Synergy pathway innovation of carbon emission trading, tradable green certificate and green power trading policies: achieving China’s dual carbon goals

Abstract

The carbon emission trading scheme (ETS), tradable green certificate (TGC) and green power trading (GPT) policies are vital for promoting energy transformation and carbon reduction under the dual carbon goals. However, the effects of and relationships among multiple policies urgently need to be studied. In this work, the panel data of 30 provinces in China from 2010 to 2023 are used. First, through the multiperiod difference-in-differences (DID) method, fixed effect models and mediating effect models, the carbon reduction effects of the pilot and national ETS policies, the renewable energy development effects of the TGC and GPT policies, and the multipolicy synergy effect are examined. A dual machine learning model is innovatively introduced to test the robustness of the results. Second, the slack-based measure–directional distance function–global Malmquist–Luenberger (SBM–DDF–GML) method is used to calculate the GTFP and investigate its transmission effect on policies. Finally, the impacts of the ETS, TGC and GPT policies on fossil fuel consumption are further analysed. The results indicate the following. (1) The pilot ETS policy reduces carbon emissions and carbon intensity, whereas the national ETS policy increases carbon emissions and carbon intensity in the short term. The TGC and GPT policies increase renewable energy generation and its proportion. (2) The synergy of the pilot ETS and GPT policies is the best for reducing carbon emissions and carbon intensity. The synergy among national ETS, TGC and GPT policies is optimal for developing renewable energy. In addition, there is redundancy between the TGC and GPT policies. (3) The pilot ETS policy inhibits GTFP, whereas the national ETS, TGC and GPT policies promote GTFP. The GTFP significantly reduces carbon emissions and carbon intensity and increases renewable energy generation and its proportion. (4) Both the pilot ETS and national ETS policies reduce the intensity of fossil fuel consumption. The GPT policy reduces the total level of fossil fuel consumption, whereas the TGC policy increases this level. In this work, innovative decarbonisation policies synergy pathways and insights into achieving green and low-carbon transitions in China and other developing countries are provided.

Introduction

Climate change and global warming pose great challenges to the world, with far-reaching impacts on ecosystems, society, and the economy. Reducing greenhouse gas (GHG) emissions is the key to addressing climate issues. China, as the largest global energy producer and carbon emitter, places significant emphasis on GHG emission reduction and the energy transition. China proposed the carbon peaking and carbon neutrality goals (abbreviated as dual carbon goals) in 2020, i.e., reaching China’s peak carbon dioxide emissions by 2030 (the total amount of carbon emissions should stop increasing and begin to decline after this point) and achieving carbon neutrality by 2060 (offsetting self-generated carbon dioxide emissions by planting trees, saving energy and reducing emissions to achieve zero carbon emissions). China has actively implemented a variety of decarbonisation policies, such as the carbon emission trading scheme (ETS), renewable energy power consumption quota (RECQ), tradable green certificate (TGC) and green power trading (GPT) policies. These policies are increasingly becoming the focus of the energy sector with direct or indirect interactions between them to decrease GHG emissions and promote energy transition.

The ETS policy, as a greenhouse gas emission reduction policy tool based on the market mechanism, combines both economic benefits and environmental benefits. This policy is regarded as an effective measure for reducing global carbon emissions and addressing climate change(Jiang et al. 2016). When the total amount of emissions of an entity is higher (or lower) than the carbon allowances allocated by the government free of charge, it can buy (or sell) carbon allowances in the ETS market for compliance. Since 2013, pilot areas have successively launched ETS, covering fields such as electricity, steel, industry and chemical engineering. To allocate of carbon quotas, the free allocation method is the primary technique adopted, and it is supplemented by paid allocation. On the basis of the phased achievements of ETS pilot projects, China has accelerated the construction of the national unified ETS market. In July 2021, the national ETS market began to trade. In the first batch, the power generation sector was included, covering approximately 4.5 billion tons of carbon dioxide. Free allocation was adopted in the initial stage to allocate quotas. In March 2025, the Chinese government approved the expansion of the sector coverage of the national ETS policy from the power sector to the iron and steel production, cement production and aluminium electrolysis sectors (MEE, 2025), which will further increase the liquidity and activity of the national ETS policy. Therefore, it is important to assess the impacts of the pilot and national ETS policies, providing policy implications to promote the expansion of the national ETS policy.

Solar and wind power generation are sources of renewable energy that are free from carbon emissions and thus help achieve the goal of global carbon neutrality. Moreover, by mitigating climate change, the power generation levels of solar power and wind power can increase. There is a mutually reinforcing connection between renewable energy development and the goal of carbon neutrality (Lei et al. 2023). Therefore, to increase the growth of renewable energy, China has introduced an electricity price subsidy policy. However, the rapid expansion of the production capacity of renewable energy has led to saturation and even oversupply in the domestic market, and the subsidy gap has expanded rapidly. To address the problem of arrears in renewable energy subsidies, China proposed the TGC policy in 2017. The National Energy Administration (NEA) issues electronic certificates for the on-grid electricity of onshore wind power and centralized photovoltaic power generation projects that benefit from subsidies. One megawatt hour of renewable energy power can be issued one TGC. The TGC policy adopts a certificate–power separation trading model, clearly stating that users can purchase TGCs as vouchers for consuming green power and that renewable power generators can obtain the benefits of the policy. In March 2025, the Chinese government issued a new policy to establish a TGC acquisition mechanism that combines mandatory and voluntary procurement to stimulate the demand for the certificates by strengthening the policy synergy between RECQ, carbon emission intensity and amount control (abbreviated as dual control for carbon emissions) and other policies (NDRC et al. 2025). Therefore, given the strong support for the TGC policy by the government, there is an urgent need to study the effects of this policy on the development of renewable energy.

Green power refers to all the electricity that is produced by renewable power generators, such as wind, photovoltaic, hydropower, biomass, geothermal and ocean energy power generation. In September 2021, China launched the GPT policy by integrating power and certificates, expanding the trading mode of TGCs and providing a Chinese solution for the global development of renewable energy. The GPT policy is a component of the medium- and long-term trading of power that refers to the transaction of both green power and the environmental benefit (i.e., TGC). In August 2022, China clearly stated that the use of renewable energy would not be counted towards the total amount or intensity control of energy consumption (abbreviated as the dual control for energy consumption). In March 2025, the China stated that increasing the scale of the GPT policy and improving its policy mechanism are necessary (NDRC et al. 2025). The support of the above policies has made the green power consumption policy system more complete, which is conducive to further increasing the market potentials of the TGC and GPT policies. Therefore, quantifying the effect of the GPT policy on the development of renewable power is highly important for further improving policy implementation and promoting a green and low-carbon transition.

Notably, since the launch of the national ETS market in July 2021, the lack of coordination among the ETS, GPT, and TGC markets has sparked widespread controversy. For example, when assessing a carbon inventory, if an entity purchases green power, the consumption of green power can be deducted from conventional power consumption to avoid doubling the payment of environmental costs. China has issued a number of policies, all of which propose improving the coordination mechanisms among the ETS, GPT, and TGC markets and effectively connecting these three polices (NDRC and NEA, 2022; NDRC et al. 2024a; 2024b). Therefore, studying the synergy between the ETS, TGC and GPT policies and determining how to enhance it are keys to promoting the development of renewable energy and the green and low-carbon transition.

Although more scholars have focused on the carbon emission reduction effect of the pilot ETS policy (Wu et al. 2021; Liu et al. 2024; Ji et al. 2025), further research is needed on the emission reduction effect of the national ETS policy, the renewable energy development effect of the TGC and GPT policies, and their influencing pathways. In addition, how can the ETS, TGC and GPT policies be coordinated to achieve carbon reduction and renewable energy development? Is there redundancy among policies? What are the transmission paths and effects of policies? What are the impacts of policies on the comprehensive transformation from the dual control of energy consumption to the dual control of carbon emissions? These questions urgently need to be addressed.

Therefore, this study makes significant contributions and innovations in the following areas. (1) The implementation time and characteristics of each policy are considered, official panel data of 30 provinces in China from 2010 to 2023 are used, and a multiperiod difference-in-difference (DID) model and fixed-effect models are adopted to explore the effects of pilot and national ETS policies on carbon emissions and intensity and the effects of TGC and GPT policies on renewable power generation and proportion. Moreover, the innovative introduction of the dual machine learning model is used to test the robustness of the results. (2) This study is different from existing studies. We innovatively design different synergistic pathways of multiple policies, construct seven regression models, and verify whether there is synergy between policies and synergistic effects. In particular, we examine whether the policies are specifically characterized as additive effects (the total effect is equal to the addition of each policy), antagonistic effects (the total effect is smaller than the additive effect) or synergistic effects (the effect is larger than the additive effect). (3) A new analytical perspective combining the slack-based measure–directional distance function–global Malmquist–Luenberger (SBM–DDF–GML) method is proposed to measure the green total factor productivity (GTFP), and more than 10 models of the mediating effect are established to assess the transmission effects of the GTFP in the realization of carbon reduction and renewable power development through the implementation of the ETS, TGC and GPT policies. (4) We consider the latest issue (the shift from the dual control for energy consumption to dual control for carbon emissions) and further analyse the effects of three policies on the transition, which is forwards-looking. (5) The results reveal that the pilot ETS promotes carbon reduction, whereas the national ETS inhibits carbon reduction in the short term. Moreover, the TGC and GPT policies can promote renewable power development. There is redundancy between the GPT and TGC policies. The national ETS, TGC and GPT policies can all contribute to carbon reduction and renewable power generation enhancement by increasing GTFP. The ETS and GPT policies help control fossil energy consumption, whereas the GPT policy increases total fossil energy consumption. (6) The results of the above innovations and their policy implications provide insights into improving decarbonisation policies practices in China and other developing countries.

The structure of this study is as follows. In “Literature review and hypotheses”, we present the literature review and hypotheses, locate an academic gap, explain the innovation of this work, and propose hypotheses through theoretical analysis. “Research design” presents a verification of the hypotheses by constructing models and explaining the data sources, sample selection and variable settings. “Results“ presents an analysis of the empirical results, including the analysis of the benchmark regression results, the parallel trend test, the robustness test, the assessment of the effect of multipolicy synergy, the analysis of the process and the impact of multipolicy on the transition from dual control for energy consumption to dual control for carbon emissions. This research justifies the hypotheses on the basis of the research results. “Conclusions and policy implications“ presents the conclusion, discussion, policy implications, and limitations of this paper and the outlook for future research.

Literature review and hypotheses

Literature review

In the following, the state of research is analysed from two perspectives: research on the impacts of the individual policies of ETS, TGC and GPT and on the impacts of synergies between these policies.

Research on the effects of single policies

Most scholars focus on researching the effects of ETS, TGC and GPT policies on carbon emission reduction and renewable energy development measures. Thus, the current effects of these policies on the environment and the economy and the transmission mechanism are analysed below.

Research on ETS policy. The ETS has become one of the most sophisticated areas of research (Sandoff and Schaad, 2009). Many scholars have adopted the DID model to test the effects of pilot ETS in terms of the environment and economy. ETSs can reduce carbon emissions (Zhang et al. 2017) and intensity (Zhou et al. 2019). The pilot ETS can increase production efficiency (Yang et al. 2021) and total factor productivity (TFP) (Wang and Qian, 2024; Dong et al. 2024), promote corporate economic development (Zhang et al. 2025; Sun et al. 2024), significantly increase corporate financialization (Zhang and Zheng, 2024), and have a significant positive effect on high-quality economic development (Zeng et al. 2024). In addition, the pilot ETS facilitates the upgrading of the green technology of firms (Zhu et al. 2019; Xi and Jia, 2025), preventing firms from relying on government subsidies while improving their green innovation performance (Wang and Zhou, 2024). The policy significantly increases the GTFP in pilot cities in the short term (Li et al. 2022); however, the long-term effects are unknown. By using the Gregory and Hansen cointegration test, Amaddeo et al. (2025) found that there is a cointegration relationship between carbon allowance prices and energy prices, shifting the tax burden from firms to consumers and hindering the ETS goal of reducing total emissions.

Research on TGC and GPT policies. To mitigate climate change, more countries with well-developed electricity markets have implemented GPT and TGC policies (Gan et al. 2007). The GPT policy can significantly promote renewable energy power consumption, and the TGC policy offers a more efficient solution for GPT (Yang et al. 2024). First, research on the TGC policy is needed. Some scholars have noted that TGCs lead to some problems, such as increasing user-side consumption costs (Bergek and Jacobsson, 2010) and causing social welfare losses (Helgesen and Tomasgard, 2018). This phenomenon does not truly reflect the actual reduction in global carbon emissions through the purchase of TGCs (Walenta, 2020). In addition, the effect of TGC policy in reducing emissions is overestimated (Bjørn et al. 2022). However, scholars hold a different view. For example, the TGC policy increases the competitiveness of renewable energy producers, effectively reduces government expenditures on renewable power subsidies (Zhang et al. 2018), advances renewable energy, and significantly decreases pollutant emissions (Zeng et al. 2022). Subsidies and grants are the most favourable strategies for incentivizing renewable energy investments, confirming once again that TGCs, a way of subsidizing renewable energy generators, can promote their investment and generation (Solangi and Magazzino, 2025).

