Does Renewable Energy Technology Innovation Enhance Carbon Productivity? Evidence from China

Abstract

Renewable energy technology innovation (RETI) is an effective means of reducing emissions without sacrificing productivity, making it a key driver of carbon productivity (CP) improvement. This study employs a dataset covering 30 Chinese provinces from 2010 to 2021 to empirically examine the impact of RETI on CP. Additionally, it explores the non-linear effects of environmental regulation (ER) and R&D investment (RD) in this relationship. The findings indicate that RETI significantly enhances CP, a result that remain robust across multiple robustness and endogeneity tests. The heterogeneity analysis reveals that RETI has the strongest impact on CP in the eastern region, followed by the central region, but has a negligible effect in the western region. Among the different types of RETI, the positive effects of solar, wind, and energy storage technology innovations on CP are more evident. Moreover, the impact of RETI on CP is stronger in economically developed provinces than in less developed ones. The mechanism analysis shows that RETI indirectly enhances CP by optimizing industrial structure, increasing renewable energy generation, and improving energy efficiency. The threshold effect analysis suggests that as ER intensifies, the positive effect of RETI on CP follows a non-linear relationship that strengthens initially but then diminishes. In contrast, as RD rises, the marginal benefits of RETI on CP are continually strengthened.

1. Introduction

Global warming caused by greenhouse gas emissions represents a critical threat to human sustainability [1]. There is a global consensus on the urgent need to control carbon dioxide emissions to mitigate climate-related risks. As the world’s largest carbon producer, China accounted for 32% of global carbon dioxide emissions in 2023. Its fossil fuel-dominated energy mix exacerbates the carbon lock-in effect, leading to dependence on high-carbon infrastructure, which hinders the low-carbon transition. This dilemma presents a significant challenge to the country’s sustainable development trajectory and creates a major obstacle to achieving global emissions reduction goals. In response, China has committed to emission reductions by setting ambitious “dual carbon” targets, aiming to peak carbon emissions by 2030 and achieve carbon neutrality by 2060 [2]. These goals are critical in the global fight against climate change. However, ${\mathrm{C}\mathrm{O}}_2$ emission levels and intensity are closely tied to economic production activities, meaning that efforts to curb emissions may disrupt domestic economic stability. As a result, exploring a synergistic approach that balances economic growth and carbon emission reduction has become a central theme of current research [3]. Carbon productivity (CP), a critical indicator of economic sustainability, measures economic output per unit of ${\mathrm{C}\mathrm{O}}_2$ emission and provides a new paradigm to address the above dilemmas [4]. Enhancing CP not only effectively reduces emission reduction costs but also optimizes resource allocation efficiency, offering a feasible solution to decouple economic growth from carbon emissions [5].

Theoretically, technology innovation (TI)—particularly in energy technology—plays a crucial role in driving economic growth, as well as energy savings and emission reductions [6]. As an emerging area in energy TI, renewable energy technology innovation (RETI) has significant potential to enhance green productivity, facilitate the clean energy transition, and achieve energy savings and emission reductions. This potential stems from RETI’s ability to enable the large-scale deployment of clean fuels and optimize the use of renewable resources. Bayer et al. argued that RETI offers the most cost-effective pathway to a low-carbon society [7]. Given its environmentally friendly nature, China is vigorously advancing RETI to replace conventional fossil fuels such as coal and oil with wind and solar power. As of September 2024, China’s total installed capacity for wind and solar power reached 1.25 billion kilowatts, surpassing the 2030 goal of 1.2 billion kilowatts ahead of schedule. Considering the rapid advancement of RETI in China, it is essential to assess whether the country can harness this innovation to achieve carbon reduction goals without compromising overall productivity. Currently, the literature mainly focuses on the emission reduction effect of RETI, with some scholars also highlighting its unique role in increasing total factor productivity. However, insufficient attention has been paid to its synergistic mechanism in promoting both economic output and environmental protection, particularly in how RETI can reduce carbon emissions while fostering stable economic recovery. A comprehensive assessment of the impact of RETI on CP, considering regional economic and environmental factors, is essential for supporting decision-making in achieving the “dual carbon” goals. To address these gaps, we integrate RETI and CP into a unified research framework and empirically explore how RETI influences CP. Moreover, the presence of “double externalities”, where market failures limit the effectiveness of RETI, diminishes its impact on CP. To address this, we analyze how environmental regulation (ER) and R&D investment (RD) affect the relationship between RETI and CP, emphasizing their potential to mitigate the challenges posed by “double externalities”.

The innovations in this study include the following: first, it addresses the scarcity of studies exploring the relationship between RETI and CP. Most existing studies focus on either the economic or environmental benefits of RETI, typically considering only one aspect. In contrast, we integrate both economic and environmental benefits, providing a comprehensive examination of the effects of RETI on CP and their underlying mechanisms. Second, this study investigates regional differences in the impact of RETI on CP and examines how variations in economic levels affect these outcomes, illustrating how different types of RETI exhibit varying effects. These findings provide valuable insights for promoting coordinated regional CP development and guiding further renewable energy investments. Third, from a non-linear perspective, this study analyzes the intrinsic relationship between RETI and CP, incorporating the influences of ER and RD, providing a conceptual basis for government agencies to design effective policies. Finally, referring to Popp [8], this study adopts a novel approach to constructing RETI levels. It considers the effects of technological depreciation and diffusion, ensuring that the index is scientifically robust and accurately reflects the dynamics of renewable energy technologies.

2. Literature Review

Given the growing urgency of climate change and the global technological revolution, the nexus between TI and CP has attracted considerable academic interest. The current literature primarily explores the following dimensions:

2.1. Study of Factors Influencing CP

Scholars widely agree that TI—especially green TI [9], low-carbon TI [10], and energy TI [11]—is an important factor influencing regional CP. However, externalities associated with knowledge diffusion and environmental protection may hinder the supply of TI. These challenges need to be addressed through government policies such as ER and RD. Existing research has shown that ER and RD not only significantly shape TI but also directly impact the development of CP. However, the specific relationship between ER and CP remains controversial, with discussions primarily revolving around the green paradox, stringent emission reduction, and non-linear effects. The green paradox, first proposed by Sinn, suggests that ER may increase the consumption of fossil fuels, thereby reducing CP [12]. Wang and Wei further found that stringent ER may lead to the green paradox, where it paradoxically increase carbon emissions instead of curbing them [13]. In contrast, Pan et al. contended that stronger ER compels high-polluting and high-energy-consuming enterprises to carry out TIs, ultimately enhancing CP [14]. Similarly, Li et al. advocated that the innovation compensation effect of ER positively contributes to CP [15]. Moreover, several studies identify a significant non-linear relationship between ER and CP. Song and Han observed that in the early stages of economic development, the effect of ER on CP is negligible, but as development progresses, ER gradually enhances CP [16]. Other scholars have argued that the impact of ER on CP varies depending on its intensity [17], type [16], and targets [18]. Huang et al. examined the impact of RD on carbon intensity at the inter-provincial level in China and revealed that sufficient RD can effectively reduce carbon intensity [19]. Similarly, Bai et al., using panel data from 88 economies worldwide, found that government subsidies encourage enterprises to pursue TI, improving cleaner production efficiency and enhancing CP [20]. Mo further demonstrated that increasing RD not only stimulates green TI but also significantly enhances CP [21]. Churchill et al. analyzed the relationship between R&D intensity and carbon emissions in G7 countries from 1870 to 2014, finding that the impact of R&D intensity on carbon emissions varied across different stages of economic development [22]. Lin and Xu explored the impact of RD on regional carbon emissions in China, identifying a “U”-shaped relationship in the central region and an inverted “N”-shaped relationship in the western region [23]. While the existing literature has confirmed the non-linear relationship between ER and RD on CP, further investigation is needed to determine whether these mechanisms, acting as corrective forces to address the “double externality” problem, exert a non-linear influence on the positive impact of RETI on CP.

