Global Environmental Change 74 (2022) 102525 Contents lists available at ScienceDirect Global Environmental Change journal homepage: www.elsevier.com/locate/gloenvcha Assessing the authenticity of national carbon prices: A comparison of 31 countries Adam Finch a, Jeroen van den Bergh a, b, c, * a Institute of Environmental Science and Technology, Universitat Autònoma de Barcelona, Spain b ICREA, Barcelona, Spain c School of Business and Economics & Institute for Environmental Studies, VU University Amsterdam, The Netherlands A R T I C L E I N F O A B S T R A C T Keywords: Many countries have carbon pricing in place, in the form of a tax and/or market. Generally, this involves low Climate policy price rates, incomplete emissions coverage, and price reductions for particular sectors. This raises the question Carbon tax whether the label “carbon price” – in the environmental-economics textbook sense – really applies. To answer it, Emissions trading we assess the authenticity of 31 national carbon prices, calculating average carbon prices and their gap with Energy prices Average carbon price advertised prices, at both national and sector levels. The results indicate a poor level of authenticity. This means that the carbon prices published by sources such as the World Bank provide a misleading representation of the actual national policy pressure on emissions. Countries show considerable differences regarding the average carbon price level and the gap with advertised prices. Moreover, there is not a one-to-one relationship between advertised and average carbon prices, suggesting the former are not a good basis for international comparison of policy effectiveness. Across countries, the mean carbon price equals €7.90/ton of CO2 while the mean price gap is 57.7%. Most noticeably, the highest advertised price for Sweden should be interpreted with care as it goes along with a price gap of almost €100 to the average price. In addition, Switzerland and Finland show relatively high price gaps. To illustrate the relevance and non-triviality of our indicators, note that Sweden occupies a 3rd position in terms of average carbon price (after Norway and Switzerland), 27th in terms of price gap, and 16th in terms of effective rate (i.e. sum of implicit and explicit carbon prices). We further find that implicit carbon prices dominate explicit ones for most countries, notably in road transport, whereas the reverse holds for industrial and electricity sectors. Combining our findings with recent empirical evidence for carbon-pricing effectiveness highlights the potential of the instrument to combat climate change, provided implementation is improved and internationally harmonized. Shifting the attention from advertised to average carbon prices might help in this regard. 1. Introduction leakage (Baranzini et al., 2017). The most common figure referenced with regards to the strength of According to textbooks in environmental economics, a carbon price pricing instruments is the upper most price rate implemented on emis­ should be uniform across all polluters (Perman et al., 2011; Stern, 2007; sions covered in a jurisdiction, before discounts are permitted. This is Aldy et al., 2010; Cramton et al., 2017). The rationale behind such an the nominal price used by the World Bank for their carbon pricing approach is that the emission of any carbon-dioxide molecule at a dashboard (World Bank, 2020) and is referred to by the OECD as the certain time has the same impact on global warming and associated ‘standard rate’ (OECD, 2016). In this paper we call it the “advertised environmental, economic and health damages, regardless of where it is price”, to distinguish it from an “average price” which we argue to emitted. A uniform price thus results in effective emissions reduction provide a better basis for an aggregate-level comparison of countries. We against minimal costs or welfare losses. In addition, if price uniformity will show that the advertised price can be highly misleading about the extends to all sectors and countries, carbon pricing will not cause unfair actual policy pressure on emissions. While the authors of carbon pricing competition, in turn limiting carbon rebound and preventing carbon databases and assessments (such as the World Bank’s annual “State and * Corresponding author at: ICTA-UAB, Edifici Z, UAB Campus, 08193 Bellaterra, Spain. E-mail address: jeroen.bergh@uab.es (J. van den Bergh). https://doi.org/10.1016/j.gloenvcha.2022.102525 Received 19 May 2021; Received in revised form 11 April 2022; Accepted 26 April 2022 Available online 30 April 2022 0959-3780/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). A. Finch and J. van den Bergh Global Environmental Change 74 (2022) 102525 Trends of Carbon Pricing”) are probably aware of the incomplete picture first is the average price paid per ton of CO2 in an economy, which takes provided by the nominal carbon prices they advertise, this is generally into account all price and emission-coverage levels used. It thus accounts not