Second, research on the GPT policy is needed. GPT can reduce air pollutant levels in regions that purchase green power (Guo et al. 2022). Tang et al. (2023) built a DID model to examine the impact of the GPT policy and noted that it greatly reduces the indebtedness of policy-covered generation companies. Kartal et al. (2024) estimated that renewable energy generation is more effective in reducing the marginal effect of CO₂ based on the KRLS methodology and suggested that policy-makers in China should improve the efficiency of renewable energy generation. Furthermore, Solangi et al. (2024) suggested that government green innovation initiatives, consumer initiatives and industry initiatives are the most important strategies for deploying renewable energy technologies in China. In addition, consumers with high power consumption are more willing to purchase green power than other consumers are (Calikoglu and Aydinalp Koksal, 2022). However, due to cost impacts, high-energy-consuming enterprises are negatively affected in terms of their willingness to participate in GPT (Kong et al. 2025).

Research on the synergistic effects of multiple policies

Multiple policy combinations can solve problems from different fields (Wilts and O’Brien, 2019) and are more effective than individual policies are (Wang et al. 2020). The current state of research on the effects of multipolicy synergies is analysed below in terms of the theoretical mechanisms of synergies, the environmental and economic effects, and the transmission mechanisms.

In terms of synergistic theoretical mechanisms, the market-based carbon price established by the EU ETS market and the adjustment costs caused by renewable energy development are passed on to users through the electricity price (Zhao et al. 2012). When accounting for carbon emissions from purchased electricity in California, electricity can be considered green with zero carbon emissions if the corresponding renewable energy certificates (RECs) are submitted for the amount of electricity (California air resources board, 2015). However, the transmission of carbon prices to electricity prices is weak in China, and the zero-carbon property of green power has not been popularized nationwide. Chinese scholars have attempted to achieve electricity–ETS–TGC synergy in many ways. For example, Shang et al. (2023) proposed strengthening the synergy of multiple markets in terms of the market scope, price system, product system and governance system. Through carbon accounting, Shang et al. (2024) subsequently analysed the mechanism of mutual recognition of environmental rights and interests among multiple markets. Some scholars have analysed multipolicy synergy from the perspectives of carbon accounting and price. For example, Zhang et al. (2024) set up an environmental rights and benefits conversion mechanism for the ETS, TGC and GPT markets on the basis of carbon emissions and average transaction prices.

Multiple policy synergies with respect to environmental impacts, the RECQ mechanism and TGC trading can drive the development of renewable energy and decrease carbon emissions (Bird et al. 2011; Ma et al. 2024). Liu et al. (2025) reveal that synergizing the GPT and TGC policies can safeguard the stable operation of the power system while promoting the development of renewable energy. Gawel et al. (2014) analysed the interactions between the EU ETS and support for renewable energy, indicating that the policy combination may improve the overall efficiency of climate and energy policies. The ETS and TGC policies can be used as complementary policies to curb carbon emissions (Cui et al. 2020; Guo et al. 2023; Wei et al. 2023), reduce the carbon intensity of the power sector (Zhang et al. 2024), and modify electricity usage composition (Zhou et al. 2024). Jia and Wen (2024) revealed that a combination of the pilot ETS and low-carbon city policies significantly enhances the emission reduction effect and green innovation power of the pilot ETS policy. The introduction of pilot TES and TGC markets can increase renewable energy production and reduce carbon emissions (Hou et al. 2024; Gong et al. 2025).

In terms of the economic impact of multipolicy synergies, Corradini et al. (2018) simulated the effects of the EU’s low-carbon policy combination and proposed that improving energy efficiency and utilizing renewable energy can effectively reduce electricity prices and economic losses. Schusser and Jaraitė (2018) reported that carbon price growth has a short-term positive effect on TGC prices. Under the policy combination of the ETS and RECQ mechanisms, a carbon price increase can lead to an increase in the electricity price and a decrease in the TGC price(Ma et al. 2024). The coordinated development of TGCs and pilot ETS reduces the pressure of renewable energy fiscal deficits and supports the energy transition (Chang et al. 2023) while increasing the market share and profits of renewable power producers(Wang et al. 2022; Guo et al. 2025). Jia et al.(2025) revealed that by introducing pilot ETS and TGCs, market players transform carbon reduction potential into economic benefits while enhancing synergies among energy markets. However, whether there is redundancy between TGC and ETS policies depends on the method used to set carbon quotas and RECQ (Feng et al. 2021). Therefore, there is an urgent need to scientifically and rationally design a synergistic mechanism for ETS, TGC and GPT policies to promote a green and low-carbon transition.

In summary, most existing studies have been focused on assessing the effects of the pilot ETS policy and the synergistic effect of the ETS and TGC policies, whereas the effects of the national ETS, TGC and GPT policies and their transmission mechanisms need to be further explored. Specifically, the previous research has four gaps:

• (1) First, from the perspective of a single policy, scholars mainly evaluate the pilot ETS policy, while the effect of the national unified ETS policy needs in-depth evaluation. There are gaps in research on the extent to which the GPT policy promotes energy transition and on which policy is more effective between the GPT and TGC policies.

• (2) Second, for multipolicy synergy, further research on the relationships among the pilot and national ETS, GPT and TGC policies is needed. This research should cover whether there is redundancy or synergy among policies and how to achieve synergy to promote carbon reduction and renewable energy development.

• (3) Third, research on the transmission paths and effects of single policies and multiple policies is urgently needed.

• (4) Fourth, the impacts of the pilot and national ETS, GPT and TGC policies on the comprehensive transformation from the dual control of energy consumption to the dual control of carbon emissions urgently need to be studied.

Therefore, to bridge the above gaps, this paper uses provincial public panel data from 2010 to 2023, constructs multi-period DID model and fixed effect models. It evaluates the effect of pilot ETS and national ETS policies on carbon emissions and carbon intensity, the effect of TGC and GPT policies on renewable energy generation and its share, and the synergistic effect of multiple policies. It constructs mediation effect models to analyse the transmission mechanisms of individual policies and multiple policy interactions. The impact of the three policies on the dual control of energy consumption is further explored. Finally, innovative policy recommendations are proposed to achieve green and low-carbon transition in China and other developing countries.

Research hypotheses

In 2021, China’s Central Economic Work Conference proposed realizing the shift from the dual control of energy consumption to the dual control of carbon emissions as early as possible. Unlike the dual control for energy consumption, the dual control for carbon emissions can account for the differences in the carbon intensities of different types of energy sources, exclude renewable energy consumption in terms of coverage, and achieve the dual carbon target more accurately. Therefore, promoting a comprehensive transition from the dual control of energy consumption to the dual control of carbon emissions can actively and steadily promote the achievement of the dual carbon goal.

Therefore, this study is based on the context of the dual carbon goals and the shift from the dual control of energy consumption to that of carbon emissions. The corresponding theoretical framework is shown in Fig. 1. (1) We analyse the effects and transmission mechanisms of the pilot ETS and national ETS policies on carbon emissions and carbon intensity at the theoretical level, explore the effect of the ETS policy on the dual control of carbon emissions, and propose theoretical hypotheses H1, H2 and H8. (2) We analyse the effects and transmission mechanisms of TGC and GPT policies on renewable energy generation and its proportion and formulate theoretical hypotheses H3, H4 and H9. (3) Through the synergistic theory, we analyse the effects and transmission mechanisms of ETS, TGC and GPT policies on the dual control of carbon emissions and renewable energy development and propose theoretical hypotheses H5, H6, H7 and H10. (4) Starting from the sources affecting total carbon emissions and carbon intensity, we further analyse the effects of ETS policies on fossil energy intensity and the of TGC and GPT policies on total fossil energy consumption, and we propose research hypotheses H11 and H12, respectively. To verify the reasonableness of the hypotheses, the research hypotheses are proposed in “Carbon reduction and renewable energy development effects of individual policies“–“Effects of policies on the dual control of energy consumption“. The specific theoretical analyses and research hypotheses are as follows.

Fig. 1: Theoretical framework.

The figure shows the study variables, the study hypotheses and the relationships between the variables. a shows the hypothesis of the ETS policy on the dual control for carbon emissions; b shows the hypothesis of the TGC and the GPT policies on the renewable energy development; c shows the hypothesis of the synergy of the ETS, TGC, and GPT policies on the dual control for carbon emissions and the renewable energy development; d shows the hypothesis of the ETS, TGC, and GPT policies on the dual control for energy consumption. Source: Decarbonisation policies synergy pathway innovation: Achieving dual carbon goals.

Carbon reduction and renewable energy development effects of individual policies

Carbon reduction effect of the ETS policy

The ETS policy refers to the fact that if a company’s total carbon emissions exceed (or fall below) the carbon quota allocated free of charge by the government, it can buy (or sell) carbon quotas in the ETS market to meet the assessed level. The policy internalizes the external cost of carbon emissions into a carbon price signal by establishing a market-based emission reduction mechanism based on quota constraints and price signals; thus, this policy encourages industries to optimize their energy structure and improves their emission reduction efficiency (Ovaere and Proost, 2022). China’s pilot ETS policy follows the theoretical frameworks of total carbon emission market regulation and intensity control market regulation, and it covers multiple industries, such as electricity, cement, and chemical engineering. The baseline method is adopted to formulate free carbon quotas to restrict the volume of carbon quota trading and reduce the carbon intensity (Winkler et al. 2021; Eslahi and Mazza, 2023). The pilot ETS policy promotes carbon reduction (Wu et al. 2021; Duan et al. 2023), and the carbon reduction effects among the pilot areas are not completely consistent due to policy stringency and industry coverage (Wen et al. 2021). However, the implementation of the pilot ETS can simultaneously reduce carbon emissions and carbon intensity in pilot regions (Wu et al. 2021; Duan et al. 2023). Therefore, we propose Hypothesis 1 and test it further through a multiperiod DID model (benchmark regression model) of the pilot ETS policy.

Currently, the proportion of paid quotas in China’s pilot ETS markets is overly cautious. If the national ETS market refers to the pilot markets to set the proportions of free and paid quotas, it is feared that China will be unable to achieve the emission reduction target by 2030 (Li and Jia, 2016). In addition, the overall operating efficiency of the national ETS market is relatively low and not yet effective (Tang and Liao, 2024). During the research period of this paper, compared with the pilot ETS policy, the national ETS policy includes the power industry only; other high-emission industries lack corresponding carbon emission constraints and emission reduction incentives. Thus, these industries may continue to produce in high-carbon-emission manners and increase the consumption of high-carbon electricity, leading to increases in carbon emissions and carbon intensity. Therefore, we propose Hypothesis 2 and further test it using the fixed effect model (benchmark regression model) for the national carbon trading policy.

Hypothesis 1: The pilot ETS policy can reduce carbon emissions and carbon intensity and has a positive effect on achieving the dual control of carbon emissions.

Hypothesis 2: The national ETS policy that only covers the electricity sector increases carbon emissions and carbon intensity and has a negative effect on achieving the dual control of carbon emissions.

Renewable energy development effect of the TGC and GPT policies

The TGC policy is the only proof of the environmental attributes of renewable power in China, and it is the only credential that recognizes the production and consumption of renewable power. The TGC policy incentivizes the construction of renewable energy facilities, has a positive impact on the development of renewable energy (Gan et al. 2007; Zeng et al. 2022), improves the competitiveness of renewable energy generators in the electricity market (Helgesen and Tomasgard, 2018) and provides a more flexible option for green electricity trading (Yang et al. 2024). TGC, as a supporting policy of the RECQ, is an important method for promoting better development, construction, consumption and use of renewable energy (Song et al. 2021). Accordingly, we propose Hypothesis 3 and further verify it through the fixed effect model (benchmark regression model) of the green certificate trading policy.

Green power is traded using the TGC and electricity bundle model, which reduces the risk of economic loss by ensuring environmental attributes through participation in medium- and long-term electricity trading. GPT can increase the production and use of renewable energy power, significantly reduce air pollutant emissions in importing regions, and facilitate the low-carbon transformation of the power industry (Guo et al. 2022). In addition, GPT can reduce the debt of renewable energy producers and achieve green and sustainable development (Tang et al. 2023). Before 2030, GPT should be implemented by bundling electricity and certificates to foster the growth of renewable energy (Hu et al. 2024). Accordingly, we propose Hypothesis 4 and further test it through the fixed-effect model (benchmark regression model) of the GPT policy.

