1. Introduction
Emissions of carbon dioxide (CO2), a key driver of global warming and climate change, have continued to increase globally in recent years. If current climate policies are unchanged, standard climate change scenarios predict an increase of around 3°C in global temperatures compared with pre-industrial levels over the course of the century (Group of Reference Moessner30, 2020; United Nations Environment Programme, 2024). This could have catastrophic consequences.Footnote 1 To avoid such a scenario, climate policies need to be expanded in order to reduce the emission of CO2 and other greenhouse gases (IPCC, Reference Pörtner, Roberts, Tignor, Poloczanska, Mintenbeck, Alegría, Craig, Langsdorf, Löschke, Möller, Okem and Rama2022). Such policies include carbon taxes and emissions trading systems (ETS), but also broader changes in regulation (Stern, Reference Stern2007, Reference Stern2008).
To be consistent with global emissions that limit an overshoot of the goal from the Paris Agreement to 1.5°C global warming above pre-industrial levels, global net anthropogenic CO2 emissions would need to decline by about 45% from their 2010 level by 2030, reaching net zero around 2050. To cap global warming at 2°C, CO2 emissions need to decrease by about 25% from the 2010 level by 2030 and reach net zero around 2070. Yet, according to recent estimates, the total global greenhouse gas emission level in 2030 is expected to be 16% above the 2010 level (United Nations, 2021).
Carbon pricing can be an effective policy to reduce carbon emissions. Higher carbon prices incentivise carbon emitters to develop and use economical ways of reducing carbon emissions (OECD, 2021), which encourages the development of low-carbon technologies (United Nations Framework Convention on Climate Change (UNFCCC), 2025). Higher carbon prices make low-carbon energy more competitive, provide incentives to reduce emissions, and reduce demand for carbon-intensive fuels (Arlinghaus, Reference Arlinghaus2015; Martin et al., Reference Martin, Muûls and Wagner2016; OECD, 2021). Moreover, a strong commitment to higher carbon prices by governments provides incentives for investors to invest in the expansion and development of low-carbon technologies (OECD, 2021) and to shift away from high-carbon-emission fossil-fuels-based technology (UNFCCC, 2025). Furthermore, carbon pricing generates revenue that can be used by the government to support research and development of low-carbon technologies (UNFCCC, 2025). The revenue raised from carbon taxes also allows the government to provide options for firms and households to switch more easily to renewable energy and increase energy efficiency, for example by providing government subsidies for district heating, public transport and housing insulation.
In this article, we examine the effects of climate policies on CO2 emissions. We provide ex post empirical analysis of the effects of carbon pricing on carbon emissions at the aggregate national level, based on a comprehensive database of 121 countries. We rely on carbon emissions data and macroeconomic variables over the 1971–2016 period, as well as data on climate policies. As climate policies, we consider national and supranational carbon taxes and ETS. We also consider a broad index that measures the overall stringency of climate policies for OECD and major emerging economies. This index captures regulatory responses that go beyond carbon pricing. We use dynamic panel regressions, to account for the large degree of persistence in emissions. We also control for macroeconomic factors such as economic development, GDP growth, urbanisation and the composition of the electricity mix based on Kohlscheen et al. (Reference Kohlscheen, Moessner and Takats2021).
Overall, we find statistical evidence that higher carbon taxes and prices of permits in ETS have significantly reduced carbon emissions. An increase in carbon taxes by $10 per ton of CO2 equivalents (tCO2) reduces CO2 emissions per capita by 1.3% in the short run and by 4.6% in the long run. This effect is statistically significant in all econometric specifications, with p-values that are always below 0.01. The same increase in the prices of ETS permits also reduces CO2 emissions per capita by 1.4% in the short run and 5.0% in the long run. The magnitude of this effect, however, varies more across specifications and is statistically significant in five out of nine specifications (with p-values below 0.05). Furthermore, more stringent climate policies as measured by a broad index for OECD and major emerging economies also significantly reduce carbon emissions, with a standard deviation increase in the index reducing CO2 emissions per capita by around 1.5% in the short run and 6% in the long run (p-values below 0.05 in five out of six specifications). The estimates are robust to the inclusion of controls for the overall quality of governance in the respective countries, as proxied by an index for control of corruption.