clear to a wide group of social scientists, the media, political for the total charge of carbon price rates implemented to the total carbon stakeholders and the broader public. Hence, our analysis fills a gap and emissions generated in that economy, giving the average charge per unit contributes to better informed debate about actual implementations and of emissions. The higher the average price rate, the greater the emissions comparisons of carbon pricing around the world. Note that Fig. 3 of reduction capability of the carbon price. The second criterion measures World Bank (2020) (and Figure 2.3 of World Bank, 2021) – probably the the difference between average and advertised carbon prices, referred to most influential overview of carbon prices worldwide – plots advertised as the price gap. This captures the combined effect of partial emissions prices against emission coverage. But the report does not offer graphs coverage by an instrument and the use of price discounts in certain with average prices and price gaps. As an illustration, this results in the sectors. A lower price gap indicates that the advertised carbon price is World Bank information presenting Sweden as having the highest car­ associated with a higher level of emissions coverage and price unifor­ bon price worldwide, while in terms of average carbon price we find it is mity. A smaller gap between advertised and average carbon price, and a only in third position (our Fig. 1), in price gap it is 27th (our Fig. 2), and higher average carbon price indicate more authenticity. Appendix A2 in effective rate (sum of implicit and explicit carbon prices) it is 16th provides more details about these criteria and their motivation, and (our Fig. 5). This illustrates the relevance and non-triviality of our Appendix A3 about the sample of countries and data. analysis. The advantage of an average carbon price is that it accounts for all We define a carbon price with total authenticity as one that applies the information about the advertised price, emissions coverage within uniformly, i.e. without any price differentiation, to 100% of the emis­ and between sectors, and price differentiation or discounts. This pro­ sions within a jurisdiction. A smaller gap between advertised and vides a reasonable measure of the average regulatory pressure to reduce average carbon price, and a higher average carbon price indicate more emissions in the country (or sector, if applied to a specific sector). authenticity. Our study assesses the extent to which existing carbon Indeed, if in some (sub)sector prices are high and in another low, then in prices – whether implemented through emissions trading systems (ETS), the first emissions reduction will be high and in the second low, so carbon taxes or a combination – differ from the authentic carbon price. moderate overall – which an average price that is also moderate in this Deviations from this authentic price are caused by low price rates, un­ case would reflect well.1 covered emissions, and price discounts or differentiation in the carbon Fig. 1 shows the drastic difference between the average carbon prices price. Previous studies have documented existing carbon prices at the measured in this study and the advertised rates, demonstrated by the time, but not answered, nor asked, this critical question (Haites, 2018). grey section of the bar. The difference is particularly significant for We should note that OECD (2016, 2018) calculates average effective countries operating a combination of carbon tax and ETS pricing. It carbon rates, which differ from our average carbon prices as they means that advertised prices are providing a misleading representation include non-carbon energy taxes. Hence, these rates are comparable of the actual price rates implemented. Sweden demonstrates a price gap with the implicit carbon prices we plot in our Fig. 5. However, the of almost €100, indicating that its high advertised price (€127.04) essence and originality of our figure is that it offers a comparison be­ should be interpreted with care. In addition, Switzerland and Finland, tween explicit and implicit average carbon price at a country level, and to a lesser extent Norway and France, show considerable price gaps. which OECD does not do – it instead compares with the advertised The reason for the observed price gaps in Fig. 1 differs between in­ carbon prices. Hence, we go a step further than OECD by comparing struments, i.e. carbon tax or ETS. Regarding the first, countries tend to average implicit and explicit carbon prices. In addition, neither OECD charge multiple rates in the implementation of the tax, often associated nor World Bank calculate price gap indicators as we do. So while we with specific sectors. In sources such as the World Bank Carbon pricing recognize that OECD has done important groundwork in this respect, the dashboard it is suggested that the advertised price represents the average topic has never received focused attention in concise and accessible form carbon price