2.2. Relationship Between TI and CP

The existing literature rarely examines the direct impact of RETI on CP, with most studies focusing on the broader relationship between TI and CP. Scholars generally agree that TI is a critical driver of improvements in CP. For example, Cheng et al. found that TI enhances CP by improving energy efficiency and reducing costs, both globally and at the provincial level in China [24]. Similarly, Lu et al. indicated that TI can significantly enhance regional CP, though the extent of this positive effect varies across different regions [25]. Chen et al. emphasized that promoting low-carbon TI in enterprises is a crucial strategy for reducing pollution, developing a green economy, and enhancing regional CP [26]. Guo et al. also found that energy-related TI improves energy efficiency and significantly contributes to reducing regional carbon emissions [27]. However, some scholars argue that the concentration of economic activities driven by TI may lead to higher carbon emissions, thereby inhibiting CP. Yan et al. suggested that while TI improves energy efficiency, it paradoxically increases energy consumption due to higher production efficiency. This rebound effect implies that energy saving from TI may be offset by increased energy demand, ultimately diminishing CP [28]. Similarly, Jaffe et al. found that although TI reduces carbon emissions, it may create scale effects that increase carbon emissions and exacerbate environmental burdens [29]. Zhang et al. reported a negative correlation between TI and CP in western China, attributing this outcome to the region’s less developed economy and limited technological capabilities [30]. Du and Li further noted that green TI primarily enhances CP in high-income economies, as low-income countries often lack the complementary support required to maximize the benefits of TI [31]. Finally, Yan et al., analyzing patent data for low-carbon technologies across 15 global economies, found no conclusive evidence that low-carbon TI influences carbon emissions [32].

2.3. Relationship Between RETI and Carbon Performance

There is no clear academic consensus on the impact of RETI on carbon emissions. The primary debates center on three aspects: First, RETI is widely recognized for helping reduce CO₂ emissions. Zhao et al. empirically indicated that RETI alleviates climate risks by improving energy efficiency [33]. Similarly, Lin and Zhu found that RETI significantly inhibits carbon emissions, with its impact depending on the energy structure [34]. Wang and Zhu highlighted the spatial spillover effects of RETI. They noted that it improves environmental performance in both local and neighboring regions, unlike fossil energy TI, which has a more limited impact [35]. Wang et al. advocated that RETI reduces carbon accumulation, with its emission reduction effect strengthening as renewable energy adoption increases [36]. Among the G7 countries, Wang et al. demonstrated that renewable energy consumption, combined with RETI, plays a crucial role in reducing emissions [37]. Second, several studies reveal a non-linear nexus between RETI and carbon emissions. Li et al. found that the emission reduction effects of RETI vary with shifts in the energy consumption structure. In coal-dominated energy systems, RETI’s impact is minimal, but it strengthens as the demand for green energy grows [38]. Yan et al. argued that the benefits of RETI emerge only when income levels surpass a threshold, indicating that its effectiveness depends on economic development [39]. Liu et al. reported sectoral variations, with RETI significantly reducing carbon intensity in energy-intensive industries while having negligible effects in sectors related to residential life [40]. Cheng and Yao highlighted the time lag in RETI’s commercial application, suggesting that its positive effects on carbon performance manifest gradually over time, leading to long-term benefits [41]. Third, some studies find no clear effect of RETI on carbon emissions. Analyzing data from 30 countries, Chen and Lei found no significant correlation between RETI and pollutant reductions. They attributed this to the persistent dominance of coal in global energy systems, which limits renewable energy’s substitution potential [42]. Similarly, Gu et al. suggested that while RETI lowers energy service costs, it may trigger a “rebound effect” that increases energy demand, offsetting its emission reduction benefits [43].

In summary, scholars have studied the intrinsic relationship between TI and CP, as well as RETI’s impact on environmental performance. These studies provide a solid foundation for understanding the potential effects of RETI on CP. However, several gaps in the research remain. First, while RETI has significant potential to enable new energy generation, promote energy savings, and support green growth, few studies examine its micro-level impacts on CP, which encompasses both environmental and economic dimensions. Furthermore, differences in evaluation indicators, sample compositions, and measurement models have led to conflicting conclusions about the effects of TI on CP. More importantly, since TI is a broad concept, its general conclusions may not be directly applicable to RETI, highlighting the need for more research on this specific relationship. Second, much of the existing literature examines the linear effects of ER and RD on CP, with some exploring their non-linear relationship. However, few studies have investigated whether and how ER and RD create a threshold effect on the positive impact of RETI on CP in the context of “double externality”. Understanding how environmental policy instruments, such as ER and RD, influence the effectiveness of RETI in enhancing CP is essential for real-world applications. To bridge these gaps, this study uses two-way fixed effects, mediation effects, and threshold models to clarify the specific pathways through which RETI affects CP, providing actionable policy recommendations.

3. Theoretical Studies and Hypothesis Formulation

3.1. The Direct Impact of RETI on CP

RETI serves as a catalyst for sustainable development by enhancing CP, boosting regional economic output, and reducing carbon emissions. RETI can lower the costs of producing, utilizing, and consuming renewable energy, making it more competitive. This, in turn, increases the willingness to consume renewable energy, facilitating its large-scale incorporation into the energy system and gradually displacing fossil fuel consumption [44]. In China, where coal is the primary energy source, expanding the share of green energy consumption will help control carbon emissions at the source and enhance regional environmental performance. In addition, RETI stimulates the production of eco-friendly products, which promote the concept of sustainable environmental protection while generating market demand for these products. This change in consumption patterns helps reduce the carbon footprint of daily activities, further enhancing CP. Schumpeter’s theory of economic development asserts that technological progress, driven by innovation, is a key driver of productivity growth. RETI continually strengthens the substitution effect of innovative factors over traditional ones by fostering green preferences, optimizing the allocation of innovative resources. This dynamic improves production processes within enterprises, boosts operational efficiency across the industrial chain, and supports the rapid growth of the regional economy [45]. RETI also generates economies of scale, facilitates the growth of new energy sectors, and promotes the agglomeration of green production factors in the local economy. This can expand the renewable energy industry and its related upstream and downstream sectors while creating green employment opportunities in manufacturing and research. As a result, regional CP can be significantly enhanced. In summary, this paper proposes the following:

H1. 

 RETI can effectively enhance regional CP.

3.2. The Indirect Impact of RETI on CP

3.2.1. Industrial Structure

RETI indirectly enhances CP by reducing the scale of pollution-intensive industries while expanding green and high-end industries, thus facilitating the transition to a low-carbon and high-value-added industrial structure. Specifically, RETI reduces the demand for fossil fuels and high-emission production processes while transforming resource-intensive production models. Traditional high-carbon industries, such as coal-fired power generation, steel production, and cement manufacturing, are the first to be affected. Under the phase-out mechanism driven by renewable energy technologies, these high-polluting industries must adopt cleaner alternatives, downsize operations, or invest in RETI to achieve a green transition. This restructuring not only mitigates carbon-intensive production activities but also enhances productivity and competitiveness. In addition, RETI promotes the growth of multiple emerging sectors, including renewable energy (wind power, photovoltaics, and energy storage), high-end manufacturing (new energy vehicles, advanced materials, and renewable energy equipment), green services (green finance, carbon trading, and sustainable resource management), and smart energy management (digital energy platforms and smart grids). The expansion of these technology- and service-intensive industries helps reduce carbon emissions while generating substantial economic benefits, ultimately enhancing regional CP. In summary, this paper proposes the following:

H2a. 

RETI enhances CP by promoting industrial structural upgrading.

3.2.2. Renewable Energy Generation

In reality, nearly half of China’s carbon emissions originate from power generation, with coal-fired power accounting for approximately 90% of total electricity production [46]. Renewable energy generation, known for its cleanliness and low emissions, effectively mitigates the high carbon emissions associated with coal-fired power generation. RETI steadily enhances the generation of electricity from renewable sources and continuously increases the share of clean electricity in overall energy consumption. This transition helps meet society’s growing energy needs, reduces the accumulation of carbon emissions from coal-fired power generation, and ultimately enhances regional CP. Specifically, RETI has advanced renewable energy generation technologies, improving the efficiency, stability, and availability of the electricity supply, thereby ensuring a reliable and sustainable clean energy source. For example, the latest generation of wind turbines is equipped with advanced blade designs, taller tower structures, and optimized variable-speed drive systems to ensure efficient power generation even at low wind speeds. These improvements minimize power outages caused by wind speed fluctuations and enhance the reliability of the energy supply. Moreover, as renewable energy technologies continue to mature, the costs associated with energy production, conversion, transmission, and utilization have decreased significantly. Lower costs facilitate the large-scale deployment of renewable energy production, leading to increased overall generation capacity. Consequently, RETI reduces reliance on traditional coal-fired power generation by expanding clean electricity production, enhancing carbon efficiency per unit of energy output, and reducing carbon emissions. This conclusion is consistent with the findings of Saidi et al. [47]. In summary, this paper proposes the following:

H2b. 

