Carbon Trading Price and the Quantity and Quality of Green Technological Innovation: A Sustainability Perspective

Abstract

Sustainable development has become an important global goal for environmental protection and economic growth. Promoting environmental sustainability and green development has become an inevitable trend for global economic transformation. The carbon emission trading market (carbon market) is a crucial market-based mechanism for pricing greenhouse gas emissions, where carbon trading prices signal the costs of emission reduction and drive firms to engage in green technology innovation for a low-carbon transition. Using a sample of A-share listed companies in China’s eight carbon pilot regions from 2013 to 2024, this study employs a two-way fixed effects model to examine how carbon prices affect both the quantity and quality of corporate green technological innovation. Baseline regressions show that a one-unit increase in carbon prices significantly boosts green patent quantity (GreNum) by 0.018 and quality (GreQua) by 0.361, both at the 1% significance level. Mechanism analysis reveals that financing constraints act as a partial mediator, while environmental regulation and media attention further amplify the positive impact of carbon prices on corporate green technological innovation. Heterogeneity analysis further indicates that this positive effect is more pronounced in non-high-tech enterprises, China’s northern and eastern regions, and state-owned enterprises. This study provides empirical evidence for optimizing carbon market allocation, supporting corporate low-carbon transition, and promoting global environmental sustainability and green development.

1. Introduction

Global warming poses formidable challenges to natural ecosystems that are essential for human survival and development. The primary cause of this warming trend is the significant increase in global carbon emissions. To address this urgent issue, the international community has implemented a series of regulatory strategies, ranging from government-mandated measures to market-based carbon emissions trading schemes. However, government-imposed measures often suffer from high implementation costs, strong corporate resistance, and limited emission-reduction efficiency [1]. In contrast, market-based carbon emissions trading conveys cost signals for carbon abatement through pricing mechanisms, thereby guiding enterprises to actively adopt effective low-carbon transition strategies [2]. China launched pilot carbon emissions trading programs in eight provinces and municipalities in 2013 and later expanded the system nationwide in 2021. Carbon price signals incentivize corporate emission reductions, while quota trading optimizes resource allocation [3]. This approach promotes low-carbon transition pathways, such as green technological innovation, to achieve carbon reduction targets [4].

Green technological innovation serves as the cornerstone of corporate low-carbon transformation and is crucial for achieving China’s “dual-carbon” goals. In recent years, China has continuously strengthened policy support for green technological innovation. In 2021, the State Council issued the Action Plan for Carbon Peaking Before 2030, emphasizing that green and low-carbon technological innovation is a central task in achieving carbon peaking. The plan highlights the importance of enhancing the role of enterprises as the primary drivers of innovation, supporting firms in undertaking major national green and low-carbon science and technology projects, and accelerating the green and low-carbon technological revolution [5]. On the one hand, rising carbon prices increase the operating costs associated with traditional high-carbon technologies, forcing enterprises to shift toward low-carbon technologies in order to reduce compliance costs [6]. On the other hand, carbon pricing generates carbon asset returns, providing additional financial support for corporate green technological innovation [7]. The price signals from carbon emissions trading reshape firms’ cost–benefit structures, transforming low-carbon R&D from passive compliance into proactive investment [8]. Moreover, market expectations further guide enterprises toward continuous innovation [9].

Against this backdrop, it is essential to investigate the relationship between carbon emissions trading prices and corporate green technological innovation. This study puts forward three key research questions: how carbon emissions trading prices affect the quantity and quality of corporate green technological innovation, whether environmental regulations and media attention play moderating roles in this relationship, and to what extent financing constraints mediate the above impact. The relationship between carbon emission trading prices and corporate green technological innovation is affected by both governmental institutional factors and societal market factors. At the governmental level, environmental regulations raise carbon emission permit prices by imposing stricter emission constraints, thus pushing firms to conduct green technological innovation to reduce emissions [10,11]. At the societal level, media attention amplifies market signals through reputation mechanisms, increasing firms’ sensitivity to carbon emission cost fluctuations and accelerating green technological innovation [12,13]. Furthermore, financing constraints act as an important transmission channel. Rising carbon prices may increase corporate operation costs and financing risks, thus aggravating financing constraints, while carbon pricing can also send positive signals to the capital market and ease financing constraints, indicating that the impact of carbon prices on green innovation is partly transmitted through corporate financial conditions.

Given this context, this paper takes carbon price as the core variable and constructs the annual weighted average carbon price indicator based on daily trading data. This study examines the impact of carbon emissions trading prices on green technological innovation using data of listed firms in eight national pilot regions from 2013 to 2024. Most existing studies focus on the policy dummy of carbon trading pilots instead of real carbon prices, while this study uses actual transaction prices to capture market signals and identify the governance effect of carbon market mechanisms more accurately. The contributions of this paper are threefold: first, it expands carbon trading research by analyzing the effect of actual carbon price on the quantity and quality of corporate green patents; second, it adopts the weighted average carbon price to reflect real market signals with more accurate and credible empirical results; third, it introduces environmental regulation, media attention and financing constraints into the analytical framework, clarifying the internal transmission mechanism of carbon price affecting corporate green technological innovation.

The remainder of this paper is organized as follows. First, we review the relevant literature and develop theoretical hypotheses regarding the impact of carbon emissions trading prices on corporate green technological innovation, as well as the moderating and mediating mechanisms underlying this relationship. Next, we define the key variables and specify the empirical models, including the benchmark regression model, moderation models, and mediation models. Subsequently, we present the empirical results and analyses, including descriptive statistics, baseline regression results, endogeneity tests, robustness checks, tests of the moderating effects of environmental regulation and media attention, an examination of the mediating role of financing constraints, and heterogeneity analyses based on technological capability and regional differences. Finally, we conclude with a summary of the key findings and corresponding policy implications.

2. Literature Review and Research Hypotheses

2.1. Carbon Emission Trading Prices and Corporate Green Technological Innovation

Corporate green technological innovation is shaped by a complex interplay of organizational and contextual factors. Internally, scholars such as Chenavaz et al. (2023) [14] find that digital inclusive finance can improve firms’ corporate green technological innovation by easing financing constraints. Similarly, Han et al. (2021) [15] highlight that inclusive digital finance enhances the innovation performance of high-tech enterprises by alleviating both debt and equity financing constraints, while Li et al. (2023) [16] confirm that digital financial inclusion improves financial efficiency and thus exerts a positive effect on firms’ corporate green technological innovation capabilities. However, most of these studies concentrate on the independent effects of financial resources or digital tools, while neglecting the coupling effect between carbon market signals and financial transmission channels. Externally, Liu et al. (2024) [17] reveal that China’s Green Finance Reform and Innovation Pilot Zones (PZGFRI) policy significantly reduces corporate carbon emissions by enhancing corporate green technological innovation and optimizing energy structures. Cifuentes-Faura et al. (2026) [18] further demonstrate that digital financial inclusion promotes regional carbon neutrality by supporting green technological innovation and facilitating low-carbon industrial development, with a particularly pronounced effect in China’s central regions. Nevertheless, the existing literature rarely integrates carbon price fluctuations, which represent core market signals of the emission trading system, into the analytical framework of green finance, and insufficiently explores their synergistic governance effects on micro-enterprise innovation.

While existing studies have examined the macro-level policy effects of carbon emissions trading systems (ETS) and the synergy between green finance and green development, systematic research on how carbon price dynamics directly drive corporate green technological innovation through the carbon finance transmission mechanism remains relatively scarce. Existing studies have provided empirical evidence for the innovation-driven effect of carbon pricing. Cui et al. (2018) [19] conducted a quasi-natural experiment based on China’s carbon trading pilots and confirmed that carbon pricing mechanisms can significantly induce low-carbon technological innovation at the empirical level, providing core evidence from China. Lin et al. (2018) [20] used energy prices as a proxy for carbon prices and found that higher carbon prices effectively incentivize clean technology R&D in ETS-covered sectors, promoting the growth of related patents. Although carbon prices have a slight negative impact on overall R&D activities, the effect is relatively small. Nevertheless, these studies only focus on the direct impact of carbon prices, while neglecting the role of financial channels in the transmission process. Ding et al. (2022) [21] find that digital finance can curb regional carbon emissions by promoting corporate green technological innovation, thereby providing a foundation for integrating carbon pricing with financial support mechanisms. Martin et al. (2016) [22] further show that carbon pricing policies exert a significant positive impact on low-carbon innovation, with stronger effects observed in countries with more developed financial systems. However, prior studies have not systematically uncovered the micro-transmission paths through which carbon price signals guide financial resource allocation and stimulate green innovation. They also fail to clarify the interactive logic between carbon prices, financing constraints, and green finance policies, leaving the core mechanism of carbon finance still to be clarified. In addition, most studies focus on the policy impact of ETS implementation rather than the price signal effect, which is the core of market-based environmental regulation. Few studies have revealed how continuous changes in carbon prices affect enterprises’ expected returns, financing availability, and innovation investment decisions.