charged on emissions covered by the tax. However, we find in the peer-reviewed literature. that this does not hold true for any of the countries with a carbon tax. In order to test for carbon-price authenticity, an economy-wide The price rate charged through ETS pricing is far more transparent, average price is calculated for each country. This is the average price because each permit has a single price governed through a market sys­ paid per ton of CO2 in an economy. It accounts for relevant price dif­ tem. But as Fig. 1 shows, while the price charged per ton of CO2 is the ferentiation and incomplete coverage of total carbon emissions gener­ same in all EU countries, the average price outcome varies greatly be­ ated in that economy. In addition, a similar sector-level average price is tween each country. For example, Luxembourg displays the lowest calculated from carbon-pricing data for six sectors in each study country: average rate of €1.95, while Greece’s average rate is at €8.40. This is due ‘road transport’, ‘other transport’, ‘industrial’, ‘agriculture & fishing’, to a number of factors, such as free allocation of permits and differences ‘residential & commercial’ and ‘electricity’. From the advertised and in the number of large emitters (and sector structure) required to pay for average prices we derive a percentual price gap indicator which serves permits under the EU-ETS. as an original, transparent indicator of the authenticity of carbon prices. Appendix A1 provides more details about the nature of these 3. Economy-wide carbon pricing calculations. The remainder of this article is organized as follows. Section 2 The results for the average price and price gap criteria show that the compares economy-wide advertised and average carbon prices between carbon pricing implemented around the world today is not very 31 countries. Section 3 examines the price-gap performance of economy- authentic. Of the two criteria, the average price rates are particularly wide carbon prices, relating it to the average price rate. Both sections poor across all countries. Fig. 2 shows that most countries have low show clear differences between three groups of countries, with carbon average prices. The majority is even below the €10 and the highest of tax, ETS and a combination of the two. Section 4 then contrasts sector- Norway is still below €40. Countries operating combined pricing level average and advertised carbon prices. Section 5 broadens the scope by comparing (explicit) carbon pricing with implicit carbon prices, at both country and sector levels. Section 6 concludes and draws policy 1 One commentator suggested that only concrete prices provide incentives, conclusions. questioning the usefulness of average prices. Apart from this questioning the general usefulness of statistical aggregation and analysis in scientific research, 2. Comparing average and advertised carbon price rates one should realize that advertised prices are not generally applied because of exceptions (price differentiation and uncovered emissions), and hence cannot The authenticity of a carbon price will be based on two criteria. The be seen as concrete prices either. 2 A. Finch and J. van den Bergh Global Environmental Change 74 (2022) 102525 Fig. 1. Comparison between economy-wide advertised and average carbon prices per ton of CO2. Fig. 2. Average price rate and price-gap performance of economy-wide carbon prices. generally have a higher price than countries operating a single instru­ countries demonstrating a price gap of 50% to 70% between their ment, which may seem obvious but is not a necessary outcome as a average and advertised price. The highest price gaps are found in single instrument might be relatively ambitious. The high combined Luxembourg, Austria and Argentina, while the by far best performing price is likely due to a high level of commitment to carbon pricing countries for this indicator are New Zealand, Korea and Japan. It cannot amongst these countries, reflected by next to participating in EU-ETS be concluded that utilising ETS, carbon tax or combined pricing led to also implementing a national carbon tax. countries having a lower price gap as the variation in each group is fairly The price gap indicator shows that many countries implement a price large. with poor emissions coverage and price uniformity, with the majority of Across all countries, the mean economy-wide price per ton of CO2 3 A. Finch and J. van den Bergh Global Environmental Change 74 (2022) 102525 Fig. 3. Sector-level average and advertised carbon prices per ton of CO2. *Mean value across countries weighted for emission levels. Fig. 4. Average price rate and price-gap performance of sector-level carbon prices. equals €7.90, which is considerably lower than the mean advertised than in any other and is implemented with a much smaller price gap price of €18.68. Hence, the mean price gap is €10.78 or 57.7% of the (€2.86). This is less than one fourth of the price gaps measured for the mean advertised price. other