RETI enhances CP by increasing renewable energy generation.

3.2.3. Energy Efficiency

CP is closely linked to carbon emissions per unit of energy produced and consumed and is an important indicator of the decarbonization level of an energy system. RETI indirectly enhances CP by improving energy efficiency. First, RETI increases energy conversion efficiency, enabling the same input to generate more low-emission electricity. Increased energy conversion efficiency reduces the demand for raw materials and resources per unit of energy, thus lowering carbon emissions from production [48]. For example, photovoltaic technology converts solar energy into electricity, avoiding energy losses and carbon emissions typical of coal-fired generation. Second, RETI improves energy storage and dispatch efficiency. Advanced storage technologies, such as solid-state batteries, hydrogen storage, and pumped storage, store surplus electricity during peak demand, reducing power interruptions and minimizing energy waste from renewable energy intermittency. Additionally, smart grid technology improves energy consumption by enabling real-time monitoring and data analysis. It optimizes power flow and enhances transmission efficiency, reducing carbon emissions from grid overloads and transmission losses. Finally, RETI supports the synergistic integration of diverse energy sources and facilitates the creation of a multi-source, complementary urban energy system, optimizing the allocation and use of various sources of energy through integrated energy management. The combined use of renewable energy sources enhances overall energy efficiency and reduces reliance on high-carbon energy. Power–heat cogeneration, using biomass and geothermal energy for both electricity and heating, is widely applied in domestic heating and industrial processes, reducing carbon emissions and minimizing energy waste. In summary, this paper proposes the following:

H2c. 

RETI enhances CP by increasing energy efficiency.

3.3. The Threshold Effects of RETI on CP

3.3.1. Environmental Regulation

As a government tool aimed at overcoming negative externalities, ER is an effective initiative for balancing economic growth and environmental protection. However, existing studies indicate that ER influences the impact of RETI on CP both positively and negatively. The positive effect of ER is primarily explained by Porter’s Hypothesis, which suggests that reasonable and moderate ER encourages enterprises to invest in technological innovation and energy efficiency improvements. This creates a “compensation mechanism” that alleviates the financial burden of compliance costs [49]. From a rational decision-making perspective, enterprises adopt cleaner, low-carbon renewable energy technologies to reduce the marginal cost of pollution control, improving productivity while reducing carbon emissions. Conversely, the negative effects of ER are linked to the “compliance cost” perspective [50]. As ER intensity increases, the government imposes stricter limits on energy consumption and carbon emissions, forcing enterprises to internalize environmental externalities. As a result, enterprises must allocate more limited resources to pollution control and regulatory compliance, diverting funds from technological advancements and equipment upgrades, which hinders the development of RETI. In response to rising compliance costs, some short-sighted enterprises may lack sufficient incentives to invest in RETI. Instead, they prioritize lower-cost process modifications that offer immediate economic benefits. While this strategy may yield short-term gains, it could also diminish the long-term environmental returns of RETI, weakening its ability to enhance CP. China’s “Double Control of Energy Consumption” policy has successfully motivated enterprises to adopt low-carbon TIs. However, excessively stringent emission reduction targets have discouraged some high-energy-consuming enterprises from investing in RETI due to uncertain regulatory risks and high costs. This phenomenon suggests that increased ER intensity may shift enterprises’ focus from proactive technological innovation to passive compliance strategies. Furthermore, enterprises with higher risk tolerance may continue expanding production despite regulatory constraints. This can worsen pollutant emissions and defy administrative directives. In sum, these findings indicate that the net effect of RETI on CP exhibits a non-linear relationship with ER intensity. In summary, this paper proposes the following:

H3. 

As ER intensity rises, the positive effect of RETI on CP initially increases and eventually diminishes.

3.3.2. R&D Investment

RD is a key resource to incentivize green innovation activities and promote the low-carbon transformation of industries. RD’s influence on the relationship between RETI and CP is phase-dependent, with different RD levels leading to differentiated effects of RETI on CP. Existing studies suggest that RD has a threshold effect on the role of green technological innovation in enhancing environmental performance, requiring RD inputs to surpass a minimum scale for a noticeable promotion effect [51]. TI activities are speculative and long-term in nature. When RD is low, the RETI of enterprises is constrained by underinvestment, increasing the risk of transformation failure and reducing potential economic returns. This limits the promotional effect of RETI on CP. However, as RD increases, the positive impact of RETI on CP is further strengthened. First, higher RD balances the risks and rewards of innovation, fostering greater tolerance for failure and stimulating sustained innovative efforts, thereby improving the effectiveness of RETI in enhancing CP. Second, securing R&D funding generates political pressure for enterprises to maintain close informal connections with the government, motivating them to align their innovation performance with government policy goals. Third, government investment in R&D encourages financial institutions to support RETI projects by offering more financial options, alleviating enterprises’ financial constraints. Finally, subsidies for RETI send positive signals to the market, boosting investor confidence in the renewable energy sector and attracting further investment. In summary, this paper proposes the following:

H4. 

As RD grows, it further strengthens the positive effect of RETI on CP.

4. Methodology and Data

4.1. Econometric Model

To investigate the direct impact of RETI on CP, we construct the following baseline model:

$${\mathrm{C}\mathrm{P}}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{I}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_2{\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\delta }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(1)

where ${\mathrm{C}\mathrm{P}}_{\mathrm{i}\mathrm{t}}$ represents CP, ${\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{I}}_{\mathrm{i}\mathrm{t}}$ denotes the level of RETI, and ${\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{X}}_{\mathrm{i}\mathrm{t}}$ reflects the control variables, ${\mathsf{\mu }}_{\mathrm{i}}$ represents the province fixed effects, ${\mathsf{\delta }}_{\mathrm{t}}$ represents the year fixed effects, and ${\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$ captures the unobserved factors affecting CP.

To examine the mechanism through which RETI influences CP, particularly through the channels of industrial structure, renewable energy generation, and energy efficiency, the following mediation effect model is proposed:

$${\mathrm{M}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{o}\mathrm{r}}_{\mathrm{i}\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{I}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_2{\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\delta }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(2)

where ${\mathsf{\beta }}_1$ is the coefficient that measures the impact of RETI on the mediator variables.

The following threshold model is constructed to examine the effects of ER and RD on the enhancement of CP driven by RETI:

$${\mathrm{C}\mathrm{P}}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{I}}_{\mathrm{i}\mathrm{t}}\mathrm{I}\left({\mathrm{T}}_{\mathrm{i}\mathrm{t}}\le {\mathsf{\gamma }}_1\right)+{\mathsf{\alpha }}_2{\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{I}}_{\mathrm{i}\mathrm{t}}\mathrm{I}\left({{\mathsf{\gamma }}_1<\mathrm{T}}_{\mathrm{i}\mathrm{t}}\le {\mathsf{\gamma }}_2\right)+{\mathsf{\alpha }}_3{\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{I}}_{\mathrm{i}\mathrm{t}}\mathrm{I}\left({\mathrm{T}}_{\mathrm{i}\mathrm{t}}>{\mathsf{\gamma }}_2\right)+{\mathsf{\theta }\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathsf{\delta }}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(3)

where ${\mathrm{T}}_{\mathrm{i}\mathrm{t}}$ represents the threshold variables, ${\mathsf{\gamma }}_1$ and ${\mathsf{\gamma }}_2$ are the estimated threshold values, and I(∙) is the indicative function.