This paper adds to the growing literature that examines the impact of carbon taxes and ETS on CO2 emissions, reviewed in Green (Reference Green2021) and Döbbeling-Hildebrandt et al. (Reference Döbbeling-Hildebrandt, Miersch, Khanna, Bachelet, Bruns, Callaghan, Edenhofer, Flachsland, Forster, Kalkuhl, Koch, Lamb, Ohlendorf, Steckel and Minx2024). As Green (Reference Green2021) noted, relatively few papers have analysed the ex post effects of carbon pricing on carbon emissions at the aggregate national level, and these papers mostly focus on a few specific countries and then extrapolate the results to a broader set of countries. We contribute to this literature by analysing a wide range of countries. This allows us to obtain more precise estimates of the impact of policies.
Moreover, as noted by Döbbeling-Hildebrandt et al. (Reference Döbbeling-Hildebrandt, Miersch, Khanna, Bachelet, Bruns, Callaghan, Edenhofer, Flachsland, Forster, Kalkuhl, Koch, Lamb, Ohlendorf, Steckel and Minx2024), there is a critical evidence gap in the literature regarding ex post estimates of the carbon price elasticity of emissions reductions, rather than just studying the effect of introducing a carbon pricing scheme. We contribute to filling this gap in the literature by providing ex post estimates of the carbon price elasticity of emissions reductions for both carbon taxes and ETS.
Our findings are relevant for future climate policies. More specifically, the finding that higher carbon taxes and prices of permits in ETS have reduced carbon emissions provides evidence that such tools could speed up the necessary transition to a world with much lower emissions. Additionally, our findings suggest that policymakers can rely on a wider range of climate policies to speed up the transition to lower carbon emissions, ideally to net-zero emissions.
The remainder of the paper proceeds as follows. Section 2 discusses the related literature. Section 3 summarises the data. Section 4 presents the methodology. Section 5 presents our empirical estimates. Section 6 concludes.
2. Literature
This paper mainly builds on two strands of the literature. First, it builds on the literature examining the effects of climate policies, such as carbon taxes and ETS, on carbon emissions. Second, it builds on the literature studying macroeconomic determinants of carbon emissions.
The effects of climate policies, such as carbon taxes and ETS, on carbon emissions have been studied in several papers. A comprehensive review of the literature on the effects of carbon pricing on carbon emissions was carried out by Green (Reference Green2021). She highlighted that surprisingly few papers have conducted an ex post empirical analysis of how carbon pricing has actually affected CO2 emissionsFootnote 2 and that the vast majority of these papers are focused on Europe. She noted that in most cases, studies have estimated emissions reductions in the sectors covered by the carbon pricing policy, although some extrapolate to broader jurisdictional effects (e.g. Murray and Maniloff, Reference Murray and Maniloff2015; Bayer and Aklin, Reference Bayer and Aklin2020; Rafaty et al, Reference Rafaty, Dolphin and Pretis2020).Footnote 3 She concluded that the majority of the studies suggest that the aggregate reductions from carbon pricing on emissions are limited, generally between 0% and 2% per year, with considerable variation across sectors. She also concluded that in general, carbon taxes perform better than ETS and that studies of the European Union’s ETS indicate limited average annual reductions in carbon emissions of 0% to 1.5%.
Our paper contributes to filling this gap in the literature by providing ex post empirical analysis of the effects of carbon pricing on carbon emissions at the aggregate national level, for a very broad sample of countries. There is also little empirical literature on the effects of broader climate policies on emissions, and we contribute to filling this gap by studying the effects of an index of broad climate policies on CO2 emissions.
The recent systematic review and meta-analysis of ex post evaluations of the effects of carbon pricing by Döbbeling-Hildebrandt et al. (Reference Döbbeling-Hildebrandt, Miersch, Khanna, Bachelet, Bruns, Callaghan, Edenhofer, Flachsland, Forster, Kalkuhl, Koch, Lamb, Ohlendorf, Steckel and Minx2024) highlighted a critical evidence gap regarding the carbon price elasticity of emissions reductions (i.e. the effect of a marginal change in the carbon price on emissions). It noted that only nine primary studies estimated such a carbon price elasticity, four of which only studied the carbon tax in the Canadian province of British Columbia to estimate elasticities for the transport and buildings sectors there. They concluded that having only nine price elasticity studies provided them with too few effect sizes for meta-analysing these price elasticities separately. Instead, they only conducted a meta-analysis of the effect of introducing carbon pricing, i.e. of the treatment effect (Döbbeling-Hildebrandt et al., Reference Döbbeling-Hildebrandt, Miersch, Khanna, Bachelet, Bruns, Callaghan, Edenhofer, Flachsland, Forster, Kalkuhl, Koch, Lamb, Ohlendorf, Steckel and Minx2024). Our paper also contributes to filling this gap in the literature of estimating carbon price elasticities of carbon emissions reductions, by providing evidence on the effectiveness of carbon pricing relative to the level of the carbon price, rather than just on the effect of introducing a carbon pricing scheme.