Hypothesis 3: Green certificate trading policies can increase renewable energy production and its proportion, promote renewable energy development and contribute to achieving the dual carbon goals.

Hypothesis 4: GPT policies can increase renewable energy production and share, promote renewable energy development and contribute to achieving the dual carbon goals.

Effect of policy synergy on the dual control of carbon emissions and renewable energy development

Synergistic effect of the pilot ETS, TGC and GPT policies

When different policy combinations produce synergistic effects, such as the linkage between the electricity sector and the pilot ETS market, there is a transmission link between carbon prices and electricity prices, which further reduces the total amount of carbon emissions (Li et al. 2022). In addition to onshore wind power, the promotion of offshore wind power development reduces the amount of carbon dioxide globally from 2020 to 2040 by 2.6–3.6 Gt (Li et al. 2022). However, when different policy combinations have antagonistic effects, such as when the pilot policies for innovative cities and low-carbon cities are combined, due to the substitutability of policy effects, there is mutual crowding out (Zhang et al. 2023). In addition, on the basis of institutional boundary theory (Vargo and Lusch, 2004, 2008), there is a functional conflict between the pilot ETS and the TGC or GPT policies in terms of environmental equity accounting rules. Specifically, the pilot CET policy aims to curb carbon emissions, whereas the TGC and GPT policies aim to offset the carbon dioxide that has been emitted. However, the offset ratio is limited, resulting in limited development space and application scenarios of the TGC and GPT policies in the pilot ETS market. Given this information, we present primary Hypotheses 5a and 5b and test them through fixed effect models of synergies between the pilot ETS policy and the TGC and GPT policies.

Hypothesis 5a: The synergy between the pilot ETS and TGC policies can reduce carbon emissions and carbon intensity, but it inhibits the increase in renewable energy generation and its proportion.

Hypothesis 5b: The synergy between the pilot ETS and GPT policies can reduce carbon emissions and carbon intensity, but it inhibits the increase in renewable energy generation and its proportion.

Synergistic effect of the national ETS, TGC and GPT policies

There is a lack of research on the synergies between the national ETS policy and the TGC and GPT policies, and existing research has focused on pilot ETS; for example, under the policy objective of promoting renewable energy development, the introduction of TGCs and pilot ETS can optimize the power source structure (Feng et al. 2018). The coordinated implementation of the two policies can increase enthusiasm for TGCs (Meng et al. 2024). Moreover, this process can more effectively achieve high-quality renewable energy development and promote the low-carbon transformation of the power industry (Su et al. 2024). In addition, the ideal share of renewable energy power varies under different policies (Ma et al. 2024b). The environmental equity conversion mechanisms among the ETS, TGC and GPT markets increase the market share and environmental benefits for renewable energy generators (Zhang et al. 2024). The analysis above shows that the application of the national ETS can temporarily lead to increases in carbon emissions and carbon intensity. If the national ETS is limited to the power industry, then carbon emissions in other sectors that have not implemented the policy can increase. However, green power has almost zero carbon emissions and can promote renewable energy development by offsetting carbon emissions. Thus, we introduce the main Hypotheses 6a and 6b and test them using the fixed-effect model of the synergies between the national ETS policy and the TGC and GPT policies.

Hypothesis 6a: When the national ETS policy (only controlling the power industry) synergizes with the TGC policy, it not only increases carbon emissions and carbon intensity but also increases renewable energy generation and its proportion.

Hypothesis 6b: When the national ETS policy (only controlling the power industry) synergizes with the GPT policy, it not only increases carbon emissions and carbon intensity but also increases renewable energy generation and its proportion.

Synergistic effect of the TGC and GPT policies

The TGC policy is the only credential for the environmental attributes of green power. The GPT policy includes the TGC policy, both of which have the same environmental rights and interests in terms of meeting the requirements of the RECQ, carbon emission dual control assessments, and carbon emissions while accounting for key products. Therefore, the synergistic effects of the TGC and GPT policies on enhancing renewable energy power generation and its proportion are consistent with the effects of implementing the GPT policy alone. Accordingly, we propose Hypothesis 7 and verify it in the fixed effect model of the synergy between the TGC and GPT policies.

Hypothesis 7: The synergy of the TGC and GPT policies can increase the level of renewable energy generation and its proportion, and the effect of the synergy between the two is consistent with that of the GPT policy alone.

Mediating effect of GTFP

Mediating effect of GTFP on the ETS policy

Porter’s hypothesis suggests that environmental regulation can stimulate innovation and improve resource productivity (Porter and van der Linde, 1995). On the basis of this hypothesis, the limitations on energy savings and carbon reduction enhance GTFP, and there is no lag effect in this impact (Mao and Wang, 2022). At the policy level, there are controversial comments on whether the pilot ETS policy can promote carbon reduction by increasing GTFP. Although some studies have shown that the pilot ETS policy promotes regional carbon equality by enhancing the growth of GTFP (Zhang et al. 2021), the pilot ETS policy has a notable adverse effect on both GTFP and enterprise TFP, with GTFP playing a mediating role (Hu and Ding, 2020). Through heterogeneity analysis, the pilot ETS policy is shown to inhibit GTFP in the central region of China (Xu et al. 2022). However, scholars have noted that this policy has an overall facilitating effect on the increase in GTFP (Xu et al. 2022). Therefore, the national ETS policy can trade freely and allocate resources optimally, thereby improving overall GTFP. Accordingly, we propose Hypotheses 8a and 8b and test them in the mediating effect model of the pilot carbon trading policy.

Hypothesis 8a: During the study period, the pilot ETS policy suppresses GTFP, but GTFP can reduce carbon emissions and carbon intensity.

Hypothesis 8b: During the study period, the national ETS policy reduces carbon emissions and carbon intensity by increasing GTFP.

Mediating effect of GTFP on the TGC and GPT policies

Environmental regulations and green technological innovation directly influence energy efficiency and increase GTFP, thereby fostering renewable energy growth (Jiang et al. 2024). The TGC policy can provide additional income to develop new technologies, encourage innovation, improve GTFP (Lin and Zhu, 2019), and benefit the environment (Yan et al. 2020). The GPT policy can promote the development of negative carbon technologies, optimize the power source structure, and enable the power industry to reach its carbon peak before 2029 (Hu et al. 2024). GTFP can increase renewable energy consumption, and its positive impact on the renewable energy transition exceeds that of TFP (Zhang et al. 2024). On this basis, we propose Hypotheses 9a and 9b and verify them in the mediating effect models of TGC and GPT policy.

Hypothesis 9a: The TGC policy promotes renewable energy development by enhancing GTFP.

Hypothesis 9b: The GPT policy promotes renewable energy development by enhancing GTFP.

On the basis of the theoretical analysis of Hypotheses 8 and 9, Hypotheses 10a and 10b are proposed and verified in the mediating effect models of multi-policy synergy.

Hypothesis 10a: The pilot ETS policy, in synergy with the TGC and GPT policies, inhibits GTFP, but GTFP reduces carbon emissions and carbon intensity and promotes renewable energy development.

Hypothesis 10b: The national ETS policy, in synergy with the TGC and GPT policies, reduces carbon emissions and carbon intensity and promotes renewable energy development by increasing GTFP.

Effects of policies on the dual control of energy consumption

Energy consumption intensity control provides important support for the realization of carbon intensity targets. The implementation of the TGC and GPT policies can increase the renewable energy power consumption and is not included in the assessment of total energy consumption. Therefore, the indicator associated with the ETS policy is the energy consumption intensity, and the indicator associated with the TGC and GPT policies is the total energy consumption. We analyse these two indicators theoretically and propose research hypotheses.

Effect of the ETS policy on the energy consumption intensity

Carbon intensity is closely related to the fossil energy consumption intensity (Wang and Dong, 2023) because, in the current energy structure, many carbon emissions are directly sourced from the consumption of fossil energy (Zaman et al. 2016; Raza et al. 2019), and there is a positive correlation between them (Esso and Keho, 2016). This positive correlation implies that as the fossil energy consumption intensity increases, the carbon intensity increases accordingly. However, the ETS policy endows the carbon emission rights with the economic value, forcing enterprises to optimize their energy consumption structures and reduce their dependence on fossil energy when facing carbon emission constraints, thereby reducing fossil energy consumption intensity. On this basis, we propose Hypotheses 11a and 11b and test them in fixed effect models of the impacts of the pilot ETS and national ETS policies on the dual control of energy consumption.

Hypothesis 11a: The pilot ETS can effectively reduce the energy intensity.

Hypothesis 11b: The national ETS can effectively reduce the energy intensity.

Effects of the TGC and GPT policies on the total energy consumption

In actual development, high energy-consuming enterprises can offset their total energy consumption by purchasing TGC to meet the indicators without focusing on measures to actually transform their energy structure, reduce their energy consumption, or reduce their emissions. According to the statistical data, in actual operation, such enterprises only work to meet the indicator requirements but fail to truly optimize and save energy. The total energy consumption indicator deducts the consumption and absorption of renewable energy power, which adds to enterprises’ demand for green power consumption and increases the share of renewable energy in the energy supply (Li and Guo, 2024). Therefore, eliminating control over renewable energy can significantly advance the shift in the dual control of energy consumption (An et al. 2024). On this basis, we introduce the main Hypotheses 12a and 12b and verify them in fixed-effect models of the impacts of TGC and GPT policies on the dual control of energy consumption.

Hypothesis 12a: The TGC policy increases the total fossil energy consumption.

Hypothesis 12b: The GPT reduces the total fossil energy consumption.

In summary, on the basis of the above theoretical analysis and research hypotheses, the research framework of this paper is designed as shown in Fig. 2.

Fig. 2: Research framework.

The figure shows the research idea, research methodology and overall framework. Step 1 is the theoretical analysis and research hypothesis; step 2 is the model construction; step 3 is the model testing and result analysis; step 4 is the policy implications. ***, **, * indicate significant at the 1%, 5% and 10% levels respectively. Source: Decarbonisation policies synergy pathway innovation: Achieving dual carbon goals.

Research design

Model construction

Benchmark regression model

Multiperiod DID model for the pilot ETS

In this paper, the pilot ETS policy is regarded as a quasinatural experiment, and the DID method is utilized to test the impact of the pilot ETS policy on carbon emissions and carbon intensity. Since the policy was implemented in phases in different pilots, we draw on the methodology of previous research(Beck et al. 2010) and adopt the multiperiod DID method, controlling for fixed effects based on the province and year, to test its carbon emission reduction effect and thus Hypothesis 1. In addition, the literature has investigated the carbon emission reduction of the pilot ETS policy by adopting this methodology (Bai et al. 2024; Zhou et al. 2024). Therefore, it is reasonable to choose this method. The specific model settings are as follows:

$${Y}_{it,pilotETS}={\alpha }_{0}+{\alpha }_{1}DI{D}_{pilotETS}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}$$

(1)

The dependent variable ${Y}_{it,pilotETS}$ represents the carbon emissions and carbon intensity of province i in year t when the pilot ETS is implemented. $DI{D}_{it,pilotETS}$ represents the core explanatory variable; that is, the multiperiod DID variable is $DI{D}_{it,pilotETS}=trea{t}_{i,pilotETS}\times pos{t}_{it,pilotETS}$. $trea{t}_{i,pilotETS}$ indicates whether it is a treatment group, and $pos{t}_{it,pilotETS}$ represents the policy implementation time. If the pilot ETS policy significantly reduces the local carbon emissions or carbon intensity, then ${\alpha }_{1}$ is significantly negative. $Contro{l}_{it}$ represents the control variables that may influence carbon reduction in the pilot areas. ${\mu }_{i}$ and${\nu }_{t}$ represent the province fixed effect and the year fixed effect, respectively, and ${\varepsilon }_{it}$ represents the disturbance term.

The value-taking rule¹ for $trea{t}_{i,pilotETS}$ is as follows. When i represents Beijing, Tianjin, Shanghai, Chongqing, Guangdong, Hubei or Fujian, then $trea{t}_{i,pilotETS}=1$; otherwise, $trea{t}_{i,pilotETS}=0$. The seven pilot regions launched ETSs successively in the following order: November 2013 (Beijing), December 2013 (Tianjin, Shanghai, Guangdong), April 2014 (Hubei), June 2014 (Chongqing), and December 2016 (Fujian). Therefore, the value-taking rule for $pos{t}_{it,pilotETS}$ is as follows: when $i$ represents Beijing, Tianjin, Shanghai and Guangdong and $t\ge 2013$, when $i$ represents Chongqing, Hubei and $t\ge 2014$, or when $i$ represents Fujian and $t\ge 2016$, then $pos{t}_{it,pilotETS}=1$; otherwise, $pos{t}_{it,pilotETS}=0$.