4.2. Variables Definition

(1)

Dependent variable

Following the approach of Kaya [52], this paper defines carbon productivity (CP) as the ratio of a region’s GDP to its ${\mathrm{C}\mathrm{O}}_2$ emissions over a given time period:

$$\mathrm{C}\mathrm{P}={\displaystyle \frac{\mathrm{G}\mathrm{D}\mathrm{P}}{\mathrm{E}\mathrm{C}}}$$

(4)

where EC denotes the total ${\mathrm{C}\mathrm{O}}_2$ emissions derived from fossil energy consumption, as described in detail by Lin and Zhu [34].

(2)

Core independent variable

Previous research commonly uses indicators such as the count of renewable energy patents or R&D expenditure to measure RETI levels [34,53]. However, these methods often overlook critical dynamics, including knowledge diffusion, time lags, and technology depreciation. As a result, they fail to accurately capture the accumulation process of innovation inputs. To address these limitations, this study adopts the knowledge stock approach, drawing on Popp’s findings [8]. By incorporating the effects of knowledge diffusion and technology depreciation, the stock of renewable energy-related patents is used to estimate RETI levels. This approach provides a more comprehensive and nuanced representation of innovation dynamics:

$${\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{I}}_{\mathrm{i}\mathrm{t}}=\sum _{\mathrm{j}=0}^{\mathrm{t}}{\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{I}}_{\mathrm{i}\mathrm{j}}\exp \left[-{\mathsf{\mu }}_1\left(\mathrm{t}-\mathrm{j}\right)\right]\{1-\exp \left[-{\mathsf{\mu }}_2\left(\mathrm{t}-\mathrm{j}\right)\right]\}$$

(5)

Popp’ empirical analysis of U.S. patent citation data found that setting the depreciation rate at 0.22 and the diffusion rate at 0.03 more accurately reflects the real value of patents, capturing the dynamics of knowledge depreciation and diffusion in technological research. This parameter ratio has been widely adopted and validated in subsequent studies on technological innovation [34,54,55,56]. Based on Popp’s setting [8], this study sets the depreciation rate (${\mathsf{\mu }}_1$) at 0.22 and the diffusion rate (${\mathsf{\mu }}_2$) at 0.03. The renewable energy patent data were sourced from the Patent Search and Analysis System of the State Intellectual Property Office of China, covering 30 provinces from 2010 to 2021. Patents were identified through a systematic search process, utilizing applicant’s address, IPC codes, and grant dates. The IPC codes were selected based on their relevance to renewable energy categories, including solar, wind, hydropower, ocean energy, biomass, and energy storage technologies.

(3)

Mediator variables

The optimization of the industrial structure (IS) reflects the transition from energy-intensive industries to higher value-added sectors. Industrial structure optimization is measured by the ratio of the value added in the tertiary sector to the value added in the secondary sector, a common indicator of shifts toward advanced and service-oriented industries. Renewable energy generation (REG) is quantified by the ratio of electricity generated from renewable sources to total electricity generation. This measure reflects the incorporation of renewable energy into the energy system, signaling progress toward a cleaner and more sustainable energy mix. Energy efficiency (EE) is expressed as the ratio of GDP to energy consumption, providing a measure of how effectively an economy converts energy inputs into economic output.

(4)

Threshold variables

Environment regulation (ER) is measured as the proportion of investment in industrial pollution control to the industrial value added. This indicator reflects the extent to which industries are investing in pollution mitigation relative to their economic output, capturing the stringency of environmental regulatory policies. R&D investment (RD) is defined as the ratio of government science and technology expenditures to total fiscal expenditures. This measure highlights the government’s prioritization of TI with its overall budget, reflecting financial support for R&D activities.

(5)

Control variables

To control for socio-economic and energy-related factors influencing CP, we include the following control variables based on Yang et al. [57], Su et al. [58], and Liu et al. [59]: urbanization rate (UR), measured as the ratio of the urban population to the total resident population of the region, energy structure (ES), measured as the share of coal consumption in total energy consumption, trade openness (OPEN), expressed as the ratio of total import and export trade to GDP, economic development level (PGDP), measured as the ratio of GDP to the total resident population, and energy intensity (EI), measured as the ratio of total energy consumption to GDP.

4.3. Data Sources

To ensure data continuity and consistency, this study focuses on a sample range covering 30 provinces in China from 2010 to 2021. Data are sourced from the China Statistical Yearbook, China Energy Statistical Yearbook, and respective provincial statistical yearbooks. For GDP-related indicators, 2010 is set as the base year, and real values are calculated using the regional GDP deflator. Certain variables are log-transformed, and missing data are handled through interpolation where necessary. The descriptive summary of all variables is presented in

Table 1. Descriptive summary.

Variables	Obs	Mean	S.D	Min	Max
CP	360	0.821	0.542	0.097	3.703
RETI	360	3.335	1.285	0.109	6.259
UR	360	0.589	0.125	0.338	0.896
ES	360	0.386	0.150	0.006	0.687
OPEN	360	0.291	0.326	0.007	1.659
PGDP	360	5.273	2.646	1.323	15.983
EI	360	0.817	0.481	0.247	3.767
IS	360	0.901	0.549	0.490	3.624
REG	360	0.266	0.238	0.016	0.919
EE	360	1.560	0.677	0.265	4.044
ER	360	0.003	0.003	0.000	0.026
RD	360	0.021	0.016	0.004	0.130

:

5. Empirical Results

5.1. Identification of Key Variables Features

Figure 1 shows the spatial distribution of CP across 30 provinces in China for 2010 and 2021. Darker colors indicate higher CP levels. The highest CP levels are mainly found in the eastern provinces, including Beijing, Shanghai, Jiangsu, Zhejiang, and Guangdong. In contrast, lower CP levels, shown in lighter colors, are observed in northeast China (Heilongjiang, Liaoning), north China (Hebei, Shanxi, Inner Mongolia), and northwest China (Shaanxi, Gansu, Ningxia). Provinces with higher CP levels are clustered in a pattern extending from the southeastern coastal regions to the central provinces. This regional disparity is driven by differences in economic structures and industrial compositions. The eastern provinces, with a focus on service-oriented industries and high-end manufacturing, have developed a comprehensive green industrial structure. This has led to higher economic returns and lowers fossil fuel consumption, significantly reducing regional carbon intensity. In contrast, the western provinces have retained and even expanded energy-intensive industries, including traditional sectors such as chemical and steel production, which have been transferred from the eastern region. This industrial composition intensifies pollution and carbon emissions, hindering improvements in CP. Furthermore, the western economy’s heavy reliance on resource extraction, coupled with outdated production technologies and high energy consumption, worsens environmental degradation and results in persistently lower CP levels.

Figure 2 shows the spatial distribution of RETI across 30 provinces in China for 2010 and 2021, revealing a clear nationwide increasing trend. The eastern provinces of Beijing, Jiangsu, Shandong, Shanghai, Zhejiang, and Guangdong have the highest RETI levels (darker colors), while the northwestern provinces of Qinghai, Xinjiang, Gansu, and Ningxia exhibit relatively low levels (lighter colors). This spatial disparity highlights the impact of regional economic development and industrial structure on RETI. The eastern coastal provinces, benefiting from financial resources, skilled human capital, and advanced energy infrastructure, have accelerated the adoption of renewable energy technologies. These provinces, as frontline windows of trade openness, foster international cooperation and cross-border knowledge spillovers in RETI research. The flow of foreign investment strengthens RETI by enabling domestic enterprises to acquire advanced renewable energy technologies, equipment, and professional experience through observation, learning, and knowledge-sharing. Conversely, the less developed northwest provinces, despite their abundant wind and solar energy resources, face significant RETI challenges. High initial investment costs and weak economic support slow the large-scale deployment of renewable energy technologies in these provinces. Notably, southwest provinces show relatively high RETI levels, largely due to their abundant hydropower resources. The implementation of the Western Development Strategy has driven population influx, industrial expansion, and growing energy demand in these provinces. Renewable energy, especially hydropower, has helped bridge the energy demand gap, creating a demand-pull effect that has boosted RETI efforts in southwestern provinces, further enhancing their RETI levels.

5.2. Baseline Regression Analysis

The Hausman test yields a p-value below 0.01, suggesting that the fixed effects model is preferred for the analysis.

