Policy Instruments Against Climate Change: A Panel Data Analysis of Carbon Taxation and Emissions Trading in OECD Countries

Abstract

Since the Industrial Revolution, the increase in greenhouse gas emissions has led to a significant rise in global temperatures compared to the pre-industrial period. This development has heightened the importance of carbon pricing policies in combating climate change. This study aims to examine the effects of carbon pricing instruments, carbon taxes and emissions trading systems (ETS) on carbon dioxide (CO₂) emissions in OECD countries. A panel data analysis covering the period 2002–2023 was conducted, taking into account structural differences across countries as well as shared economic dynamics. The findings indicate that both carbon taxes and ETS mechanisms are effective in reducing CO₂ emissions in the long run. Moreover, while increased industrial activity contributes to higher emissions, a greater share of renewable and nuclear sources in the energy mix is found to support emission reduction. The study demonstrates that carbon pricing policies exert limited short-term effects but generate structural and lasting impacts in the long term. The findings are consistent with the existing literature and theoretical framework. Achieving permanent reductions in emissions requires a comprehensive policy approach that not only implements carbon pricing, but also strengthens energy efficiency and fuel substitution in the industrial sector while continuously increasing the share of clean sources in the energy supply. The analysis shows that carbon taxes and emissions trading systems (ETS) are effective in reducing emissions over the long run in OECD countries, and that their success varies depending on countries’ energy profiles and policy designs. These results underline that a well-designed and complementary carbon pricing framework is critical for achieving a sustainable transition.

1. Introduction

Since the Industrial Revolution, the amount of greenhouse gases released into the atmosphere has increased, causing the global average temperature to rise by approximately 1.2 °C compared to pre-industrial levels (NASA, 2025). Over time, it has become evident that carbon dioxide and similar pollutants are major drivers of global warming (Stern, 2007). Not only have the effects of these emissions been recognized relatively late, but it has also taken considerable time for countries to acknowledge greenhouse gas emissions as a serious threat to humanity. According to the United Nations (2024), fossil fuels particularly coal, oil, and natural gas are responsible for 68% of global greenhouse gas emissions and approximately 90% of carbon dioxide emissions. This underscores the significant impact of fossil fuels on climate change. Between 2005 and 2020, global greenhouse gas emissions increased by 18.2%, accelerating an unprecedented warming process on the planet and marking one of the fastest recorded periods of temperature rise in history.

Paris on 12 December 2015, and signed the Paris Agreement, which aims to limit the global temperature increase to 1.5 °C (Türkoğlu Üstün, 2021). Although this consensus was reached relatively late, 194 countries agreed on a gradual action plan to curb global temperature rise. Major greenhouse gas emitters, including China, the United States, and India, also became parties to the Paris Agreement and committed to reducing their emissions below the prescribed levels by 2030. For instance, Turkey aims to reduce emissions by 21% below 2005 levels by 2030, while Canada initially targets a 30% reduction, followed by 40–45%, with the goal of achieving net-zero emissions by 2050 (Karakaya, 2016).

The European Union (EU) has emerged as one of the pioneering institutions to identify climate change as a primary threat and has taken significant steps accordingly. Leveraging public support, the EU announced the European Green Deal in 2019, a policy regarded as a long-term transformation strategy aiming to achieve net-zero anthropogenic greenhouse gas emissions by 2050 (Küçük & Dural, 2022). Despite these international initiatives and the implementation of various preventive measures, voluntary compliance regarding pollutants contributing to global warming has not reached the desired level (Nordhaus, 2019). Consequently, in addition to punitive measures, the deployment of financial instruments has become necessary. Within this framework, the carbon tax mechanism, which envisions the taxation of firms emitting greenhouse gases, has come to the forefront (Tietenberg, 2006; Metcalf, 2009). The purpose of a carbon tax is to levy charges on activities that pollute the atmosphere in proportion to the environmental costs they cause. The carbon tax is one of the primary economic instruments aimed at reducing greenhouse gas emissions. Under this taxation system, a specific cost is imposed on emissions resulting from the combustion of fossil fuels or other greenhouse gas–producing activities. In this way, the cost of polluting activities increases, thereby encouraging businesses and consumers to shift toward cleaner energy sources. At this point, the carbon tax functions not only as a financial regulation but also as a market mechanism evaluated within the broader framework of carbon pricing policies (Çelikkaya, 2023).

The literature on carbon pricing policies indicates that both taxation and emissions trading systems produce environmental as well as economic outcomes (Metcalf, 2009; Jeffrey & Perkins, 2015; Al-Abdulqader et al., 2025). However, existing studies primarily examine the singular or country-specific effects of these instruments, with limited attention to comparative cost-effectiveness and sectoral differentiation. Empirical evidence generally suggests that carbon pricing policies have a significant and statistically negative impact on emissions (L. Cui et al., 2021; Jeffrey & Perkins, 2015; Al-Abdulqader et al., 2025). Nevertheless, the effectiveness of these policies varies according to energy intensity, income level, and sectoral structure, with stronger effects particularly observed in economies with a high industrial share (Guo & Xiang, 2022; T. Sunanda et al., 2025; Karapınar et al., 2019b). However, comprehensive empirical panel data studies that examine the simultaneous effects of carbon taxes and emissions trading systems in OECD countries together with key determinants such as energy use patterns and economic scale remain limited. In the existing literature, the macro-level outcomes of jointly implementing these two policy instruments are predominantly framed through modeling or simulation-based analyses (Karapınar et al., 2019b; Çelikkaya, 2023). Therefore, the present study analyzes the impact of carbon pricing instruments on emissions in a sample of OECD countries using panel data methods; the macro-level effects of the carbon tax and emissions trading system variables are tested alongside economic and sectoral control variables. This approach aims to empirically measure the effectiveness of these policy tools and to identify how their impacts vary depending on income levels and structural characteristics. In doing so, the study seeks to contribute to the carbon pricing literature by validating theoretical findings at the country-group level and informing the determination of appropriate policy mixes.

2. Literature Review

Carbon pricing is regarded as one of the most widely used market-based policy instruments for combating climate change. Among these instruments, the carbon tax and the emissions trading system (ETS) aim to reduce greenhouse gas emissions while maintaining economic efficiency. Existing studies generally highlight the positive impact of carbon pricing on emission reductions, yet they indicate that the magnitude of these effects is sensitive to country-specific conditions, sectoral structure, and policy design (Baranzini et al., 2000; Schmalensee & Stavins, 2017; I. W. H. Parry et al., 2022). In an analysis of China’s rate-based emissions trading system, Goulder et al. (2022) note that the system provides flexibility to balance differences in income and sectors, although this flexibility reduces overall cost-effectiveness. Similarly, Cao et al. (2019) show that the hybrid system covering the electricity and cement sectors in China achieves emission reductions at lower cost losses compared to individual instruments.

Research conducted within the context of the European Union suggests that carbon taxes and ETS implementations are largely complementary rather than substitutive. Jeffrey and Perkins (2015) examined EU panel data covering the period 1996–2009 to investigate the relationship between energy taxes, ETS participation, and carbon intensity, finding that the intensity-reducing effect of energy taxes persisted post-ETS through efficiency channels. Similarly, Charitou (2015) emphasized the robustness of the tax-efficiency relationship and highlighted the determining role of investment incentives. An effectiveness indicator developed by Di Foggia and Beccarello (2024) demonstrated that, in the EU manufacturing sector, the ETS provides comparability in terms of policy fine-tuning. Candelon and Hasse (2023), using frequency-domain causality tests in the case of Sweden, showed that over the long term, only the carbon tax has a causal impact on emission reductions. In contrast, Wittneben (2009) argued that, at the international level, carbon taxes possess the potential to achieve faster and lower-cost reductions compared to ETS.

In Asia, China and Indonesia have emerged as particularly prominent cases. Guo and Xiang (2022), using a market simulation analysis, found that an increase in the carbon tax rate raised the market-clearing price and enhanced the profitability of non-coal producers. H. Zhang and Zhang (2021) compared carbon tax, ETS, and hybrid policy scenarios in the electricity market, reporting that all three systems reduced emissions, although the hybrid system yielded more balanced outcomes in terms of welfare gains. W. Sunanda et al. (2025), employing a TIMES model for the Indonesian electricity sector, demonstrated that high tax levels achieved deep emission reductions but increased the investment burden, whereas a hybrid approach maintained cost-effectiveness while ensuring implementability. Furthermore, Duan et al. (2024), using a supply chain network equilibrium model, showed that high carbon prices reduced emissions, with reductions being more pronounced when the carbon tax level exceeded the subsidy threshold.