Benchmark regression model for the national ETS policy

The national ETS policy is a nationwide policy, and all provinces are in the treatment group after policy implementation. There is no control group, making it unsuitable for the use of a DID model. Therefore, by referring to previous studies (Li et al. 2022; Wei et al. 2023), we build a province fixed effect model to analyse the impacts of the national ETS policy on carbon emissions and carbon intensity and to test Hypothesis 2, as shown in the following equation. In addition, the policy generates full covariance with the year fixed effect. Thus, in the robustness tests in “Endogeneity problems”, we delve deeper into the endogeneity effect caused by not controlling for the year fixed effect.

$${Y}_{it,nationalETS}={\alpha }_{0}+{\alpha }_{1}pos{t}_{nationalETS}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}$$

(2)

${Y}_{it,nationalETS}$ represents the carbon emissions and carbon intensity of province $i$ in year $t$ when the national ETS was implemented. $pos{t}_{nationalETS}$ represents the core explanatory variable and indicates the national ETS policy implementation time. When $t\ge 2021$, $pos{t}_{it,nationalETS}=1$; otherwise, $pos{t}_{it,nationalETS}=0$. If the national ETS significantly reduces local carbon emissions or carbon intensity, then ${\alpha }_{1}$ is significantly negative.

Benchmark regression model for the TGC policy

The TGC policy is treated in the same way as the national ETS policy; thus, the information will not be repeated. We construct a province fixed effect model to analyse the impacts of the TGC policy on renewable energy generation and its proportion and to test Hypothesis 3, as shown in the following equation:

$${Y}_{it,TGC}={\alpha }_{0}+{\alpha }_{1}pos{t}_{TGC}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}$$

(3)

${Y}_{it,TGC}$ represents the power generation volume of renewable energy and its proportion of province $i$ in year $t$ when the TGC policy is implemented. $pos{t}_{TGC}$ represents the core explanatory variable and indicates the TGC policy implementation time. When $t\ge 2017$, $pos{t}_{it,TGC}=1$; otherwise, $pos{t}_{it,TGC}=0$. If the TGC policy significantly increases the local power generation volume of renewable energy and its proportion, then ${\alpha }_{1}$ is significantly positive.

Benchmark regression model for the GPT policy

The GPT policy is treated in the same way as the national carbon trading policy; thus, the information will not be repeated. We construct a province fixed effect model to analyse the impacts of the GPT policy on renewable energy generation and its proportion and to test Hypothesis 4, as shown in the following equation:

$${Y}_{it,GPT}={\alpha }_{0}+{\alpha }_{1}pos{t}_{GPT}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}$$

(4)

${Y}_{it,GPT}$ represents the power generation volume of renewable energy and its proportion of province $i$ in year $t$ when the GPT policy is implemented. $pos{t}_{it,GPT}$ represents the core explanatory variable and indicates the GPT policy implementation time. When $t\ge 2021$, $pos{t}_{it,GPT}=1$; otherwise, $pos{t}_{it,GPT}=0$. If the GPT policy significantly increases the local power generation volume of renewable energy and its proportion, then ${\alpha }_{1}$ is significantly positive.

Notably, since $pos{t}_{nationalETS}$, $pos{t}_{TGC}$ and $pos{t}_{GPT}$ generate perfect collinearity with the year fixed effect, the year fixed effect is not controlled in the baseline regression model. However, if the year dummy variables are not controlled, some important unobservable factors may be omitted. For this reason, the endogenous impact caused by the lack of control over the year fixed effect will be discussed in depth later in the text.

Multipolicy synergy model

To further test whether there is synergy among the pilot ETS, national ETS, TGC and GPT policies, we refer to existing studies (Wu et al. 2021) to construct a multipolicy interaction model to analyse the carbon emission reduction and renewable energy development effects of multiple policy interactions and to test Hypotheses 5–7, as shown in the following equation:

$$\left\{\begin{array}{ll}{Y}_{it,pilotETS,TGC}={\alpha }_{0}+{\alpha }_{1}DI{D}_{it,pilotETS}\times pos{t}_{it,TGC}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {Y}_{it,nationalETS,TGC}={\alpha }_{0}+{\alpha }_{1}pos{t}_{it,nationalETS}\times pos{t}_{it,TGC}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {Y}_{it,pilotETS,GPT}={\alpha }_{0}+{\alpha }_{1}DI{D}_{it,pilotETS}\times pos{t}_{it,GPT}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {Y}_{it,nationalETS,GPT}={\alpha }_{0}+{\alpha }_{1}pos{t}_{it,nationalETS}\times pos{t}_{it,GPT}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {Y}_{it,TGC,GPT}={\alpha }_{0}+{\alpha }_{1}pos{t}_{it,TGC}\times pos{t}_{it,GPT}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {Y}_{it,pilotETS,TGC,GPT}={\alpha }_{0}+{\alpha }_{1}DI{D}_{it,pilotETS}\times pos{t}_{it,TGC}\times pos{t}_{it,GPT}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {Y}_{it,nationalETS,TGC,GPT}={\alpha }_{0}+{\alpha }_{1}pos{t}_{it,nationalETS}\times pos{t}_{it,TGC}\times pos{t}_{it,GPT}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\end{array}\right.$$

(5)

$DI{D}_{it,pilotETS}\times pos{t}_{it,TGC}$, $pos{t}_{it,nationalETS}\times pos{t}_{it,TGC}$, $DI{D}_{it,pilotETS}\times pos{t}_{it,GPT}$, $pos{t}_{it,nationalETS}\times pos{t}_{it,GPT}$, $pos{t}_{it,TGC}\times pos{t}_{it,GPT}$, $DI{D}_{it,pilotETS}\times pos{t}_{it,TGC}\times pos{t}_{it,GPT}$ and $pos{t}_{it,nationalETS}\times pos{t}_{it,TGC}\times pos{t}_{it,GPT}$ represent the dummy variables indicating whether the pilot ETS and TGC policies are synergistic, whether the national ETS and TGC policies are synergistic, whether the pilot ETS and GPT policies are synergistic, whether the national ETS and GPT policies are synergistic, whether the TGC and GPT policies are synergistic, whether the pilot ETS, TGC, and GPT policies are synergistic, and whether the national ETS, TGC, and GPT polices are synergistic, respectively.

Mediating effect model

To further study the transmission paths of pilot ETS and national ETS policies on carbon emissions and carbon intensity and the transmission paths of the TGC and GPT policies on renewable energy generation and its proportion, we refer to a previous study (Baron and Kenny, 1986) and construct mediating effect models to test the GTFP transmission effects between the ETS policy and carbon emission reduction level and between the TGC and GPT policies and the renewable energy development level; in addition, we test Hypotheses 8–10.

The mediating effect model for individual policies is as follows:

$$\left\{\begin{array}{ll}{M}_{it}={\beta }_{0}+{\beta }_{1}DI{D}_{it,pilotETS}+{\beta }_{2}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {Y}_{it,pilotETS}={\gamma }_{0}+{\gamma }_{1}DI{D}_{it,pilotETS}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {M}_{it}={\beta }_{0}+{\beta }_{1}pos{t}_{it,nationalETS}+{\beta }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {Y}_{it,nationalETS}={\gamma }_{0}+{\gamma }_{1}pos{t}_{it,nationalETS}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {M}_{it}={\beta }_{0}+{\beta }_{1}pos{t}_{it,TGC}+{\beta }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {Y}_{it,TGC}={\gamma }_{0}+{\gamma }_{1}pos{t}_{it,TGC}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {M}_{it}={\beta }_{0}+{\beta }_{1}pos{t}_{it,GPT}+{\beta }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {Y}_{it,GPT}={\gamma }_{0}+{\gamma }_{1}pos{t}_{it,GPT}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\end{array}\right.$$

(6)

The mediating effect model for the synergy of multiple policies is as follows:

$$\left\{\begin{array}{ll}{M}_{it}={\beta }_{0}+{\beta }_{1}DI{D}_{it,pilotETS}\times pos{t}_{it,TGC}+{\beta }_{3}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {Y}_{it,pilotETS,TGC}={\gamma }_{0}+{\gamma }_{1}DI{D}_{it,pilotETS}\times pos{t}_{it,TGC}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {M}_{it}={\beta }_{0}+{\beta }_{1}pos{t}_{it,nationalETS}\times pos{t}_{it,TGC}+{\beta }_{3}Contro{l}_{it}+{\mu }_{it}+{\varepsilon }_{it}\\ {Y}_{it,nationalETS,TGC}={\gamma }_{0}+{\gamma }_{1}pos{t}_{it,nationalETS}\times pos{t}_{it,TGC}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {M}_{it}={\beta }_{0}+{\beta }_{1}DI{D}_{it,pilotETS}\times pos{t}_{it,GPT}+{\beta }_{3}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {Y}_{it,pilotETS,GPT}={\gamma }_{0}+{\gamma }_{1}DI{D}_{it,pilotETS}\times pos{t}_{it,GPT}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {M}_{it}={\beta }_{0}+{\beta }_{1}pos{t}_{it,nationalETS}\times pos{t}_{it,GPT}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{it}+{\varepsilon }_{it}\\{Y}_{it,nationalETS,GPT}={\gamma }_{0}+{\gamma }_{1}pos{t}_{it,nationalETS}\times pos{t}_{it,GPT}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {M}_{it}={\beta }_{0}+{\beta }_{1}pos{t}_{it,TGC}\times pos{t}_{it,GPT}+{\beta }_{2}Contro{l}_{it}+{\mu }_{it}+{\varepsilon }_{it}\\ {Y}_{it,TGC,GPT}={\gamma }_{0}+{\gamma }_{1}pos{t}_{it,TGC}\times pos{t}_{it,GPT}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {M}_{it}={\beta }_{0}+{\beta }_{1}DI{D}_{it,pilotETS}\times pos{t}_{it,TGC}\times pos{t}_{it,GPT}+{\beta }_{2}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {Y}_{it,pilotETS,TGC,GPT}={\gamma }_{0}+{\gamma }_{1}DI{D}_{it,pilotETS}\times pos{t}_{it,TGC}\times pos{t}_{it,GPT}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {M}_{it}={\beta }_{0}+{\beta }_{1}pos{t}_{it,nationalETS}\times pos{t}_{it,TGC}\times pos{t}_{it,GPT}+{\beta }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {Y}_{it,nationalETS,TGC,GPT}={\gamma }_{0}+{\gamma }_{1}pos{t}_{it,nationalETS}\times pos{t}_{it,TGC}\times pos{t}_{it,GPT}+{\gamma }_{2}{M}_{it}+{\gamma }_{3}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\end{array}\right.$$

(7)

Where ${M}_{it}$ is the mediating variable, ${\alpha }_{1}$ is the total effect of the policy, ${\gamma }_{1}$ is the direct effect, and ${\beta }_{1}$ ${\gamma }_{2}$ are the mediating effects of the mediating variable. If ${\alpha }_{1}$ is not significantly equal to 0, ${\beta }_{1}$ is not significantly equal to 0. If ${\gamma }_{2}$ is not significantly equal to 0, then the mediating effect is significant. If ${\gamma }_{1}$ is not significantly equal to 0, it is defined as a partial mediating effect; otherwise, it is referred to as a complete mediating effect. We use the Sobel–Goodman method to test the significance of the mediating effect (Li et al. 2023).

Fixed effect model for the dual control of energy consumption

To further investigate the impacts of the pilot ETS and national ETS policies on fossil energy intensity and of the TGC and GPT policies on total fossil energy consumption, with reference to (Li et al. 2022), we construct fixed effect models to test research Hypotheses 11 and 12, as shown in the following equation:

$$\left\{\begin{array}{ll}{Y}_{it,FFCI}={\alpha }_{0}+{\alpha }_{1}DI{D}_{it,pilotCET}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {Y}_{it,FFCI}={\alpha }_{0}+{\alpha }_{1}pos{t}_{it,nationalCET}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\varepsilon }_{it}\\ {Y}_{it,\mathrm{ln}FFC}={\alpha }_{0}+{\alpha }_{1}pos{t}_{it,TGC}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\\ {Y}_{it,\mathrm{ln}FFC}={\alpha }_{0}+{\alpha }_{1}pos{t}_{it,GPT}+{\alpha }_{2}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}\end{array}\right.$$

(8)

Where ${Y}_{it,FFCI}$ and ${Y}_{it,\mathrm{ln}FFC}$ denote fossil energy intensity and total fossil energy consumption, respectively.

