and households). This flat carbon tax of EUR 80 per tCO2e could generate up to EUR 208 billion annually for the EU, more than double the projected budget for the Social Climate Fund. Our results suggest that if these revenues were used to compensate households with a lump-sum cash transfer, the increase in inequality would be fully reverted. However, the practical and political feasibility of these cash transfers is somehow limited. This is why we explore ways of making the carbon tax less regressive and, with it, improving the distributional outcomes without relying on such compensatory measures. In this sense, we show that progressive tax structures, with tax allowances and (although to a lesser extent) differentiating tax rates across products (like VAT), perform relatively well in preventing large inequality-increasing effects and, at the same time, boosting government revenues. These are also quite attractive since they do not depend on a complementary policy (such as the cash transfer), which could make their implementation more challenging, and hamper their public acceptability. These results underscore the potential for carbon taxes to serve as effective tools for both environmental and fiscal policy, provided they are carefully designed to address equity concerns. References Akoğuz, E.C., Capéau, B., Decoster, A., De Sadeleer, L., Güner, D., Manios, K., Paulus, A., Vanheukelom, T., 2020. A new indirect tax tool for EUROMOD: final report. Technical Report. https://euromod-web.jrc.ec.europa.eu/sites/default/files/2021-03/A Amores, A.F., Basso, H.S., Bischl, J.S., De Agostini, P., De Poli, S., Dicarlo, E., Flevotomou, M., Freier, M., Maier, S., García-Miralles, E., et al., 2023a. Inflation, fiscal policy and inequality. ECB Occasional Paper . Amores, A.F., Christl, M., De Agostini, P., De Poli, S., 2023b. Limiting prices or transferring money? JRC Working Papers on Taxation and Structural Reforms . 29 Amores, A.F., De Poli, S., Maier, S., Wood, R., 2024. GHG footprints in the eu: mapping EXIOBASE to COICOP. Internal European Commission report. Amores, A.F., Maier, S., Ricci, M., 2023c. Taxing household energy consumption in the EU: The tax burden and its redistributive effect. Energy Policy 182, 113721. doi:https://doi.org/10.1016/j. enpol.2023.113721. Andersson, J.J., 2019. Carbon taxes and CO2 emissions: Sweden as a case study. American Economic Journal: Economic Policy 11, 1–30. doi:10.1257/pol.20170144. Barrios, S., Pycroft, J., Saveyn, B., 2013. The marginal cost of public funds in the EU: the case of labour versus green taxes. Fiscal Policy and Growth 403. Böhringer, C., Rosendahl, K.E., Storrøsten, H., 2021. Smart hedging against carbon leakage. Economic Policy 36, 439–484. Böhringer, C., Rosendahl, K.E., Storrøsten, H.B., 2017. Robust policies to mitigate carbon leakage. Journal of Public Economics 149, 35–46. Bovenberg, L., 1999. Green tax reforms and the double dividend: an updated reader’s guide. International Tax and Public Finance 6, 421–443. URL: https://doi.org/10.1023/A:1008715920337. Chancel, L., 2022. Global carbon inequality over 1990–2019. Nature Sustainability 5, 931–938. Chancel, L., Piketty, T., 2015. Carbon and inequality: From Kyoto to Paris Trends in the global inequality of carbon emissions (1998-2013) & prospects for an equitable adaptation fund world inequality lab. Chancel, L., Piketty, T., Saez, E., Zucman, G., 2022. World inequality report 2022. Harvard University Press. Cludius, J., de Bruyn, S., Schumacher, K., Vergeer, R., 2020. Ex-post investigation of cost pass-through in the eu ets - an analysis for six industry sectors. Energy Economics 91, 104883. URL: https: //www.sciencedirect.com/science/article/pii/S0140988320302231, doi:https://doi.org/10. 1016/j.eneco.2020.104883. De Poli, S., Gil-Bermejo Lazo, C., Leventi, C., Maier, S., Papini, A., Ricci, M., Serruys, H., Almeida, V., Christl, M., Cruces, H., et al., 2023. EUROMOD baseline report. Technical Report. JRC Working Papers on Taxation and Structural Reforms No 5/2023. Dechezleprêtre, A., Nachtigall, D., Venmans, F., 2023. The joint impact of the european union emissions trading system on carbon emissions and economic performance. Journal of Environmental Economics and Management 118, 102758. URL: https://www.sciencedirect.com/science/article/ pii/S0095069622001115, doi:https://doi.org/10.1016/j.jeem.2022.102758. Decoster, A., Loughrey, J., O’Donoghue, C., Verwerft, D., 2010. How regressive are indirect taxes? A microsimulation analysis for five european countries. Journal of Policy Analysis and Management 29, 326–350. Döbbeling-Hildebrandt, N., Miersch, K., Khanna, T.M., Bachelet, M., Bruns, S.B., Callaghan, M., Edenhofer, O., Flachsland, C., Forster, P.M., Kalkuhl, M., et al., 2024. Systematic review and meta-analysis of ex-post evaluations on the effectiveness of carbon pricing. Nature Communications 15, 4147. Douenne, T., Fabre, A., 2022. Yellow vests, pessimistic beliefs, and carbon tax aversion. American Economic Journal: Economic Policy 14, 81–110. 