There is limited empirical evidence that higher carbon prices reduce carbon emissions (Arlinghaus, Reference Arlinghaus2015; Martin et al., Reference Martin, Muûls and Wagner2016; OECD, 2021). Sen and Vollebergh (Reference Sen and Vollebergh2018) found that for OECD economies, an increase in a broad-based tax on energy consumption of €10/tCO2 is expected to lead to a 7.3% reduction in carbon emissions from fossil fuel consumption in the long run. Metcalf and Stock (Reference Metcalf and Stock2023) found that Europe’s carbon taxes led to a cumulative reduction of around 4% to 6% for a $40/tCO2 tax covering 30% of emissions. They argued that emissions reductions would likely be larger for a broad-based US carbon tax, since European carbon taxes do not include sectors with the lowest marginal costs of carbon pollution abatement in the tax base. Related recent papers are Best et al. (Reference Best, Burke and Jotzo2020), D’Arcangelo et al. (Reference D’Arcangelo, Pisu, Raj and van Dender2022) and Schroeder and Stracca (Reference Schroeder and Stracca2023). Based on sectoral analysis, Rafaty et al. (Reference Rafaty, Dolphin and Pretis2020) found an imprecisely estimated semi-elasticity of a 0.5% reduction in carbon emissions growth per average $10/tCO2 carbon price.Footnote 4 An overview of the results from 24 ex ante models in the IPCC’s AR6 Scenario Database for the effects of carbon taxes on CO2 emissions, in comparison with the results from ex post empirical models, was provided in Tol (Reference Tol2023).
The review of Döbbeling-Hildebrandt et al. (Reference Döbbeling-Hildebrandt, Miersch, Khanna, Bachelet, Bruns, Callaghan, Edenhofer, Flachsland, Forster, Kalkuhl, Koch, Lamb, Ohlendorf, Steckel and Minx2024) noted that the contributions of carbon pricing schemes to carbon emissions reductions remain a subject of heated debate in science and policy. It mentioned that there is a critical evidence gap with regard to dozens of unevaluated carbon pricing schemes. It concluded from its meta-analysis of ex post studies that introducing a carbon price led to substantial emission reductions for at least 17 of 21 carbon pricing schemes studied, with statistically significant emissions reductions ranging between −5% and −21% across the schemes (or −4% and −15% after correcting for publication bias).
The effects of macroeconomic variables on carbon emissions were studied in Raupach et al. (Reference Raupach, Marland, Ciais, Quere, Canadell, Klepper and Field2007), Sadorsky (Reference Sadorsky2014), Kasman and Duman (Reference Kasman and Duman2015), Menyah and Wolde-Rufael (Reference Menyah and Wolde-Rufael2010), Gonzalez-Sanchez and Martin-Ortega (Reference Gonzalez-Sanchez and Martin-Ortega2020), Feng et al. (Reference Feng, Davis, Sun and Hubacek2015), Peters et al. (Reference Peters, Marland, Quere, Boden, Canadell and Raupach2012), Doda (Reference Doda2014), Wang (Reference Wang2012) and International Energy Agency (2020), among others. Our approach for controlling for macroeconomic determinants of carbon emissions follows the study of Kohlscheen et al. (Reference Kohlscheen, Moessner and Takats2021). For a review of the literature studying the macroeconomic determinants of carbon emissions, readers are referred to that study.