Studies in Latin America and South Africa provide further insights into the regional applicability and effectiveness of carbon pricing instruments. In Chile, Mardones (2021) analyzed the industrial sector and found that, under a 30% emission reduction target, the ETS significantly reduced total costs compared to a carbon tax, although the combined implementation of both instruments generated additional costs. Similarly, in Mexico’s electricity sector, Barragán-Beaud et al. (2018) report that the ETS is more advantageous in terms of cost-effectiveness, while emphasizing that the cap level constitutes a critical threshold for policy acceptability. Extending the discussion to South Africa, Abiodun et al. (2024) suggest that blockchain-based monitoring and automation in carbon tax collection can enhance transparency and improve policy effectiveness. Collectively, these studies highlight the importance of context-specific design features, sectoral considerations, and technological enablers in determining the success of carbon pricing mechanisms across diverse regions.

Multi-country analyses across the OECD indicate that the combined implementation of carbon taxes and ETS exerts a significant mitigating effect on emissions. Al-Abdulqader et al. (2025), using a global sample of thirty countries, found that carbon pricing reduced emissions by approximately 10%, with the effect of ETS being slightly stronger than that of carbon taxes. In Portugal, Pereira and Pereira (2019) demonstrated that a carbon tax indexed to the ETS price, when paired with a well-designed revenue recycling mechanism, can simultaneously support economic growth and deliver environmental gains. Van der Ploeg et al. (2025), employing a microsimulation model incorporating household heterogeneity, showed that as sensitivity to inequality increases, carbon taxes align more closely with a progressive tax structure while preserving environmental benefits. Such micro-level analyses are crucial for understanding the underlying drivers of macro-level emission trends. Accordingly, a comprehensive assessment of carbon pricing policies requires examining the temporal dynamics of emissions across countries to capture their broader effects (see

Table 1. Total CO₂ Emissions by Year in Leading Countries (1990–2023, Million Metric Tons).

Country	1990	1995	2000	2005	2010	2015	2020	2021	2022	2023
United States of America	4983.8	5271.7	5929.0	5888.3	5521.8	5133.2	4466.0	4755.2	4786.6	4682.0
Russia	2436.3	1765.2	1681.1	1739.3	1746.8	1757.1	1828.8	1957.9	2025.1	2069.5
China	2414.3	3365.8	3667.0	6264.7	9134.1	10,776.9	12,022.4	12,621.6	12,526.8	13,259.6
Japan	1166.8	1238.6	1248.8	1283.0	1220.5	1235.4	1072.2	1084.9	1010.0	944.8
Germany	1013.0	910.8	871.7	842.1	815.4	785.4	642.6	677.8	659.5	583.0
Ukraine	785.5	462.6	360.0	356.1	314.3	226.0	198.7	194.1	138.3	136.2
India	600.7	796.5	995.7	1216.5	1743.7	2260.1	2318.9	2548.5	2740.8	2955.2
United Kingdom	582.3	548.1	551.7	559.3	500.6	415.9	319.0	339.5	327.5	302.1
Canada	440.5	475.9	543.0	581.1	561.6	577.1	546.7	561.6	575.3	575.0
Italy	426.4	435.3	454.7	493.5	424.0	354.0	295.9	334.3	332.8	305.5
France and Monaco	385.3	382.1	401.2	408.2	377.8	333.0	287.3	319.4	310.5	282.4
Poland	371.5	356.4	313.0	317.1	329.4	305.7	296.9	323.8	317.3	286.9
South Africa	314.2	325.2	347.3	436.3	464.3	460.1	418.8	428.5	409.4	397.4
Mexico	289.7	326.2	396.7	447.3	476.9	490.3	427.8	452.9	465.9	487.1
Australia	277.7	302.7	353.9	384.1	415.3	400.5	392.5	384.7	374.9	373.6
South Korea	272.1	404.7	474.2	515.4	598.1	642.7	614.2	631.1	587.8	573.5
Kazakhstan	248.6	180.3	132.2	187.6	249.9	205.4	224.7	244.9	240.2	239.9
Spain and Andorra	231.8	256.4	313.2	370.4	291.2	275.1	218.7	237.6	235.5	217.3
Brazil	226.9	280.0	349.4	369.2	439.8	517.6	447.4	506.0	478.9	479.5
Iran	208.5	283.4	353.9	469.0	576.6	629.3	712.4	738.6	760.6	778.8
Romania	187.7	133.4	97.4	104.5	83.3	80.3	75.7	80.4	76.0	70.8
Saudi Arabia	173.9	217.2	265.2	346.4	486.9	611.4	570.9	584.5	605.3	622.9
Czech Republic	165.5	133.6	132.3	129.0	121.0	108.3	94.4	99.1	99.4	90.5
Netherlands	164.9	177.7	176.9	182.2	186.2	169.4	141.9	146.3	133.0	122.9
Indonesia	161.8	239.1	299.1	362.6	443.7	512.4	594.3	619.3	656.7	674.5
Turkey	155.2	178.4	227.0	246.1	312.1	367.8	422.8	460.7	433.7	438.3
North Korea	132.0	89.1	73.8	79.7	69.0	32.7	69.1	57.3	59.5	64.3
Taiwan	130.7	176.7	238.0	280.6	284.3	296.6	289.5	299.8	287.6	279.8

Source: EDGAR-Emissions Database for Global Atmospheric Research.

).

Globally, carbon dioxide (CO₂) emissions have exhibited a sharp increase over the past thirty years, particularly in economies experiencing rapid industrialization. In 1990, the United States, Russia, and China accounted for the majority of global emissions. By 2023, China’s emissions had reached 13.2 billion tons, approximately three times those of the United States and four and a half times those of India. In contrast, developed economies such as European countries, Japan, and Canada have achieved sustained emission reductions through carbon pricing, the transition to renewable energy, and energy efficiency policies. Overall, the period 1990–2023 demonstrates that while policy interventions have driven a declining trend in developed countries, economic growth and energy dependence have been the primary drivers of emission increases in developing countries. These patterns underscore the necessity of supporting carbon pricing policies with long-term structural transformations to ensure their effectiveness.

This study aims to examine the effects of carbon pricing instruments specifically carbon taxes and emissions trading systems on CO₂ emissions in OECD countries. In the analysis, the annual total CO₂ emission level is employed as the dependent variable, while the carbon tax and the emissions trading system constitute the main independent variables. The carbon tax is measured by the ratio of fossil fuel–based tax revenues to GDP, whereas the emissions trading system is represented by a dummy variable indicating the year in which it was introduced. To account for economic and structural differences across countries, the model additionally incorporates variables such as real GDP per capita, agricultural and industrial value added, and the share of alternative and nuclear energy. All data are compiled on a country-year basis and obtained from the World Bank and OECD databases. The constructed model seeks to analyze the macro-level effects of carbon pricing instruments on emissions and to reveal how these effects vary in accordance with countries’ economic structures. Since the 1970s, the development of environmental taxation in OECD countries has gradually become institutionalized. The process, initiated with the adoption of the “polluter pays” principle, advanced further after the 1990s with the establishment of comprehensive environmental tax mechanisms such as carbon taxes and emissions trading systems. In these countries, environmental taxes have become an important policy instrument that not only improves environmental quality but also supports economic efficiency. The literature emphasizes that tax instruments and emissions trading systems should be evaluated not only in terms of their environmental impacts but also within their multidimensional context, including revenue allocation, competitiveness, and international compatibility (Çelikkaya, 2023, 2024; T. He & Guo, 2023). The use of carbon pricing revenues is also a critical component of policy design. Birinci (2020) notes that these revenues may be directed toward various categories such as tax reductions, the general budget, climate investments, direct transfers, and development financing and that transparency in revenue allocation and adherence to environmental objectives enhance policy acceptance.

Studies assessing the comparative effectiveness of carbon taxes and emissions trading systems increasingly underscore that the determining factor is not the inherent superiority of one instrument over the other, but rather the quality of their design frameworks. Çelikkaya (2023) asserts that carbon pricing instruments must be evaluated not only in terms of their categorical distinctions but also with respect to their operational modalities, emphasizing that design quality constitutes a central determinant of environmental effectiveness. Given that the emissions trading system (ETS) operates as a market-based mechanism, concerns related to non-compliance risks and the robustness of verification procedures are frequently highlighted in the literature. Ubay and Bilgici (2021) contend that although the ETS represents a strong market instrument grounded in the polluter-pays principle, deficiencies in monitoring, reporting, and verification processes may distort price signals and diminish environmental efficacy. Within this framework, risks such as underreporting of emissions by firms, unregistered carbon credit transactions, and weaknesses in regulatory oversight emerge as critical considerations in the design and implementation of ETS policies.