30 Eichner, T., Pethig, R., 2015. Unilateral consumption-based carbon taxes and negative leakage. Resource and Energy Economics 40, 127–142. European Commission, 2021. European Green Deal: the Commission proposes transformation of EU economy and society to meet climate ambitions. Press Release. European Commission, 2022. Report from the Commission to the European Parliament and the Council. COM(2022) 516 final. URL: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=COM:2022: 516:FIN. European Commission, 2024. Impact Assessment Report Part 1: Accompanying the document Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions: Securing our future Europe’s 2040 climate target and path to climate neutrality by 2050 building a sustainable, just and prosperous society. Commission Staff Working Document SWD/2024/63 final. European Commission. Fabra, N., Reguant, M., 2014. Pass-through of emissions costs in electricity markets. American Economic Review 104, 2872–99. URL: https://www.aeaweb.org/articles?id=10.1257/aer.104.9.2872, doi:10.1257/aer.104.9.2872. Feindt, S., Kornek, U., Labeaga, J.M., Sterner, T., Ward, H., 2021. Understanding regressivity: Challenges and opportunities of European carbon pricing. Energy Economics 103, 105550. Girod, B., van Vuuren, D.P., Hertwich, E.G., 2014. Climate policy through changing consumption choices: Options and obstacles for reducing greenhouse gas emissions. Global Environmental Change 25, 5–15. Gore, T., 2021. Carbon inequality in 2030: Per capita consumption emissions and the 1.5°C goal. Commission Staff Working Document. Institute for European Environmental Policy, Oxfam. Hassler, J., Krusell, P., Nycander, J., 2016. Climate policy. Economic Policy 31, 503–558. doi:10.1093/ epolic/eiw007. Hepburn, C., Stern, N., Stiglitz, J.E., 2020. Carbon pricing. European Economic Review 127, 103440. Holland, S.P., 2012. Emissions taxes versus intensity standards: Second-best environmental policies with incomplete regulation. Journal of Environmental Economics and Management 63, 375–387. Ivanova, D., Stadler, K., Steen-Olsen, K., Wood, R., Vita, G., Tukker, A., Hertwich, E.G., 2016. Environmental impact assessment of household consumption. Journal of Industrial Ecology 20, 526–536. Ivanova, D., Wood, R., 2020. The unequal distribution of household carbon footprints in Europe and its link to sustainability. Global Sustainability 3, e18. Kakwani, N.C., 1977. Measurement of tax progressivity: an international comparison. The Economic Journal 87, 71–80. King, M.A., 1983. Welfare analysis of tax reforms using household data. Journal of Public Economics 21, 183–214. Klenert, D., Mattauch, L., 2016. How to make a carbon tax reform progressive: The role of subsistence consumption. Economics Letters 138, 100–103. doi:10.1016/j.econlet.2015.11.019. Klenert, D., Mattauch, L., Combet, E., Edenhofer, O., Hepburn, C., Rafaty, R., Stern, N., 2018. Making carbon pricing work for citizens. Nature Climate Change 8, 669–677. 31 Köppl, A., Schratzenstaller, M., 2023. Carbon taxation: A review of the empirical literature. Journal of Economic Surveys 37, 1353–1388. Labandeira, X., Labeaga, J.M., López-Otero, X., 2017. A meta-analysis on the price elasticity of energy demand. Energy Policy 102, 549–568. URL: https://www.sciencedirect.com/science/article/ pii/S0301421517300022, doi:https://doi.org/10.1016/j.enpol.2017.01.002. Lamarche, P., Oehler, F., Rioboo, I., 2020. European household’s income, consumption and wealth. Statistical Journal of the IAOS 36, 1175–1188. Lambert, P.J., 1992. The distribution and redistribution of income. Springer. Lin, B., Li, X., 2011. The effect of carbon tax on per capita CO2emissions. Energy Policy 39, 5137–5146. Linden, J., O’Donoghue, C., Sologon, D.M., 2024. The many faces of carbon tax regressivity—why carbon taxes are not always regressive for the same reason. Energy Policy 192, 114210. doi:10.1016/j.enpol. 