Table 2. Baseline regression results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
RETI	0.308 ***	0.423 ***	0.393 ***	0.385 ***	0.369 ***	0.375 ***
	(0.103)	(0.098)	(0.089)	(0.086)	(0.083)	(0.084)
UR		−5.749 ***	−6.123 ***	−5.440 ***	−4.769 ***	−4.760 ***
		(1.537)	(1.378)	(1.257)	(1.298)	(1.296)
ES			−1.187 ***	−1.174 ***	−1.151 ***	−1.122 ***
			(0.257)	(0.259)	(0.275)	(0.272)
OPEN				−0.246 ***	−0.258 ***	−0.267 ***
				(0.083)	(0.080)	(0.085)
PGDP					0.030 ***	0.032 ***
					(0.009)	(0.009)
EI						−0.046 **
						(0.021)
Constant	−0.114	2.621 ***	3.422 ***	3.158 ***	2.739 ***	2.747 ***
	(0.235)	(0.726)	(0.719)	(0.655)	(0.769)	(0.763)
N	360	360	360	360	360	360
Province fixed	YES	YES	YES	YES	YES	YES
Year fixed	YES	YES	YES	YES	YES	YES
R2	0.640	0.780	0.811	0.813	0.815	0.817

Note: *** and ** indicate significance levels of 1% and 5%, respectively; standard errors are reported in parentheses.

presents the results of examining the effect of RETI on CP. In Column (1), excluding control variables, RETI shows a significant positive effect on CP, with a coefficient of 0.308 (p-value < 0.01). After incorporating all control variables, the coefficient of RETI stabilizes at 0.375 (p-value < 0.01). The results indicate a consistently positive relationship between RETI and CP, regardless of the inclusion of control variables, thus supporting the validity of H1. First, RETI reduces the cost of renewable energy production while improving the efficiency of clean energy conversion [60]. The increased renewable energy supply enhances the fossil fuel substitution effect, thereby reducing carbon emissions from heat production, heavy industries, and the construction sector. Additionally, RETI increases the share of electricity generated from wind, solar, and hydropower, reducing reliance on coal-fired power generation and increasing its positive impact on CP. Second, RETI promotes the development of a comprehensive industrial ecosystem by driving the synergistic growth in upstream raw materials, midstream equipment manufacturing, and downstream application industries. Beyond building industrial chains, RETI drives the expansion of related sectors into high-value-added areas such as R&D and design, system integration, and carbon asset management. This process enhances the spatial agglomeration of green production factors, making RETI a key engine in promoting green employment and driving sustainable regional economic growth. As a result, regional CP has been significantly enhanced.

It is necessary to explain the effects of the control variables on CP. The results indicate that UR has a significant negative effect on CP, with a 1% significance level, indicating that urbanization reduces CP. It is likely due to the high concentration of population and economic activities in urban areas, which results in increased energy consumption and higher carbon emissions. ES also exhibits a significantly negative coefficient, suggesting that a higher reliance on high-carbon energy sources worsens carbon emissions and hinders regional CP. Similarly, the coefficient of OPEN is significantly negative, implying that higher levels of international trade may reduce CP. The finding can be attributed to certain regions engaging in trade by hosting highly polluting manufacturing activities, creating a “pollution haven effect”. Conversely, PGDP has a significantly positive effect on CP, implying that as regional economic development progresses, there is a growing demand for improved environmental quality. Consequently, production activities tend to adopt cleaner technologies, leading to enhancing CP. However, EI shows a significant negative impact on CP, highlighting that the current economic development model is overly dependent on energy inputs, which reduces carbon efficiency and impedes progress toward sustainability.

5.3. Robustness Tests

This study employs five approaches to test the robustness of the empirical results:

(1)

Alternative explanatory variables. Following Lin and Zhu [34], this study uses the count of renewable energy patent applications (RETI_2) as an alternative proxy for RETI. To recalibrate the knowledge stock associated with RETI, the parameters μ₁ and μ₂ in Equation (5) are adjusted to 0.36 and 0.3, respectively. As shown in column (1) of

Table 3. Robustness test results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
RETI_2	0.177 ***
	(0.069)
RE		0.304 ***
		(0.111)
RETI			0.372 ***	0.360 ***	0.337 ***
			(0.081)	(0.094)	(0.076)
L.RETI						0.373 ***
						(0.086)
FDI			2.742 ***
			(1.071)
Control	YES	YES	YES	YES	YES	YES
N	360	360	360	312	360	330
Province fixed	YES	YES	YES	YES	YES	YES
Year fixed	YES	YES	YES	YES	YES	YES
R2	0.782	0.800	0.825	0.774	0.835	0.819

Note: *** indicates a significance level; standard errors are reported in parentheses.

, the sign and significance of the coefficient are consistent with the baseline regression, confirming the robustness of the results. Meanwhile, R&D expenditures reflect overall technological input and are commonly used to measure the level of RETI [53]. Therefore, this study further uses the share of science and technology expenditures in regional GDP as an alternative proxy for RETI. As shown in column (2), the regression result of R&D expenditure (RE) is significantly positive, proving the accuracy of H1.

(2)

Adding control variable. Foreign direct investment (FDI) can have dual effects on regional environment performance. While FDI may cause pollution by relocating energy industries, it also facilitates knowledge transfer, technological innovation, and financial support. To capture this effect, we include FDI as a control variable in Equation (1). As shown in column (3), including FDI does not change the sign or significance of the RETI coefficient, maintaining the expected results.

(3)

Excluding specific samples. Directly governed municipalities have distinct advantages over regular provinces in terms of economic structure, industrial composition, and the concentration of scientific resources, which could influence the regression results. Therefore, we exclude four directly governed municipalities from the full sample. From column (4), the coefficient for RETI remains significantly positive.

(4)

Removing outliers. Extreme values can distort regression outcomes. To mitigate this, we truncate all variables at the 1% level on both ends. From column (5), the results for RETI are similar to the baseline regression, confirming the robustness of the results.

(5)

Explanatory variable lags. TI often requires time to transition from research outcomes to practical implementation, leading to a delayed impact on economic and environmental performance. To account for this inherent lag, this study incorporates the one-period lagged term of RETI (L.RETI) as an explanatory variable. In column (6), the coefficient of RETI remains significantly positive, indicating that even with the lagged explanatory variable, the results continue to support the hypothesis that RETI enhances CP.

5.4. Endogeneity Test

In this paper, we use IV-2SLS and SYS-GMM estimation to mitigate potential endogeneity issues arising from omitted variables bias, reverse causality, and measurement error. First, following Yang et al. [61], we use the history of the opening of the commercial port as the proxy for RETI. China maintained a closed-door policy from the Qing Dynasty until the mid-19th century when foreign invasions and Western industrial advancements gradually opened the country to international trade. The early establishment of commercial ports fostered a favorable business environment for technological innovation, which facilitated knowledge spillovers supporting RETI, thereby satisfying the relevance condition of IV. Moreover, as demonstrated by Yan et al. [28], improvements in CP do not affect a province’s historical trade openness, ensuring that the IV meets the exogeneity and exclusivity requirements. Accordingly, the number of years from the opening of the commercial port of its capital city to 2021 is used as the instrumental variable for RETI. Second, we choose the number of research institutes in each province in 1997 as an additional proxy for RETI. The number of research institutes is the main source of technological innovation, which reflects the regional scientific research foundation and technological development level, and can have a significant positive impact on the RETI. Given that the number of research institutes in 1997 is a historical variable that has no direct impact on current CP, it meets the exogeneity requirement. Since the number of research institutes in 1997 is cross-sectional data, we construct an IV by interacting the lagged number of renewable energy patents with the number of research institutes in 1997. Columns (1) and (3) of

Table 4. Endogeneity test results.