The literature on the economic impacts of carbon policies has also become increasingly diverse, highlighting the need to account for factors such as competitiveness, inter-sectoral transition costs, and economic stability in policy design. Recent studies underscore that carbon pricing instruments cannot be evaluated in isolation; rather, they must be assessed in conjunction with the broader economic structure, administrative capacity, and prevailing market dynamics (Karakaya et al., 2023). Assessments of environmental tax systems further indicate that existing tax categories provide only indirect contributions to environmental objectives and often fall short of meeting the definitional and functional criteria of a true carbon tax (Yalçın, 2013). Practices such as exemptions granted to energy-intensive industries or the use of carbon revenues primarily for budget compensation are identified as key factors limiting the effectiveness of carbon energy taxes (Uğur, 2014). Additionally, Oğuz and Yıldız (2024) note that, although environmental taxes can serve as an effective instrument for emission reduction, implementation has progressed slowly in many countries due to legislative and design-related shortcomings.

When the broader literature is examined, it becomes evident that holistic frameworks integrating environmental, fiscal, and competitiveness-related dimensions have gained prominence in the field of carbon pricing, and that comprehensive, institutionalized policy designs play a critical role in achieving emission reduction targets. The prevailing trend suggests that carbon taxes and emissions trading systems yield effective outcomes when implemented in a manner that aligns with design quality and the structural characteristics of the economy.

A carbon tax is a market-based fiscal policy instrument designed to internalize the environmental externalities associated with carbon dioxide (CO₂) emissions resulting from the use of fossil fuels. Developed within the framework of the “polluter pays” principle, this type of tax is applied at varying rates according to the carbon content of fossil fuels such as coal, oil, and natural gas. Coal possesses the highest carbon intensity, containing approximately 22.5 kg of carbon per unit of energy, followed by oil with 19.6 kg and natural gas with 14.2 kg of carbon. These differences lead to heterogeneous cost impacts across energy sources when a carbon tax is implemented. The carbon tax is a quantity-based specific tax that aims both to promote environmental sustainability and to generate fiscal revenue. One of the key advantages of this system is its so-called “double dividend” effect in the literature, whereby it simultaneously contributes to emission reductions and allows the revenues generated to be directed toward low-income households or environmental investments (Baranzini et al., 2000). OECD countries, particularly those in Europe, have long implemented carbon taxes and emissions trading systems (ETS) as institutionalized and effective climate policy instruments. The integration of market-based mechanisms such as the EU Emissions Trading System together with carbon taxes applied to energy consumption and sector-specific emissions has significantly enhanced the scope and maturity of carbon pricing across the OECD. In these countries, carbon pricing has become a central component of climate policy, exhibiting an advanced institutional structure in terms of coverage, transparency, and market integration.

In OECD countries particularly in Europe carbon taxes and emissions trading systems (ETS) have long been institutionalized and implemented as effective climate policy instruments. The integration of carbon taxes targeting energy consumption and sector-based emissions with market-based mechanisms such as the EU Emissions Trading System has significantly expanded the scope and maturity of carbon pricing across the OECD. In these countries, carbon pricing has become a fundamental component of climate policy, demonstrating an advanced institutional structure in terms of coverage, transparency, and market integration. OECD data show that, over the long term, countries implementing carbon pricing exhibit stronger emission-reduction performance when evaluated alongside their energy supply structure, share of renewable energy, and sectoral transformation. In this context, it is evident that carbon pricing policies in OECD countries have reached an advanced level in terms of maturity, scope, and market integration, and that institutional processes aimed at enhancing the effectiveness of climate policies have largely been completed (Çelebi-Boz & Örs-Onur, 2024).

Many countries are seeking to accelerate the energy transition through market-based instruments such as coal taxes and carbon pricing. These mechanisms encourage low-emission production; however, because they can increase energy costs, they must be designed in a socially equitable manner. For this reason, support systems that protect low-income groups and emission-based differentiated obligations are of particular importance. Such an arrangement makes it possible to achieve both emission reductions and the preservation of energy supply security (Baranzini et al., 2000). While energy demand in advanced economies typically increases by less than 1% annually, developing and industrializing countries experience an increase of around 2–4% (IEA, 2024).

Carbon taxes and emissions trading systems have long been institutionalized and implemented as effective climate policy instruments in OECD countries, particularly across Europe. The expansion of carbon taxes targeting energy consumption and sector-specific emissions, together with the integration of advanced market-based mechanisms such as the EU Emissions Trading System, has substantially increased the scope and maturity of carbon pricing policies. While carbon taxes generate direct price signals, emissions trading systems provide market-based flexibility for compliance, thereby reinforcing the central role of these instruments within climate policy frameworks across the OECD. The literature indicates that OECD countries employing carbon pricing exhibit more stable long-term emission reduction trajectories, and that energy efficiency improvements, low-carbon technologies, and sectoral transformation processes are supported through these policy tools. Nevertheless, differences in institutional capacity, the quality of market regulations, the effectiveness of monitoring–reporting–verification processes, and the sectoral coverage of schemes emerge as key determinants shaping the degree of integration and maturity of carbon pricing policies across countries. Overall, the findings demonstrate that carbon taxes and emissions trading systems deliver high mitigation effectiveness when implemented within a robust institutional framework and with a comprehensive policy design (Çelebi-Boz & Örs-Onur, 2024). At the global level, carbon taxes are implemented in four primary forms: direct carbon taxation, capacity or performance-based taxation, border carbon adjustments, and emissions trading systems (Ubay & Bilgici, 2021). Direct carbon taxes are levied at a fixed rate per ton of CO₂ equivalent emitted. Capacity-based models impose taxes according to firms’ production levels or carbon efficiency. Border carbon adjustments aim to prevent carbon leakage by accounting for the carbon footprint embedded in imported goods. This system, which the European Union has begun to implement gradually since 2023, seeks to ensure environmental competitiveness by increasing the import costs of high-carbon-intensive products (European Commission, 2023).

The second principal instrument of carbon pricing policies is emissions trading systems (ETS). Based on a “cap-and-trade” logic, ETS sets an upper limit on total emissions and allocates emission allowances to firms, which can then be bought and sold in the market, thereby creating an economic value for emission reductions. Compared to carbon taxes, ETS provides greater certainty in terms of emission quantities, while carbon taxes offer price stability and administrative simplicity (Schmalensee & Stavins, 2017). The effectiveness of both systems depends on their design principles: setting carbon prices in alignment with emission reduction targets, directing revenues toward environmental investments, and ensuring social equity are all critical considerations (Çelikkaya, 2023; I. W. H. Parry et al., 2022). Carbon pricing represents an environmental policy instrument based on the principle of assigning a direct cost to greenhouse gas emissions (PMR, 2019). The direct pass-through of carbon pricing to energy costs can hinder its social acceptability; consequently, alternative instruments that are relatively less effective but enjoy broader public support are often preferred. Therefore, a balanced approach is required between effective yet socially sensitive carbon pricing and other emission reduction policies that, although less impactful, have higher feasibility. For a successful climate policy, it is crucial that carbon pricing is integrated with complementary environmental measures to maximize both effectiveness and public acceptance (Black et al., 2022; Borenstein & Kellogg, 2022). Carbon pricing continues to serve as a significant source of public revenue for countries, which can be directed toward general budgetary objectives. Despite global challenges such as high inflation, fiscal pressures, and energy crises, revenues from carbon pricing reached a record level of approximately USD 95 billion in 2023 (WBG, 2023). This development is particularly noteworthy for countries with large informal economies and limited tax collection capacity. Remarkably, even during periods of significant economic hardship, high-income countries have prioritized carbon pricing policies to support emission reductions. Although the share of global greenhouse gas emissions covered by carbon pricing remains relatively low, it has increased substantially over the past decade, rising from 7% to 23% (Timilsina et al., 2021). The high share of the population exposed to local air pollution in OECD countries clearly reveals the significant impact of air pollution on public health. International estimates indicate that deteriorating air quality poses substantial health risks, particularly in the form of premature deaths and chronic diseases (I. Parry et al., 2023). In this context, the implementation of an effective carbon pricing policy has the potential to generate positive health outcomes by contributing to improvements in air quality-especially for populations living in densely urbanized areas and for groups more vulnerable to changes in environmental conditions.

One of the most significant challenges of carbon pricing is the sudden and rapid increase in energy prices. Over the past two years, global prices for natural gas, coal, and oil have risen by approximately 700%, 180%, and 110%, respectively (I. Parry et al., 2023). However, such pricing, which will be reflected in energy costs, is likely to encounter strong resistance, particularly among households with incomes below 60% of the national median disposable income, that is, those below the relative poverty line, commonly referred to as “low-income” or “income-at-risk” households (Odekon, 2021). Examples of such social opposition have been observed multiple times in the past: in Switzerland, proposed energy taxes in 2000 and 2015 were rejected by the public; in the U.S. state of Washington, carbon tax proposals in 2016 and 2018 were similarly defeated; and most recently, the “Yellow Vest” protests in France in 2018 led to the suspension of planned fuel tax increases (Black et al., 2022; Çelikkaya, 2024).