Ellis et al. (Reference Ellis, Nachtigall and Venmans2019) reviewed ex post empirical assessments on the impact of carbon pricing on competitiveness in OECD and G20 countries in the electricity and industrial sectors. They concluded that there are no significant effects on competitiveness. In turn, Lilliestam et al. (Reference Lilliestam, Patt and Bersalli2021) reviewed the empirical evidence available in academic ex post analyses of the effectiveness of carbon pricing schemes in promoting technological change necessary for full decarbonisation. They considered the European Union, New Zealand and Scandinavia. They found that there is little empirical evidence in this direction so far, with most of the papers on the topic being theoretical. Recent papers studying the macroeconomic effects of carbon taxes include Metcalf and Stock (Reference Metcalf and Stock2023, Reference Metcalf and Stock2020).
3. Data
In order to assess the empirical drivers of CO2 emissions, we collected data from several sources. Data on CO2 emissions per capita, measured in metric tons per capita, were from the World Development Indicators (WDI) of the World Bank. They measure CO2 emissions stemming from the burning of fossil fuels and the manufacture of cement and include CO2 produced during the consumption of solid, liquid and gas fuels and gas flaring. Demographic, economic and energy use data were from the World Bank (GDP per capita, GDP growth, the urbanisation rate, the share of manufacturing in GDP and the share of coal, oil and renewables in electricity generation).Footnote 5 Our database spanned the 1971–2016 period and 121 countries. The country selection was based solely on data availability, and the list of countries in the sample is presented in the Appendix.
Our main variables of interest are CO2 emissions and climate policies.
Carbon emissions emanate mostly from the burning of fossil fuels (to heat, transport goods and people or generate electricity) and the manufacture of steel and cement. The level of carbon emissions, following that of economic development, is highly uneven across countries. Highly developed advanced economies tend to emit large quantities of carbon per capita, while less developed economies, particularly in Africa, tend to emit less (Figure 1). However, there is no definite advanced–emerging economy divide. Due to fast economic development, many emerging market economies (including energy producers and fast-developing East Asian economies) already have high carbon emission levels per capita. Besides, given their larger populations, their contribution to global CO2 emissions has grown very rapidly.

Figure 1. Carbon dioxide emissions per capita.
Notes: CO2 emissions per capita in metric tons per capita. CO2 emissions measure carbon dioxide emissions stemming from the burning of fossil fuels and the manufacture of cement and include carbon dioxide produced during the consumption of solid, liquid and gas fuels and gas flaring. Source: World Development Indicators (WDI) of World Bank, code EN.ATM.CO2E.PC.
Primary climate policies are carbon taxes and ETS. The first carbon tax was introduced in Finland in 1990. In turn, emissions trading has been considered a possible tool for mitigating greenhouse gas emissions since the early 1990s and formed a key part of the Kyoto Protocol agreement (Philibert and Reinaud, Reference Philibert and Reinaud2004). Our database contains information on carbon taxes and ETS implemented at the national and supranational levels since 1990 from the World Bank (World Bank, 2021). Country coverage and prices of these policies are shown in Figures 2 and 3 separately for carbon taxes and ETS, for the start and the end of our sample period in 2016. The share of global greenhouse gas emissions covered by the carbon taxes and ETS at the national and supranational levels is shown in Figure 4. As the figure shows, there has been a huge increase in this proportion.

Figure 2. Prices of carbon taxes as of 2016 and 1971.
Note: Carbon taxes implemented at the national and supranational levels. Nominal prices as of 1 April each year in $/tCO2 (US dollars per metric ton carbon dioxide emissions) equivalents. In the case of the UK, the number refers to the carbon price floor. Source: Carbon Pricing Dashboard, World Bank (2021).

Figure 3. Prices of carbon ETS as of 2016 and 1971.
Note: Prices of the carbon emission trading system (ETS) implemented at the national and supranational levels. Nominal prices as of 1 April each year in $/tCO2 (US dollars per metric ton carbon dioxide emissions) equivalents. EU: European Union. Source: Carbon Pricing Dashboard, World Bank (2021).

Figure 4. Share of global greenhouse gas emissions covered by carbon taxes and ETS at the national and supranational levels.
Notes: Share of global greenhouse gas emissions covered by carbon taxes and emissions trading systems (ETS) at the national and supranational levels, in percentage. The coverage of each carbon pricing initiative is presented as a share of annual global GHG emissions for 1990–2015 based on data from the Emission Database for Global Atmospheric Research (EDGAR) version 5.0, including biofuels emissions. From 2015 onwards, the share of global GHG emissions is based on 2015 emissions from EDGAR. The greenhouse gas emissions coverage for each jurisdiction is based on official government sources and/or estimates.