To address this research gap, this study develops a theoretical framework for the carbon finance transmission mechanism based on green finance theory and the Porter Hypothesis, and clarifies the micro-level transmission pathway through which carbon emission trading prices influence corporate green technological innovation. Specifically, rising carbon prices not only increase the opportunity cost of high-carbon production and enhance the expected returns of corporate green technological innovation projects, but also improve the financial support environment for corporate green technological innovation through the carbon finance transmission mechanism. First, financial institutions adjust their credit allocation strategies in response to carbon price signals, providing preferential lending conditions for green research and development (R&D) projects [23]. Second, carbon price signals improve the environmental, social, and governance (ESG) performance of green enterprises, facilitating their access to green bond financing and expanding funding sources for corporate green technological innovation [24]. Third, green finance policies (such as PZGFRI) and digital financial inclusion act as dual bridges connecting carbon price signals with micro-level enterprises, reducing information asymmetry in corporate green technological innovation financing and improving the efficiency of capital allocation to green technology R&D [17,18].

Under the combined effects of carbon price-induced cost constraints and the financial support provided by the carbon finance mechanism, enterprises tend to reallocate both internal and external resources toward green technological innovation activities. This process accelerates the research, development, and commercialization of clean technologies, ultimately improving the level of corporate green technological innovation. Within the ETS framework, carbon prices function as dynamic market signals, with their fluctuations directly reflecting the scarcity of emission allowances and the cost of emission reduction. These signals are transmitted in real time through market transactions and carbon finance systems, guiding enterprises’ strategic decisions regarding production optimization and R&D investment. As a result, a positive feedback loop emerges among carbon pricing, financial resource allocation, and corporate green technological innovation. Consequently, increases in carbon prices not only impose cost pressures on high-carbon production but also activate the carbon finance transmission mechanism, channeling more financial resources into corporate green technological innovation. This process ultimately encourages enterprises to accelerate the transition toward low-carbon technologies and achieve sustainable development goals. Based on the above analysis, this study proposes the following hypothesis:

Hypothesis 1:

Carbon emissions trading prices promote both quantitative and qualitative improvements in corporate green technological innovation.

2.2. The Moderating Effect of Environmental Regulation

To further explore the boundary conditions of the impact of carbon emissions trading prices on corporate green technological innovation, this study chooses environmental regulation as a moderating variable. Environmental regulation, as a typical formal institutional factor, shapes the institutional environment in which carbon price signals operate and affects firms’ incentives and capabilities to conduct green innovation. It can influence the intensity of firms’ responses to carbon prices and thus plays an indispensable moderating role. Based on Porter hypothesis and institutional theory, appropriate environmental regulation can form effective institutional constraints, enhance the guiding effect of carbon market signals, stimulate enterprises’ innovative initiative, and strengthen the promoting effect of carbon prices on green innovation, which constitutes the important theoretical basis of this study.

Institutional economics theory suggests that environmental regulation constrains corporate environmental behavior through formal institutional arrangements and market-based policy instruments, such as emission standards and environmental taxation. According to Porter and van der Linde (1995) [5], stringent environmental regulations may generate an “innovation compensation” effect, whereby firms increase R&D investment and strengthen their corporate green technological innovation capabilities in order to offset regulatory compliance costs. Subsequent empirical studies by Johnstone et al. (2010) [25] and Ambec et al. (2013) [26] provide further evidence supporting this argument, demonstrating that appropriately designed environmental policies can stimulate technological progress while simultaneously enhancing firms’ competitiveness. Within carbon emissions trading markets, environmental regulations influence carbon prices through mechanisms such as setting emission caps, allocating emission allowances, and establishing trading systems. Meanwhile, regulatory pressure encourages enterprises to optimize resource allocation and increase investment in green technological innovation. As environmental regulations become more stringent and carbon prices rise, firms face stronger incentives to reduce emission costs through technological upgrading and corporate green technological innovation activities, thereby promoting sustainable development.

Although a substitution effect between environmental regulation and carbon prices may exist under extreme regulatory pressure, such situations are not common in China’s current institutional context. At present, China’s environmental regulation is in a stage of gradual improvement and appropriate enhancement, rather than excessive restriction. In most cases, environmental regulation and carbon trading markets work together to strengthen the incentive effect on green innovation. Therefore, this study focuses on the synergistic effect between the two. Based on the above analysis, this study proposes the following hypothesis:

Hypothesis 2:

Environmental regulation positively moderates the relationship between carbon emissions trading prices and corporate green technological innovation.

2.3. The Moderating Effect of Media Attention

To further clarify the boundary conditions of the impact of carbon emissions trading prices on corporate green technological innovation, this study selects media attention as another moderating variable. As an important informal governance mechanism, media attention shapes the external supervision environment and public pressure faced by enterprises, affects the transmission and perception of carbon price signals, and thus plays a crucial moderating role in the relationship between carbon prices and green technological innovation. Based on media governance theory, media attention can strengthen external public supervision, reduce information asymmetry, and amplify the incentive effect of carbon price signals, which provides sufficient theoretical support for the selection of this moderating variable.

Under the framework of green development, media attention to corporate environmental behavior can function as an important form of informal regulation. As an external governance mechanism, media coverage helps reduce information asymmetry, increases public scrutiny of corporate environmental performance, and incentivizes firms to engage in corporate green technological innovation in order to maintain organizational legitimacy [27]. When carbon prices increase, enterprises face higher emission compliance costs. Media attention may further amplify the visibility of firms’ environmental performance, thereby strengthening external pressure on enterprises to adopt green technologies, reduce emissions, and enhance their environmental reputation [28]. In this context, media attention reinforces the incentive effect of carbon pricing on corporate green technological innovation. Conversely, when media attention is relatively weak, the signaling role of carbon prices may be attenuated, reducing firms’ responsiveness to market signals.

Although some may argue that media attention and carbon market signals may have overlapping governance functions, substitution effects between them are unlikely to occur in practice. Media attention mainly improves information transparency and forms external public supervision, while carbon prices provide real economic incentives for enterprises. The two belong to different governance mechanisms: one is external supervision, the other is market incentive. They complement rather than substitute each other. Therefore, this study focuses on the synergistic effect between media attention and carbon prices. Based on the above analysis, this study proposes the following hypothesis:

Hypothesis 3:

Media attention positively moderates the relationship between carbon emissions trading prices and corporate green technological innovation.

2.4. The Mediating Effect of Financing Constraints

While prior studies have identified multiple potential mechanisms linking carbon prices to green innovation, such as cost pressure [29], strategic motivation [30], and market signaling [31], we focus on financing constraints as the core mediating mechanism in this study. This choice is grounded in evidence highlighting the central role of financial frictions in transmitting carbon price signals to corporate innovation [32]. This setting is mainly based on the typical characteristics of green innovation projects, such as large capital investment, long payback period and high R&D risks. Financing constraints are the key factor restricting corporate innovation decisions and R&D investment, which can accurately identify the core channel through which carbon market signals affect corporate green technological innovation. By concentrating on this pathway, we can conduct a rigorous and in-depth analysis of the carbon finance transmission mechanism, avoiding overcomplicating the empirical framework and ensuring the robustness of our findings.

Signaling theory and green finance theory provide theoretical explanations for how rising carbon prices may alleviate firms’ financing constraints and thereby promote corporate green technological innovation. When the price of carbon emission allowances increases, firms face substantially higher compliance costs associated with emission reduction, as well as stronger incentives to pursue low-carbon transformation. On the one hand, rising carbon prices transmit market signals regarding the long-term value of low-carbon development, which can enhance external investors’ confidence in firms’ green transformation prospects. This signaling effect helps reduce information asymmetry between firms and capital markets, thereby alleviating financing constraints [33,34]. On the other hand, in the context of increasing carbon prices, the preferential allocation of green financial instruments—such as green credit and green bonds—provides firms with more favorable financing channels. This institutional support further reduces financing difficulties and lowers capital costs [35,36]. Recent cross-regional evidence further supports this view. Cifuentes-Faura et al. (2026) [18] found that digital financial inclusion can improve access to capital for green projects and support carbon neutrality targets, particularly in certain regions of China. Their study highlights that aligning digital financial services with sustainable development goals can effectively reduce firms’ financing constraints for low-carbon transformation, which is consistent with the green finance mechanism discussed in this section. The alleviation of financing constraints can provide sufficient financial resources for firms’ green technological innovation, enabling them to overcome capital bottlenecks and increase investment in R&D equipment acquisition, technology patent introduction, and the recruitment of R&D personnel. These investments ultimately contribute to improvements in corporate green technological innovation output. This transmission pathway suggests that rising carbon prices can indirectly promote corporate green technological innovation through the mediating mechanism of alleviating financing constraints. Based on the above analysis, this study proposes the following hypothesis:

Hypothesis 4:

Financing constraints mediate the relationship between carbon emission allowance trading prices and corporate green technological innovation.