Increasing the social acceptability of carbon pricing requires the implementation of complementary policies, such as targeted support for low-income households, workers, and regions, revenue redistribution, and sectoral tax-subsidy programs (Timilsina et al., 2021, pp. 4–5). In this context, carbon taxes emerge as a particularly suitable instrument. For instance, in Scandinavian countries, carbon taxation has contributed to alleviating the tax burden on workers (Borenstein & Kellogg, 2022). Similarly, emissions trading systems (ETS) can be used to reduce the tax burden on employment; however, in practice, free allocation of allowances is common. In ETS programs with auctioned allowances, the resulting revenues are generally directed toward climate action rather than the general budget. Climate action in this context includes measures such as emission reductions, support for renewable energy technologies, afforestation and carbon capture and storage projects, transitions to low-emission transport, research and development activities, regional heating initiatives, and social assistance programs (Çelikkaya, 2024). Carbon pricing is widely recognized as one of the most effective economic instruments in combating climate change. Within this framework, carbon taxes and ETS constitute the two primary policy tools. Carbon taxes provide price stability and administrative simplicity, whereas ETS ensure certainty in emission quantities and often receive greater political support through free allocation of allowances (Çelikkaya, 2023; I. W. H. Parry et al., 2022).

In conclusion, the success of carbon pricing policies depends less on the choice of a specific instrument and more on the balance that the implemented mechanism achieves between economic efficiency, environmental effectiveness, and social equity. Carbon taxes and ETS are not mutually exclusive but complementary tools; an effective climate policy requires the coordinated design and implementation of both mechanisms.

3. Variables and Hypotheses

The impact of instruments developed within the framework of carbon pricing on emissions has been examined through multi-layered approaches across different economies and sectors. In the literature, macro-level panel data and quasi-experimental designs, policy simulations (including computable general equilibrium and energy system models), network- and game-theory-based microdecision models, and difference-in-differences analyses on firm-level data have been employed in combination. This methodological diversity indicates that, beyond the individual effects of carbon taxes and emissions trading systems, design features such as hybrid schemes, revenue recycling, and allocation methods also play a critical role in determining policy outcomes.

Energy system models are among the most advanced analytical tools used to evaluate the impacts of carbon pricing policies, offering a comprehensive framework for examining energy supply demand balances, sectoral transition costs, and technological change. These models classified as partial equilibrium, general equilibrium, optimization-based, and simulation-based integrate a wide range of technical inputs such as technology options, fuel prices, demand elasticities, emission coefficients, and investment and operating costs to estimate long-term policy effects. Optimization models (e.g., MARKAL/TIMES) identify the least-cost technology mix under a given carbon price, whereas general equilibrium models internalize the interactions between energy markets and the broader macroeconomic structure. More recently developed multi-temporal and scenario-based models demonstrate that mechanisms such as carbon taxes and emissions trading systems generate limited short-term but strong and persistent long-term emission reductions. Indeed, energy system model applications by Zhu et al. (2024) and Zhu and Gao (2023) show that carbon pricing accelerates the transformation of the energy mix and increases the share of renewables. Accordingly, energy system models are regarded as critical analytical tools for incorporating technological change, energy-efficiency investments, and sectoral transition costs into policy design.

The empirical literature indicates that increases in carbon taxation have a statistically significant and negative effect on emissions. A causality analysis focused on Sweden has shown that carbon taxes serve as a causal determinant of long-term emission reductions (Candelon & Hasse, 2023). Similarly, studies conducted across European countries demonstrate that increases in implicit energy tax rates reduce carbon intensity and enhance energy efficiency (Jeffrey & Perkins, 2015; Charlier et al., 2023). Computable general equilibrium and simulation studies in China indicate that higher tax rates lead to significant emission reductions, although they may impose short-term pressures on economic growth (Gao & Zhang, 2023). The allocation of revenues generated from carbon taxes to environmental investments is considered a fundamental factor in enhancing policy effectiveness. In this context, channeling revenues into renewable energy infrastructure (solar, wind, biomass), energy efficiency measures, clean and low-carbon production technologies, carbon capture and storage (CCUS) systems, and R&D activities accelerates the technological and structural transformation of the emission reduction process. Studies indicate that the gradual and flexible implementation of carbon taxes strengthens mitigation performance and particularly encourages investments in energy efficiency and low-carbon technologies (Zhou et al., 2023; Pereira & Pereira, 2019). Such environmental investments not only reduce carbon intensity but also support long-term productivity gains, thereby contributing to economic growth. Accordingly, the findings (Bovenberg & de Mooij, 1997) suggest that the carbon tax–investment nexus constitutes a strategic policy mechanism for sustainable development (Zhou et al., 2023; Pereira & Pereira, 2019). These findings collectively support the emission-reducing impact of carbon taxation. Accordingly, the following hypothesis is proposed:

H1: 

An increase in the carbon tax is expected to reduce CO₂ emissions.

Extensive evidence indicates that emission levels have decreased significantly in countries implementing emissions trading systems (ETS). Difference-in-differences analyses of China’s regional ETS pilots reveal reductions of 16.7% in total emissions and 9.7% in emission intensity (J. Cui et al., 2021). Policy analyses of the European Union sample similarly show that ETS, compared to carbon taxes, produces broader sectoral effects, contributing to declines in manufacturing sector emissions (Di Foggia & Beccarello, 2024). Likewise, modeling studies of the energy markets in Mexico and Chile indicate that emissions trading achieves more cost-effective outcomes than carbon taxes under equal reduction targets (Mardones, 2021; Barragán-Beaud et al., 2018). Panel analyses conducted by Al-Abdulqader et al. (2025) on a global sample demonstrate that ETS results in an average reduction of approximately 12% in emissions, with effects more pronounced than those of carbon taxation. These findings collectively suggest that countries implementing ETS experience systematically lower emissions. Accordingly, the second hypothesis of this study is as follows:

H2: 

An increase in the scope of the emissions trading system is expected to reduce CO₂ emissions.

The effects of carbon pricing policies are observed to be limited in the short term but exhibit a stable and increasing mitigation trend over the long term. A causality analysis conducted in Sweden found no significant relationship between carbon prices and emissions in the short run, whereas a strong negative relationship emerged in the long run (Candelon & Hasse, 2023). Similarly, multi-country panel analyses covering the period 2002–2023 indicate that the impact of carbon pricing strengthens over time, becoming statistically significant particularly in periods of five years or longer (Tello, 2025; Al-Abdulqader et al., 2025). Simulations using energy system models corroborate these findings, showing that phased and multi-temporal carbon trading systems generate stable and persistent emission reduction effects (Zhu et al., 2024; Zhu & Gao, 2023). Based on these findings, the observation that policy effects are limited in the short term but significantly strengthened in the long term has led to the formulation of the following hypothesis:

H3: 

The effect of the policy is expected to be limited in the short term but to become more pronounced in the long term.

Table 2. Variables.

Variable Type	Abbreviation	Description	Unit	Data Source
Dependent Variable	CO2	CO2 Emissions	Mt (Million tons)	World Bank (WB-World Development Indicators Database)
Independent Variables	CTX *	Carbon Tax	%	Organization for Economic Co-operation and Development (OECD- Taxing Energy Use)
	ETS *	Emissions Trading System	-	World Bank Carbon Pricing Dashboard
Control Variables	GDP	Gross Domestic Product	United States Dollar	World Bank (WB-World Development Indicators Database)
	AGR	Agricultural Value Added	United States Dollar	World Bank (WB-World Development Indicators Database)
	IND	Industrial Value Added	United States Dollar	World Bank (WB-World Development Indicators Database)
	ENG	Use of Alternative or Nuclear Energy	%	World Bank (WB-World Development Indicators Database)

* The variables marked in the table were constructed by the author.

summarizes the variables and data sources used in the study. The dependent variable is the annual total carbon dioxide (CO₂) emissions of countries. The independent variables represent carbon pricing instruments and include the carbon tax (CTX) and the emissions trading system (ETS). The carbon tax is defined as the ratio of tax revenues derived from fossil fuels to GDP, while ETS is represented by a dummy variable indicating the year of system implementation. To control for economic and structural differences, real GDP per capita (GDP), agricultural value added (AGR), industrial value added (IND), and the share of alternative and nuclear energy (ENG) are included as control variables. All variables are compiled at the country-year level, with data sourced from the World Bank and OECD databases. The model is designed to test the macro-level effects of carbon pricing instruments on emissions and to capture variations arising from differences in energy structure and income levels.