2024.114210. Maier, S., Ricci, M., 2024. The redistributive impact of consumption taxation in the EU: Lessons from the post-financial crisis decade. Economic Analysis and Policy 81, 738–755. Martin, R., Muûls, M., Wagner, U.J., 2016. The impact of the european union emissions trading scheme on regulated firms: What is the evidence after ten years? Review of environmental economics and policy 10, 129–148. Merciai, S., Schmidt, J.H., 2018. Methodology for the construction of global multi-regional hybrid supply and use tables for the exiobase v3 database. Journal of Industrial Ecology 22. URL: https://api. semanticscholar.org/CorpusID:158339115. Murray, B., Rivers, N., 2015. British Columbia’s revenue-neutral carbon tax: A review of the latest “grand experiment” in environmental policy. Energy Policy 86, 674–683. Murray, C.K., 2013. What if consumers decided to all ‘go green’? Environmental rebound effects from consumption decisions. Energy Policy 54, 240–256. Nachtigall, D., Ellis, J., Errendal, S., 2022. Carbon pricing and COVID-19: Policy changes, challenges and design options in OECD and G20 countries. OECD. Nordhaus, W.D., 1991. A sketch of the economics of the greenhouse effect. The American Economic Review 81, 146–150. Ohlendorf, N., Jakob, M., Minx, J.C., Schröder, C., Steckel, J.C., 2021. Distributional impacts of carbon pricing: A meta-analysis. Environmental and Resource Economics 78, 1–42. Parry, I., Black, S., Zhunussova, K., 2022. Carbon taxes or emissions trading systems?: instrument choice and design. International Monetary Fund Washington, DC, USA. Pearce, D., 1991. The role of carbon taxes in adjusting to global warming. The Economic Journal 101, 938–948. Pigou, A., 1920. The Economics of Welfare. Macmillan, London. Reynolds, M., Smolensky, E., 1977. Post-fisc distributions of income in 1950, 1961, and 1970. Public Finance Quarterly 5, 419–438. 32 Rubin, A.J., Sengupta, S., 2018. Yellow vest’protests shake France. Here’s the lesson for climate change. The New York Times 6. Sommer, S., Mattauch, L., Pahle, M., 2022. Supporting carbon taxes: The role of fairness. Ecological Economics 195, 107359. Stadler, K., Wood, R., Bulavskaya, T., Södersten, C.J., Simas, M., Schmidt, S., Usubiaga, A., AcostaFernández, J., Kuenen, J., Bruckner, M., et al., 2018. EXIOBASE 3: Developing a time series of detailed environmentally extended multi-regional input-output tables. Journal of Industrial Ecology 22, 502–515. Stiglitz, J.E., Stern, N., Duan, M., Edenhofer, O., Giraud, G., Heal, G.M., La Rovere, E.L., Morris, A., Moyer, E., et al., 2017. Report of the High-Level Commission on Carbon Prices . Sutherland, H., Figari, F., 2013. EUROMOD: the European Union tax-benefit microsimulation model. International Journal of Microsimulation 6, 4–26. Thomas, A., 2022. Reassessing the regressivity of the VAT. Fiscal Studies 43, 23–38. Tol, R.S., 2023. Social cost of carbon estimates have increased over time. Nature Climate Change , 1–5. Tukker, A., De Koning, A., Wood, R., Hawkins, T., Lutter, S., Acosta, J., Rueda Cantuche, J.M., Bouwmeester, M., Oosterhaven, J., Drosdowski, T., et al., 2013. EXIOPOL–development and illustrative analyses of a detailed global MR EE SUT/IOT. Economic Systems Research 25, 50–70. UNFCCC, 2015. Paris agreement, in: Report of the Conference of the Parties to the United Nations Framework Convention on Climate Change (21 session, 2015: Paris), United Nations. Valencia, F.M., Mohren, C., Ramakrishnan, A., Merchert, M., Minx, J.C., Steckel, J.C., 2023. Public support for carbon pricing policies and different revenue recycling options: a systematic review and meta-analysis of the survey literature 5. doi:https://doi.org/10.21203/rs.3.rs-3528188/v1. Vandyck, T., Weitzel, M., Wojtowicz, K., Los Santos, L.R., Maftei, A., Riscado, S., 2021. Climate policy design, competitiveness and income distribution: A macro-micro assessment for 11 EU countries. Energy Economics 103, 105538. Wang, Q., Hubacek, K., Feng, K., Wei, Y.M., Liang, Q.M., 2016. Distributional effects of carbon taxation. Applied Energy 184, 1123–1131. Wood, R., Stadler, K., Bulavskaya, T., Lutter, S., Giljum, S., De Koning, A., Kuenen, J., Schütz, H., Acosta-Fernández, J., Usubiaga, A., et al., 2014. Global sustainability accounting—developing EXIOBASE for multi-regional footprint analysis. Sustainability 7, 138–163. World Bank, 2023. State and Trends of Carbon Pricing 2023. Technical Report. World Bank. URL: http://hdl.handle.net/10986/39796. 33 A Appendix A: data and imputations A.1 Approaches on the implementation of the Polluter Pays Principle Figure A.1: Producer Based Approach vs Consumer Based Approach Source: Own elaboration. A.2 Imputation of HBS consumption expenditures into SILC This subsection describes the imputation method used to impute households consumption expenditures reported in HBS to SILC’s surveyed households. Since the households interviewed in these two surveys are not the same, merging information from one to another requires an imputation method.30 To identify similar households across surveys, we adopt the semi-parametric procedure developed by Akoğuz et al. (2020), which combines the estimation of Engel curves (see, for instance, Decoster et al., 2010) with matching techniques. Below we provide a brief step-by-step description of this procedure. 