sectors (ranging between €14.15 and €17.31). The smaller price gap represents the strong emissions coverage and price uniformity in the 4. Sector-level carbon pricing ‘electricity’ sector, as well as its lower advertised price rate. This is the result of a small proportion of emissions coming from countries with Fig. 3 shows that sector-level average prices across countries are all combined pricing, as these tend to be the nations with the highest below €22. However, while all sectors are categorised in the same advertised rates. In this sector they are only responsible for 10.5% of threshold, there is a clear difference between pricing in the ‘electricity’ emissions, compared to contributions ranging from 19.6% to 24.6% in sector compared to all others. all other sectors. The average price in the ‘electricity’ sector (€10.67) is far greater Fig. 3 also suggests that ETSs and carbon taxes tend to price sectors 4 A. Finch and J. van den Bergh Global Environmental Change 74 (2022) 102525 Fig. 5. Average explicit carbon prices and implicit carbon prices in each country. differently. ETSs are shown to have an extremely low contribution to fuels with a high carbon intensity can be taxed at lower rates than low- pricing in the ‘road transport’, ‘agriculture & fishing’ and ‘residential & intensity alternatives. This translates then into a relative advantage for commercial’ sectors, where carbon taxes are most prevalent. the carbon-intensive option, which goes against the idea of carbon Conversely, for the ‘industrial’ and ‘electricity’ sectors, where the ETS pricing.2 (Lachapelle, 2011). Nevertheless, a full perspective on the contribution is greatest, the carbon tax contribution is far lower. The implications of carbon prices requires that implicit prices are accounted ‘other transport’ sector is unique in that there are similar levels of for as well. Another reason to assess the sum of the two is that it makes a contribution from both instruments. This reflects that ETSs and carbon difference whether countries substitute implicit by explicit carbon prices taxes are implemented in a complementary fashion between sectors. or instead add the second type to the first. Because in the latter case, the The results from the average price and price-gap criteria show that regulatory effect will be stronger. sector-level pricing averaged among emissions across all countries is With this in mind, Fig. 5 shows the average effective carbon rates in inauthentic. The average price rate in each sector is very low: Fig. 4 each of our study countries and compares these with the explicit average shows that all sectors have a price which is near or below €10. As has carbon prices (so the difference is the implicit average carbon price). been highlighted already in this section, by far the highest authenticity is The implicit rates are measured using fuel excise tax rates in each found in the electricity sector, where the average price is almost €3 economy. The key takeaway from this figure is that the implicit prices higher and the price gap at least three times lower than in any other tends to dominate the explicit ones in most countries. A similar sector. conclusion was drawn by the OECD (2016) but based on advertised instead of average carbon prices. South Korea, Finland, Sweden, France 5. Comparing carbon pricing with implicit carbon prices and Norway are the only countries where carbon price rates are greater than 25% of effective carbon rates. The figure shows that there is no It is not only carbon pricing which influences emission levels in an clear correlation between the explicit carbon price and the overall economy. Other energy taxes indirectly price emissions. The sum of effective carbon rate, i.e. countries with high explicit carbon prices do explicit and implicit prices results in a so-called effective carbon rate, not necessarily have high effective rates. For example, whereas which is the total charge per ton of CO2 (OECD, 2016). Although both Luxembourg has one of the lowest average explicit prices, its average types of pricing are capable of incentivising carbon emissions reduction, effective rate is the third highest in the sample. Conversely, Sweden has it is important to distinguish between them because their accuracy of one of the highest average explicit price rates, but its average effective incentivizing such reduction differs. Unlike carbon pricing, general en­ rate is mid-range relative to other countries in the dataset. Such ten­ ergy taxes provide an imprecise incentive for carbon emissions reduc­ dencies may indicate that countries such as Sweden are transforming tion because their tax levels are not set in proportion to the carbon implicit carbon prices to explicit prices. However, it is also possible that content of associated energy sources. In fact, under general energy taxes the large gap for Sweden is due to tax base erosion, i.e. past effectiveness 2 To illustrate, diesel produces more CO2 per litre than gasoline when com­ busted, which means that its