Methodology and Method

This study examines the effects of carbon taxes and emissions trading systems (ETS) on CO₂ emissions at the OECD country level using panel data methods. The analysis evaluates how these policy instruments influence emission levels through channels such as economic structure and energy intensity. The hypotheses of the study are formulated as follows:

H1: 

An increase in the carbon tax is expected to reduce CO₂ emissions.

H2: 

An increase in the scope of the emissions trading system is expected to reduce CO₂ emissions.

H3: 

The effect of the policy is expected to be limited in the short term but to become more pronounced in the long term.

The study employs an annual panel dataset for OECD countries. The sample covers the period 2002–2023, depending on data availability, and exhibits an unbalanced panel structure at the country level. All variables are matched at the country-year level, and series requiring monetary values are expressed in constant 2015 U.S. dollars. South Korea was excluded from the sample due to missing data. All other countries are included in the analysis and are listed as follows: Australia, Austria, Belgium, Canada, Chile, Colombia, Costa Rica, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Latvia, Lithuania, Luxembourg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey, the United Kingdom, and the United States.

Due to the lack of directly available carbon tax data for most countries, a proxy variable for the independent variable, carbon tax, was constructed based on insights from the literature. Accordingly, the formulation of the carbon tax variable is as follows:

$$\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n} \mathrm{T}\mathrm{a}\mathrm{x} = {\displaystyle \frac{\mathrm{F}\mathrm{o}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{l} \mathrm{E}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{T}\mathrm{a}\mathrm{x} \mathrm{R}\mathrm{e}\mathrm{v}\mathrm{e}\mathrm{n}\mathrm{u}\mathrm{e}\mathrm{s}}{\mathrm{G}\mathrm{r}\mathrm{o}\mathrm{s}\mathrm{s} \mathrm{D}\mathrm{o}\mathrm{m}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{c} \mathrm{P}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}}} \times 100$$

(1)

The carbon tax is expressed as the percentage ratio of tax revenues derived from fossil fuels-including coal, petroleum products, LPG, and natural gas to Gross Domestic Product (GDP). In this way, for countries lacking direct data on carbon taxes per ton, the measure captures the effective tax burden on carbon-based energy in the economy and, consequently, the intensity of actual carbon pricing. The other independent variable, the Emissions Trading System (ETS), is constructed as a dummy variable. For each country, the starting year of ETS implementation was identified from the carbon pricing dashboard. Accordingly, the variable is coded as 0 for years prior to the introduction of ETS and 1 for the years following its implementation.

Based on the variables, the model developed for this study is specified as follows:

$$\ln \left({{CO}_2}_{it}\right)= {\alpha }_i+ {\beta }_1\left({CTX}_{it}\right)+{\beta }_2\left({ETS}_{it}\right)+ {\gamma }_{1}ln\left({GDP}_{it}\right)+ {\gamma }_{2}ln\left({AGR}_{it}\right)+{\gamma }_{3}ln\left({IND}_{it}\right)+{\gamma }_4\left({ENG}_{it}\right)+{\mu }_{it}$$

(2)

Definitions (i = country, t = year):

• 〖CO₂〗_it: Total CO₂ emissions (Mt). (Log)
• 〖CTX〗_it: Carbon tax (or carbon-based energy tax/GDP, %).
• 〖ETS〗_it: Emissions Trading System dummy (1 if implemented, 0 otherwise).
• 〖GDP〗_it: Real GDP per capita (USD, log).
• 〖AGR〗_it: Agricultural value added (USD, log).
• 〖IND〗_it: Industrial value added (USD, log).
• 〖ENG〗_it: Share of alternative and nuclear energy (as a percentage of total energy).
• α: Constant term.

4. Findings and Discussion

The descriptive statistics of the series, which allow us to observe the scale and dispersion of the series, identify outliers or data errors, assess the closeness of the distribution to normality, and consider the potential need for transformation, are presented in

Table 3. Descriptive Statistics of the Variables for OECD Countries (2002–2023).

Statistics	LN_CO2	CTX	ETS	LN_GDP	LN_AGR	LN_IND	ENG
Mean	4.474	2.227	0.633	10.239	22.59	25.121	17.654
Median	4.209	2.286	1.0	10.463	22.396	25.14	12.19
Maximum	8.681	5.078	1.0	11.629	26.124	29.252	89.72
Minimum	1.108	0.135	0.0	8.309	18.513	21.711	0.01
Standard Deviation	1.571	0.843	0.482	0.717	1.495	1.591	17.545
Skewness	0.244	0.232	−0.55	−0.356	−0.22	−0.034	1.781
Kurtosis	2.861	3.451	1.303	2.283	2.914	2.645	6.854
Jarque–Bera	8.754	14.201	138.78	34.632	6.806	4.425	934.397
Probability	0.013	0.001	0.0	0.0	0.033	0.109	0.0
Sum	3641.611	1812.487	515.0	8334.776	18,387.87	20,448.16	14,370.4
Sum of Squares	2005.978	577.647	189.17	418.027	1817.919	2057.242	250,257.5
Number of Observations	814.0	814.0	814.0	814.0	814.0	814.0	814.0

.

According to the results presented in Table 3, the skewness values indicate that the variables of CO₂ emission absorption and industrial value added are nearly symmetric, while GDP per capita and agricultural value added exhibit a slight left skew. Alternative energy refers to energy sources that are used in place of limited and carbon-intensive fuels, such as fossil fuels, and have lower environmental impacts. In the panel data, the alternative energy variable exhibits the highest volatility, indicating substantial differences in alternative energy usage among countries (Table 3). In this context, renewable energy sources such as solar, wind, hydroelectric, and biomass are included within the definition of alternative energy. In some studies, nuclear energy is also considered as part of alternative energy (Panwar et al., 2011). The alternative energy variable, on the other hand, is distinctly right-skewed. Moreover, alternative energy displays the highest volatility, suggesting substantial differences among countries in terms of alternative energy use. The carbon tax variable has, on average, a low rate and limited variability. The results of the Jarque–Bera test reveal that most variables deviate from the normal distribution. To determine the direction and strength of the relationships among the variables, the correlation matrix is presented in

Table 4. Correlation Matrix.

	LN_CO2	CTX	ETS	LN_GDP	LN_AGR	LN_IND	ENG
LN_CO2	1.0
CTX	−0.311	1.0
ETS	−0.273	0.278	1.0
LN_GDP	0.079	0.027	0.237	1.0
LN_AGR	0.895	−0.331	−0.326	−0.035	1.0
LN_IND	0.936	−0.346	−0.212	0.259	0.898	1.0
ENG	−0.34	−0.05	0.142	0.223	−0.132	−0.184	1.0

.

According to the correlation matrix presented in Table 4, the emission indicator exhibits a strong comovement with production components. There is a positive relationship between CO₂ emissions and both industrial value added and agricultural value added. However, the association between CO₂ emissions and GDP per capita appears to be weak. The relationships between CO₂ and policy variables are in the expected direction. Specifically, carbon tax shows a moderate negative correlation with CO₂ emissions, agricultural value added, and industrial value added. Furthermore, emissions trading also displays a negative correlation with CO₂ emissions and sectoral indicators. The relationship between emissions trading and GDP per capita is positive but weak. The relationship between alternative energy use and CO₂ emissions is clearly negative, as expected. The correlation matrix indicates that emissions tend to increase with the scale of industrial and agricultural activity, while their association with per capita income remains weak. Negative correlations between carbon tax, ETS, and emissions suggest that pricing-based policy instruments are effective in reducing emissions. Additionally, the evident negative relationship between the share of alternative/nuclear energy and emissions indicates that higher adoption of such energy sources contributes to lower emission levels. The weak yet positive link between ETS and GDP per capita may imply that this system has been adopted earlier in countries with higher institutional capacity.

In the empirical literature, findings generally indicate that carbon pricing whether implemented through carbon taxes or emissions trading schemes has been effective in reducing emissions in the long run, although the magnitude of this effect varies depending on a country’s structural characteristics, sectoral composition, and policy design. In this regard, the patterns observed in the correlation table are consistent with the existing literature. For example, Jeffrey and Perkins (2015) and Charitou (2015) reported that energy taxation within the European Union significantly reduced carbon intensity, particularly in the industrial sector, and that this effect persisted through productivity channels following the introduction of emissions trading. Similarly, Al-Abdulqader et al. (2025) found that carbon taxes were associated with an approximately nine percent reduction in emissions across a global sample, while emissions trading achieved an even greater decline. Studies such as Di Foggia and Beccarello (2024) and Candelon and Hasse (2023) emphasized that the long-term impacts of carbon pricing are more pronounced than short-term fluctuations, which may explain the weak yet negative correlations between policy variables and emissions observed in the present analysis. Moreover, Gao and Zhang (2023) and Huang and Pan (2024) demonstrated that the joint design of tax and trading systems yields superior outcomes in terms of both environmental and economic efficiency.