2.5. Theoretical Framework of the Carbon Finance Transmission Mechanism

As a core market-based environmental policy instrument, carbon emission allowance trading prices exert an indirect yet significant influence on corporate green technological innovation through the carbon finance transmission mechanism. This mechanism operates by reshaping firms’ cost–benefit structures, transmitting signals through financial markets, and interacting with external governance factors. Specifically, carbon price signals directly alter firms’ production cost structures. High-carbon production activities incur higher marginal costs due to mandatory purchases of emission allowances, while low-carbon technologies generate additional economic benefits through the saving or trading of emission permits. This dual mechanism fundamentally reshapes firms’ cost–benefit calculations and creates strong incentives for enterprises to reallocate resources from carbon-intensive production toward green R&D and sustainable innovation. These price signals are further transmitted through financial markets. Rising carbon prices increase the environmental risk premium associated with high-carbon assets, thereby raising capital costs and tightening financing conditions for heavily polluting firms. In contrast, enterprises that actively engage in corporate green technological innovation are more likely to obtain preferential access to green financial instruments, such as green credit and green bonds. This preferential access reduces financing costs and strengthens firms’ capacity and willingness to invest in R&D activities. During this transmission process, environmental regulation and media attention function as important moderating factors. Stricter environmental regulation amplifies the cost pressures generated by carbon pricing, making firms more responsive to carbon price signals and accelerating their transition toward greener production practices. Meanwhile, increased media scrutiny raises the reputational consequences of environmental performance, transforming carbon prices into an important indicator of corporate environmental responsibility. Under such reputational pressure, firms are further incentivized to accelerate green technological innovation in order to maintain legitimacy and protect their corporate reputation.

3. Research Design

Definitions and measurements of all variables are summarized in

Table 1. Key Variables and Definitions.

Variable Type	Variable Name	Variable Symbol	Variable Description
Dependent Variable	Number of Green Technology Innovations	GreNum	Proportion of Green Patent Applications to Total Patent Applications
	Quality of Green Technology Innovation	GreQua	Number of citations for green invention patents (log-scale)
Explanatory Variable	Carbon emission trading price	CPrice	Annual weighted average of carbon emission trading prices across pilot provinces/cities/100
Moderator Variable	Environmental Regulation Intensity	ERS	Frequency of Environmental Terms in Each City Proportion of Sentences Containing Terms to Total Word Count in Government Work Report
	Media attention	Media	Total media coverage/10,000
Control Variables	Board size	Board	Total Liabilities at Year-End/Total Assets at Year-End
	Tobin’s Q ratio	TobinQ	(Market Value of Floating Shares + Market Value of Non-Floating Shares × Book Value per Share + Book Value of Liabilities)/Total Assets
	Price-to-Book Ratio	PB	Price per share/Net asset value per share
	Book-to-market ratio	BM	Book Value/Total Market Value
	Total Asset Turnover Ratio	ATO	Revenue/Average Total Assets
	Cash Flow Ratio	Cash Flow	Net Cash Flow from Operating Activities/Total Assets
	Quick Ratio	Quick	(Current Assets − Inventory)/Current Liabilities
	Return on Assets	ROA	Net Profit/Total Assets

.

3.1. Data Sources and Sample Selection

This study focuses on A-share listed companies located in the eight pilot regions of China’s carbon emissions trading system—Beijing, Shanghai, Guangdong, Shenzhen, Hubei, Chongqing, Fujian, and Tianjin—covering the period from 2013 to 2024. Firms located in these pilot regions are subject to mandatory carbon allowance allocation and compliance requirements, and are directly involved in carbon emissions trading [19]. Therefore, they are directly affected by carbon market regulations and price fluctuations, which provides a clear institutional foundation for examining the impact of carbon prices on corporate green innovation. Notably, the Guangdong Carbon Exchange and the Shenzhen Carbon Exchange operate as separate carbon trading platforms with distinct institutional designs. Moreover, the Guangdong carbon market covers all regions within Guangdong Province except Shenzhen. Therefore, carbon trading data for these two regions are collected and processed separately.

Following the standard sample selection procedures in empirical corporate finance research, financial and insurance firms are excluded due to their unique financial structures. In addition, ST-listed companies (firms subject to special treatment due to financial distress), PT-listed companies (firms under particular transfer arrangements), and firms with missing key variables are also removed. After applying these screening criteria, the final sample consists of 14,437 firm-year observations.

The data used in this study are obtained from multiple authoritative sources, including the Wind Database, the CSMAR (Guotai An) Database, the official website of the China National Carbon Emissions Trading Exchange, and the Green Patent Research Database. The empirical analysis is conducted using Stata 17 (StataCorp LLC, College Station, TX, USA).

3.2. Variable Definitions

3.2.1. Dependent Variable

The dependent variable in this study is corporate green technological innovation, denoted as Green. To comprehensively capture firms’ green technological innovation activities, we measure this variable from two dimensions: the quantity and quality of green patents.

Given that the authorization of green patents typically requires three to five years, following the method of Chen and Chen (2021) [37], this study uses the ratio of green patent applications to total patent applications in the same year to measure the quantity of green technological innovation, denoted as GreNum. This measurement is reasonable because green patent applications can timely reflect firms’ real-time investment and strategic orientation in green technology R&D, avoiding the lag bias caused by the long authorization cycle of patents, and the ratio form effectively eliminates the scale difference in patent output among enterprises with different sizes, enhancing the comparability of cross-firm data. In addition, following the method of Du et al. (2019) [38], the quality of corporate green technological innovation is proxied by the logarithm of the number of citations received by green invention patents, denoted as GreQua. This indicator is scientifically justified as patent citations directly reflect the technological influence, knowledge spillover effect, and practical value of green innovation outputs, and taking the logarithm can mitigate the skewness of citation data, making it more compatible with the assumptions of subsequent econometric models. This indicator reflects the technological influence and knowledge value of corporate green technological innovation outputs.

3.2.2. Explanatory Variables

This study constructs the annual emissions-weighted carbon price following the method proposed by Dolphin and Merkle (2024) [39]. Specifically, the annual weighted average price is calculated using daily average trading prices weighted by daily trading volumes. This method can effectively eliminate short-term disturbances and abnormal fluctuations in the carbon market, and accurately reflects the stable and continuous price level, which is consistent with the long-term decision-making characteristics of corporate green technological innovation. The annual weighted average carbon price is adopted because corporate green technological innovation represents a long-term strategic decision, which tends to respond to persistent and stable market signals rather than short-term fluctuations in carbon prices. To reduce potential scale differences between the explanatory variable and the dependent variables and to facilitate the interpretation of regression coefficients, the annual weighted average carbon price is scaled by dividing it by 100. This standardization can unify variable dimensions and enhance the stability and reliability of regression estimation results. Each firm is then matched with the carbon price of the province in which its headquarters is registered. For firms operating across multiple regions, carbon emission regulation and compliance are primarily supervised by the local authorities in the headquarters region. This allocation is consistent with the actual environmental supervision mechanism, which can ensure the accuracy and rationality of variable matching.

$$\mathit{CPrice}={\displaystyle \frac{{\mathit{\sum }}_{i=1}^n\left(P_i \times V_i\right)}{{\sum }_{i=1}^n{V}_i}}/100$$

(1)

Here, n represents the total number of trading days in a year, $P_i$ denotes the average carbon emission trading price on day i, and $V_i$ indicates the carbon emission trading volume on day i.

3.2.3. Moderating Variable

To further clarify the potential boundary conditions of the relationship between carbon pricing and green technological innovation, this study introduces two moderating variables: environmental regulation and media attention. These moderating variables are incorporated to examine whether and how the baseline relationship varies under different external institutional constraints and public supervision environments. This research design enables the identification of the heterogeneous effects and applicable contexts of carbon pricing, thereby improving the robustness and explanatory power of the empirical findings. The specific measurement methods for these variables are defined as follows:

• Environmental Regulation Intensity (ERS)

Environmental regulation intensity (ERS) is measured using a textual analysis approach based on the Government Work Report (GWR). The GWR is an annual official document released by local municipal governments that outlines policy priorities, development objectives, and key tasks for the current year.

Following the method developed by Li et al. (2022) [40], we construct the ERS index by first collecting the full text of the GWR for each city during the sample period. Next, we identify a set of core environmental keywords that reflect governmental commitment to environmental governance, including “carbon emission reduction,” “energy conservation,” “low-carbon development,” “environmental protection,” “pollution control,” “green transformation,” and “ecological civilization.” This measurement method is highly representative and authoritative, which can truly reflect the focus and policy orientation of local governments in environmental governance. The ERS indicator is then calculated as the ratio of the number of words contained in sentences including these environmental keywords to the total word count of the GWR. The word proportion measurement can effectively avoid the interference of regional economic scale and document length, ensuring the objectivity and comparability of the measurement results. A higher value of ERS indicates a stronger intensity of local environmental regulation.