Source: Carbon Pricing Dashboard, World Bank (2021).
In the case of carbon taxes, governments set the price of carbon emissions and let private agents determine emissions reductions. ETS have two main forms, cap-and-trade and baseline-and-credit ETS. For cap-and-trade ETS, governments set a limit on emissions, and allowances up to this limit are auctioned or allocated according to certain criteria. These permits are then traded, and carbon prices are determined by supply and demand in the market. For baseline-and-credit ETS, baselines for emissions are set for regulated emitters. Emitters with emissions above their designated baseline need to give up credits to make up for these emissions, while those with emissions below their baseline receive credits for these reductions, which they can sell to other emitters (World Bank, 2021).
Besides carbon taxes and ETS, we also evaluated the effectiveness of climate policies using a broader index for the overall stringency of climate policies for OECD member countries and major emerging economies. This was based on the Environmental Policy Stringency Index (EPS), which is compiled by the OECD and available from 1990 to 2015. The index covers a much wider range of climate policies, also considering other regulatory policies (Botta and Kozluk, Reference Botta and Kozluk2014). Generally, stringency is defined as the degree to which environmental policies put an explicit or implicit price on polluting or, more generally, on environmentally harmful behaviour.Footnote 6
We conducted several checks of the robustness of our baseline results. In particular, to control for the quality of governance in a country, we used a measure of the control of corruption from the World Bank. It reflects the perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as “capture” of the state by elites and private interests.Footnote 7
4. Methodology
We quantified the effects of climate policies on carbon emissions through a dynamic panel model. The dynamic specification accounts for the high degree of persistence in CO2 emissions. Throughout the analysis, we controlled for macroeconomic factors, such as economic development (GDP per capita), GDP growth, urbanisation, the share of manufacturing in total output and the energy mix used in electricity production. Note that we included only the respective shares of the energy mix, not the intensity of use. Formally, we explored the effects of climate policies (denoted by CP) on the logarithm of CO2 emissions per capita (denoted by lnCO2) according to the following equation:
$$ {\displaystyle \begin{array}{c}\mathrm{lnCO}{2}_{i,t}={\alpha}_i+{\beta}_t+\rho \hskip0.32em \mathrm{lnCO}{2}_{i,t-1}+\lambda \hskip0.32em {\mathrm{CP}}_{i,t-1}+\gamma \hskip0.32em {\mathrm{GDPpc}}_{i,t}+\theta \hskip0.32em {\mathrm{growth}}_{i,t}\\ {}\hskip2.6em +\hskip0.32em \unicode{x03C9} \hskip0.32em {\mathrm{urbanisation}}_{i,t}+\vartheta \hskip0.32em {\mathrm{manufacturing}}_{i,t}+{\mu}_1{\mathrm{share}}_{\mathrm{oil},i,t}\\ {}\hskip-5.2em +\hskip0.32em {\mu}_2\hskip0.32em {\mathrm{share}}_{\mathrm{coal},i,t}+{\mu}_3\hskip0.32em {\mathrm{share}}_{\mathrm{renewables},i,t}+{\varepsilon}_{i,t}\end{array}} $$
Besides the lagged dependent variable, we included climate policies as key explanatory variables of interest. To address potential endogeneity concerns, we used lagged climate policy variables, which minimised the risk of reverse causality. As climate policies, we considered carbon taxes and prices of permits in ETS implemented at the national and supranational levels, as well as the broader index for the stringency of climate policies described in Section 2. We also included a number of macroeconomic variables as controls.
Our models included country fixed effects to capture unobserved heterogeneities across countries that might affect the rate of CO2 emissions. These included fixed institutional factors such as enforcement of environmental laws. They also included natural factors such as average median temperatures, which tend to correlate with heating or cooling needs. We also included the full set of yearly time dummies to control for the effects of global factors. These subsumed, for instance, technological advances that may reduce environmental effects, as well as other global trends or global shocks.
In this analysis, we used fixed effect panel estimations and based inference on cluster robust standard errors. As a robustness check, we controlled additionally for the quality of governance in a country, using the measure of the control of corruption as described in Section 3.