Nevertheless, it should be noted that correlation does not imply causation, and the strong comovement observed between industrial–agricultural variables and emissions must be carefully addressed in the modeling process. Therefore, in the subsequent stage, tests that account for cross-country differences and common shocks will be crucial.

Table 5. Cross-Sectional Dependence Test.

Test	Statistic	Probability
Breusch-Pagan LM	7273.189	0.000 ***
Pesaran scaled LM	181.0359	0.000 ***
Pesaran CD	56.04447	0.000 ***

Note: *** indicates a statistical significance level of 0.01.

presents the results of the cross-sectional dependence test conducted to assess the extent to which shocks in one country affect other countries within the panel.

As shown in Table 4, the probability values of the Breusch-Pagan LM, Pesaran scaled LM, and Pesaran CD tests are all below the critical significance level. This indicates the presence of cross-sectional dependence within the panel, meaning that the countries do not act independently of one another. It suggests that economic or environmental shocks occurring in one country may influence others, implying that carbon emissions cannot be fully isolated within national boundaries. This finding underscores the need for global climate policies to be implemented not only at the national level but also through international coordination. Moreover, the existence of cross-sectional dependence implies that countries in the panel may respond similarly to policy changes, making it essential to employ econometric methods that account for heterogeneous tendencies rather than a single, uniform model. Consequently, the analysis proceeds with a unit root test that considers cross-sectional dependence-specifically, the CIPS (Cross-sectionally Augmented Im-Pesaran-Shin) test. The slope heterogeneity test developed by Pesaran and Yamagata (2008) examines whether the model coefficients are identical across countries. The null hypothesis (H₀) states that “slope coefficients are homogeneous,” meaning that the relationships among variables are the same for all countries, while the alternative hypothesis (H₁) posits that “slope coefficients are heterogeneous,” indicating differences across countries. The results of this test are presented in

Table 6. Slope Homogeneity Test.

Test	Statistic	Probability
Delta	13.045	0.000 ***
Adjusted Delta	16.352	0.000 ***

Note: *** indicates a statistical significance level of 0.01.

.

As shown by the results in Table 6, both the Delta (13.045) and Adjusted Delta (16.352) statistics have probability values of 0.000. Since these values are well below the 1% significance level, the null hypothesis is rejected. Consequently, the alternative hypothesis is accepted, confirming the presence of slope heterogeneity in the panel. This indicates that the model coefficients are not identical across countries in the panel.

Table 7. CIPS Unit Root Test.

Model	Level I(0)		Difference I(1)
Variables	CIPS Test		CIPS Test
LN_CO2	−1.505		−4.251 ***
CTX	−1.387		−3.998 ***
LN_GDP	−2.000		−2.723 ***
LN_AGR	−2.579***
LN_IND	−1.600		−3.531 ***
ENG	−1.902		−4.057 ***
Critical Values	%1	−2.29	%1	−2.30
	%5	−2.14	%5	−2.14
	%10	−2.06	%10	−2.06

Note: *** indicates a statistical significance level of 0.01.

presents the CIPS unit root test applied to assess the stationarity of the series.

According to the results presented in Table 6, the CIPS statistics indicate that at the level, the CO₂ emissions, carbon tax, GDP per capita, industrial value added, and alternative energy variables do not exceed the absolute value of the critical thresholds required for stationarity. Therefore, these series are not stationary at the level I(0). In contrast, agricultural value added appears to be stationary at the level. When first differences are taken, the CIPS statistics for CO₂ emissions, carbon tax, GDP per capita, industrial value added, and alternative energy exceed the critical values, indicating that these variables become stationary at first difference, I(1). Overall, the CIPS unit root test results suggest that the panel variables are generally non-stationary at the level but become stationary after first differencing. This implies that countries respond differently to short-term fluctuations, but there exists a potential convergence toward a long-term equilibrium. In other words, although there are short-term volatilities among carbon taxes, emissions trading, and economic indicators, a process of adjustment occurs over time. Slope heterogeneity and cross-sectional dependence tests indicate that the short-term effects of carbon pricing policies vary across OECD countries. The Delta and Adjusted Delta tests presented in Table 5 show that the coefficients in the panel are not identical for all countries (p < 0.01), while the cross-sectional dependence tests in Table 4 reveal that countries respond to economic and environmental shocks in an interdependent manner. Furthermore, the correlation results (Table 3) confirm the existence of negative effects of carbon taxes and the emissions trading system on CO₂ emissions, although the magnitude of these effects differs across countries. These findings suggest that carbon pricing policies have limited and heterogeneous short-term impacts across countries, whereas in the long term they produce stable and pronounced effects, consistent with studies emphasizing the strengthening of long-term impacts in the literature. This finding aligns with the literature, which emphasizes that the effects of carbon pricing are limited in the short run but become more pronounced in the long run. For instance, Candelon and Hasse (2023), in their frequency-domain analysis, found that the carbon tax has a significant effect on emissions only in the long run, while in the short run, the effect remains weak due to the adjustment costs of the economy. Similarly, Jeffrey and Perkins (2015) report that, in the European Union, energy taxes and emissions trading reduce carbon intensity in the long term, although short-term price volatilities delay the manifestation of these effects. Karapınar et al. (2019a) and Gao and Zhang (2023) also emphasize the long-term equilibrium of carbon tax and emissions trading systems, showing that environmental effectiveness increases once the economic adjustment process is completed. These findings are consistent with the “stationarization process” indicated by the CIPS test in this study. Additionally, Di Foggia and Beccarello (2024) and W. Sunanda et al. (2025) demonstrate that the persistent effects of carbon pricing instruments are strengthened over time through market structure and investment behavior. Therefore, the stationarity of the series at first difference is not merely a statistical outcome but also a structural finding that can be explained by the economic “adjustment” process of climate policies. Countries that respond differently to policy shocks in the short term tend to converge toward a similar long-term equilibrium, suggesting that carbon pricing can operate as an integrated mechanism on a global scale. The Pedroni panel cointegration test examines whether the I(1) panel data series move together in the long run. The results of this test are presented in

Table 8. Pedroni Panel Cointegration Test.

Test	Statistic	Probability
Modified Phillips–Perron test	6.9122	0.000 ***
Phillips–Perron test	−6.1761	0.000 ***
Augmented Dickey–Fuller test	−4.7177	0.000 ***

Note: *** indicates a statistical significance level of 0.01.

.

According to the Pedroni panel cointegration test results presented in Table 8, the null hypothesis (H₀) of “no cointegration” is rejected. All within-dimension statistics of the panel-Modified Phillips–Perron, Phillips–Perron, and Augmented Dickey–Fuller are statistically significant at the 1% level (p = 0.000 ***).

The results of the Pedroni cointegration test indicate the existence of a long-term relationship among the variables in the panel. This finding suggests that the variables included in the model move together over the long run and share a common equilibrium relationship. The Pedroni cointegration test provides the basis for the subsequent estimation of long-run coefficients. Moreover, the previously conducted cross-sectional dependence and slope heterogeneity tests revealed significant differences and interactions among countries in the panel. These results imply that, although the short-term effects of policies may be limited, carbon pricing instruments establish a stable mechanism for reducing emissions in the long run. This is consistent with the literature, which demonstrates that carbon pricing policies achieve long-term equilibrium with economic and environmental indicators. For instance, Candelon and Hasse (2023) and Jeffrey and Perkins (2015) emphasize that carbon taxes and emissions trading have limited short-term effects but produce stable long-term impacts, with emissions gradually responding to price signals. Furthermore, Gao and Zhang (2023) and Huang and Pan (2024) find that hybrid carbon pricing systems exhibit stronger long-term environmental impacts, which are also compatible with economic stability. In line with this literature, the Pedroni cointegration test results in this study indicate that carbon pricing is not merely a temporary fiscal instrument, but a permanent policy mechanism that becomes integrated with the economic structure and energy markets over time.

To reliably estimate the long-run relationship, an estimation method that accounts for both heterogeneity and cross-sectional dependence is required. Accordingly, the study employs the Mean Group (MG) estimator to examine the long-run relationship while simultaneously incorporating common shocks and structural differences among countries into the model.

According to the results presented in

Table 9. Mean Group Results.