2.

Media Attention (Media)

Media attention (Media) is measured by the total number of media reports related to a firm’s environmental and low-carbon activities, following the method developed by Deng et al. (2025) [41]. This measurement can directly reflect the public attention and external supervision imposed on enterprises’ environmental behaviors, and fully conforms to the logical orientation of media governance effect. To mitigate the influence of scale differences, the raw number of media reports is scaled by dividing it by 1000, thereby constructing the final media attention index. This processing can effectively weaken the heteroscedasticity and numerical dispersion caused by extreme values, and improve the stability of empirical results.

3.2.4. Mediating Variable

According to the theoretical analysis presented above, financing constraints may serve as an important transmission channel through which carbon pricing influences corporate green technological innovation. On the one hand, stricter carbon pricing policies increase firms’ environmental compliance costs and financing risks. On the other hand, they may also transmit positive signals to the capital market regarding firms’ commitment to green development. Therefore, introducing financing constraints as a mediating variable helps to clarify the mechanism through which carbon pricing affects corporate green technological innovation, that is, the pathway through which this impact is realized. Identifying this mediating effect contributes to a deeper understanding of the internal transmission mechanism between the core variables and further enhances the explanatory power of the empirical findings. The measurement of financing constraints is defined as follows.

The SA index is calculated using the formula proposed by Hadlock and Pierce: $SA= -0.737\times Size+0.043\times Siz{e}^2 -0.04\times Age$ [32]. This measurement is unaffected by endogenous interference and has strong exogeneity and stability, which can accurately reflect the financing constraints of enterprises. This index is determined by two exogenous variables: company size (Size) and company age (Age). The SA value is typically less than zero, and a higher absolute value of SA indicates a greater degree of financing constraints for the enterprise.

3.2.5. Control Variables

The following variables, representative of corporate characteristics, were selected as control variables: Board size (Board), Tobin’s Q ratio (TobinQ), Price-to-book ratio (PB), Book-to-market ratio (BM), Total asset turnover (ATO), Cash flow ratio (CashFlow), Quick ratio (Quick), Return on total assets (ROA). These control variables fully cover corporate governance, growth ability, profitability, operation capacity, solvency and other important dimensions, which can effectively exclude the interference of other factors on green technological innovation and ensure the accuracy and reliability of the regression results.

3.3. Model Specification

3.3.1. Benchmark Regression Model

To examine the relationship between carbon emission trading prices and green technological innovation, the following benchmark regression model is constructed:

$${\mathit{Green}}_{i,t} = a_0 + a_1{\mathit{CPrice}}_{i,t} +a_2{\mathit{Controls}}_{i,t} + \mathit{Ind} + \mathit{Year} + {\epsilon }_{i,t}$$

(2)

where i denotes industry; t denotes year; the dependent variable ${\mathit{Green}}_{i,t}$ represents corporate green technology innovation in industry i during year t, comprising both quantity and quality dimensions; ${\mathit{CPrice}}_{i,t}$ denotes carbon emission trading price, serving as the explanatory variable in this study; $a_1$ represents the regression coefficient of the independent variable, reflecting the impact of carbon emission trading prices on corporate green technology innovation; ${\mathit{Controls}}_{i,t}$ constitutes a set of control variables; Year denotes the fixed effect for year; Ind denotes the fixed effect for industry; ${\epsilon }_{i,t}$ represents the random disturbance term.

In the baseline model, we control for industry fixed effects and year fixed effects, rather than firm fixed effects. The core explanatory variable, carbon price, primarily varies across industries and over time, while its within-firm variation is relatively limited. Controlling for industry fixed effects helps absorb time-invariant industrial heterogeneity, whereas year fixed effects capture common macroeconomic shocks. This specification allows us to better identify the impact of carbon prices on corporate green technological innovation. In all baseline regressions, we report robust standard errors to account for potential heteroskedasticity in the error term. This adjustment is important because the variance of the error term may not be constant across observations in practice, which can invalidate conventional standard errors and lead to biased statistical inference. By employing robust standard errors, the estimated t-statistics and p-values remain valid even when the classical homoskedasticity assumption is violated, thereby strengthening the reliability of the empirical findings.

3.3.2. Moderated Effects Model

To examine the moderating effects of environmental regulation (ERS) and media attention (Media), the following moderation models are constructed:

$${\mathit{Gretotal}}_{i,t} ={ \beta }_0 +{ \beta }_1{\mathit{CPrice}}_{i,t} + \beta 2{\mathit{ERS}}_{i,t} + {\beta }_3{\mathit{CPrice}}_{i,t} \times { \mathit{ERS}}_{i,t} + {\beta }_4{\mathit{Controls}}_{i,t }+ \mathit{Ind}+\mathit{Year} +{ \epsilon }_{i,t}$$

(3)

$${\mathit{Gretotal}}_{i,t}={ \delta }_0+{ \delta }_1{\mathit{CPrice}}_{i,t}+{ \delta }_2{\mathit{Media}}_{i,t}+{\delta }_3{\mathit{CPrice}}_{i,t} \times { \mathit{Media}}_{i,t}t+{\delta }_4{\mathit{Controls}}_{i,t}+\mathit{Ind}+\mathit{Year}+{ \epsilon }_{i,t}$$

(4)

Here, β₃ and δ₃ represent the observed coefficients of the interaction terms, with our primary focus being on whether these variables are statistically significant. In Equations (3) and (4), ${\mathit{ERS}}_{i,t}$ and ${\mathit{Media}}_{i,t}$ denote the levels of environmental regulation and media attention for industry i in year t. ${\mathit{CPrice}}_{i,t} \times {\mathit{ERS}}_{i,t}$ and ${\mathit{CPrice}}_{i,t} \times {\mathit{Media}}_{i,t}$ are the interaction terms between carbon price and environmental regulation, and between carbon price and media attention, respectively. ${\beta }_0$ and ${\delta }_0$ are the intercept terms. ${\beta }_1$ to ${\beta }_4$ and ${\delta }_0$ to ${\delta }_4$ represent the coefficient values for the respective variables, with the meanings of the remaining variables consistent with Equation (2). Notably, if the coefficient values for β₃ and δ₃ significant, it indicates that environmental regulations and media attention exert a moderating effect on the relationship between carbon emission trading prices and corporate green technological innovation.

3.3.3. Mediation Effect Model

Mediation Effect Model: Drawing on the research of Baron and Kenny (1986) [42], this study employs the mediation effect model to examine the transmission path of carbon emission trading price to green technological innovation through recursive Equations (2), (5) and (6).

$${SA}_{i,t}={ \gamma }_0+{\gamma }_1{CPrice}_{i,t}+{\gamma }_2{Controls}_{i,t}+\sum \mathit{Industry}+\sum \mathit{Year}+{ \epsilon }_{i,t}$$

(5)

$${\mathit{Green}}_{i,t}={\chi }_0+{ \chi }_1{\mathit{CPrice}}_{i,t}+{\chi }_2{SA}_{i,t}+{\chi }_3{\mathit{Controls}}_{i,t}+\sum \mathit{Industry}+\sum \mathit{Year}+{\epsilon }_{i,t}$$

(6)

in which SA is the mediator variable; if both coefficients ${\gamma }_1$ and ${\chi }_2$ are significant, and ${\chi }_1$ is also significant, with ${\gamma }_1$ and ${\chi }_2$ sharing the same sign as ${\chi }_1$, this indicates a partial mediating effect; if both coefficients ${\gamma }_1$ and ${\chi }_2$ are significant but ${\chi }_1$ is not, it suggests a full mediating effect.

4. Empirical Findings and Analysis

4.1. Descriptive Statistics

Table 2. Descriptive Statistics of Key Variables.