Variables	The First Stage	The Second Stage	The First Stage	The Second Stage	SYS-GMM
	(1)	(2)	(3)	(4)	(5)
IV	1.147 ***		0.135 ***
	(0.345)		(0.041)
L.CP					0.954 ***
					(0.070)
RETI		0.329 ***		0.021 ***	0.212 **
		(0.124)		(0.008)	(0.099)
Control	YES	YES	YES	YES	YES
N	360	360	360	360	330
Province fixed	YES	YES	YES	YES	YES
Year fixed	YES	YES	YES	YES	YES
Kleibergen–Paap rk LM statistic		10.241		9.774
Chi-sq (1) p-value		0.0029		0.0014
Cragg–Donald Wald F statistic		34.004		68.753
AR(1)					0.004
AR(2)					0.982
Sargan					0.131

Note: *** and ** indicate significance levels of 1% and 5%, respectively; standard errors are reported in parentheses.

show the first-stage regressions for the two IVs, both yielding significantly positive coefficients, proving their relevance. Columns (2) and (4) display the second-stage regressions, where the Kleibergen–Paap rk LM test yields p-values of 0.0029 and 0.0014, passing the unidentifiable test. The Cragg–Donald Wald F statistics exceed the critical threshold of 16.38 at the 10% significance level, eliminating concerns regarding weak instruments. The sign and significance of the RETI on CP remain unchanged, indicating that the introduction of IVs mitigates endogeneity issues.

In addition, this paper adopts one-period lagged explanatory variables as IV and applies the SYS-GMM model to deal with the possible endogeneity between RETI and CP. As shown in column (5), the p-value of Sargan’s test is 0.131, which is not statistically significant at the 10% level, confirming that the IV satisfies the exogeneity condition. The p-value of the AR(2) test is 0.982, which is not statistically significant at the 10% level, indicating no presence of second-order autocorrelation. The regression results show that the lagged term of CP has a coefficient of 0.954, which is significant at the 1% level, suggesting strong time dependence in CP. The coefficient of RETI is significantly positive, demonstrating that the positive effect of RETI on CP is robust across different estimation methods. This finding further validates the robustness of the regression results in addressing endogeneity issues.

5.5. Heterogeneity Analysis

5.5.1. Regional Heterogeneity

Given the variations in economic strength and resource endowment across different provinces in China, analyzing the full sample may result in offsetting internal effects, potentially distorting the overall results. To address this, we divide the 30 provinces into east, central, and west to explore the regional heterogeneity in the impact of RETI on CP. The results are shown in columns (1)–(3) of

Table 5. Regional and economic development level heterogeneity analysis results.

Variables	Eastern Region	Central Region	Western Region	Developed Provinces	Less Developed Provinces
	(1)	(2)	(3)	(4)	(5)
RETI	0.298 ***	0.238 **	0.104	0.392 ***	0.252
	(0.090)	(0.107)	(0.095)	(0.066)	(0.156)
Control	YES	YES	YES	YES	YES
N	156	72	132	180	180
Province fixed	YES	YES	YES	YES	YES
Year fixed	YES	YES	YES	YES	YES
R2	0.893	0.813	0.740	0.876	0.758

Note: *** and ** indicate significance levels of 1% and 5%, respectively; standard errors are reported in parentheses.

.

In column (1), the coefficient of RETI is 0.298 (p-value < 0.01), indicating that RETI significantly enhances CP in the eastern provinces. First, the eastern region, especially the coastal provinces, enjoys a robust economic foundation, high foreign investment, and a well-developed talent pool, providing essential financial, managerial, and human resources for R&D in renewable energy technologies. This favorable environment stimulates local technological innovation and accelerates the transformation of research outcomes into practical applications, thereby improving long-term carbon efficiency. Second, the eastern region has well-developed industrial clusters, supported by mature manufacturing chains and strong upstream and downstream linkages. These factors create economies of scale and collaborative innovation networks, which reduce the costs of developing and applying renewable energy technologies, thus strengthening the positive effect of RETI on CP. Additionally, the region has been an early pilot area for market-based mechanisms such as carbon trading, green certificate trading, and green finance. Regional integration strategies, including those in Beijing–Tianjin–Hebei and the Yangtze River Delta, have further promoted collaborative innovation and the sharing of renewable energy technologies, providing strong institutional support for RETI. Finally, the eastern region has relatively limited fossil fuel reserves, and the mismatch between low fossil energy endowment and high electricity demand accelerates the adoption of renewable energy. The southeastern coastal provinces have abundant renewable resources and a well-established power grid infrastructure, which facilitates the large-scale deployment of wind, solar, hydropower, and tidal energy projects. The combination of strong market demand and a well-integrated regional energy layout reinforces RETI’s role in expanding renewable energy development, transmission, and utilization, ultimately enhancing CP.

In column (2), the coefficient of RETI in the central region is 0.238 (p-value < 0.05), indicating a significant but smaller effect compared to the eastern region. As China’s energy hub and a high-tech industrial base, and with the support of the national strategy for the revitalization of central China, the central region enjoys favorable policy frameworks, financial resources, and a skilled workforce, which collectively create conducive conditions for advancing renewable energy technologies. The central region is actively pursuing an energy transition, diversifying its energy mix by incorporating renewable sources such as wind, biomass, and geothermal energy. However, the full potential of RETI in this region is constrained by its continued reliance on traditional energy systems and the slower pace of infrastructure development compared to the eastern region. Despite these challenges, as the central region gradually adopts renewable energy technologies, the growing acceptance of these technologies is expected to intensify the crowding-out effect of R&D in fossil fuel-based energy technologies, spurring the growth of renewable energy consumption and, consequently, enhancing CP.

In column (3), the coefficient for RETI on CP in the western region is positive but statistically insignificant, indicating that RETI has a limited impact on CP in this region. Despite the potential for wind energy in the northwest and hydropower in the southwest, the efficient utilization of these renewable resources requires mature technical support and advanced energy infrastructure, both of which demand extensive capital investment. Indeed, the western region faces significant income disparity, which creates a major barrier to the transformation and application of RETI outcomes. While these provinces have abundant renewable resources, they also face high production costs and intermittent power generation, especially in wind and solar energy. These challenges drive increased demand for fossil energy consumption, creating a “rebound effect” that diminishes the potential benefits of RETI. In less developed provinces, citizens often prioritize higher income levels over environmental concerns. To stimulate economic growth, these provinces frequently invest in heavy industries. In some cases, local governments may even subsidize fossil fuel consumption to reduce industrial production costs, further increasing reliance on non-renewable energy sources. These policies reduce the incentive for enterprises to invest in renewable energy technologies, diminishing the impact of RETI on CP.

5.5.2. Heterogeneity in the Level of Provincial Economic Development

It has been shown that the carbon reduction effect of RETI materializes only when income reaches a certain level, with its marginal effect increasing as income grows. To test whether RETI has a heterogeneous effect on CP across regions with different economic levels, this study adopts GDP per capita as an indicator of economic development. Provinces are categorized into developed and less developed groups based on whether their GDP per capita is above or below the median.

As shown in column (4) of Table 5, the coefficient of RETI for developed provinces is 0.392 and is significantly positive at the 1% level, indicating a more pronounced impact of RETI on CP in high-income provinces. Column (5) shows that the coefficient of RETI for less developed provinces is 0.252, but it does not pass the significance test, indicating that the positive impact of RETI on CP is not evident in underdeveloped provinces. Several factors may explain this disparity. First, developed provinces generally have advanced industrial structures, characterized by a larger share of service and high-end manufacturing industries. These provinces also exhibit a greater capacity for adopting and integrating RETI resources, effectively reducing carbon emissions and enhancing CP. In contrast, less developed provinces are in the industrialization stage, with an industrial structure dominated by high-pollution and high-energy-consumption industries, relying more on high-carbon energy sources to meet economic growth needs. In these regions, the adoption and development of clean energy technologies are relatively difficult, and deployment and application bottlenecks make it difficult to significantly enhance CP in the short term. Second, developed provinces benefit from robust market incentives and comprehensive policy frameworks for green development. Carbon pricing mechanisms, green bonds, and strict environmental regulations facilitate the market-oriented application of RETI, improving the return on technological investment and accelerating CP growth. Conversely, less developed provinces lack sufficient policy incentives and immature green financial markets, leading to weaker investment motivation for RETI, thus weakening its impact on CP over time. Finally, developed provinces have more advanced energy infrastructures, including new energy grid integration, power dispatch, and smart grids, and developed grid facilities enhance the capacity of renewable energy acceptance, enabling innovation results to be rapidly converted into real productivity. In contrast, less developed provinces face energy infrastructure deficiencies, leading to frequent occurrences of wind and solar curtailment, which restricts the positive contribution of RETI to CP.