5. Estimation results
5.1. Effects of climate policies
We first present baseline estimates of our dynamic panel equation (1). The estimations for the effects of carbon taxes, prices of permits in the ETS and the EPS policy index separately are shown in columns I, II and IV of Table 1, respectively. The results when including carbon taxes and prices of permits of ETS together are shown in column III, and those when including all three climate policies together in column V.
Table 1. Effects of climate policies on CO2 emissions

Note: Sample period: From 1971 to 2016, annual data. Cluster-robust standard errors reported in the second line are clustered at the country level. ***/**/* denote statistical significance at the 1%/5%/10% level. (I): nominal price of first carbon tax in USD/tCO2 equivalents; (II): nominal price of ETS in USD/tCO2 equivalents; (IV): EPS index.
Overall, the specifications describe the country-specific evolution of carbon emissions quite well. Our models with climate policies are able to explain 97%–98% of the variation in per capita CO2 emissions across countries and across time (Table 1). Importantly, this is driven by variation both within and between countries, as R2 within ranges from 0.82 to 0.88 and R2 between ranges from 0.97 to 0.99. The lagged dependent variable has a coefficient of just above 0.7 and is clearly statistically significant, which confirms that a dynamic panel specification is indeed appropriate. Put differently, more than 70% of CO2 emissions can be explained by previous year emissions alone.Footnote 8 Most of the macroeconomic control variables are statistically significant with the expected signs.
Consistently, we found that higher carbon taxes significantly reduce carbon emissions (p-values always below 0.01). On average, an increase in carbon taxes by $10 per ton of CO2 (tCO2) reduces CO2 emissions per capita by 1.3% in the short run and by 4.6% in the long run (baseline model in column III of Table 1).Footnote 9 Importantly, this result is robust to controlling for ETS prices (see column III).
We also found that higher prices of ETS permits reduce carbon emissions (p-values below 0.01 in columns II and III of Table 1). An increase in prices of ETS permits by $10/tCO2 reduces CO2 emissions per capita by 1.4% in the short run and by 5.0% in the long run, when carbon taxes are controlled for (column III).
Moreover, we found that the broad EPS index of climate policy stringency has a negative effect on carbon emissions at the 10% significance level (column IV of Table 1). A one standard deviation increase in the EPS index reduces CO2 emissions per capita by 1.6% in the short run and 6.2% in the long run. It is striking that these results are broadly similar to the results we obtained for carbon taxes, for which a standard deviation increase leads to a reduction in carbon emissions of 1.1% in the short run and 3.9% in the long run.
When we included all three measures together, the coefficients on carbon taxes and the EPS index remained significantly negative, with a somewhat smaller magnitude (column V). While the coefficient on the prices of ETS permits remained negative, it lost significance. This is most likely due to the much smaller sample size, as we lost more than 80% of the sample when including the EPS index. Furthermore, the EPS index also reflects carbon prices, which introduces the problem of multicollinearity of the variables—which is particularly acute for small sample sizes.
5.2. Robustness
We performed several robustness checks. First, we controlled for the quality of governance in a country, proxied by a measure of control of corruption described in Section 3 (Table 2).Footnote 10 We found that higher carbon taxes and more stringent climate policy based on the broader EPS policy index significantly reduce carbon emissions when controlling for governance in all specifications (Table 2). By contrast, the coefficient on the prices of ETS permits remained significant only in one specification (column II). These results imply that the effects of carbon taxes and the broad climate policy index on carbon emissions are generally more robust than those for the effects of ETS permit prices.
Table 2. Effects of climate policies on CO2 emissions: with control of corruption

Note: Sample period: From 1971 to 2016, annual data. Cluster-robust standard errors reported in the second line are clustered at the country level. ***/**/* denote statistical significance at the 1%/5%/10% level. (I): nominal price of first carbon tax in USD/tCO2 equivalents; (II): nominal price of ETS in USD/tCO2 equivalents; (IV): EPS index.
Furthermore, we also estimated the effects of carbon policies when using real carbon prices instead of nominal ones; that is, we considered prices obtained by deflating the nominal USD carbon price by the US consumer price index (Table 3). The results for the significantly negative effects of higher carbon taxes, of higher prices of ETS permits and of more stringent climate policy based on the EPS index on CO2 emissions are robust to using these alternative measures.