Variable	Observed Value	Mean	Standard Deviation	Minimum	Median	Maximum
Grenum	14,437	0.056	0.146	0.000	0.000	1.000
Grequa	14,437	0.546	1.133	0.000	0.000	7.893
CPrice	14,437	0.448	0.279	0.000	0.416	1.085
Media	14,437	0.357	1.064	0.000	0.156	43.223
ERS	14,437	1.004	0.212	0.441	0.992	2.577
SA	14,437	−3.390	1.289	−5.980	−3.803	0.000
Board	14,437	2.091	0.203	1.609	2.197	2.708
TobinQ	14,437	2.070	1.263	0.795	1.689	17.676
PB	14,437	3.672	3.025	0.387	2.817	41.587
BM	14,437	0.606	0.247	0.057	0.592	1.258
ATO	14,437	0.586	0.398	0.047	0.500	2.645
CashFlow	14,437	0.048	0.066	−0.202	0.047	0.267
Quick	14,437	2.461	2.735	0.149	1.504	18.073
ROA	14,437	0.035	0.065	−0.556	0.038	0.222
BM	14,437	0.606	0.247	0.057	0.592	1.258

reports the descriptive statistics of the key variables used in this study. For the dependent variables, the maximum values of green technological innovation quantity (GreNum) and green technological innovation quality (GreQua) are 1.000 and 7.893, respectively, while the minimum values of both variables are 0.000. This indicates substantial heterogeneity in innovation capabilities among listed firms in the carbon emissions trading pilot regions. While some firms exhibit relatively high levels of corporate green technological innovation, others show no green innovation output during the sample period. For the core explanatory variable, the carbon emission allowance trading price (CPrice), the mean value is 0.448 and the standard deviation is 0.280, with a maximum of 1.085 and a minimum of 0.000. Since the carbon price variable is scaled by dividing the original price by 100, the reported values reflect the transformed data. The original carbon price data suggest that firms located in different regions face varying carbon prices, and the relatively large variation indicates that carbon pricing policies may exert heterogeneous effects across regions and firms. The mediating variable, financing constraints (SA), has a mean value of −3.390 and a standard deviation of 1.289, with values ranging from −5.980 to 0.000. This suggests that sample firms generally face a certain degree of financing constraints, while the level of constraint varies considerably across firms. Regarding the moderating variables, media attention (Media) has a mean value of 0.357, ranging from 0 to 43.223, indicating substantial variation in the degree of media coverage related to firms’ environmental activities. Environmental regulation intensity (ERS) has a mean of 1.004, with values ranging from 0.441 to 2.577, suggesting that the stringency of local environmental regulation—reflected in municipal government work reports—also differs significantly across cities. Among the control variables, the mean value of Board (2.091) indicates a moderate level of board size among the sample firms, while the relatively large standard deviation of TobinQ (1.263) suggests substantial differences in firm market valuation. In addition, the wide ranges of PB (0.387–41.587) and BM (0.057–1.258) further reflect the heterogeneity in firms’ financial characteristics.

Overall, these descriptive statistics confirm the considerable variation in the key variables across the sample. Such variability provides a suitable empirical basis for the subsequent regression analysis and supports further investigation into the relationships between carbon pricing, financing constraints, and corporate green technological innovation.

4.2. Baseline Regression Results

The benchmark regression results presented in

Table 3. Baseline regression results.

	(1)	(2)	(3)	(4)
Variable	GreNum	GreNum	GreQua	GreQua
CPrice	0.017 ***	0.018 ***	0.339 ***	0.361 ***
	(3.363)	(3.504)	(9.037)	(9.849)
Board		0.005		0.429 ***
		(0.782)		(8.847)
TobinQ		−0.004 **		0.130 ***
		(−2.258)		(10.179)
PB		−0.001		−0.032 ***
		(−1.403)		(−8.556)
BM		−0.020 **		0.947 ***
		(−2.308)		(13.281)
ATO		−0.004		0.010
		(−1.091)		(0.390)
CashFlow		−0.048 **		0.210
		(−2.332)		(1.495)
Quick		−0.002 ***		−0.046 ***
		(−4.417)		(−17.983)
ROA		0.030		−0.048
		(1.484)		(−0.340)
Constant	0.048 ***	0.067 ***	0.395 ***	−1.133 ***
	(19.200)	(4.366)	(22.325)	(−9.903)
Industry	YES	YES	YES	YES
Year	YES	YES	YES	YES
Adj.R2	0.124	0.126	0.186	0.224

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

reveal some interesting findings. When control variables are excluded from the analysis, the coefficients indicating the impact of carbon emission trading prices on the quantity and quality of corporate green technological innovation are 0.017 and 0.339, respectively. Both coefficients are statistically significant at the 1% level. After including control variables in the model, these coefficients shift to 0.018 and 0.361, respectively, while still remaining significant at the 1% level. We observe a nuanced pattern in the control variables: some exhibit a negative association with the quantity of corporate green technological innovation (GreNum) but a positive association with its quality (GreQua), while others display consistent signs across both measures. This pattern reflects the heterogeneous nature of firms’ green innovation strategies. For variables such as Tobin’s Q, ROA, and BM, the negative coefficient on GreNum suggests that firms with stronger market valuation, higher profitability, or lower asset tangibility may reduce incremental, quantity-oriented green innovation activities in order to reallocate resources toward more strategic priorities. By contrast, the positive coefficient on GreQua indicates that such firms are better positioned to invest in exploratory and high-quality green technological innovation, which typically requires long-term R&D investment and the development of core patents. These firms generally possess stronger financial buffers and greater risk tolerance, enabling them to undertake more ambitious innovation projects. In contrast, variables such as CashFlow and Quick show consistent positive associations with both innovation quantity and quality. This finding suggests that stronger liquidity and short-term solvency provide firms with the necessary resources to support both incremental improvements and exploratory advancements in green technology. Overall, this dual relationship highlights that corporate green technological innovation is not a homogeneous activity. Instead, firms make strategic trade-offs between increasing the quantity of innovation outputs and enhancing the quality of innovation outcomes, depending on their available resources, risk tolerance, and long-term strategic objectives. These findings are consistent with the literature distinguishing between exploitative (incremental) and exploratory (radical) innovation strategies. These results suggest that carbon pricing plays a significant role in promoting both the quantity and the quality of corporate green technological innovation. Notably, the incentive effect appears to be more pronounced for innovation quality. One possible explanation is that rising carbon prices not only encourage firms to increase the number of green patent applications but also motivate them to increase investment in research and development (R&D) and the development of core invention patents. As a result, firms tend to shift from merely expanding the scale of green innovation to enhancing its technological depth and value. Overall, carbon pricing stimulates simultaneous improvements in both the quantity and quality of green technological innovation, thereby providing preliminary empirical support for Hypothesis 1.

4.3. Endogeneity Test

To address the endogeneity bias arising from potential reverse causality, we use the mean value of corporate carbon emission trading prices as an instrumental variable (IV) by employing the two-stage least squares (2SLS) method. As shown in columns (1) to (4) of

Table 4. Endogeneity Test Results.

	(1)	(2)	(3)	(4)
Variable	CPrice	GreNum	CPrice	GreQua
IV	0.936 ***		0.936 ***
	(201.213)		(201.213)
CPrice		0.030 **		0.379 ***
		(2.570)		(3.282)
Controls	YES	YES	YES	YES
Industry	YES	YES	YES	YES
Year	YES	YES	YES	YES
K-P rk LM		639.576 [0.0000]		639.576 [0.0000]
K-P Wald rk F		4.0 × 104		4.0 × 104
		{16.38}		{16.38}
Constant	0.044 ***		0.044 ***
	(2.772)		(2.772)
N	14,437	14,437	14,437	14,437

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

, the coefficient of the instrumental variable is significantly positive, indicating its relevance to the endogenous variable. The p-value of the Kleibergen–Paap rk LM statistic is 0.0000, which rejects the null hypothesis of underidentification. The Kleibergen–Paap Wald rk F statistic is 4.0 × 10⁴, far exceeding the critical value of 16.38 at the 10% significance level, indicating no weak instrument problem. Theoretically, this instrumental variable is determined by exogenous regional carbon trading policies and market rules, and it affects corporate green technological innovation only through the channel of carbon emission trading prices (CPrice), without directly correlating with unobserved regional economic or regulatory conditions that may influence corporate green technological innovation, thus satisfying the exclusion restriction. The coefficients of CPrice remain significantly positive in all specifications, further confirming the robustness of our main findings.

4.4. Robustness Test

To ensure the robustness of our empirical findings, this study employs several complementary approaches to re-estimate the econometric model.

4.4.1. Replacing the Dependent Variable

To mitigate potential measurement bias in the green patent quantity indicator (GreNum), this study follows the approach of Yu et al. (2022) [43] and uses two alternative measures to evaluate corporate green technological innovation: the total number of green patent applications (Green), and the natural logarithm of one plus the total number of green invention and utility model patent applications (GreTotal). These indicators comprehensively reflect the actual output level of green innovation and avoid estimation bias caused by single-indicator measurement. As shown in Columns (1) and (2) of

Table 5. Part of the Robustness Test Results.