1. Calculate the source income shares: For each household h’s in the source dataset, s(HBS), the expenditure on a good i,es hp, is converted into a share of disposable income, ys h: ws hp=es hp ys h , p ∈P(9) 30In general, we use HBS 2015, with a few exceptions. One is Germany, because a consistent dataset was not available at the time of calculation. Another is Italy, where the process is further complicated due to the absence of net household income data in the HBS surveys. To address this, we first impute income data to HBS using a third survey, the 2010 Survey on Household Income and Wealth (SHIW), before merging HBS with SILC. Finally, another exception is Austria, which is based on proprietary national data from 2010. For more details, see Akoğuz et al. (2020). 34 These shares are calculated for the Pgoods available in the HBS (around 200 for most countries). 2. Aggregate the source income shares: These income shares are aggregated under broader categories, indexed by X=A, B, .... Thus, the income share of expenditure category X, Ws hX: Ws hX≡X p∈PX ws hp(10) 3. Regression analysis: Source income shares Ws hXare regressed against a set of covariates common to both the HBS (source, s) and SILC (recipient, r) datasets. This regression draws on the specification of Engel curves.31 Since aggregated categories Xmay still have zero expenditures, a two-step regression is performed: (a) Probit model: The probability of positive expenditure on category Xis modelled using a probit model: Pr Ws hX>0= 1 −ϕ−γ ′ Xxs h=ϕ−γ ′ Xxs h(11) where ϕ(·)is the standard normal distribution function, xs hdenotes the vector of covariates for household hin the source dataset s, and γ′ Xcontains parameters to be estimated. (b) Regression model: For households with positive expenditures, a continuous regression model is estimated: Ws hX=β ′ XXs h+ϵhX, Ws hX>0(12) 4. Fitted Values for income shares: The estimated parameters bγand b βare used to generate fitted values for the income shares of expenditures on the aggregated categories Xfor all households in both HBS and SILC: Wd hX=ϕ−bγ ′ Xxd hb β ′ XXd h, d =s, r (13) where sand rdenote respectively HBS and SILC. 5. Distance calculation: Using the vectors of fitted shares, c Wd, calculated in the previous step, the Mahalanobis distance between a household hin HBS, and a household gin SILC is defined as: dist (h, g) = dist Wr g, Ws h=rc Wr g−c Ws h′ Σ−1c Wr g−c Ws h(14) where Σis the covariance matrix of the vectors c Wd, using data from both datasets. 6. Matching households: The household hin HBS that has the smallest distance to a household gin the SILC is selected as the match. 7. Detailed expenditures: For each match paired (h, g), the income shares of expenditures at the most detailed level of products p∈Pfor the EU-SILC household g, are obtained from the corresponding values of the HBS household h: wr gp=ws hp(15) 31The covariates include a third-degree polynomial in the log of incomes and detailed household composition characteristics, such as number of household members by gender, labour market status, and age. 35 A.3 Green EUROMOD: bridging EXIOBASE to HBS The step-by-step methodology, coefficients and validation results are reported in Amores et al. (2024). Here we summarize the main steps and methodological choices. EXIOBASE (v3.8 database) has environmental extensions listed individually in 1707 categories. Most uses of these extensions involve an aggregation of, e.g., different material quantities. Carbon footprints were calculated as the amount of the different GHG in kgCO2-equivalents per yearusing the Global Warming Potential 100 (GWP100), which equivalences are CO2: 1, CH4: 28, N2O: 265, SF6: 23,500. 1. Direct and indirect intensities: In this step, we estimate direct and indirect intensities separately. The direct emissions are only relevant for burnable fuels in EXIOBASE. One point of confusion is the treatment of biomass, which is partly treated as an agricultural product in the EXIOBASE energy accounts. The combustion of biomass results in non-CO2air emissions (biomass is assumed carbon neutral in the CO2accounts in EXIOBASE). Household emissions are stored in a 3 dimensional matrix in EXIOBASE (extension ×product ×final demand category). These intensities refer to the GHG emissions or other environmental pressures per unit of consumption. Note, total expenditure is used, as these emissions occur on the domestic territory, but can result from the purchase of either domestically produced or imported fuels. For indirect emissions, GHG intensities (emissions per EUR) are obtained from an environmentally extended input-output model built over EXIOBASE at basic prices. All emission sources are allocated to different types of final demand. However, at the sector and transaction level, significant noise may occur in the data, as many small values are estimated with a very low level of accuracy but have an indiscernible impact on total results. 