carbon tax should be higher. But in many countries the energy tax per litre of diesel is in fact lower (of course, fuel taxes may have other motivations, such as tax revenues or local particulate contamination). To further complicate, since the fuel efficiency of diesel is considerably higher than of gasoline, diesel cars tend to emit less CO2 per km than comparable gasoline cars. 5 A. Finch and J. van den Bergh Global Environmental Change 74 (2022) 102525 of the high carbon tax in terms of emissions reduction in sectors where it large. Indeed, across all countries, the mean economy-wide price per ton applied. This would represent a positive step forward for carbon pricing of CO2 is only €7.90, while the price gap is more than half (57.7%) of the in securing consistency and accuracy in national carbon prices, as well as mean advertised price. Two key insights of our analysis are as follows. contributing to a harmonisation of carbon prices at an international First, advertised prices should be interpreted with caution as they tend level. But it also means that the stringency of a country’s climate policy to exaggerate the carbon price pressure on economic activities. Instead, can be easily overestimated. average carbon prices are a more reliable measure as they account for Undertaking a similar analysis at a sector level, Fig. 6 shows that emissions coverage and price differentiation. Second, countries show implicit prices are implemented unevenly across sectors. It is centred in considerable differences with respect to the average carbon price level the road transport sector, contributing 96% of the effective carbon rate as well as to the gap between advertised and average carbon prices. of €177.02 in this sector. This explains the low reactivity of carbon Moreover, these differences do not have a one-to-one relation with di­ emissions in this sector to explicit carbon pricing efforts, as its influence versity in advertised carbon prices. Hence, if one wants to compare is so much lower than that of implicit charges (Tvinnereim and Mehling, effectiveness among countries, the advertised prices form a unreliable 2018). The second highest effective rate is as low as €28.89, for the basis, and one would do better to use the average prices for international agriculture & fishing sector. comparison of policy effectiveness. Although implicit prices are not so high in the ‘other transport’, Similar issues were found at a sector level too. Average sector prices ‘agriculture & fishing’ and ‘residential & commercial sectors’, they are measured across all countries were well below the Paris’ €35 benchmark still dominant over explicit prices in the composition of the effective with the ‘agriculture & fishing’ and ‘residential & commercial’ sectors rate, contributing 72%, 90% and 80%, respectively. Meanwhile, the low demonstrating particularly high price gaps. Only pricing in the ‘elec­ implicit prices in the ‘industrial’ and ‘electricity’ sectors mean that these tricity’ sector was found to be far superior to any other sector, with a are the only ones where explicit pricing is dominant, but they also have price gap of only 21%, indicating strong emissions coverage and price the lowest effective carbon rates, €12.19 and €13.38, respectively. In uniformity. these cases, it is likely that carbon prices have a clearer impact on The previous section put the average pricing calculations in the emissions reduction relative to other sectors. context of implicit carbon price rates in each country and sector ana­ This section has shown that taking into account implicit prices both lysed. A general dominance of implicit over explicit carbon pricing was nuances and complicates the understanding of overall carbon-pricing found in most countries. However, the sectoral analysis demonstrated performance of countries and sectors. that the majority of implicit prices is focused in the ‘road transport’ sector, where rates are six or more times higher than in any other sector. 6. Conclusion and policy implications In contrast, implicit prices were shown to be much lower in the ‘in­ dustrial’ and ‘electricity’ sectors, where explicit carbon prices dominate Our study shines a fresh light on carbon prices worldwide and can implicit ones. help to improve the quality of the debate about what has been achieved While we find low average carbon prices, recent empirical studies so far, and what are the most important areas requiring improvement. indicate these have already been quite effective in reducing emissions or We proposed the average carbon price charged across total emissions in curtailing emissions growth. One study estimated for OECD countries the economy as a clear and fair approach to summarize implemented that one € increase in energy taxes reduces carbon emissions from fossil carbon prices. In fact, this average carbon price could serve as a stand­ fuel consumption by 0.73% in the long run (Sen and Vollebergh, 2018). ardised way of publicising carbon prices as it succinctly identifies the