	Replacing the Dependent Variable			Adding Control Variables		Lagged Explanatory Variable		Alternative Estimation Method
Variable	(1) GreTotal	(2) Green	(3) GreenIndex	(4) GreNum	(5) GreQua	(6) GreNum	(7) GreQua	(8) Grequa
CPrice	0.110 ***	14.557 ***	0.018 ***	0.019 ***	0.364 ***			0.625 ***
	(3.717)	(8.073)	(2.656)	(3.711)	(10.034)			(8.529)
Lev				0.098 ***	1.126 ***
				(6.500)	(10.747)
EM				−0.009 ***	0.045 **
				(−3.805)	(2.546)
ROE				−0.028	0.461 ***
				(−1.267)	(3.730)
GrossProfit				−0.027 ***	0.295 ***
				(−2.873)	(4.509)
L.CPrice						0.023 ***	0.375 ***
						(3.949)	(8.882)	(0.001)
Constant	−0.041	−39.071 ***	0.195 ***	0.062 ***	−1.544 ***	0.066 ***	−1.195 ***	−26.574
	(−0.467)	(−7.714)	(9.759)	(3.818)	(−13.033)	(3.954)	(−9.324)	(−0.015)
lnalpha								0.219 ***
								(4.991)
Controls	YES	YES	YES	YES	YES	YES	YES	YES
Industry	YES	YES	YES	YES	YES	YES	YES	YES
Year	YES	YES	YES	YES	YES	YES	YES	YES
N	14,437	14,437	13,252.000	14,437	14,437	11,930	11,930	14,437
Adj.R2	0.149	0.137	0.245	0.129	0.245	0.125	0.230	0.137

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values. For column (6), we report McFadden’s Pseudo R² = 0.137, as the negative binomial regression does not produce a conventional R².

, the core explanatory variable carbon price (CPrice) is significantly positive at the 1% level in both alternative indicator models, consistent with the baseline regression results, indicating that the core conclusion is not affected by the measurement method of green innovation.

4.4.2. Constructing a Comprehensive Green Innovation Index

To compensate for the limitations of relying solely on patent-based indicators, this study constructs a comprehensive corporate green technological innovation index (GreenIndex) using the entropy weight method, which integrates four dimensions of green innovation performance: R&D intensity, weighted patent applications, breakthrough innovation, and innovation efficiency, to capture firms’ green innovation performance more comprehensively. As shown in Column (3) of Table 5, carbon price (CPrice) remains significantly positive at the 1% level in the GreenIndex model, suggesting that the promoting effect of carbon price on corporate green technological innovation is still robust under the comprehensive green innovation evaluation system, and the conclusion is not constrained by single patent-based indicators.

4.4.3. Adding Control Variables

To test the sensitivity of the results to the selection of control variables, this study adds four additional control variables to the baseline model: the debt-to-equity ratio (Lev), equity multiplier (EM), return on equity (ROE), and gross profit margin (GrossProfit), to further control for the interference of factors such as corporate capital structure and profitability. As shown in Columns (4) and (5) of Table 5, after adding the control variables, carbon price (CPrice) is significantly positive at the 1% level in both the GreNum and GreQua models, with no substantial changes in the core coefficients and significance levels, indicating that omitted variable bias does not materially affect the core conclusion.

4.4.4. Lagging the Explanatory Variable

To alleviate endogeneity concerns such as potential reverse causality between carbon price and green innovation, this study re-estimates the model by lagging the core explanatory variable carbon price (CPrice) by one period (L.CPrice) to test the robustness of the results. As shown in Columns (6) and (7) of Table 5, the one-period lagged carbon price (L.CPrice) is significantly positive at the 1% level in both the GreNum and GreQua models, highly consistent with the baseline regression results, further verifying the positive promoting effect of carbon price on corporate green technological innovation and alleviating endogeneity concerns.

4.4.5. Alternative Estimation Method

Considering the right-skewed and zero-inflated distribution of green patent citations (GreQua), this study employs a negative binomial regression model, which is well-suited for overdispersed count data, to re-estimate the baseline results and verify their robustness. As shown in Column (8) of Table 5, in the negative binomial regression model, carbon price (CPrice) is still significantly positive at the 1% level, and the core conclusion remains stable, indicating that the difference in estimation methods does not change the positive impact of carbon price on corporate green technological innovation, and the results are highly reliable.

4.4.6. Subsample Regression

To further ensure the rationality of including the full sample and address the concern that “not all enterprises are affected by the carbon market”, this study divides the sample into key emission units and non-key emission units for subsample regression, aiming to examine the heterogeneous impact of carbon prices on green innovation across different types of enterprises. As shown in

Table 6. Robustness Test Results: Subsample Regression.

	(1) Key Emission Units	(2) Non-Key Emission Units	(3) Key Emission Units	(4) Non-Key Emission Units
Variable	GreNum	GreNum	GreQua	GreQua
CPrice	0.013 *	0.024 ***	0.384 ***	0.296 ***
	(1.708)	(3.217)	(7.987)	(5.362)
Constant	0.090 ***	0.057 ***	−1.162 ***	−1.032 ***
	(3.990)	(2.695)	(−7.376)	(−6.224)
Controls	YES	YES	YES	YES
Industry	YES	YES	YES	YES
Year	YES	YES	YES	YES
Adj.R2	0.143	0.116	0.283	0.206
N	6310	8127	6310	8127

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

, carbon prices exert a significant positive incentive effect on green innovation for both key emission units and non-key emission units: in the green patent quantity model, the coefficient of carbon price is significant at the 10% level for key emission units and at the 1% level for non-key emission units; in the green patent quality model, the coefficients of carbon price for both types of enterprises are significant at the 1% level. This result is consistent with the findings of existing literature. Calel & Dechezleprêtre (2016) [7], based on a study of the EU Emissions Trading System (ETS), found that carbon policies generate significant spillover effects through technology diffusion, promoting the green transformation of enterprises not directly regulated. Lai & Chen (2023) [44] also confirmed the existence of innovation spillover effects from the carbon market on non-covered enterprises through empirical research on China’s pilot carbon emission trading policy. Specifically, the carbon market not only exerts a direct constraint effect on key emission units but also forms significant policy spillover effects on non-key emission units through channels such as supply chain transmission and policy signal transmission. Non-key emission units are not irrelevant noise in the sample but important carriers of the overall green transformation effect of the carbon market. Therefore, including all listed enterprises in pilot regions in the sample can more comprehensively and unbiasedly identify the real impact of the policy, avoid sample selection bias, and the empirical design is supported by sufficient rationality and literature.

4.5. Moderation Effect Test

4.5.1. Environmental Regulation

The carbon emissions trading market is established under the framework of environmental regulations, which not only restrict corporate emissions but also encourage green technological innovation through market incentives, making environmental regulation a key moderating variable. The results in columns (2) and (4) of

Table 7. Moderation Effect Test Results: Environmental Regulations.

Variable	(1)	(2)	(3)	(4)
	GreNum	GreNum	GreQua	GreQua
CPrice	0.018 ***	0.018 ***	0.361 ***	0.373 ***
	(3.491)	(3.253)	(9.849)	(10.010)
ERS		−0.003		0.096 **
		(−0.524)		(2.313)
CPrice × ERS		0.053 ***		0.585 ***
		(3.207)		(4.626)
Constant	0.069 ***	0.071 ***	−1.139 ***	−1.205 ***
	(4.476)	(4.448)	(−9.814)	(−10.113)
Controls	YES	YES	YES	YES
Industry	YES	YES	YES	YES
Year	YES	YES	YES	YES
N	14,437	14,437	14,437	14,437
Adj.R2	0.132	0.133	0.230	0.230

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

show a significantly positive coefficient for the interaction between carbon price and environmental regulation, indicating that stricter environmental policies amplify the positive impact of carbon prices on both the quantity and quality of green technological innovation. By imposing stringent emission standards, taxes, and other measures, environmental regulations drive firms toward low-carbon production. As governments tighten these policies, carbon supply becomes constrained, pushing prices up and increasing firms’ carbon procurement costs. This, in turn, motivates firms to reduce emissions through innovations like process improvements, clean energy adoption, and energy efficiency enhancements, which lower both carbon emissions and production costs. This cost-saving effect further fuels firms’ enthusiasm for corporate green technological innovation, thereby validating the moderating role of environmental regulations and confirming Hypothesis 2.

4.5.2. Media Attention

Media attention serves as an important external supervisory mechanism that can shape the impact of carbon trading prices on corporate green technological innovation by influencing both firms’ internal decision-making and external market dynamics. As reported in columns (2) and (4) of

Table 8. Moderation Effect Test Results: Media Attention.