2. GHG intensities into purchasers’ prices: They need to be transformed into purchasers’ prices. To do so, we use the tables of trade and transport margins, and taxes less subsidies on products to relocate the trade and transport services and their associated emissions to the final products consumed by households. 3. Bridge matrix: We build bridging matrices to estimate the GHG intensities in COICOP from the EXIOBASE results. We follow a very similar methodology than Ivanova and Wood (2020) but at a much more product granularity (COICOP 5 digits). Bridge matrices are usually constructed to bridge between two common observations in two different classifications. For the purpose here, the “observation” is the expenditure of households. The expenditure of households is available in the EXIOBASE classification and is also available in the EUROMOD classification. Given the availability of expenditure in both classifications, bridge matrices can then be estimated in value terms that show the value of products in one classification in terms of the other classification at teh same time. While the mapping of EXIOBASE products to EUROMOD products could be done with a manual approximate mapping based on expert choice, the use of “bridge matrices” enhances the accuracy of the accounting as balances are kept: i) all emissions from the source data are transferred to the model provided that bridges converged to the aggregate data during balancing, ii) bridges are calibrated for every country and year aggregate data, what no expert could do manually, and iii) the correspondence between classifications is often such that many products of EXIOBASE products correspond to many products in EUROMOD, so it is not only a matter of aggregation in order to keep the balances mentioned above. 4. Reconcile the bridge: The procedure to reconcile the bridge to the EXIOBASE and EUROMOD data is based on the bi-proportional balancing technique RAS commonly used in the 36 Figure B.4: Distributional effect: change in S8020 ratio Figure B.5: Carbon tax (S1.1) vs consumption taxes: average tax rate and redistributive effect 43 Figure B.6: Share of GHG emissions taxed and impact on Gini coefficient: decomposition Food Figure B.7: Share of GHG emissions taxed and impact on Gini coefficient: decomposition Energy 44 Figure B.8: Share of GHG emissions taxed and impact on Gini coefficient: decomposition Transport 45 Getting in touch with the EU In person All over the European Union there are hundreds of Europe Direct centres. You can find the address of the centre nearest you online (european-union.europa.eu/contact-eu/meet-us_en). On the phone or in writing Europe Direct is a service that answers your questions about the European Union. You can contact this service: — by freephone: 00 800 6 7 8 9 10 11 (certain operators may charge for these calls), — at the following standard number: +32 22999696, — via the following form: european-union.europa.eu/contact-eu/write-us_en. Finding information about the EU Online Information about the European Union in all the official languages of the EU is available on the Europa website (european-union.europa.eu). EU publications You can view or order EU publications at op.europa.eu/en/publications. Multiple copies of free publications can be obtained by contacting Europe Direct or your local documentation centre (europeanunion.europa.eu/contact-eu/meet-us_en). EU law and related documents For access to legal information from the EU, including all EU law since 1951 in all the official language versions, go to EUR-Lex (eur-lex.europa.eu). EU open data The portal data.europa.eu provides access to open datasets from the EU institutions, bodies and agencies. These can be downloaded and reused for free, for both commercial and non-commercial purposes. The portal also provides access to a wealth of datasets from European countries. The Joint Research Centre (JRC) provides independent, evidence-based knowledge and science, supporting EU policies to positively impact society EU Science Hub Joint-research-centre.ec.europa.eu