5.5.3. Technological Innovation Heterogeneity

Given the different levels of innovation and application scales among renewable energy technologies, we classify RETI into six subtypes: solar, wind, hydropower, ocean, biomass, and energy storage. We then examine the individual impacts of these six subtypes on CP, as shown in columns (1)–(6) of

Table 6. Technological innovation heterogeneity analysis results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
SETI	0.319 ***
	(0.066)
WETI		0.181 ***
		(0.068)
HETI			0.069
			(0.048)
OETI				0.051
				(0.053)
BETI					0.066
					(0.060)
ESTI						0.084 *
						(0.045)
Control	YES	YES	YES	YES	YES	YES
N	360	360	360	360	360	360
Province fixed	YES	YES	YES	YES	YES	YES
Year fixed	YES	YES	YES	YES	YES	YES
R2	0.821	0.791	0.775	0.773	0.775	0.783

Note: *** and * indicate significance levels of 1% and 10%, respectively; standard errors are reported in parentheses.

.

While all types of RETI consistently contribute to enhancing CP, only solar, wind, and energy storage TIs have a significant positive impact. Specifically, solar and wind TIs significantly enhance CP at the 1% level, reflecting their technological maturity and strong market demand. Both solar and wind technologies have achieved great breakthroughs in reliability, power generation costs, infrastructure completeness, and industrialization. As the main sources of renewable energy in China, these technologies are crucial to the national power supply. Solar and wind TIs are more readily translated into emission reduction benefits, resulting in a substantial impact on CP.

The impact coefficient of ESTI on CP is 0.084 (p-value < 0.1), indicating that energy storage TI has a certain positive but relatively limited impact on CP, which may stem from the early stage of energy storage technology development in China. Theoretically, energy storage TI can enhance CP by improving the stability of renewable energy integration into the power grid, optimizing power dispatch, and increasing the utilization of wind power and photovoltaic power generation [62]. However, energy storage technology remains in its early commercialization stage and faces significant uncertainties, including technological challenges, subsidy constraints, and limited market acceptance. In particular, the energy density, charging–discharging efficiency, cycle life, and cost-effectiveness of current energy storage systems lag behind international standards, limiting their economic feasibility and large-scale deployment. Despite its limited current impact on CP, recent major advancements in solid-state batteries, lithium-ion batteries, pumped storage, heat pump storage, and hydrogen storage indicate strong long-term potential. As technological breakthroughs accelerate, costs decline, and financial support grows, energy storage is likely to play an increasingly important role in enhancing CP in the future.

The impact of other energy TIs (hydropower, ocean, and biomass) on CP is statistically insignificant. This is likely due to technological bottlenecks, uneven resource distribution, and a limited market, which currently constrain their contribution to CP.

5.6. Mechanism Test

This study explores whether industrial structure, renewable energy generation, and energy efficiency are the channels through which RETI impacts CP. The findings are shown in

Table 7. Mechanism test results.

Variables	(1)	(2)	(3)	(4)
	CP	IS	PEG	EE
RETI	0.375 ***	0.039 **	0.054 ***	0.284 ***
	(0.084)	(0.017)	(0.021)	(0.097)
Control	YES	YES	YES	YES
N	360	360	360	360
Province fixed	YES	YES	YES	YES
Year fixed	YES	YES	YES	YES
R2	0.817	0.851	0.945	0.900

Note: *** and ** indicate significance levels of 1% and 5%, respectively; standard errors are reported in parentheses.

.

In column (2), the coefficient of the impact of RETI on IS is 0.039 (p-value < 0.05), indicating that RETI significantly promotes industrial structure. It has been demonstrated that industrial structure upgrading positively contributes to regional CP [63,64]. The evolution towards knowledge-intensive and cleaner sectors not only reduces the negative environmental impacts of traditional sectors but also enhances green productivity. This transition maximizes economic benefits while reducing environmental costs, thereby enhancing CP. Consequently, RETI enhances CP by driving industrial structural transformation, confirming H2a. RETI facilitates cleaner production pathways for polluting enterprises, steadily enhancing CP by introducing carbon-free energy sources and green processes. Furthermore, RETI stimulates the growth of green manufacturing and service industries. The increasing economic weight of these sectors directly reduces the carbon intensity of production activities, further enhancing regional CP.

In column (3), the coefficient of the impact of RETI on REG is 0.054 (p-value < 0.01), suggesting that RETI significantly promotes electricity generation from wind, solar, and hydroelectric sources. The existing literature supports that increased renewable energy generation reduces the carbon intensity of power systems, positively contributing to CP [65]. Therefore, RETI enhances CP by increasing the share of clean electricity generation, confirming H2b. RETI effectively reduces the cost of renewable energy generation while simultaneously increasing output. As a result, renewable energy increasingly substitutes coal-fired power generation, reduces regional carbon emissions, and further enhances CP.

In column (4), the coefficient of the impact of RETI on EE is 0.284 (p-value < 0.01), indicating that RETI significantly enhances energy utilization efficiency. The existing literature generally recognizes that improving energy efficiency enhances CP [66]. Specifically, a reduction in energy consumption per unit of economic output reduces ${\mathrm{C}\mathrm{O}}_2$ emissions during energy consumption, thereby enhancing CP. This confirms H2c. RETI minimizes energy losses during the conversion and consumption of renewable electricity and lowers ${\mathrm{C}\mathrm{O}}_2$ emissions without compromising production, especially through the application of advanced energy management systems, which further enhances CP.

5.7. Threshold Effect Analysis

This study uses ER and RD as threshold variables to further examine the effect of RETI on CP. The results in

Table 8. Threshold effect test results.

Threshold Variables	Threshold Number	F-Value	p-Value	Threshold Value	95% Confidence Intervals
ER	Single	68.39 ***	0.000	0.0004	[0.0004, 0.0005]
	Double	19.70 *	0.057	0.0015	[0.0015, 0.0016]
RD	Single	60.04 ***	0.000	0.0146	[0.0139, 0.0151]

Note: *** and * indicate significance levels of 1% and 10%, respectively; standard errors are reported in parentheses.

are based on a test using 300 bootstrap samples. The threshold analysis shows that ER passes the double-threshold test, with the first and second thresholds at 0.0004 and 0.0015, respectively. In contrast, RD passes only the single-threshold test, with a threshold of 0.0146.

Table 9. Threshold analysis regression results.

Threshold Variables	Threshold Interval	Estimated Coefficient	Control	Constant	N	R2
ER	ER ≤ 0.0004	0.356 ***	YES	2.785 ***	360	0.855
		(0.070)		(0.601)
	0.0004 < ER ≤ 0.0015	0.374 ***
		(0.071)
	ER > 0.0015	0.349 ***
		(0.068)
RD	RD ≤ 0.0146	0.279 ***	YES	3.556 ***	360	0.844
		(0.073)		(0.705)
	RD > 0.0146	0.328 ***
		(0.074)

Note: *** indicates a significance level; standard errors are reported in parentheses.

presents the results of the threshold analysis. The threshold analysis results clarify the non-linear relationship between RETI and CP under varying levels of ER and RD.

When ER is below the first threshold of 0.0004, the coefficient of RETI is 0.356, significantly positive at the 1% level, indicating that RETI positively impacts CP in regions with low ER. As ER increases to between 0.0004 and 0.0015, the coefficient rises to 0.374, suggesting that moderate ER further strengthens the role of RETI in enhancing CP. This finding supports the view that moderate ER activates an innovation compensation mechanism. In this context, enterprises are encouraged to invest in RETI and adopt more sustainable production practices. Moreover, moderate ER promotes the efficient allocation of green innovation resources, which flow more freely across industries, boosting RETI development and intensifying its positive impact on CP. However, when ER surpasses 0.0015, the coefficient decreases to 0.349, indicating that overly stringent ER may dampen the beneficial impact of RETI on CP. Stringent ER policies, such as steep carbon taxes, can increase operational costs for enterprises, diverting resources away from technological R&D and thereby diminishing the overall effectiveness of RETI in enhancing CP.