Data sources and sample description

On the basis of the availability of data, we select panel data from 30 provinces in China (excluding Tibet, Hong Kong, Macao and Taiwan) from 2010 to 2023 to analyse the carbon emission reduction effect and mechanism of the ETS policy and the renewable energy development effects and mechanisms of the TGC and GPT policies. We further analyse the effects of the three policies on the comprehensive transition from the dual control of energy consumption to that of carbon emissions. To eliminate the influence of variable magnitude, the method for taking the natural logarithm is used for variables with large absolute values. With reference to existing studies (Noor et al. 2014), missing data are filled in using linear interpolation. In addition, to ensure that the amount of data in the sample can support the subsequent analysis of the mechanisms by which the policies impact carbon emissions, carbon intensity, renewable power generation and its proportion, we no longer match the propensity scores of the treatment group with those of the control group (Li et al. 2023). For the treatment of data outliers, 1% of the samples of carbon emissions, carbon intensity, and other variables are subjected to a shrinking tail treatment in the other robustness test of “Other robustness tests“. The abbreviations and brief overview of the dataset can be found in Tables A1 and A2 in Appendix A.

Variable setting and descriptive statistics

Explanatory variables

The dependent variables for the pilot ETS are the provincial carbon emissions (CE) and carbon intensity (CI), respectively, in the form of logarithms (lnCEa and lnCIa). Since the pilot ETS covers industries such as electricity, heat, chemical engineering, and cement, the carbon emissions are taken as the total regional carbon emissions. After the launch of the national ETS policy in 2021, the carbon emissions within the pilot areas were considered the emissions after those of the power industry were deducted. The dependent variables for the national ETS are provincial carbon emissions and carbon intensity, which are logarithms (lnCEb and lnCIb). As the national ETS policy encompasses only the power industry, the carbon emissions are sourced from the power industry. The carbon intensity is calculated on the basis of the actual regional gross domestic product (GDP) at constant prices in 2010 and the actual carbon emissions.

For the TGC and GPT policies, the dependent variables are the provincial-level renewable energy power generation (REG) and the proportion of renewable energy generation (PREG), respectively. Both the REG and PREG are presented in logarithmic form, namely, lnREG and lnPREG. In the initial stage, the issuance of TGC and GPT are only targeted at wind and photovoltaic power generation. Hence, the renewable energy power generation discussed within this work is denoted as wind power and photovoltaic power generation².

The explained variables for the dual control of energy consumption are the total provincial fossil energy consumption and the fossil energy intensity, where the fossil energy intensity is the share of the fossil energy consumption in the GDP (at constant 2010 prices).

The core explanatory variables are whether to launch the pilot ETS ($DI{D}_{pilotETS}$), the national ETS ($pos{t}_{nationalETS}$), the TGC policy ($pos{t}_{TGC}$), the GPT policy ($pos{t}_{GPT}$), whether the pilot ETS and TGC policies are synergistic ($DI{D}_{pilotETS}\times pos{t}_{TGC}$), whether the national ETS and TGC policies are synergistic ($pos{t}_{nationalETS}\times pos{t}_{TGC}$), whether the pilot ETS and GPT policies are synergistic ($DI{D}_{pilotETS}\times pos{t}_{GPT}$), whether the national ETS and GPT policies are synergistic ($pos{t}_{nationalETS}\times pos{t}_{GPT}$), whether the GPT and TGC policies are synergistic ($pos{t}_{TGC}\times pos{t}_{GPT}$), whether the pilot ETS, TGC, and GPT policies are synergistic ($DI{D}_{pilotETS}\times pos{t}_{TGC}\times pos{t}_{GPT}$), and whether the national ETS, TGC, and GPT policies are synergistic ($pos{t}_{nationalETS}\times pos{t}_{TGC}\times pos{t}_{GPT}$).

Control variables

For comparison of the carbon emissions, carbon intensity, renewable energy power generation and its proportion in the treatment group and the control group, we select the following control variables:

The economic development level, i.e., the logarithm of the actual GDP per capita (lnPGDP), is calculated on the basis of the constant price in 2010. The urbanization level is the logarithm of the urbanization rate (lnUR). The intensity of innovation is the logarithm of the total number of valid invention patents (lnRD). The degree of economic concentration is represented by the population density (PD) and the aggregate population at year-end (lnPS). The fiscal support level is the proportion of local general budgetary expenditures to the regional GDP (FS). The industrial structure (is) is expressed as the proportion of the added value of the secondary industry. The advanced industrial structure (ais) is expressed as the ratio of the added value of the tertiary industry to that of the secondary industry. The high industrial structure (his) is expressed as the proportion of the primary industry *1 + the proportion of the secondary industry *2 + the proportion of the tertiary industry *3.

Intermediary variables

The mediator variable is selected as GTFP, and its calculation formula and input and output indicators are shown in Appendix B.

Descriptive Statistic

The data cover the time period from 2010 to 2023, and the research objects are 30 provinces in China, with a total of 420 sample data points. The descriptive statistics of each variable are presented in Table 1.

Table 1 Descriptive statistics of each variable.

Diagnostic tests

Drawing on the literature(Hausman, 1978; Zhao et al. 2023), we clarify the exact form of the model through diagnostic tests, the results of which are shown in Table 2. First, the F test is used, and the results show that the fixed effect model is superior to the ordinary least squares (OLS) model. Next, the Hausman test is used, and the results show that the fixed effect (FE) model is superior to the random effect (RE) model. This finding proves the correctness of the choice of model in this paper.

Table 2 Diagnostic test results.

In summary, we solve the multiperiod DID model of the pilot ETS policy and the fixed effect model of the national ETS, TGC and GPT with the help of StataMP 17 software.

Results

Benchmark regression results and analysis

Regression results and analysis of ETS

Analysis of the ETS regression results

Table 3 presents the average treatment effects of the pilot ETS and national ETS on carbon emissions and carbon intensity, respectively. For the pilot ETS, models (1)–(2) are the estimated results when the explained variable is lnCEa, and models (3)–(4) are the estimated results when the explained variable is lnCIa. By focusing on the coefficients of $DI{D}_{pilotETS}$, we find that when adding control variables and replacing the explained variables with the carbon intensity, the coefficients are always significantly negative. Overall, after the pilot ETS is implemented, the carbon emissions of the provinces in the treatment group decrease by 15.06%, and the carbon intensity decreases by 16.60% relative to that of the provinces in the control group. The pilot ETS has a significant inhibitory effect on carbon emissions and carbon intensity. Therefore, the implementation of a pilot ETS can effectively control carbon emissions. Thus, Hypothesis 1 is proved.

Table 3 Impacts of the ETS on CE and CI.

For the national ETS policy, models (5)–(6) and (7)–(8) are the estimated results, with the explained variables being lnCEb and lnCEb, respectively. The regression results indicate that, whether the control variables are added or the explained variable is replaced with the carbon intensity, the coefficient of $pos{t}_{nationalETS}$ is significantly positive. A possible reason for this phenomenon is that although the national ETS market represents a nationwide fusion of carbon quotas for the power sector, the initial quotas are allocated uniformly by the Ministry of Ecology and Environment of China, leading to significant differences in quota scarcity between regions. Such differences have prompted high-energy-consuming firms to shift their production capacity to quota-relaxed regions, resulting in carbon leakage. Therefore, Hypothesis 2 is verified.

Regression results and analysis of TGC

Table 4 presents the average treatment effects of the TGC policy on renewable energy generation and its proportion. After adding control variables and replacing the explained variable with the proportion of renewable energy power generation, the coefficient of $pos{t}_{TGC}$ remains significantly positive. Overall, after the TGC policy is implemented, renewable energy generation increases by 26.40%, and the proportion of renewable energy power generation increases by 11.74%. Therefore, the TGC policy effectively promotes renewable energy generation and its proportion, and Hypothesis 3 is verified.

Table 4 Effects of TGC on REG and PREG.

Regression results and analysis of the GPT

Table 5 shows the average treatment effects of the GPT policy on the renewable energy power generation and proportion. The regression results indicate that when the control variables are added or the explained variable is replaced with the proportion of renewable energy power generation, the coefficient of $pos{t}_{GPT}$ is significantly positive, and the renewable energy generation and its proportion increase by 44.71% and 21.15%, respectively. The results indicate that the GPT policy can promote the development of renewable energy. Compared with the TGC policy, the GPT policy better facilitates the development of renewable energy. Therefore, Hypothesis 4 is verified.

Table 5 Impacts of GPT on REG and PREG.

In summary, in terms of the dual control of carbon emissions, the pilot ETS policy has been implemented for a longer period and covers a wider range of industries; thus, it has significant advantages and can effectively reduce carbon emissions and carbon intensity. During the study period, the national ETS policy only covers the power industry, and if the national ETS policy significantly achieves the dual control of carbon emissions, it is imperative to expand the coverage of the industry. From the perspective of renewable energy development, the implementation of TGC and GPT policies facilitates the development of renewable energy, and the GPT policy is relatively superior. This result arises because the GPT policy involves the actual physical delivery of electricity, and users who buy green power directly consume the electricity generated by renewable energy. This policy has a more direct effect on reducing carbon emissions and promoting renewable energy.

Parallel trend test

We perform parallel trend tests for the pilot ETS policy with lnCEa and lnCIa as the explained variables, for the national ETS with lnCEb and lnCIb as the explained variables, and for the TGC and GPT policies with lnREG and lnPREG as the explained variables. A dynamic changing trend model with fixed effects is constructed to observe whether the ETS, TGC, and GPT policies conform to the parallel trend assumption. The specific model settings are as follows:

$${Y}_{it}={\alpha }_{0}+\mathop{\sum}\limits_{j}{\alpha }_{j}x{h}_{j}+{\alpha }_{k}Contro{l}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}$$

(9)

Where ${Y}_{it}$ is the explained variable, $x{h}_{j}(j < 0)$ is the period before the policy implementation, $x{h}_{j}(j=0)$ is the current period of the policy, and $x{h}_{j}(j > 0)$ is the period after the policy implementation.

The results of the pilot ETS illustrate that before implementing the pilot ETS policy, the coefficients of the dummy variables are not significant. After implementing the pilot ETS policy, the regression coefficients of the explained variables of lnCEa and lnCIa are significantly negative in the first year. The results of the national ETS show that the regression coefficients of lnCEb and lnCIb as the explained variables are significant in the second year after implementation. The results of the TGC and GPT policies show that with lnREG and lnPREG as the explained variables, the regression coefficients are significantly positive after the implementation of the policies.

In summary, the changes in the pilot ETS, national ETS, TGC, and GPT policies satisfy the parallel trend assumption (Fig. C1–C4 in Appendix C).

Robustness test

Endogeneity problems

Problem of omitted variables

In testing the effects of the national ETS, TGC and GPT policies, the year fixed effects are not controlled because of the multiple covariance of the year dummy variables with carbon emissions, carbon emissions intensity, renewable energy generation, and renewable energy share. Although we add many macro variables, such as the economic level, industrial structure, innovation degree and fiscal dependence, the control variables may still be unable to eliminate the impacts of time effects, resulting in omitted variables. On this basis, we further mitigate the endogeneity problem by controlling for other policy uncertainties, the time-varying features of the region, and the correlation of variables in the time dimension.

When testing the effects of the national ETS, TGC and GPT policies, since the year dummy variables have multicollinearity with carbon emissions, carbon intensity, renewable energy power generation, and renewable energy proportion, the year fixed effects are not controlled for in the benchmark regression. Although many macro variables, such as economic level, industrial structure, innovation degree, and fiscal dependence, are controlled for, the influence of time effects on the regression results may still be unable to be removed, resulting in omitted variable problems. On this basis, this paper further alleviates the endogeneity problem from the aspects of controlling other policy uncertainties, the time-varying characteristics of regions, and variable correlations in the time dimension.

First, other policy uncertainties should be controlled. Before and after the implementation of other emission reduction and carbon reduction policies, there was often an increase in the uncertainties of ETS, TGC and GPT policies. In September 2016, China released pilot projects on energy use rights in Zhejiang, Henan, Fujian and Sichuan. To exclude the interference of the policy, this paper measures the policy uncertainty of the energy use rights policy and puts the dummy variable $DI{D}_{energy}$, which indicates whether different provinces implement energy use rights, into the regression model as a control variable.

Second, the time-varying characteristics of regions are considered. Different regions have different resource endowments, and dual control of carbon emissions and renewable energy advancement may also lead to deviations in the regression results. After considering the macro factors of each province, we need to separate the factors influencing the dual control of carbon emissions and renewable energy development in the region as much as possible. We incorporate the time-varying features of provinces in the form of fixed effects in the benchmark model.

Third, considering the correlation of variables in the time dimension, with reference to the research method (Nguyen and Phan, 2017), the standard deviation of the regression results is adjusted by clustering in the time dimension to obtain more reliable regression results.

The regression results after adding the above three omitted variables (Table D1 in Appendix D) illustrate that the coefficients of carbon emissions, carbon intensity, and renewable energy power generation and its proportion are still significant, and the signs are the same as those of the benchmark regression results, proving that the results of this study are robust.