Table 3. Effects of climate policies on CO2 emissions: using real carbon prices

Note: Sample period: From 1971 to 2016, annual data. Cluster-robust standard errors reported in the second line are clustered at the country level. ***/**/* denote statistical significance at the 1%/5%/10% level. (I): real price of first carbon tax in USD/tCO2 equivalents (deflated by US CPI); (II): real price of ETS in USD/tCO2 equivalents (deflated by US CPI); (IV): EPS index.
Finally, we excluded countries whose economic development, as measured by GDP per capita, is below that of the country with the lowest GDP per capita among those that have implemented carbon taxes or ETS. We did so by restricting GDP per capita to above $1500 in constant 2010 dollars in columns I to III of Appendix Table A1. Very similar results were obtained when this restriction was imposed.
Our main takeaway from the robustness analysis is that the results for the carbon price elasticities of emission reductions are generally robust to the different empirical models, which enhances confidence in these results. Moreover, the results of this paper for the ex post effects of carbon pricing on carbon emissions are quantitatively similar to those obtained by D’Arcangelo et al. (Reference D’Arcangelo, Pisu, Raj and van Dender2022), Metcalf and Stock (Reference Metcalf and Stock2023) and Rafaty et al. (Reference Rafaty, Dolphin and Pretis2020) (see Table 1 of Tol, Reference Tol2023), even though these papers have used different methods. The results of Rafaty et al. (Reference Rafaty, Dolphin and Pretis2020) are, for example, based on sectoral analysis. This consistency of our results with those of other papers with different methodologies also enhances confidence in our results.
6. Conclusion
We used a database of 121 countries to study how climate policies have affected CO2 emissions ex post. Our findings were based on comprehensive data on carbon emissions between 1971 and 2016, and we controlled for macroeconomic developments. As climate policies, we considered carbon taxes and ETS, as well as a broad index for the stringency of climate policies. Overall, we found that higher carbon taxes and prices of permits in the ETS are associated with significant reductions in carbon emissions. Furthermore, more stringent climate policies, as measured by a broader index for OECD and major emerging economies, also reduced carbon emissions.
Overall, an increase in carbon taxes by $10/tCO2 reduces CO2 emissions per capita by 1.3% in the short run and by 4.6% in the long run. This negative effect on emissions is statistically significant for all nine specifications that were used, with p-values that are always below 0.01. The same increase in the prices of ETS permits reduces CO2 emissions per capita by 1.4% in the short run and 5.0% in the long run, although this effect is found to be less robust to alternative specifications. More stringent climate policies as measured by a broad index for OECD and major emerging economies also significantly reduce carbon emissions, with an increase of one standard deviation in the index reducing CO2 emissions per capita by around 1.5% in the short run and 6.0% in the long run.
Our findings are relevant for the design of climate policies. The fact that higher carbon tax rates and prices of permits in the ETS rates have reduced carbon emissions suggests that further increases in these and expansion to more countries, thus covering a greater share of global carbon emissions, are promising avenues to speed up the necessary transition towards lower carbon emissions economies. Furthermore, the finding that more broadly, stringent climate policies have reduced carbon emissions indicates that future enhancements in a wider range of climate policies can also be helpful for a speedy transition towards much lower carbon emissions, ideally to net-zero emissions.
The results of this paper for the ex post effects of carbon pricing in reducing carbon emissions are lower than those assumed in most ex ante models in the IPCC’s AR6 Scenario Database for the effects of carbon taxes on carbon emissions (see Table 1 of Tol, Reference Tol2023). By contrast, Metcalf and Stock (Reference Metcalf and Stock2023), Rafaty et al. (Reference Rafaty, Dolphin and Pretis2020) and D’Arcangelo et al. (Reference D’Arcangelo, Pisu, Raj and van Dender2022) obtained quantitatively similar empirical results to ours. In order to meet the goals of the Paris Agreement, carbon taxes and carbon ETS need to be increased by more and more quickly than currently planned, and they should be complemented with other climate policies (Dubash et al., Reference Dubash, Mulugetta and Pauw2024). Other important climate policies include green technology support policies, such as subsidies and tax incentives for low-carbon research and development, and support for the adoption of solar and wind energy. For example, green technology policies have been implemented in the USA recently (White House, 2023). Recent ex post cross-country empirical analysis has shown that green technology policies have a significant effect on reducing carbon emissions per capita (Moessner, Reference Moessner2024).