LN_CO2	Coefficient	Standard Deviation	z	p > |z|	[%95 Conf. Interval]
LR
CTX	−0.0123	0.0353	2.90	0.004 ***	[0.0331, 0.1715]
ETS	−0.0192	0.0078	2.47	0.013 **	[0.0040, 0.0344]
LN_GDP	−0.2300	0.1642	−1.40	0.161	[−0.5518, 0.0918]
LN_AGR	−0.0442	0.0607	−0.73	0.467	[−0.1632, 0.0749]
LN_IND	0.3478	0.1129	3.08	0.002 ***	[0.1264, 0.5693]
ENG	−0.0362	0.0069	−5.20	0.000 ***	[−0.0498, −0.0225]
SR
_EC	−0.0586	0.0330	1.77	0.076 *	[−0.0062, 0.1233]
CTX	−0.0057	0.0472	−0.13	0.893	[−0.0895, 0.0781]
ETS	0.0516	0.0815	0.63	0.526	[−0.1077, 0.2108]
LN_GDP	−0.2209	0.4919	−0.45	0.653	[−1.1852, 0.7432]
LN_AGR	−0.0346	0.1444	−0.30	0.763	[−0.2588, 0.1897]
LN_IND	−0.0844	0.2846	−0.30	0.767	[−0.6422, 0.4734]
ENG	−0.0073	0.0130	−0.56	0.575	[−0.0181, 0.0327]
_CONS	−0.9433	2.3237	−0.41	0.685	[−5.4976, 3.6110]

Note: * 0.1; ** 0.05; *** 0.01 denote statistical significance at the 10%, 5%, and 1% levels, respectively. LR: Long Run; SR: Short Run.

, the Mean Group estimates indicate that, in the long run, the effect of the carbon tax on CO₂ emissions is negative and statistically significant. The relationship between CO₂ and industrial value added is positive, as expected, indicating that increases in industrial production significantly raise CO₂ emissions. The emissions trading system also has a negative and statistically significant impact on CO₂ emissions in the long run. Thus, the effects of carbon tax and emissions trading systems-developed as policies to combat climate changes-on CO₂ emissions are confirmed as anticipated. The use of alternative energy negatively and significantly affects CO₂ emissions, consistent with the definition of the variable: as the share of alternative, nuclear, or clean energy increases, CO₂ decreases. GDP per capita and agricultural value added are observed to be statistically insignificant in the long run. In summary, the long-run results indicate that a 1% increase in industrial value added raises CO₂ emissions by approximately 0.35%. A one-percentage-point increase in the share of alternative and nuclear energy reduces CO₂ emissions by approximately 3.6%. A one-unit increase in the carbon tax level reduces CO₂ emissions by roughly 1.23%. Although the coefficients for GDP and agricultural value added are negative, they are not statistically significant, and therefore no definitive conclusions can be drawn regarding their effects. Examining the short-run results, the short-run error correction coefficient (ec) is 0.0586, which is negative and close to significance at the 10% level. This indicates that the system adjusts slowly and weakly toward the long-run equilibrium. The finding suggests that CO₂ emissions exhibit strong persistence and inertia in the short run, implying that past emission levels significantly influence current emissions. Moreover, the short-run coefficients for CO₂ emissions, carbon tax, GDP per capita, industrial value added, agricultural value added, and alternative energy are statistically insignificant, indicating that the effects of policy and structural variables do not materialize immediately but accumulate over time. The short-run coefficients for the carbon tax, the emissions trading system, per capita income, agricultural value added, industrial value added, and the share of alternative/clean energy are all statistically insignificant. This outcome suggests that, given the annual frequency of the data and the characteristics of the panel structure, the impacts of policy-related and structural variables on emissions predominantly materialize in the long run. Consequently, the analysis does not provide robust evidence for short-term dynamics. Accordingly, the short-run results should not be interpreted as an absence of causal influence; rather, they indicate that the current model specification and data structure have a limited capacity to capture short-term adjustments.

The hypotheses presented in

Table 10. Hypothesis Table.

Hypotheses	Expected	Observed	Status
H1: As the carbon tax increases, CO2 emissions decrease.	(-)	(-)	Accepted
H2: As the coverage of the Emissions Trading System (ETS) increases, CO2 emissions decrease.	(-)	(-)	Accepted
H3: The policy effect is limited in the short run and becomes significant in the long run.	Significant	Significant	Accepted

The hypotheses marked in the table were formulated by the author.

are largely consistent with the literature emphasizing the effectiveness of market-based policy instruments, such as carbon taxes and emissions trading systems, in combating climate change. The findings indicate that these instruments have a significant and negative long-term impact on CO₂ emissions, thereby supporting emission reduction. This aligns with studies by Andersson (2019), Dong et al. (2017), and J. Zhang and Zhang (2018), which observe notable emission reductions across various sectors in response to higher tax levels. Similarly, Pretis (2022) and Colmer et al. (2025) demonstrate that carbon pricing fosters mitigation by inducing structural transformations in production dynamics, while Dechezleprêtre et al. (2018) show that ETS implementation reduces emissions without adversely affecting economic performance. In the Turkish context, findings by Uğur (2014), Çelikkaya (2024), and Kumbaroğlu et al. (2017) also support the suitability of carbon pricing as a tool for achieving sustainability objectives. Within this framework, the empirical results of the study corroborate policy-oriented environmental efficiency approaches highlighted in the literature.

The main assumption of this study is that key policy instruments developed to combat climate change, namely, the carbon tax and emissions trading, exhibit a negative and significant relationship with CO₂ emissions, one of the most important factors affecting climate change. Examining the results obtained for carbon tax-based policies, it is evident that taxation reduces CO₂ emissions in the long run. These findings are consistent with Andersson (2019), who observed that the carbon tax in Sweden significantly reduced transport-related emissions. Andersson (2019) reported a substantial decline of approximately 11% in emissions from transportation following the implementation of the tax. Moreover, Dong et al. (2017), using CGE scenario analyses for China, found that higher carbon tax levels lead to notable reductions in industrial CO₂ emissions, particularly identifying energy-intensive sectors such as electricity, metals, and chemicals as “priority reduction areas.” Therefore, their findings are aligned with the results of this study. J. Zhang and Zhang (2018), applying a computable general equilibrium (CGE) model to China’s tourism sector, demonstrate that the carbon tax not only reduces CO₂ emissions but also allows for an assessment of economic welfare effects while accounting for inter-sectoral interactions. Their results indicate significant decreases in tourism-related emissions with higher tax levels, although certain subsectors experienced short-term effects, and limited welfare losses in employment were observed.

Our panel analysis indicates that carbon taxation reduces emissions in the long run, while its short-term effect is weak or heterogeneous. However, no significant impact on per capita income, i.e., welfare, was observed. Pretis (2022) evaluated British Columbia’s 2008 carbon tax and found a statistically significant decrease in total CO₂ emissions. Transportation-related emissions, in particular, decreased by 19%. In the long run, similar to Andersson (2019), Pretis also observed reductions in the transportation sector, which aligns with our findings. No significant long-term effect was found for per capita income or agricultural value-added at the panel average. These findings suggest that emissions are primarily shaped by the production structure and energy mix, while price-based instruments can support reductions in the long term. Colmer et al. (2025) show that the EU ETS reduced CO₂ emissions by an average of 14–16% in manufacturing firms between 2005 and 2012 at the firm level, without adverse effects on value-added, employment, or productivity. The reductions were achieved through investments that decreased emissions intensity rather than leakage. Finally, Dechezleprêtre et al. (2018) find that ETS-covered plants in the European Union reduced CO₂ emissions, while employment, revenue, and profitability indicators did not experience significant negative effects.

5. Conclusions

Carbon pricing has emerged in recent years as an increasingly popular policy instrument for combating climate change and is expected to become widespread and regulated globally in the future to reduce emissions. In this context, two of the most widely implemented forms, carbon taxes and the Emissions Trading System (ETS), aim to reduce greenhouse gas emissions while maintaining economic efficiency.

This study examines the effects of carbon pricing instruments, namely, carbon taxes and the Emissions Trading System (ETS), on carbon dioxide (CO₂) emissions in OECD countries using an OECD panel dataset. The unbalanced panel covers the period 2002–2023, with the dependent variable being annual total CO₂ emissions. The main policy variables are the carbon tax (carbon-related energy taxes as a share of GDP) and an ETS dummy, while GDP per capita, agricultural and industrial value added, and the energy mix (share of alternative/nuclear energy) are included as control variables. For countries lacking direct data on carbon taxes, a proxy indicator based on tax revenues was constructed. Initially, significant cross-sectional dependence and slope heterogeneity were detected in the panel, indicating the presence of common shocks and structural differences across countries, which may render classical homogeneous estimation approaches misleading. Accordingly, stationarity tests (CIPS) and cointegration tests (Pedroni) were conducted using procedures that account for cross-sectional dependence and allow for heterogeneity. The results generally indicate that most series are non-stationary at the level but become stationary after first differencing, and that there exists a long-run cointegration relationship between the policy and control variables.