Instrumental variable analysis

Endogeneity is a problem that cannot be ignored. Specifically, the endogeneity problem arises from the reverse causal relationship between the dependent and independent variables. Instrumental variable tests can alleviate the aforementioned omitted variable problem to some extent and address the reverse causality issue. Therefore, we adopt the instrumental variable method to verify the robustness of the results (Hering and Poncet, 2014). Under the dual carbon goals, an ETS requires emission reduction and temperature control, and renewable energy generation is affected by temperature. Therefore, the annual average temperature of Chinese provinces is chosen as the instrumental variable. These data are from the National Centres for Environmental Information (NCEI) of the National Oceanic and Atmospheric Administration (NOAA) of the United States. The empirical results (Table D2 in Appendix D) show that there is no weak instrumental variable problem in Columns (1)–(4), indicating that temperature is a reasonable instrumental variable. After the instrumental variable is used, the main regression results remain stable.

Moreover, the Hausman test demonstrates that there is no endogeneity problem for the ETS pilot, TGC and GPT policies when the share of renewable energy generation is the explained variable. Therefore, no instrumental variable test is performed for them.

Placebo test

To strengthen the credibility of the regression results, we follow existing studies and conduct placebo tests (Topalova, 2010; Li et al. 2023). The samples and policy implementation times are randomly selected from the treatment group provinces, and 1,000 benchmark regression calculations are performed. The test results for the random policy shocks of the pilot ETS, national ETS, TGC, and GPT policies (Fig. D1–D3 in Appendix D) indicate that the p values are mostly greater than 0.1. Thus, it is verified that the impacts of the random policy shocks of these policies are not significant. The previous research results are reliable and do not result from other random factors.

Dual machine learning model test

Traditional multiple linear regression models have limitations in modelling and covariate selection. If the relationship between variables is nonlinear, the estimation is biased, and excess covariates cause problems such as multicollinearity and dimensionality. In the current context of big data, dual machine learning can effectively compensate for the shortcomings of traditional research in variable selection and model estimation (Chernozhukov et al. 2018; Yang et al. 2020). In addition, current empirical application results provide strong support for the feasibility of dual machine learning (Wang et al. 2022; Duhirwe et al. 2024).

Therefore, we adopt a dual machine learning model to reestimate the impacts of the pilot ETS and national ETS policies on carbon emissions and carbon intensity and of the TGC and GPT policies on renewable energy power generation and its proportion. With the help of Python 3.13 and StataMP 17 software, the random forest algorithm is used to solve the prediction problem of the benchmark model with a sample split ratio of 1:4. The regression results are shown in Table D3 in Appendix D. The pilot ETS policy can still reduce carbon emissions and carbon intensity. The national ETS policy significantly contributes to the increase in carbon intensity, but the effect on carbon emissions is not significant. Both the TGC and GPT policies significantly promote renewable energy generation and its proportion. This finding suggests that the results of the test remain robust after the model is replaced.

Regional heterogeneity analysis

The results of the regional heterogeneity analysis are shown in Tables D4–D7 in Appendix D. The regression results show that the pilot ETS policy can significantly reduce carbon emissions and carbon intensity in the eastern, central and western regions, which is consistent with the baseline results of the pilot carbon trading policy. The national ETS policy significantly increases the carbon emissions and carbon intensity in the eastern, western and northeastern regions, whereas the effect is not significant in the central region. This finding is consistent with the benchmark regression results of the national ETS policy. The TGC policy significantly enhances renewable energy power generation in the eastern and central regions, whereas for the western and northeastern regions, the effect of the TGC policy on both renewable energy power generation and its proportion is not significant. This finding indicates an imbalance in the distribution of renewable power consumption weights, i.e., the east is the load centre, but the assessment pressure is insufficient, resulting in the western TGC not being able to circulate effectively. The GPT policy significantly increases renewable power generation and its proportion in the eastern, central, western and northeastern regions, which is consistent with the baseline regression of the GPT policy and proves the stability of the study results.

In summary, the response to policies varies among regions due to differences in their economic development levels and resource endowments, among other factors. Therefore, it is necessary to build a regional synergistic system, and the development of additional targeted policy implementation programs is considered key to achieving the dual carbon goal.

Other robustness tests

Eliminating the influences of some specific samples

We further examine the robustness of the benchmark regression results while excluding specific samples. In addition to implementing the ETS policy, in the 12th Five-Year Plan period, Beijing, Shanghai and Guangdong might have adopted policies such as industrial structure adjustment, energy structure adjustment, energy savings and emission reduction. These additional policies can interfere with the accurate identification of ETS, TGC and GPT policies. Among the pilot areas, Chongqing is the sole city located in the western region. The unique characteristics of economic development in the western region may affect the results of baseline regression. Owing to the distinctiveness of geographical location, Inner Mongolia is divided into eastern Inner Mongolia and western Inner Mongolia, and Hebei is divided into Hebei proper and northern Hebei. Such divisions might influence the benchmark regression results. The regression results (Table D8 in Appendix D) indicate that after excluding some of the above specific samples, the regression coefficients remain significant and are in line with the benchmark regression findings. Consequently, this result validates the robustness of the benchmark regression results presented in this paper.

Excluding the effects of outliers

To mitigate the outliers’ impact, we first apply a 1% indentation to the explained variables, including the carbon emissions, carbon intensity, renewable energy power generation, and renewable energy proportion. Then, a benchmark regression analysis is conducted. The robustness test results (Table D9 in Appendix D) demonstrate that after indentation, the regression results remain unchanged.

Excluding unexpected events

During the sample period from 2010 to 2023 selected for this work, the COVID-19 pandemic occurred. Hence, to eliminate the impact caused by the pandemic, we exclude the relevant data from 2020 to 2022. The regression coefficients obtained are significant, and their direction is in line with that of the benchmark regression results (Table D10 in Appendix D).

Analysis of the effects of multiple policy synergies

ETS–TGC synergy

Table 6 presents the average treatment effects resulting from the synergy among the pilot ETS, national ETS, and TGC policies regarding carbon emissions, carbon intensity, renewable energy generation, and renewable energy proportion in each province across China.

Table 6 Effects of ETS–TGC synergy on the CE, CI, REG and PREG.

Regarding the regression results of the synergy between the pilot ETS and TGC policies, when there is the synergy between these two policies, the carbon emissions and carbon intensity decrease by 22.74% and 23.60%, respectively. Compared with the situation in which the pilot ETS policy is implemented independently, the synergy with the TGC policy can strengthen the carbon reduction effect. Nevertheless, the regression results concerning renewable energy generation and its proportion are significantly negative. The probable reason lies in the fact that the measure of utilizing the TGC policy to offset carbon emissions has not been put into practice in all pilot projects. Presently, only Tianjin has explicitly stated that the TGC policy can be employed to offset carbon emissions. Consequently, the development space for TGCs within the pilot ETS framework is rather limited, which in turn impacts the enthusiasm for investment in renewable energy power generators. This conclusion aligns with existing research (Lin and Jia, 2020), thereby verifying Hypothesis 5a.

According to the regression results of the synergy between the national ETS and TGC policies, the carbon emissions and carbon intensity increase by 19.70% and 19.67%, respectively. Compared with the implementation of the national ETS alone, the regression results do not change. In the national ETS market, the utilization of TGCs to offset carbon emissions is not permitted, and this regression result attests to the rationality of the actual measures in place. Renewable energy power generation and its proportion increase by 44.71% and 21.15%, respectively, which is in line with the outcomes obtained from the implementation of the GPT policy alone. This finding suggests that the synergy between the national ETS policy and the TGC policy, although not effective in reducing carbon emissions and carbon intensity, can increase renewable power generation and its proportion and promote the development of renewable energy. A possible reason is that the national ETS market increases the cost of power generation in the thermal power sector by tightening the carbon emission quotas, prompting power companies to shift to renewable energy investments. Moreover, TGC provides a channel for renewable energy projects to realize environmental premiums, which directly increase the income of power generators. This two-way regulation of ETS costs and TGC revenues strengthens the incentive for companies to invest in renewable energy projects. Therefore, this finding verifies Hypothesis 6a.

ETS–GPT synergy

Table 7 presents the regression results of the synergistic effects of the pilot ETS, national ETS and GPT policies on carbon emissions, carbon intensity, renewable energy power generation, and renewable energy proportion in each province of China. The regression results of the synergy between the pilot ETS and GPT indicate that after synergy is established, carbon emissions and carbon intensity decrease by 56.54% and 56.63%, respectively. Compared with the regression results of the pilot ETS alone and the synergy between the pilot ETS and TGC policies, the synergy between the pilot ETS and GPT policies can enhance the emission reduction effect. Although it restrains renewable energy development, the degree of restraint is weakened. A possible reason is that the measure of using green power to offset carbon emissions has been gradually implemented in various pilot projects. The use of green power reduces carbon emissions, but the proportion of carbon emissions offset by green power is limited, which still affects the development potential of renewable energy to some extent. Therefore, Hypothesis 5b is confirmed.

Table 7 Effects of ETS–GPT synergy on CE, CI, REG and PREG.

For the synergy between the national ETS and GPT policies, compared with the results of implementing the national ETS policy alone and those of the synergy between the national ETS and TGC policies, no changes in carbon emissions and carbon intensity occur. When the explained variables are the renewable energy generation and the renewable energy proportion, the regression result is consistent with that of the implementation of the GPT policy alone. This finding indicates that there is currently a lack of connection between green power and the national ETS policy. Therefore, Hypothesis 6b is verified.

TGC‒GPT synergy

In terms of REG and PREG, the regression results of the synergy between the TGC and GPT policies are in line with those of implementing the GPT policy alone (Table E1 in Appendix E), thus validating Hypothesis 7. This result shows better results than the implementation of the TGC alone because the GPT is conducted in the form of a bundle of electricity and certificates, through which the dual value of green power is directly reflected, not only in the form of its value as electricity energy but also in the environmental protection benefits it brings. The benefits are reflected in the price of the TGC.

ETS–TGC–GPT synergy

The regression results of the synergies among ETS, TGC and GPT policies regarding carbon emissions, carbon intensity, renewable energy power generation, and renewable energy proportion indicate that the effect of the three-synergy system is consistent with that of the ETS–GPT synergy system (Table E2 in Appendix E). The likely reason for this phenomenon is that ETS and GPT policies play more direct roles in promoting the transformation of mixed forms of energy. The ETS policy incentivizes firms to reduce carbon emissions and switch to relatively clean energy sources by reducing free carbon allowances. GPT encourages more investment in renewable energy by increasing the demand for green power through market mechanisms. In contrast, while the TGC policy promotes renewable energy development, its main function is to certify and track the environmental attributes of green power, and its impact on the actual emission reduction effect is relatively indirect.

In summary, according to the regression results of synergistic effects between the pilot ETS, national ETS, TGC and GPT policies, we find that the synergistic carbon emission reduction effect of the pilot ETS and GPT policies is optimal and that the synergistic effect of the national ETS policy with the TGC or GPT policy is better able to promote renewable energy development; however, there is a phenomenon of policy redundancy in synergistic TGC and GPT policies. Therefore, it is necessary to implement a differentiated synergy strategy for policies.

Analysis of the mediating effect of GTFP

Analysis of the mediating effect of GTFP on individual policies

Analysis of the mediating effect of GTFP on the ETS policy

The specific results are presented in Table 8. For the pilot ETS policy, in model (2), by taking GTFP as the explained variable, the coefficient value of $DI{D}_{pilotETS}$ is −0.0834, which passes the significance test. Specifically, when other factors remain unchanged, in contrast with that in the control group, the GTFP in the treatment group decreases by 8.34%. Although the pilot ETS policy is aimed at promoting green technology innovation among emission control enterprises through the offering of economic incentives, in long-term practice, emission control enterprises may be cautious due to concerns about the hazards of green technology innovation, thereby causing the speed and effectiveness of green technology innovation to fall short of expectations and limiting the improvement in the GTFP. When the mediating variable GTFP is added to models (3) and (5), according to the Sobel test results, the intermediary effects with respect to carbon emissions and carbon intensity are significant. The coefficients of GTFP are −0.3268 and −0.3325, respectively. Specifically, as the GTFP increases by 1%, the carbon emissions and carbon intensity decrease by 32.68% and 33.25%, respectively, on average. The GTFP can significantly reduce carbon emissions and intensity. Additionally, the direct effect is significantly negative; thus, the mediating effects of GTFP on decreasing carbon emissions and carbon intensity are only partial. The p value of the Sobel test is less than 0.1, and the effect of GTFP as the mediating variable is significant, and the proportions of the mediating effect are −18.11% and −16.71% for carbon emissions and carbon intensity, respectively. Therefore, Hypothesis 8a is verified; that is, during the study period, the pilot ETS policy suppresses the GTFP, but the GTFP reduces the carbon emissions and carbon intensity. A likely reason for this phenomenon is that energy-intensive firms need to purchase additional quotas when they face a shortage of quotas. This cost pressure forces firms to compress their R&D investment, which inhibits energy utilization and green technological innovation and thus negatively affects the increase in the GTFP.