Increasing carbon prices by more need not be held back by concerns about large overall inflationary effects, since recent ex post empirical work has shown that overall effects of carbon pricing on consumer price inflation have been small so far (Konradt and Weder di Mauro, Reference Konradt and di Mauro2023; Moessner, Reference Moessner2025). There is a potential for international leakages associated with higher carbon pricing, with carbon-intensive production being moved abroad (Schroeder and Stracca, Reference Schroeder and Stracca2023). Such carbon leakages can be mitigated with carbon border adjustment mechanisms. This mechanism has been phased in recently in the EU as “the EU’s tool to put a fair price on the carbon emitted during the production of carbon intensive goods that are entering the EU, and to encourage cleaner industrial production in non-EU countries” (Directorate-General for Taxation and Customs, 2025). The carbon border adjustment mechanisms may also lead other countries to increase their own carbon taxes, in order to decarbonise their economies more quickly and capture revenues from these carbon taxes themselves, instead of having their exporters make payments to foreign countries for the high carbon content of their products.
There are some concerns about adverse effects on growth from higher carbon prices. While some concerns are valid, it needs to be recognised that supporting the green transition with more stringent climate policy can create green jobs and thus support growth, enabling domestic firms to capture more of the increasing global demand for green products (Climate Change Committee, 2025). Other climate policies also allow the government to provide options to firms and households to switch more easily to renewable energy and increase energy efficiency. Some examples are government subsidies for district heating, public transport and housing insulation. These policies can help poorer households to switch more easily to renewable energies. Additionally, governments could mandate the installation of heat pumps for new homes or when an existing heating system needs to be replaced, and by providing support for poorer households for the one-off costs of this installation (Climate Change Committee, 2025). Furthermore, there can be co-benefits between development and better climate policies for developing economies (Dubash et al., Reference Dubash, Raghunandan, Sant and Sreenivas2013; Ürge-Vorsatz et al., Reference Ürge-Vorsatz, Herrero, Tirado, Dubash and Lecocq2014).
Acknowledgements
We would like to thank the participants of the seminar at University of Venice Ca’ Foscari conducted on 21 October 2024 and SSES annual congress participants at ETH Zürich for their helpful comments and discussions. The views expressed in this paper are those of the authors and do not necessarily reflect those of the Bank for International Settlements.
Competing interests
The authors declare none.
Appendix
Table A1. Effects of climate policies on CO2 emissions: Restricted sample

Note: Sample period: From 1971 to 2016, annual data. Cluster-robust standard errors reported in the second line are clustered at the country level. ***/**/* denote statistical significance at the 1%/5%/10% level. (I): nominal price of first carbon tax in USD/tCO2 equivalents; (II): nominal price of ETS in USD/tCO2 equivalents; (IV): EPS index; for (I) to (III): restricted to GDP per capita in constant 2010 dollars of above 1500. The countries included in the full unrestricted sample of this paper were chosen solely based on data availability. The following 121 countries were included in the full unrestricted sample, with their country codes listed below according to the standard International Organization for Standardization (ISO) three-letter country codes: AGO, ALB, ARE, ARG, AUS, AUT, AZE, BEL, BEN, BGD, BHR, BIH, BLR, BOL, BRA, BWA, CAN, CHE, CHL, CHN, CIV, CMR, COD, COG, COL, CRI, CUB, CZE, DEU, DNK, DOM, DZA, ECU, EGY, ERI, ESP, EST, ETH, FIN, FRA, GBR, GEO, GHA, GRC, GTM, HKG, HND, HRV, HTI, HUN, IDN, IND, IRL, IRN, IRQ, ISR, ITA, JAM, JOR, JPN, KAZ, KEN, KGZ, KHM, KOR, KWT, LBN, LKA, LTU, LVA, MAR, MDA, MEX, MKD, MMR, MNG, MOZ, MUS, MYS, NAM, NER, NGA, NIC, NLD, NOR, NPL, NZL, OMN, PAK, PAN, PER, PHL, POL, PRT, PRY, QAT, ROU, RUS, SAU, SDN, SGP, SLV, SVK, SVN, SWE, TGO, THA, TJK, TKM, TUN, TUR, TZA, UKR, URY, USA, VEN, VNM, YEM, ZAF, ZMB and ZWE.