When examining short-run dynamics, the lack of statistical significance of policy variables in the short term indicates that the effects of carbon pricing and low-carbon transition policies materialize relatively slowly through cumulative and structural channels. Computable general equilibrium and macroeconomic simulation studies in the literature similarly suggest that the short-term impacts of carbon pricing are limited, whereas long-term effects are more pronounced and, in many cases, potentially positive (Böhringer & Rutherford, 2017). This finding is consistent with the “slow adjustment” mechanism observed in the present study. Moreover, the error correction term’s near-significant speed of adjustment toward the long-run equilibrium confirms, in line with empirical evidence of cointegration, that policy effects are shaped by long-term structural trends.

Policy implications can be summarized along three dimensions. First, the long-term negative coefficient of the carbon tax supports its potential for emission reduction. The findings of the study are interpreted not to provide detailed recommendations regarding tax design, but rather to demonstrate the existence of the policy effect. In this context, aspects such as revenue recycling, sector-specific transitions, and gradual pricing were not incorporated into the scope of the empirical model and are considered solely as prominent trends highlighted in the literature. This approach provides an important policy perspective in the context of alignment with the European Green Deal and the Carbon Border Adjustment Mechanism. Second, the positive coefficient of industrial value added underscores the need for complementary regulations in energy-intensive sectors that focus on efficiency and fuel substitution, such as performance standards, R&D incentives, and green financing. Third, increasing the share of alternative/nuclear and renewable energy sources represents the most effective channel for emissions reduction in the short and medium terms. Statements on grid flexibility, storage, and demand-side management are expected to reinforce this effect.

Differences in alternative energy usage among countries can primarily be explained by the nature of energy infrastructure and technological investments. Investments in grid flexibility help balance the production fluctuations of renewable energy sources, facilitating their integration into the energy system. Investments in clean technology and low-carbon energy infrastructure reduce carbon intensity, while environmental and climate-focused investments ensure that revenues are directed toward sustainable energy projects. In this context, investments in renewable energy sources such as solar, wind, hydroelectric, and biomass enhance the resilience of the energy system against short-term fluctuations and strengthen their long-term emission-reducing effects. The high volatility in the share of alternative energy usage indicates significant differences among countries in terms of access to and implementation of such investments (Gao & Zhang, 2023; Di Foggia & Beccarello, 2024).

Considering country-specific differences, policy design should be adapted to national conditions rather than following a one-size-fits-all approach. Economies with coal-dominated energy mixes and a high industrial share tend to exhibit higher marginal abatement effects, whereas countries with a high share of natural gas or renewables in electricity generation may prefer smoother transition pathways. In light of findings indicating that the ETS and carbon tax are complementary rather than substitutive under proper design, the consistency of the price signal and expectation management becomes crucial. Expanding the scope of the ETS, implementing price floors, and aligning the design of the tax and ETS can reduce uncertainty in investment decisions.

Hybrid models that combine carbon taxes with emissions trading systems (ETS) represent regulatory frameworks that integrate different elements of carbon pricing policies within a single structure. In these models, price-based carbon taxes and quantity-based emissions trading are employed simultaneously, providing a broader set of instruments in policy design. The literature highlights hybrid regulations for their advantages in policy flexibility, diversification of revenue streams, and the expansion of cross-sectoral applications. Hybrid systems have been shown to limit price volatility, enhance investment certainty, and strengthen long-term emission reductions (Pizer & Zhang, 2022).

The contribution of this study lies in three aspects: (i) evaluating both the carbon tax and ETS within the same model while accounting for the energy mix and sectoral shares, (ii) estimating long-run coefficients using the MG approach after statistically confirming cross-sectional dependence and heterogeneity, and (iii) linking the results to policy design. Nevertheless, several limitations exist: the carbon tax indicator is necessarily a proxy measure, the ETS dummy variable cannot capture design differences, and country-level averages may obscure within-sector behavioral heterogeneity. One of the major limitations of this study is the limited ability to observe short-run dynamics due to the use of annual data and the panel structure. In some countries, emissions, industrial output, and the energy mix can exhibit sharp fluctuations over very short periods as a result of policy changes, price shocks, or technological transformations. Moreover, studies employing different data frequencies or longer samples could better clarify delayed response channels. For future research, the following avenues are suggested: (1) assessing the welfare effects of integrating revenue recycling scenarios with border carbon adjustments, (2) examining the role of ETS tax hybrid schemes in price stability and cost-effectiveness, and (3) comparing marginal abatement costs through sectoral disaggregation, particularly for energy, steel, cement, and transportation sectors. The findings derived from the panel data analysis indicate that carbon pricing policies, particularly the carbon tax and the Emissions Trading System (ETS), have a significant and statistically robust impact on reducing greenhouse gas emissions in the long run. The analysis shows that both policy instruments exert a negative and significant effect on emissions over the long term, and as the coverage of the ETS expands, per capita CO₂ emissions decrease. This evidence confirms the hypotheses proposed in the study. In particular, the increased inclusiveness of the ETS strengthens emission reductions, and the policies exhibit limited effects in the short run but pronounced impacts over the long term. In the short run, emission levels remain largely determined by past production and consumption patterns, limiting the immediate effect of policy changes. However, over time, as the implementation of measures and the regulatory framework matures, both the carbon tax and the ETS enhance their effectiveness, leading to more pronounced emission reductions in the long term. This demonstrates that these policy instruments operate through a time-distributed mechanism, and when considered alongside structural transformation processes, they produce more durable and sustained outcomes.

The design of a carbon tax in accordance with the “polluter pays” principle ensures that environmental costs are internalized within the economic system. By taxing fossil fuels (coal, oil, and natural gas) at rates proportional to their carbon intensity, the carbon tax both incentivizes emissions reduction and generates a significant fiscal resource for the government. In this context, the carbon tax functions not only as a fiscal instrument but also as a market mechanism that supports environmental sustainability. Directing the resulting tax revenues toward environmental protection projects, renewable energy investments, or shielding low-income households from carbon costs enhances both environmental effectiveness and social equity. Thus, carbon tax implementation produces a “double dividend” by simultaneously reducing emissions and generating public revenue. The Emissions Trading System (ETS), on the other hand, emerges as a mechanism that directly controls the total quantity of emissions. The government sets a cap on emissions for specific sectors and allocates allowances to firms, either freely or via auction. The determination of emission caps for specific sectors, and the allocation of allowances to firms either through distribution or auctioning within these caps, falls under the responsibility of competent national or regional regulatory authorities. Setting the cap is a multi-stage policy process with legal and institutional dimensions and does not constitute a fixed standard. This process is shaped by countries’ climate legislation, long-term mitigation strategies, and regional regulations (e.g., the European Union Emissions Trading System). Cap levels are designed based on nationally determined contributions (NDCs) under the Paris Agreement and sectoral emission reduction targets, thereby serving as a regulatory parameter that reflects national policy priorities while remaining consistent with the guiding norms of the global climate regime (P. He et al., 2019; Çelebi-Boz & Örs-Onur, 2024). Firms that emit below their allocated allowances can sell the surplus on the market, whereas firms exceeding their allocation must purchase additional permits. In this way, the ETS maintains control over total emissions while allowing the carbon price to be determined by market supply and demand dynamics. From this perspective, the carbon tax represents a price-based policy instrument, whereas the ETS functions as a quantity-based mechanism.

The analysis further indicates that external policy instruments, such as the European Union’s Carbon Border Adjustment Mechanism (CBAM), have a significant impact on countries operating in carbon-intensive sectors. Sectors such as cement, iron and steel, and aluminum are expected to be particularly affected by this regulation. Overall, the findings suggest that implementing carbon pricing instruments alone may not be sufficient. An effective climate policy requires not only pricing carbon but also transitioning to clean production technologies in industry and increasing the share of renewable energy in the energy supply. Additionally, the level of the carbon tax, the stringency of the ETS cap, and the allocation of generated revenues directly influence policy effectiveness. In this regard, the study emphasizes that carbon pricing should be considered not only as an economic instrument but also in conjunction with structural, technological, and social dimensions.

In conclusion, this analysis of OECD countries empirically demonstrates that both carbon taxes and the Emissions Trading System (ETS) are effective instruments in the long run, both environmentally and fiscally. The success of these policy tools varies depending on countries’ energy structures, income levels, and the design of the policies. Therefore, designing carbon pricing policies in a complementary manner and allocating the resulting revenues to balance environmental objectives with social equity are critical for achieving a sustainable and inclusive green transition.