Table 8 Results of the mediating effect of GTFP on the ETS policy.

From the perspective of the national ETS, in model (7), when the GTFP serves as the explained variable, the coefficient value of $pos{t}_{nationalETS}$ is significant and equals 0.0508. Specifically, if other factors remain unchanged, after the launch of the national ETS, the GTFP increases by 5.08%. When the mediating variable GTFP is added to models (8) and (10), the mediating effects when the carbon emissions and carbon intensity are the dependent variables are significant. The coefficients of GTFP are −0.2635 and −0.2692 for the carbon emissions and carbon intensity, respectively. Specifically, when the GTFP increases by 1%, carbon emissions and carbon intensity decrease by 26.35% and 26.92%, respectively. The GTFP can effectively decrease the carbon emissions and carbon intensity in the national ETS market. The direct effect is significantly positive; thus, the mediating effects of GTFP on reductions in carbon emissions and carbon intensity are considered partial. The p value of the Sobel test is less than 0.1, indicating that GTFP is a significant the mediating variable and that the percentages of the mediating effect are −6.79% and −6.95%, respectively. Therefore, Hypothesis 8b is proven. Specifically, during the study period, the national ETS policy can reduce carbon emissions and carbon intensity by enhancing the GTFP. This enhancement suggests that the national ETS policy significantly improves the GTFP by forcing the power industry to adopt low-carbon technologies or clean energy through the tightening of carbon emission allocations.

Analysis of the mediating effect of GTFP on the TGC policy

The effects of the TGC policy on promoting renewable energy development through enhancements in GTFP are shown in Table 9. In Model (2), when the GTFP is the dependent variable, the coefficient value of $pos{t}_{TGC}$ is 0.0502 and thus considered significant. Specifically, if the other factors remain unchanged, the implementation of the TGC policy significantly promotes an increase of 5.02% in the GTFP. When the mediating variable GTFP is added to models (3) and (5), the mediating effect is significant with renewable energy generation as the explained variable. The coefficient of GTFP is 0.8415. When other factors remain stable, when the GTFP increases by 1%, the average renewable energy generation increases by 84.15%. The direct effect is significantly positive; thus, the mediating effect of GTFP in increasing renewable energy generation is considered partial. The p value of the Sobel test is considered significant, and the proportion of the mediating effect is 16.01%. Thus, Hypothesis 9a is proved. The green certificate trading policy promotes renewable energy development by enhancing green total factor productivity. With the share of renewable energy power generation as the explanatory variable, it is not significant that GTFP can enhance the share of renewable energy power generation. Possible reasons are that with the advancement of renewable energy power generation technology and the rapid expansion of installed capacity, the supply of the TGC market is greater than the demand, making its price trend towards lower values. The revenue from TGCs obtained by renewable energy generators cannot yet effectively reduce the adjustment costs resulting from the volatility of power generation, making the increase in the share of renewable energy generation insignificant.

Table 9 Results of the mediating effect of GTFP on the TGC policy.

Analysis of the mediating effect of GTFP on the GPT policy

The results of the GPT policy facilitating the development of renewable energy through GTFP are presented in Table 10. In model (2), with the GTFP as the dependent variable, the coefficient value of $pos{t}_{GPT}$ is 0.0508 and considered significant. Specifically, if other elements remain unchanged, the GPT policy remarkably enhances the GTFP by 5.08%. When the mediating variable GTFP is incorporated into models (3) and (5), the mediating effect with renewable energy generation as the dependent variable is prominent. The coefficient of GTFP is 0.6894. Specifically, if the other conditions remain constant, for every 1% increase in GTFP, both renewable energy power generation and its proportion experience an average increase of 68.94%. This direct effect is evidently positive; thus, the mediating effect exerted by the GTFP in augmenting renewable energy power generation is considered partial. The p value of the Sobel test is significant, with the proportion of the mediating effect amounting to 11.05%. Therefore, Hypothesis 9b is confirmed; specifically, the GPT policy promotes renewable energy development by enhancing the GTFP. Similarly, when the proportion of renewable energy generation is used as the explanatory variable, the effect of GTFP on increasing the share of renewable energy generation is not significant. A possible reason for this phenomenon is that renewable energy (such as wind power and photovoltaic energy) has significant volatility and uncertainty and needs thermal power to provide auxiliary services. Even if renewable energy power generation increases, thermal power is still needed to stably operate the power system, which makes the increase in the share of renewable energy generation not obvious.

Table 10 Results of the mediating effect of GTFP on the GPT policy.

In summary, by analysing the transmission mechanism of individual policies, the regression results show that over the study period, the pilot ETS policy suppresses GTFP and reduces both carbon emissions and carbon intensity. The national ETS policy reduces carbon emissions and carbon intensity by increasing GTFP. The TGC and GPT policies promote renewable energy development by increasing GTFP.

Analysis of the mediating effect of GTFP on multiple policies

Analysis of the mediating effect of GTFP on ETS–TGC synergy

The results of the mediating effect of GTFP on carbon emissions, carbon intensity, renewable energy power generation and its proportion under the pilot ETS-TGC synergy are presented in Table 11. In model (2), with GTFP acting as the explained variable, the coefficient value of $DI{D}_{pilotETS}\times {{\rm{post}}}_{TGC}$ is −0.1002 and considered significant. Under circumstances where other conditions remain unaltered, the pilot ETS–TGC synergy can lead to a significant reduction in GTFP of 10.02%. When the mediating variable GTFP is incorporated into models (3), (5), (7) and (9), according to the Sobel test outcomes and when taking carbon emissions and intensity as explained variables, the mediating effect of GTFP is determined to be remarkably significant, with coefficients of −0.3732 and −0.3775, respectively. When other elements remain constant, for every 1% increase in GTFP, carbon emissions and intensity decrease by 37.32% and 37.75%, respectively. Moreover, the direct effect is prominent, indicating that the mediating effect of GTFP on carbon emissions and carbon intensity is partial. The p value of the Sobel test is significant, and the percentages of the mediating effect are −16.44% and −16.02% for carbon emissions and carbon intensity, respectively.

Table 11 Results of the mediating effect test of GTFP on pilot ETS–TGC policy synergy.

When renewable energy generation is the explained variable, the mediating effect of GTFP is considered significant, and its coefficient is 0.4604. When the GTFP increases by 1%, the renewable energy generation increases by 46.04%. However, the direct effect of GTFP is not obvious, indicating that its mediating effect on renewable energy power generation is complete. The p value of the Sobel test is significant, and the proportion of the mediating effect is 23.60%.

The effects of the GTFP on carbon emissions, carbon intensity, renewable energy power generation and renewable energy proportion under the national ETS–TGC synergy are shown in Table F1 in Appendix F. According to the results of the Sobel test, with carbon emissions and carbon intensity serving as the explained variables, the mediating effect of GTFP is consistent with that of implementing the national ETS policy alone. With REG and PREG serving as the explained variables, the mediating effect of GTFP is consistent with the mediating effect of implementing the GPT policy alone. Therefore, Hypothesis 10b is proven. Specifically, the synergies of the national ETS policy with the TGC and GPT policies can enhance GTFP, thus reducing carbon emissions and promoting renewable energy development.

Analysis of the mediating effect of GTFP on ETS–GPT synergy

The results of the effects of GTFP on carbon emissions, carbon intensity, renewable energy power generation and renewable energy proportion under the pilot ETS–GPT synergy scenario are shown in Table 12. In model (2), with GTFP serving as the dependent variable, the coefficient value of $DI{D}_{pilotETS}\times pos{t}_{GPT}$ equals −0.0280 and does not pass the significance test; however, the GTFP still reduces carbon emissions and carbon intensity and promotes renewable energy development. Therefore, Hypothesis 10a is confirmed.

Table 12 Results of the test of the mediating effect of GTFP on the pilot ETS–GPT synergy.

Under the national ETS–GPT synergy (Table F2 in Appendix F), the results of the mediating effect of GTFP on carbon emissions, carbon intensity, renewable energy power generation and renewable energy proportion are in line with the mediating effects of implementing the national ETS policy alone and of the synergy between the national ETS and TGC policies. Thus, Hypothesis 10b is verified.

Analysis of the mediating effects of TGC–GPT synergy and ETS–TGC–GPT synergy

The test results of the promotion of renewable energy enhancement under the synergistic effect of TGC–GPT through GTFP are consistent with the mediating effect of implementing the GPT policy alone (Table F3 in Appendix F). This finding indicates that regardless of whether the policy is implemented in a synergistic form or alone, GTFP performs a crucial mediating function in promoting renewable energy advancement. Specifically, by enhancing the intermediate link of GTFP, the impact of the policy is transmitted to the renewable energy field, facilitating the development of renewable energy.

The mediating effect of the ETS–TGC–GPT synergy is in line with that of the ETS–GPT synergy (Tables F4–F5 in Appendix F). We will not elaborate further in this section.

Overall, a comparison of the regression results of the mediating effects reveals that, regardless of the individual implementation of the pilot ETS, national ETS, TGC and GPT policies or the synergy of multiple policies, GTFP can suppress carbon emissions and carbon intensity and promote renewable energy development. This finding demonstrates that GTFP is the core force driving carbon emission reduction and energy transformation and that through the processes of technological development, efficiency optimization and market synergy, a win‒win situation of carbon emission intensity reduction and renewable energy development can be achieved.

Analysis of the effects of policies on the dual control of energy consumption

A previous study has shown that in the short term, the national ETS market shows growth in carbon emissions and carbon intensity. For this outcome, one possible explanation is that the carbon quotas of the power industries across various regions are all traded in the unified national ETS market, leading to the expansion of the market scale and subsequently affecting the relationship between supply and demand in the market. However, this explanation may not be complete. To further test the impact of decarbonisation policies on the achievement of dual carbon goals, we evaluate the effects of the ETS, TGC and GPT policies on the dual control of energy consumption.

Table 13 presents the relevant regression results of the effects of the pilot ETS and national ETS policies on fossil fuel energy consumption intensity and of the TGC and GPT policies on the total fossil fuel energy consumption. When fossil fuel energy consumption intensity (FFCI) is taken as the explained variable, regardless of whether the pilot ETS or the national ETS are enacted, the regression results are significantly negative, and the regression coefficients equal −0.0505 and −0.0396, respectively. Therefore, the ETS significantly inhibits fossil fuel energy consumption intensity. Hypotheses 11a and 11b are confirmed; specifically, both the pilot ETS and the national ETS policies are effective in reducing energy intensity.

Table 13 Impacts of the FFC and FFCI on the CE and CI.

When the total amount of fossil fuel energy consumption (lnFFC) is used as the explained variable, the coefficient of the TGC policy is considered notably positive. A possible reason for this phenomenon is that some enterprises may deduct the total fossil fuel energy consumption by purchasing TGCs without actually using green power. Eventually, this phenomenon leads to ineffective control of the total fossil fuel energy consumption. Instead, due to increases in the demands of economic development and production, the consumption of fossil fuels has increased. However, the coefficient of the GPT policy is significantly negative because the consumption of green power is not included in the assessment of total energy consumption. By increasing the purchase volume of green power, fossil fuel energy consumption is genuinely substituted with renewable energy consumption. Thus, Hypotheses 12a and 12b are proven. Specifically, the TGC policy increases the total fossil energy consumption, and the GPT policy reduces the total fossil energy consumption.

The regression results of the synergies between the ETS policy and the TGC and GPT policies on the intensity and total amount of fossil fuel energy consumption are consistent with those of the policies when used alone, and the research conclusions remain unchanged.

Conclusions and policy implications

Conclusions

We utilize balanced panel data from 30 provinces in China from 2010 to 2023 with respect to the ETS, TGC and GPT policies as quasinatural experiments. First, we empirically examine the carbon reduction effect of the pilot ETS policy by using the multiperiod DID method and establish fixed effect models to examine the effects of the national ETS, TGC and GPT policies. Then, multiple policy interaction models are constructed, and the synergistic effects are analysed. Next, the transmission paths of single and multiple policy synergies are evaluated. Finally, the effects of the ETS, TGC and GPT policies on the dual control of energy consumption are analysed. The conclusions of this study are as follows:

• (1) In the benchmark regression of individual policies, the pilot ETS policy promotes reductions in the carbon emissions and carbon intensity, thus achieving dual control of carbon emissions, whereas the national ETS policy increases the carbon emissions and carbon intensity in the short term. The TGC and GPT policies promote increases in renewable energy power generation and its proportion. The analysis of regional heterogeneity reveals that the pilot ETS policy significantly reduces carbon emissions and intensity in different regions. The national ETS policy increases carbon emissions and intensity in the eastern, western and northeastern regions. The TGC policy significantly increases renewable power generation in the eastern and central regions. The GPT policy significantly boosts renewable power generation and its proportion in all regions.