	(1)	(2)	(3)	(4)
Variable	GreNum	GreNum	GreQua	GreQua
CPrice	0.018 ***	0.018 ***	0.361 ***	0.344 ***
	(3.491)	(3.495)	(9.849)	(9.117)
Media		0.003 ***		0.215 ***
		(3.219)		(7.903)
CPrice × Media		0.008 **		0.358 ***
		(1.990)		(3.578)
Constant	0.069 ***	0.070 ***	−1.139 ***	−1.012 ***
	(4.476)	(4.581)	(−9.814)	(−8.838)
Controls	YES	YES	YES	YES
Industry	YES	YES	YES	YES
Year	YES	YES	YES	YES
N	14,437	14,437	14,437	14,437
Adj.R2	0.132	0.133	0.230	0.266

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

, the interaction term between carbon emission trading prices and media attention is significantly positive. This result indicates that higher levels of media attention strengthen the positive effect of carbon prices on both the quantity and quality of corporate green technological innovation. One possible explanation is that increased media coverage enhances information transparency and public awareness, thereby amplifying the influence of carbon price signals on firms’ innovation decisions. Under greater media scrutiny, firms face stronger reputational pressure and are more motivated to respond proactively to carbon pricing policies by investing in green technological innovation. In addition, extensive media coverage of the carbon trading market strengthens the transmission of market signals, enabling firms to more clearly perceive changes in carbon emission costs. This heightened awareness further encourages firms to accelerate their green technology innovation activities. Overall, these findings suggest that media attention positively moderates the relationship between carbon pricing and corporate green technological innovation, thereby providing empirical support for Hypothesis 3.

4.6. Mediating Effect Test

Table 9. Mediation effect test results: financing constraints.

	(1)	(2)	(3)	(4)	(5)	(6)
Variable	GreNum	SA	GreNum	GreQua	SA	GreQua
CPrice	0.018 ***	0.088 ***	0.017 ***	0.361 ***	0.088 ***	0.325 ***
	(3.491)	(10.044)	(3.197)	(9.848)	(10.044)	(8.983)
SA			0.018 ***			0.414 ***
			(3.976)			(9.292)
Constant	0.069 ***	−3.348 ***	0.128 ***	−1.139 ***	−3.348 ***	0.248
	(4.475)	(−121.794)	(6.047)	(−9.813)	(−121.794)	(1.337)
Controls	YES	YES	YES	YES	YES	YES
Industry	YES	YES	YES	YES	YES	YES
Year	YES	YES	YES	YES	YES	YES
N	14,437	14,437	14,437	14,437	14,437	14,437
Adj.R2	0.126	0.961	0.127	0.224	0.961	0.233

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

presents the results of the mediating effect of financing constraints. Columns (1) and (4) show that the regression coefficient of carbon emission trading price is significantly positive at the 1% level, consistent with previous findings, and Path 1 passes the test. Further analysis using the recursive equation model reveals that Columns (2) and (5) present the results of Model (5), where the regression coefficient of financing constraints is 0.088, remaining significant at the 1% level. This indicates that increased carbon prices alleviate financing constraints, confirming Path 2. Columns (3) and (6) present the results of Model (6), the regression coefficient of green technological innovation remains significantly positive at the 1% level. This demonstrates that reduced financing constraints effectively promote corporate green technological innovation, as lessened constraints provide sufficient funding for sustained green R&D investments. The regression coefficient of carbon emission trading price remains significantly positive at the 1% level, validating Path 3. These results confirm that financing constraints pass the recursive equation model test, indicating that they partially mediate the impact of carbon emission trading price on green technological innovation. Thus, Hypothesis H4 is supported.

4.7. Heterogeneity Test

4.7.1. Disparities in Technical Proficiency

The results reported in columns (1)–(4) of

Table 10. Heterogeneity Test Results: Technical Level.

	(1) High-Tech Enterprises	(2) Non-High-Tech Enterprises	(3) High-Tech Enterprises	(4) Non-High-Tech Enterprise
Variable	GreNum	GreNum	GreQua	GreQua
CPrice	−0.002	0.036 ***	0.276 ***	0.434 ***
	(−0.264)	(4.965)	(4.844)	(8.891)
Constant	0.077 ***	0.050 **	−1.017 ***	−1.352 ***
	(3.369)	(2.312)	(−5.811)	(−8.249)
Controls	YES	YES	YES	YES
Industry	YES	YES	YES	YES
Year	YES	YES	YES	YES
N	5910	8399	5910	8399
Adj.R2	0.032	0.184	0.152	0.271
Chow test	0.000		0.001

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

show that carbon emission trading prices significantly increase the quantity of green technological innovation among non-high-tech enterprises. At the same time, carbon prices exert a similarly significant effect on the quality of green technological innovation across firms with different technological levels, as reflected by the consistent coefficient patterns. This heterogeneity can be explained by differences in firms’ innovation orientation and technological capabilities. High-tech enterprises typically maintain a stronger focus on innovation and R&D activities, which leads to sustained investment in green technologies regardless of changes in carbon prices. In contrast, non-high-tech enterprises are more sensitive to the profitability implications of carbon emission costs. Given their relatively lower technological barriers, these firms are more likely to respond to rising carbon prices by adopting or imitating existing green technologies, thereby increasing the quantity of green technological innovation and potentially facilitating their transition toward high-tech status. As a result, non-high-tech enterprises exhibit greater sensitivity to carbon price fluctuations, and the carbon pricing mechanism generates a stronger incentive effect on their green technological innovation activities. Furthermore, the overlapping 95% confidence intervals of the CPrice coefficients across the regressions, together with the significant differences confirmed by the Chow test, further support the robustness of these findings.

4.7.2. Regional Variability

• Southern and Northern Regions

Given the disparity in carbon quota requirements between northern and southern development, firms exhibit differing motivations for green technological innovation. Adopting the geographical median line at 35°N latitude as the demarcation between northern and southern regions, this study examines the impact of carbon emission trading prices on corporate green technological innovation.

As evident from columns (1) to (4) in

Table 11. Heterogeneity Test Results: Southern and Northern Regions.

	(1) Northern	(2) Southern	(3) Northern	(4) Southern
Variable	GreNum	GreNum	GreQua	GreQua
CPrice	0.040 ***	−0.002	0.465 ***	0.054
	(3.233)	(−0.344)	(4.388)	(1.117)
Constant	−0.015	0.101 ***	−1.809 ***	−0.742 ***
	(−0.477)	(5.709)	(−6.337)	(−6.010)
Controls	YES	YES	YES	YES
Industry	YES	YES	YES	YES
Year	YES	YES	YES	YES
N	3536	10,901	3536	10,901
Adj.R2	0.191	0.110	0.327	0.200

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

, carbon emission trading prices significantly enhance both the quantity and quality of green technological innovation among northern enterprises. The regional heterogeneity may stem from the northern region’s abundant energy resources and robust industrial foundations, coupled with severe environmental pollution and ecological degradation. Its carbon emission structure is dominated by industrial, transport, and construction sectors, with industrial emissions accounting for a high proportion. Consequently, northern enterprises face greater urgency and demand for advancing green technological innovation. This regional disparity in firms’ responses to green technological innovation is closely related to differences in carbon quota allocation policies between northern and southern China. Northern regions rely heavily on energy-intensive industries and exhibit relatively high baseline emissions, partly due to substantial winter heating demand. As a result, these regions typically face stricter carbon quota allocations, which generate stronger incentives for enterprises to invest in green technologies, comply with environmental regulations, and reduce emission-related costs. In contrast, southern regions generally have a more diversified industrial structure, lower carbon emission intensity, and greater access to clean energy resources. Carbon quota allocation policies in these regions are therefore relatively more flexible, which weakens the incentive for firms to engage in corporate green technological innovation in response to changes in carbon trading prices. This institutional and industrial background helps explain why enterprises in northern regions exhibit a stronger innovation response to carbon emission trading prices.

2.

Eastern and Central-Western Regions

China exhibits substantial regional differences in natural endowments and economic development. To further examine how carbon emission trading prices affect corporate green technological innovation across regions, this study divides China into eastern and central–western regions based on the classification of the National Bureau of Statistics.

As shown in columns (1)–(4) of

Table 12. Heterogeneity Analysis: Eastern and Central–Western Regions.

	(1) Eastern	(2) Central–Western	(3) Eastern	(4) Central–Western
Variable	GreNum	GreNum	GreQua	GreQua
CPrice	0.015 ***	0.099	0.368 ***	−0.036
	(2.771)	(1.048)	(9.793)	(−0.366)
Constant	0.066 ***	0.134 **	−1.123 ***	−0.451
	(4.141)	(1.988)	(−9.337)	(−0.872)
Controls	YES	YES	YES	YES
Industry	YES	YES	YES	YES
Year	YES	YES	YES	YES
N	13,149	1288	13,149	1288
Adj.R2	0.130	0.135	0.231	0.241

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

, carbon emission trading prices significantly promote both the quantity and quality of corporate green technological innovation among enterprises located in the eastern region. This regional disparity may stem from the eastern region’s relatively advanced level of economic development and more mature industrial system. Compared with the central–western regions, the eastern region also faces more severe environmental pollution and carbon emission pressures, which create stronger incentives for enterprises to engage in green technological innovation and address climate-related challenges. Moreover, as a market-based environmental policy instrument, the carbon emission trading system encourages firms to reduce emissions and invest in green technologies through the allocation of emission quotas and the formation of carbon trading prices. In economically developed eastern regions—where environmental regulations are generally stricter and market institutions are more developed—the carbon trading market tends to operate more actively and exert stronger influence on corporate innovation decisions.