When RD is below the threshold of 0.0146, the coefficient of RETI is 0.279, statistically significant at the 1% level. When RD surpasses 0.0146, the coefficient increases to 0.328, showing that higher RD consistently strengthens the positive effect of RETI on CP, thereby confirming H4. The findings suggest that higher RD allows enterprises to obtain more R&D subsidies, alleviating financial constraints and effectively reducing potential risks associated with RETI. Additionally, increased government financial support for RETI sends a positive signal to the market, fostering a more favorable financing environment for enterprises. As a result, enterprises are more inclined to develop and apply renewable energy technologies, which enhances the effectiveness of RETI in enhancing CP. Although some studies have found that excessive RD may lead to resource misallocation, organizational inertia, and rent-seeking behavior, weakening the positive effect of RETI on CP [67], our results do not support this assertion within the observed period. A possible explanation is that government RD in RETI has not reached the saturation point where diminishing returns or inefficiencies emerge.

6. Conclusions and Implications

This study provided a comprehensive empirical analysis of the relationship between RETI and CP using panel data from 30 Chinese provinces from 2010 to 2021. The findings are as follows: First, CP shows an increasing trend in most regions, except for the western and northeastern regions where the change is insignificant. RETI also shows an increasing trend across all regions during the study period. Second, RETI significantly enhances CP, which rises by 0.375 units for every 1% increase in RETI, confirming its potential to support sustainable economic growth and reduce carbon emissions. Third, regional differences in the impact of RETI on CP are evident, with significant improvements in the eastern and central regions, while the effect in the western region remains insignificant. The impact of RETI on CP is stronger in economically developed provinces than in less developed ones. Among different RETI types, solar, wind, and energy storage TIs contribute more to CP growth. Fourth, RETI enhances CP by promoting an advanced industrial structure, boosting renewable energy generation, and improving energy efficiency. Finally, under the double-threshold effect of ER, the promotion of RETI on CP demonstrates a non-linear relationship that strengthens initially but then weakens. Under the single-threshold effect of RD, the promotion effect of RETI on CP demonstrates a continuously increasing non-linear effect.

Based on these findings, we propose the following recommendation:

Firstly, to further amplify the positive impact of RETI on CP, national innovation strategies should prioritize policy support for renewable energy technologies. This entails providing targeted incentives for universities, research institutions, and enterprises to collaborate on R&D in key areas such as energy generation, conversion, and storage. Policy efforts should focus on expanding renewable energy subsidy schemes and advancing green financial reforms to improve energy conversion efficiency and reduce power generation costs. Accelerating renewable energy marketization through feed-in tariffs, green certificates, and carbon trading rights can strengthen its competitive advantage in the market. In addition, integrating resources and fostering collaborative innovation in renewable energy engineering will stimulate green innovation across industries, expediting the transformation of technological achievements into tangible productivity gains. Although RETI has a positive impact on CP, China’s coal-dominated energy structure constrains its full potential. Therefore, a strategic transition from coal to renewable energy consumption is essential. The government should gradually reduce coal use, restrict the approval of new coal-fired power plants, regulate the scale of coal-fired power generation, and promote “zero-carbon mines” as demonstration projects to enhance CP.

Secondly, regions should develop tailored strategies for RETI based on their unique conditions. In the eastern and central regions, the focus should be on strengthening inter- and intra-regional technological cooperation, increasing financial investment, improving infrastructure, and cultivating renewable energy technology innovators to enhance overall innovation capacity. Policymakers should adopt targeted measures to strengthen the impact of RETI on CP in the western region. First, implementing tax exemptions for renewable energy equipment purchases and promoting low-interest green credits programs can reduce cost pressures on enterprises, thereby encouraging greater investment in renewable energy technologies. Second, policymakers should enhance renewable energy infrastructure, including ultra-high-voltage transmission networks and expanded energy storage capacity, to enable cross-regional clean electricity transmission. Finally, establishing demonstration projects for renewable energy technology applications, gradually phasing out fossil energy subsidies, and directing financial resources into renewable energy R&D can create a favorable environment for RETI.

Thirdly, solar and wind energy TIs are key drivers for enhancing CP, necessitating targeted policy incentives and financial support from governments. Implementing green financial support mechanisms is essential to providing renewable energy enterprises with stable financing channels. To strengthen the role of energy storage TIs in cleaner production, governments should prioritize energy storage grid integration, subsidize peak–valley electricity price differences, and improve the cost-effectiveness of energy storage systems. Promoting market-oriented energy storage trading, encouraging energy storage enterprises to participate directly in the spot electricity market, and accelerating commercialization are also crucial. Additionally, increasing R&D in diverse energy storage technologies, developing intelligent energy storage management systems, and addressing existing technological gaps will further improve efficiency. Although the impact of hydropower, ocean energy, and biomass TIs on CP is not currently significant, their long-term potential should not be overlooked. Tailored policy incentives, aligned with regional energy profiles, can unlock the contributions of these renewable energy sources to CP.

Finally, governments should implement policy incentives for RETI through special funds, tax breaks, loan subsidies, and low-carbon innovation rewards to encourage technological advancement and capital accumulation. Furthermore, ER should be maintained within a reasonable range while activating the innovation compensation mechanism to maximize the positive impact of RETI on CP.

Although the empirical results confirm that RETI is an effective means of enhancing CP, several unintended consequences merit attention. First, the large-scale deployment of renewable energy technologies may entail significant environmental costs. For example, renewable energy infrastructure, such as wind farms and solar power plants, requires substantial land resources, potentially disrupting natural habitats and exacerbating ecological degradation. The operation of wind turbines generates low-frequency noise, which may negatively affect local ecosystems and reduce biodiversity [68]. Additionally, the manufacturing, operation, and maintenance of renewable energy equipment contribute to waste generation and air pollution, potentially offsetting the expected environmental benefits. While these environmental concerns do not negate the positive impact of RETI on CP, they underscore the necessity of conducting thorough environmental impact assessments for renewable energy projects. Future policies should encourage the adoption of circular production models within the renewable energy industry and the continuous advancement of pollution-minimizing renewable energy technologies to mitigate these negative externalities. Second, developed regions with strong capital accumulation, complete industrial chains, and highly skilled human capital can rapidly absorb and apply RETI, translating it into economic dividends and further enhancing CP. However, less developed regions, especially resource-dependent provinces, face significant challenges in technological transformation. These regions often serve as energy exporters but struggle to capture the high-value benefits of renewable energy industry upgrading, widening the “green divide”. To avoid this problem, the government should establish cross-provincial RETI cooperation platforms to promote technology transfer from developed to less developed regions. Meanwhile, it is gradually exploring incentives for local consumption while ensuring that a portion of the green power proceeds is retained to support the transformation of local industries. Finally, the expansion of RETI will likely drive greater demand for renewable energy, with fossil energy-dependent provinces, such as Shanxi, among the first to experience adverse impacts. The decline of traditional industries, structural unemployment, diminished economic resilience, and widening income inequality are likely consequences. To address these economic displacement challenges, policymakers should adopt a just transition strategy by developing industrial convergence plans, providing re-employment training, creating green industries funds, and gradually reducing fossil fuel quotas on an annual basis.

While this study provides empirical insights into the relationship between RETI and CP, several aspects require further refinement. First, the accuracy and comprehensiveness of the indicators used in this study could be improved. For instance, future studies could adopt total factor carbon productivity, a metric that accounts for both inputs and outputs, to better evaluate carbon performance. In addition, this study mainly used patent data as a proxy for RETI, which may not accurately reflect the real-world adoption and effectiveness of renewable energy technologies. Incorporating metrics such as the renewable energy technology adoption rate or technology diffusion rate as supplementary indicators could offer a more comprehensive assessment of the impact of RETI on CP. Second, while this study examines the impact of RETI on CP from 2010 to 2021, the gradual nature of technology diffusion and policy impacts suggests that a longer time horizon may be necessary to capture long-term trends and dynamic effects. Extending the study period could provide deeper insights into the evolving relationship between RETI and CP. Finally, the influence of global market dynamics has not been fully considered. In real-world scenarios, factors such as trade frictions, instability in the supply of raw materials, carbon border adjustments, and fluctuations in international energy policies may affect the cost of renewable energy technologies, corporate investment decisions, and the pace of technology diffusion. These uncertainties, in turn, complicate the relationship between RETI and CP. Future research could further investigate the moderating role of global market dynamics in this relationship to provide more targeted policy recommendations.