• (2) When multiple policies are synergistic, the synergy between the pilot ETS policy and the TGC or GPT policies can enhance the carbon reduction effect, and the effect is stronger than that of the implementation of the pilot ETS alone. The synergistic emission reduction effect of the pilot ETS and GPT policies is better than that of the pilot ETS and TGC policies. The synergistic effect of the national ETS and TGC or GPT policies can promote the development of renewable power. For renewable power development, the synergistic effect of the national ETS and TGC policies is consistent with that of the national ETS and GPT policies. Notably, the renewable power development effects of the TGC policy synergized with the GPT policy are consistent with the implementation of the GPT policy alone, and the synergy of the three policies is consistent with the synergy of the ETS and GPT policies. Thus, there is redundancy between the TGC and GPT policies.

• (3) In terms of the mediating effect, an analysis of the transmission mechanism between single policies reveals that the pilot ETS policy inhibits GTFP growth, whereas the inhibitory effect is not significant when it is coordinated with the GPT policy. The national ETS, TGC and GPT policies facilitate increases in GTFP. A comparison of the regression results of the mediating effects reveals that, regardless of whether the pilot ETS, national ETS, TGC and GPT policies are implemented individually or if multiple policies are synergized, the GTFP can reduce carbon emissions and carbon intensity and increase renewable energy generation and renewable energy proportion, thereby achieving dual carbon control while promoting renewable power development.

• (4) In terms of the dual control of energy consumption, both the pilot ETS and national ETS policies can decrease the intensity of fossil fuel consumption. Therefore, ETS policies have a significant effect on reducing the intensity of fossil energy consumption. However, although the TGC policy helps promote the investment and development of renewable energy, it does not directly substitute the user’s actual source of green power and, to a certain extent, may indirectly increase fossil energy. In contrast, the TGC policy reduces total fossil energy consumption by promoting the actual consumption of green power. This finding suggests that the pilot ETS, national ETS and GPT policies are all effective at promoting a comprehensive transition from the dual control of energy consumption to that of carbon, whereas the TGC policy must be further optimized to reduce dependence on fossil energy. The research results are shown in Fig. 3.

  Fig. 3: Research result.

  

  a illustrates the benchmark regression results of the pilot ETS, national ETS, TGC and GPT policies. b presents the policy synergy results. c describes the intermediary effect results for individual policies. d shows the intermediary effect results for multiple policy. e demonstrates the results of the effects of policies on the dual control energy consumption. ***, **, * indicate significant at the 1%, 5% and 10% levels respectively. Source: Decarbonisation policies synergies pathway innovation: Achieving dual carbon goals.

Discussion

Impact of a single policy

The weemissions and carbon intensity, which is consistent with the conclusions of past studies; for example, the pilot ETS policy significantly suppresses carbon emissions in pilot cities (Ma et al. 2023) and reduces the carbon emission intensity (Wu et al. 2021). Scholars have reported that the implementation of a total carbon emission control policy under the framework of a national ETS can achieve long-term reductions in carbon emissions and increase total output (Gu et al. 2024). In addition, we reveal that there are differences in the emission reduction effects across regions, which is consistent with the findings of previous studies (Liu et al. 2019). Moreover, another study has noted that the pilot ETS policy has excellent spillover effects and is suitable for more regions (Wang et al. 2025).

The results of this study show that the national ETS has increased carbon emissions and intensity in the short term. Currently, the main sources of carbon emissions in China include the power and heat production, industrial, construction, transportation, and agriculture sectors. Therefore, it is necessary to expand the coverage of sectors in the national ETS policy and tighten the allowance in sufficient areas to enhance their carbon emission reduction effect. With the continuous improvement and development of the national ETS policy, more energy-intensive sectors will be included in the national ETS policy in the future, and a long-term emission reduction effect should gradually appear (Wang et al. 2024).

The results show that the TGC and GPT policies facilitate the development of renewable energy. This finding is in line with the findings of previous studies, where scholars have noted that the TGC policy can promote the development of wind power, which is the most economical renewable energy technology (Zhang et al. 2018). By participating in TGC trading, environmental benefits are obtained, which in turn reduces the average LCOE of PV power generation and promotes its development (Wang et al. 2025). However, the policy is not effective in the western and northeastern regions; in the long run, with the continuous improvement in the market mechanism and its application scenarios, the effect of the TGC policy in promoting the development of renewable energy may gradually improve (Wang et al. 2024).

The GPT policy can realize the balance between renewable power supply and demand in different regions and further promote the development of renewable energy while promoting the market development of green power in China (Zhang et al. 2018b). Green power contributes positively to sustainable development, and increasing green power can increase the sustainable development goal index (Kirikkaleli et al. 2025). These studies are consistent with our findings. Compared with previous studies, we find innovative results suggesting that the GPT policy promotes renewable energy more than the TGC policy does.

Synergistic effects of multiple policies

The results show that synergizing the national ETS policy with the TGC or GPT policies is more likely to promote the development of renewable energy. This result is consistent with the findings of existing studies. Scholars assume that if ETS revenues are used for renewable energy, they will be favourable for renewable energy development (Lin and Jia, 2020). In addition, we reveal that the synergy between the pilot ETS policy and the TGC or GPT policies can enhance carbon emission reduction. Moreover, the synergy between the pilot ETS and TGC policies improves the emission reduction effect and promotes the consumption of renewable energy (Yan et al. 2023). The introduction of the GPT and ETS policies effectively reduces actual carbon emissions and improves the utilization of renewable energy (Li et al. 2025).

Compared with the existing studies, a major innovative finding of this study is the policy redundancy of the synergy between the TGC and GPT policies. TGC, as an environmental attribute, can be used independently of power trading, whereas GPT combines environmental benefits and power consumption through the integration of certificates and power.

Transmission mechanism of the policy

Research results show that the pilot ETS policy inhibits GTFP growth, and scholars note that this policy fails to significantly promote green technological innovation, such as in cities or old industrial bases in the western region, and it may have an inhibitory effect (Zhou and Wang, 2022). However, this study shows that the national ETS policy significantly promotes the growth of GTFP. Research on the national ETS policy is lacking, but conclusions can be drawn on the basis of actual operations, where entities can trade allowances freely and the allowance resources are optimally allocated nationwide, thereby improving the overall GTFP.

The results of the study show that GTFP can curb carbon emissions and carbon intensity and promote renewable energy development, regardless of whether the pilot ETS, national ETS, TGC and GPT policies are implemented individually or in synergy with multiple policies. This finding is consistent with those of existing studies, indicating that green technology innovation directly affects energy efficiency, which in turn has a positive effect on environmental sustainability (Jiang et al. 2024). Technological progress is the foundation and driving force of sustainable development, and the shift to renewable energy can help realize a relatively clean energy mix and reduce carbon intensity (Ai et al. 2025). Therefore, enhancing GTFP plays a key role in achieving carbon reduction and promoting renewable energy.

Impact of policies on the dual control of energy consumption

The results of the study show that both the pilot ETS and national ETS policies can effectively curb energy intensity. This finding is consistent with the conclusions of existing studies, which indicate that the pilot ETS policy significantly reduces the energy intensity of pilot cities (Geng and Fan, 2021). Some scholars have noted that the pilot ETS policy significantly reduces the energy intensity of entities by taking high-carbon emission companies as the object of study (Shi et al. 2024). Furthermore, these studies show that the ETS policy plays an important role in achieving the dual carbon goals.

The results of the study show that the TGC policy increases and the GPT policy decreases the total fossil energy consumption. This finding is consistent with the findings of existing studies. Since the TGC policy cannot be completely attached to power consumption, when a user buys a TGC, it is not equivalent to them using green electricity (Li et al. 2024). With improvements in the market mechanism and recognition of TGCs, the TGC policy is expected to better promote the development of renewable energy and reduce the dependence on fossil energy in the long term (Lin et al. 2025).

In contrast, the GPT policy provides a new option for promoting green and sustainable development by mandating the physical consumption of renewable power through the certificate bundling mode (Tang et al. 2023). This policy ensures the physical consumption of green power, effectively contributing to the optimization of the energy structure and the achievement of emission reduction targets. In addition, some scholars have noted that during the peak demand of power, additional demand is fulfilled mainly by supplementing renewable energy generation and have predicted that by 2050, China’s electricity mix will be dominated by renewable energy, such as wind and photovoltaic energy (Wang et al. 2025).

Policy implications

The following policy implications are proposed on the basis of the conclusions:

• (1) Expanding the sector coverage of the national ETS policy to promote synergistic regional emission reduction and renewable energy development

  First, during this study, the national ETS policy covers only the power sector, which is still dominated by coal-fired power generation. Therefore, considering factors such as dual carbon control targets, industry heterogeneity and regional heterogeneity, high-carbon emission industries and industries with high potential for emission reduction are both included in the national ETS policy. Moreover, the allowance allocation mechanism should be designed for different phases of the ETS policy according to the principles of equality, responsibility, capacity and efficiency (Wang et al. 2025). Second, according to the economic development, industrial structure and energy consumption characteristics of different regions, the cross-regional carbon allowance exchange mechanism is designed to strengthen the carbon emission reduction effect in the central and western regions, which provides experience for national ETS. Moreover, the results enrich the application scenarios of the TGC and GPT policies in the northeastern region and promote the development of renewable energy.

• (2) Implementing synergistic schemes of the ETS, TGC and GPT policies in a differentiated manner to achieve policy synergy for decarbonisation.

  The results show that the synergy between the pilot ETS and GPT policies has the best emission reduction effect, and the synergy between the national ETS and TGC policies has the best renewable energy development effect. There is redundancy between the TGC and GPT policies. Therefore, different policy synergy programs can be selected according to the carbon emission reduction and renewable energy development goals. Other developing countries can select the appropriate synergistic programs of three policies on the basis of their carbon emission reduction and renewable energy development plans to achieve the best environmental and economic benefits. In addition, to guarantee the simultaneous realization of green and low-carbon development, for the pilot ETS policy, entities are allowed to use TGCs for compliance. In response to the national ETS policy, the mandatory RECQ is assigned to entities, and rigid constraints on the use of green power and TGCs by high energy-consuming entities covered by the national ETS are strengthened.

• (3) Enhancing GTFP to strengthen its transmission effect in the ETS policy

  The results show that GTFP can reduce carbon emissions and carbon intensity through both the pilot ETS and national ETS policies. However, the pilot ETS policy inhibits the enhancement of GTFP. In addition, the national ETS policy enhances the GTFP. However, GTFP has only a partial mediating effect, and the national ETS policy still significantly increases carbon emissions and carbon intensity. Therefore, drawing from the European Union’s energy technology innovation subsidy policy (Baena-Moreno et al. 2020; Gowd et al. 2023), a budget plan for carbon revenues can be formulated, and a certain proportion of ETS revenues can be used to support low-carbon technological innovation projects and enhance competitiveness and low-carbon technological innovations, thus realizing carbon emission reductions and increasing GTFP. This phenomenon provides a way for other developing countries to effectively utilize ETS revenues to accelerate the transition to a low-carbon economy.

• (4) Improving the TGC mechanism to accelerate the comprehensive transition from the dual control for energy consumption to the dual control for carbon emissions

The results show that both the pilot ETS and national ETS policies significantly reduce the intensity of fossil energy consumption and that the GPT policy significantly reduces the total fossil energy consumption. However, the TGC policy increases the total fossil energy consumption. Therefore, it is crucial for some countries that wish to promote the use of TGC to improve the system of TGC issuance, trading, and verification to ensure that the physical power consumption scenarios are consistent with the TGC verification scenarios. This result is a solution to the problem in which the TGC policy fails to significantly reduce the total fossil energy consumption. Moreover, the GPT policy should be strengthened to promote the consumption of green power through energy substitution, promote the development of renewable energy, accelerate the comprehensive transition from the dual control for energy consumption to that of carbon, and thus realize the dual carbon goals.

Limitations and future research

Future studies should be extended in the following areas. First, the sample limitations should be overcome. Owing to data availability, only the provincial level is selected as the research object in this work. In the future, researchers should utilize the entity level to accurately reflect the effects of the ETS, TGC and GPT policies. Second, the content length is limited. Researchers should further explore the spatial spillover effects of policies. In addition, we choose only the GTPF as the mediating variable, and the transmission pathway of each policy can be studied in depth in other aspects.