Another uses data for 142 countries over two decades, 43 of which had a average strength of the price incentive for emissions reduction and al­ carbon price in place by the end of the study period (Best et al., 2020). It lows for a direct comparison between countries or sectors. finds that the average annual growth rate of CO2 emissions from fossil- Our analysis finds that carbon pricing implemented around the world fuel combustion has been around 2% lower in countries with a carbon is not very authentic. The key problems for almost all countries is that price, and that an additional € per tonne of CO2 reduces annual emis­ the average price level is low and the price gap with the advertised price sions growth by about 0.3%. A third study concluded that the EU-ETS, Fig. 6. Average explicit carbon prices and implicit carbon prices in each sector. 6 A. Finch and J. van den Bergh Global Environmental Change 74 (2022) 102525 developed to meet the Kyoto Protocol, reduced covered emissions by Finally, future research could give more attention to three issues. 7.5%. This is equivalent to 3.8% of total EU-wide emissions, repre­ First, an intertemporal analysis along the lines of our approach would be senting about half of the EU’s Kyoto aim (Bayer and Aklin, 2020). Since useful to see if average carbon prices and gap with advertised prices all these results are for rather weak carbon prices as assessed in our follow a trend and whether differences among countries become smaller study, one should expect considerable emissions reduction with higher over time. In addition, this could test if gaps get smaller over time due to carbon prices. erosion of the carbon-tax base. This might, for instance, explain why Currently low carbon prices are due to climate policy being pre­ Sweden has such a large gap. Second, the statistical comparison of dif­ dominantly unilateral in nature, despite the Paris Agreement. Indeed, ference between carbon pricing by countries, as undertaken here, could the latter focuses on voluntary emission targets and did not achieve an be usefully complemented by historical information obtained from case inter-country coordination of climate policies and their strength. studies about institutional and policy processes. This could clarify public Harmonizing explicit carbon prices among countries, such as is already support for, and political feasibility of, carbon pricing and other climate happening through ETS systems, will encourage substitution of wide­ policies in the various countries. This is an immense task given the 31 spread non-carbon energy taxes by carbon taxes, facilitating an inter­ countries involved, which merits a separate study. Third, broadening the national comparison through explicit carbon prices. This in turn will average-price and price-gap indicators to capture the implicit prices of allow for a gradual transition to coordination or even harmonization of non-price instruments could result in a metric allowing the comparison effective carbon prices around the world, notably if aided by expanding of overall climate policy among jurisdictions. coalitions and border carbon tariffs as contemplated by the EU (Al Khourdajie and Finus, 2020; van den Bergh et al., 2020). In turn, this could support rising carbon price rates necessary to achieve sufficient Declaration of Competing Interest global emissions reduction. In addition, if the advertised or visible prices would be replaced by The authors declare that they have no known competing financial the lower average prices one would expect less resistance. While there interests or personal relationships that could have appeared to influence are cases of resistance which have been repeatedly mentioned in popular the work reported in this paper. and even academic writings (such as the yellow vest protests in France or the Washington state referenda), the truth is much more balanced, Acknowledgements namely there are many carbon prices implemented (most extensively the EU-ETS in 30 countries). Moreover, many public survey studies show We are grateful for helpful comments from Stefan Drews and Eric quite some public support for carbon pricing and provide lessons for Galbraith. This research was funded by an ERC Advanced Grant from the design, notably revenue use, to maximize support (Carattini et al., 2019; European Research Council (ERC) under the European Union’s Horizon Maestre-Andrés et al., 2021). Shifting the discussion from advertised to 2020 Research and Innovation Programme [grant agreement 741087]. lower equivalent average carbon prices, as confirmed by our findings, It also has benefitted from the “María de Maeztu” Programme for Units can make it easier to garner support for higher prices and thus more of Excellence of the Spanish Ministry of Science and Innovation emissions reduction. (CEX2019-000940-M) which was awarded to our institute (ICTA-UAB). Appendices. A1. Calculation of average carbon prices In this section we demonstrate how to calculate the average price of a carbon pricing