4.7.3. Disparities in Property Rights

To examine the moderating effect of property rights on the relationship between carbon pricing and corporate green technological innovation, this paper divides the sample into state-owned enterprises (SOEs) and private enterprises for subgroup regressions and uses the Chow test to examine the significance of coefficient differences between groups. As shown in columns (1) and (2) of

Table 13. Heterogeneity Analysis: SOE and Private Enterprises.

	(1) SOE	(2) Private Enterprises	(3) SOE	(4) Private Enterprises
Variable	GreNum	GreNum	GreQua	GreQua
CPrice	0.026 ***	0.016 **	0.622 ***	0.205 ***
	(2.709)	(2.403)	(7.719)	(4.892)
Constant	0.088 ***	0.078 ***	−0.780 **	−0.126
	(2.706)	(3.977)	(−2.542)	(−1.008)
Controls	YES	YES	YES	YES
Industry	YES	YES	YES	YES
year	YES	YES	YES	YES
N	4056	9211	4056	9211
Adj.R2	0.152	0.143	0.375	0.156
Chow test	0.000		0.001

Note: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively; values in parentheses are t-values.

, regarding the quantity of corporate green technological innovation, carbon pricing has a significantly positive impact on both SOEs and private enterprises, with coefficients of 0.026 and 0.016 respectively, both significant at the 1% level, and the effect is stronger for SOEs. The Chow test indicates that the difference between groups is statistically significant (p = 0.000). As shown in columns (3) and (4) of Table 13, for the quality of corporate green technological innovation, carbon pricing also exerts a significantly positive influence on both types of firms, with coefficients of 0.622 and 0.205 respectively, both significant at the 1% level, and the improvement is much larger for SOEs. The Chow test further confirms the moderating role of property rights (p = 0.001).

The reasons for this difference are as follows. First, as important implementers of national policies, SOEs undertake stronger environmental governance responsibilities and respond more actively to carbon pricing policies. Second, SOEs have more abundant capital, technology, and human resources, and are more likely to obtain government subsidies and credit support, which enables them to more effectively transform carbon cost pressure into corporate green technological innovation momentum. Third, SOEs focus more on long-term social benefits in their business decisions and are more willing to invest in corporate green technological innovation projects with long cycles and high risks. Fourth, most SOEs are in carbon-intensive industries and are key targets of carbon pricing policies, so their corporate green technological innovation behaviors have a stronger industry demonstration effect. In summary, the promoting effect of carbon pricing on corporate green technological innovation is more significant in SOEs. This study further enriches the heterogeneous analysis of carbon pricing policies by identifying the differential impacts of carbon emission trading prices under different property rights structures. Most prior literature focuses on the average effect of carbon trading pilot policies, while neglecting the heterogeneous responses driven by actual price signals and property rights differences. The above findings help to more accurately understand the implementation effects of carbon market mechanisms and provide targeted implications for differentiated policy design.

5. Conclusions and Implications

5.1. Conclusions

This study utilizes data from A-share listed companies in eight pilot regions—Beijing, Shanghai, Shenzhen, Guangdong, Hubei, Chongqing, Fujian, and Tianjin—covering the period from 2013 to 2024. A two-way fixed effects model is employed to empirically examine the relationship between carbon emission trading prices and corporate green technological innovation, as well as the underlying mechanisms. The results indicate that carbon trading prices significantly promote both the quantity and quality of green technological innovation, confirming the incentive effect of carbon market price signals in stimulating low-carbon technological R&D. Furthermore, environmental regulations and media attention strengthen the positive impact of carbon trading prices on corporate green technological innovation, highlighting the critical role of policy coordination in facilitating green innovation. In addition, the effect of carbon trading prices on corporate green technological innovation exhibits significant heterogeneity across technological levels, regions, and ownership types. Non-high-tech enterprises, firms located in northern and eastern regions, and state-owned enterprises demonstrate greater sensitivity to carbon trading prices. This suggests that enterprises and regions respond differently to carbon market signals during the low-carbon transition process.

To better position this study and highlight its contributions, we systematically compare our empirical results with the existing literature. Our baseline finding that carbon prices significantly promote both the quantity and quality of corporate green technological innovation is consistent with mainstream studies such as Liu et al. (2021) [45] and Chen et al. (2024) [46].Compared with prior research, this study extends the literature by simultaneously identifying financing constraints as a core mediating channel, examining key moderating factors, and conducting multi-dimensional heterogeneity analysis. In particular, we clarify how carbon prices affect corporate green technological innovation by reducing financing constraints, which complements and deepens the findings of Wang and Zhou (2022) [47] and Yu and Jin (2025) [48], who focus more on policy effects rather than integrated transmission mechanisms. Moreover, our moderating analysis identifies the boundary conditions of the carbon price–innovation relationship, while our heterogeneity results further reveal differential impacts across regions, ownership structures, and industry types, which is consistent with Zhang (2024) [49]. By providing a more complete, systematic, and multi-layered analytical framework, this study offers more comprehensive and comparable evidence than previous studies, thereby enriching the understanding of how carbon pricing drives corporate green technological innovation in a more detailed and convincing manner.

5.2. Policy Implications

Based on these findings, the study offers the following recommendations. First, carbon pricing mechanisms should be further refined to strengthen market incentives. Given the significant role of carbon prices in promoting corporate green technological innovation, quota allocation methods should be optimized by gradually increasing the proportion of auctions and exploring the establishment of price floor mechanisms. This would help prevent excessive price volatility from undermining long-term innovation expectations and encourage firms to increase investment in low-carbon R&D through clearer market signals. Second, policy coordination should be improved to prevent subsidies from diluting carbon price signals. Governments should shift subsidy policies from directly compensating firms for carbon emission costs toward supporting green technology R&D and industrialization, for example by establishing dedicated low-carbon technology funds or providing tax incentives for innovation. At the same time, strengthened environmental supervision and disclosure requirements, together with media oversight, can form a multi-layered pressure mechanism involving policy, market, and societal forces to encourage voluntary corporate innovation. Third, differentiated carbon market development strategies should be implemented to promote balanced regional growth. Given the stronger responsiveness of non-high-tech enterprises, stricter quota reduction policies could be applied to traditional high-carbon industries to intensify emission-reduction pressure and stimulate innovation incentives. Northern and eastern regions, which respond more strongly to carbon prices, could take the lead in establishing cross-regional carbon market linkages to expand market scale and improve price discovery functions. In contrast, central, western, and southern regions should prioritize strengthening low-carbon technology transfer and capacity building—such as by establishing regional corporate green technological innovation alliances—to reduce the transition costs faced by enterprises. Fourth, targeted support should be provided to state-owned enterprises (SOEs) to leverage their stronger responsiveness to carbon prices. Governments can encourage SOEs to take the lead in green technology demonstration projects, such as pilot low-carbon industrial parks or large-scale carbon capture and utilization facilities, to drive industry-wide innovation. Additionally, SOEs should be required to disclose more detailed green innovation strategies and carbon reduction progress, enhancing transparency and accountability. By leveraging SOEs’ resource advantages and policy compliance, policymakers can amplify the spillover effects of carbon pricing and accelerate the low-carbon transformation of the entire economy.

5.3. Research Limitations and Future Research

Despite the significant findings of this study, several limitations should be acknowledged. First, the study focuses solely on A-share listed companies in eight pilot regions, which may limit the generalizability of the results to non-listed firms, SMEs, or enterprises located in non-pilot areas. Second, although the quantity and quality of green patents are widely accepted proxies for green technological innovation, they do not fully capture the commercialization efficiency or the actual emission-reduction contributions of these innovations. Third, the analysis of mechanisms is limited to financing constraints as a mediating variable and media attention and environmental regulation as moderating factors. Other potential channels, such as corporate governance, executives’ environmental awareness, or the development of digital finance, warrant further investigation. Finally, this study employs annual average carbon trading prices, which may overlook short-term price volatility and dynamic shocks that could influence firms’ innovation decisions.

Future research can build on this study in several meaningful ways. First, expanding the sample to include non-listed enterprises and SMEs would enhance the external validity of the findings. Second, incorporating more comprehensive measures of corporate green technological innovation, such as innovation efficiency, market adoption rates, or actual carbon reduction effects, would provide a more holistic evaluation. Third, future studies could examine additional mediating and moderating factors, as well as their interactive effects, to uncover a more comprehensive set of mechanisms through which carbon prices influence corporate green technological innovation. Finally, employing high-frequency carbon price data and longer time-series analyses would allow for a deeper understanding of the dynamic and long-term effects of carbon market signals on corporate green technological innovation.