instrument. The formula can be used for both ETSs and carbon taxes. For countries only operating one instrument, their carbon price will be equal to the average price of the pricing instrument they use. For countries using two instruments, the average price of both instruments will be summed to derive the overall carbon price in the economy. The sector-level average price per country is then calculated as follows: Pcs = Xsets Â⋅Csets + Xsctax Â⋅Csctax Here: Xsets = Average implemented prices applied in a sector (s) by an ETS. Xsctax = Average implemented prices applied in a sector (s) by a carbon tax. Cets s = Emissions coverage applied in a sector (s) by an ETS. Cctax s = Emissions coverage applied in a sector (s) by a carbon tax. Pcs = Average carbon price in a sector (s) implemented by a single country (c). Where the super-index ctax denotes “carbon tax” and ets denotes “emission trading system”. The economy-wide average price is calculated as follows: ∑ n Pc = Pcs Â⋅Esc /Ec s=1 Here: Ecs = The sector emissions (s) in a country (c). Ec = Total emissions in a country (c). Pc = Average price implemented in a country (c). The percentual price gap for the economy-wide carbon prices is calculated as follows: ( ) Ac − Pc Gc = Â⋅100 Ac Here: 7 A. Finch and J. van den Bergh Global Environmental Change 74 (2022) 102525 Ac = Advertised price in a country (c). Gc = Percentual price gap in a country (c). The sector-level average price for emissions generated across all countries is calculated as follows: ∑ n Pas = Pcs Â⋅Esc /Esa c=1 Here: Eas = The total sector emissions (s) in all countries studied (a). Pas = The sector-level average price (s) for all countries studied (a). The percentual price gap for the sector-level carbon prices is calculated as follows: ( a ) As − Pas Gas = Â⋅100 Aas Here: Aas = Mean of advertised prices for all countries studied (a), weighted for sector-level emission levels (s) in each country. Gas = The sector-level percentual price gap (s) for all countries studied (a). A2. Procedure to assess the authenticity of carbon prices There are three factors which govern the strength of a carbon price in an economy. The first is the average price implemented in relation to 100% of the emissions in the economy, not just the emissions covered by the pricing instrument(s). The second is the percentage of emissions covered in an economy, which ideally should be 100% of emissions. The third is the price uniformity (or differentiation) of the actual price levels implemented in an economy, relative to the advertised price. In the perfect scenario all prices implemented should be equal to the advertised price, as such there should be no price reductions permitted. We have designed two criteria which we will use to assess these three factors for each carbon price. (1) Average carbon price level: Calculation of the average carbon price is designed as an indicator of the influence carbon prices have on emission levels in the national economy. The greater the average price level implemented, the greater the emissions reduction capability of the carbon price. (2) Percentual price gap: The second criterion measures the difference between the average price and advertised price and is referred to as the price gap. This indicator combines information relating to the emissions coverage and price uniformity of the carbon price. The reason for combining the factors is threefold: it is difficult to accurately separate this information when two pricing instruments are in operation; deviation from carbon price uniformity is only relevant to carbon taxation, not to ETS pricing; and it simplifies graphical representation. The assumption for this indicator is that the lower the price gap, the higher level of emissions covered and the closer the uniformity of the prices implemented with the advertised rate. Now a smaller gap between advertised and average carbon price, and a higher average carbon price indicate more authenticity, meaning that the associated policy pressure on emissions reduction will be more consistent and stronger. A3. Data The data for the composition of the carbon price calculations were taken from recent OECD research projects. The data on emissions, emissions coverage and advertised prices for carbon taxation was extracted from OECD (2019) and for ETS from OECD (2018) and Postic and Métivier (2019). There are slight variations in the emissions data between the two datasets, but this has a minimal effect on the average price calculations that form the core of the analysis. The OECD datasets use the most up to date carbon emissions statistics collected by the International Energy Agency in 2016 (IEA, 2018). In comparison, emissions data used in World Bank carbon pricing dashboard dates to 2012. A further benefit of the OECD reports is that they provide in- depth sectoral information for each carbon price. Emissions and pricing levels are available for six sectors: road transport, other transport (aviation, rail, shipping), industrial, agriculture & fishing, residential & commercial, and electricity. In selecting the countries for study, our goal was to include as many national carbon pricing instruments operating around the world. The reports do not include information for countries outside of the OECD which claim to have carbon prices, such as Kazakhstan, Indonesia and Taiwan as we could not guarantee that their data are consistent with the OECD data. We also omitted countries with regional carbon pricing instruments that were included in the OECD reports, notably China, Canada and the US. This was because our focus is on national instruments and we found that analysing regional pricing based on national emissions statistics distorted results. We were also unable to find supplementary data of equal quality to the OECD reports in order to include these countries and regions into our study. For eight countries participating in EU-ETS we could not find consistent data. These are countries, and generally small emitters, who do not fall within the remit of the OECD and include: Cyprus, Lithuania, Latvia, Liechtenstein, Romania, Bulgaria, Malta, Croatia. Following these choices, we obtained a strong and consistent sample of 14 countries using only ETS, 5 countries using only carbon taxes and 12 countries using both. These countries are as follows (abbreviations appear in the figures in the main text): • Countries with an ETS: Austria (AUT), Belgium (BEL), Czech Republic (CZE), Germany (DEU), Greece (GRC), Hungary (HUN), Italy (ITA), South Korea (KOR), Luxembourg (LUX), Netherlands (NED), New Zealand (NZ), Poland (POL), Slovak Republic (SLK) and Spain (ESP). • Countries with a carbon tax: Argentina (ARG), Chile (CHL), Mexico (MEX), Colombia (COL) and Japan (JAP). • Countries with combined carbon pricing: Denmark (DNK), Estonia (EST), Finland (FIN), France (FRA), Iceland (ISL), Ireland (IRE), Norway (NOR), Portugal (PRT), Slovenia (SVN), Sweden (SWE), Switzerland (CHE) and UK (GBR). 8 A. Finch and J. van den Bergh Global Environmental Change 74 (2022) 102525 References Maestre-Andrés, S., Drews, S., Savin, I., van den Bergh, J., 2021. Carbon tax acceptability with information provision and mixed revenue uses. Nat. Commun. 12, 7017. OECD, 2016. Effective Carbon Rates: Pricing CO2 through Taxes and Emissions Trading Aldy, J., Krupnick, A., Newell, R., Parry, I., Pizer, W., 2010. Designing climate mitigation Systems. Organisation for Economic Co-operation and Development, Paris. policy. J. Econ. Lit. 48 (4), 903–934. OECD, 2018. Supplement to Effective Carbon Rates 2018: Country Profiles. Organisation Al Khourdajie, A., Finus, M., 2020. Measures to enhance the effectiveness of for Economic Co-operation and Development, Paris. international climate agreements: the case of border carbon adjustments. Eur. Econ. Perman, R., Ma, Y., Common, M., Maddison, D., McGilvray, J., 2011. Natural Resource Rev. 124, 103405. and Environmental Economics, 3rd ed. Pearson Education Limited, Harlow. Baranzini, A., van den Bergh, J., Carattini, S., Howard, R., Padilla, E., Roca, J., 2017. Postic, S., Métivier, C., 2019. Global Carbon account (2019). Organisation for Economic Carbon pricing in climate policy: seven reasons, complementary instruments, and Co-operation and Development, Paris. political-economy considerations. WIREs Clim. Change 8 (4), e462. Sen, S., Vollebergh, H., 2018. The effectiveness of taxing the carbon content of energy Bayer, P., Aklin, M., 2020. The European Union Emissions Trading System reduced CO 2 consumption. J. Environ. Econ. Manage. 92, 74–99. emissions despite low prices. PNAS 117 (16), 8804–8812. Stern, N., 2007. The Economics of Climate Change: The Stern Review. Cambridge Best, R., Burke, P.J., Jotzo, F., 2020. Carbon pricing efficacy: cross-country evidence. University Press, Cambridge, UK. Environ. Resour. Econ. 77, 69–94. Tvinnereim, E., Mehling, M., 2018. Carbon pricing and deep decarbonisation. Energy Carattini, S., Kallbekken, S., Orlov, A., 2019. How to win public support for a global Policy 121, 185–189. carbon tax. Nature 565 (7739), 289–291. van den Bergh, J., Angelsen, A., Baranzini, A., Botzen, W., Carattini, S., Drews, S., Cramton, P., MacKay, D.J.C., Ockenfels, A., Stoft, S., 2017. Global Carbon Pricing: The Dunlop, T., Galbraith, E., Gsottbauer, E., Howarth, R., Padilla, E., Roca, J., Path to Climate Cooperation. The MIT Press, Cambridge, Mass. Schmidt, R., 2020. A dual-track transition to global carbon pricing. Climate Policy 20 Haites, E., 2018. Carbon taxes and greenhouse gas emissions trading systems: what have (9), 1057–1069. we learned? Climate Policy 18 (8), 955–966. World Bank, 2020. State and Trends of Carbon Pricing 2020. World Bank Group, IEA, 2018. Extended world energy balances. IEA World Energy Statistics and Balances. Washington D.C. International Energy Agency, Paris. World Bank, 2021. State and Trends of Carbon Pricing 2021. World Bank Group, Lachapelle, E., 2011. The hidden factor in climate policy: implicit carbon taxes. Washington D.C. Sustainable Prosperity. University of Ottawa, Ottawa. 9