The Impact of Climate Risk Disclosure on Corporate Green Technology Innovation

Abstract

Amid escalating global climate challenges, the interplay between corporate climate risk disclosure and green technological innovation has become a pivotal scholarly focus in sustainability research. This study empirically examines the impact of climate risk disclosure on corporate green technology innovation and its underlying mechanisms using data from China’s A-share listed companies spanning 2004 to 2022. Key findings reveal that climate risk information disclosure significantly enhances green innovation capabilities through dual pathways: elevating media attention and reducing agency costs. Specifically, media scrutiny exerts external pressure via reputational incentives and public oversight, driving firms to accelerate green technology deployment. Concurrently, reduced agency costs mitigate information asymmetry between shareholders and management, enabling optimized resource allocation for long-term innovation investments. Heterogeneity analysis indicates that this catalytic effect is more pronounced in larger firms and those facing lower financing constraints. The research theoretically and practically elucidates the dual mechanisms through which climate disclosure propels green innovation, providing empirical support for refining corporate sustainability reporting systems and recalibrating regulatory frameworks. Policy recommendations include adopting differentiated climate disclosure standards, strengthening media and investor oversight, and incentivizing green innovation through executive performance metrics to facilitate low-carbon economic transition.

1. Introduction

Sustainable development has become an eternal issue for human society. Globally, the adverse impacts of climatic anomalies and marine inundation threats on ecosystems and human production and life have continued to increase, and the United Nations Intergovernmental Panel on Climate Change (IPCC) has pointed out in its Sixth Assessment Synthesis Report, Climate Change 2023, that the global temperature has risen by 1.1 °C compared with the pre-industrial period, which has led to frequent extreme weather events and increased global ecological and socio-economic instability. The report also noted that an increase in global temperature of more than 1.5 °C would lead to even more serious consequences, including loss of biodiversity and a significant increase in human health risks (IPCC AR6 Synthesis Report: Climate Change 2023).

China is also facing serious challenges posed by climate change. The China Blue Book on Climate Change 2023 released by the National Meteorological Administration (NMA) shows that the global warming trend has continued, and many of China’s climate change indicators have reached record highs. The report shows that from 1961 to 2022, China’s surface annual average warming of approximately 0.30 degrees Celsius per decade (https://www.cma.gov.cn/2011xwzx/2011xqxxw/2011xqxyw/202307/t20230708_5635282.html, accessed on 15 October 2024).

Climate change risk not only affects the ecosystem, but also the economic and social development of human beings. 2025, in January, affected by the low rainfall, dry weather and strong winds, the Los Angeles area of California, the United States suffered from the most serious fires in the region’s history, causing local businesses to suffer significant losses, such as: the fire caused one of the top ten U.S. power utility giants, Edison International, the stock price plummeted, and the market value evaporated more than USD 8 billion. The company responded with an emergency response saying it was actively investigating the cause of the fires, but this failed to quell the public’s criticism (https://www.thepaper.cn/newsDetail_forward_29933779?commTag=true, accessed on 1 February 2025). The Economist Intelligence Unit (EIU) estimates that future climate change could result in a loss of nearly USD 43 trillion in global stock market value (https://impact.economist.com/perspectives/sites/default/files/The%20cost%20of%20inaction.pdf, accessed on 5 December 2024). Studies have shown that climate-related natural disasters not only reduce firms’ productivity (Shun, W. et al., 2024) [1] and increase the likelihood of downstream customer termination (Pankratz and Schiller, 2024) [2], but also have a adverse effect on firms’ financial performance (Pankratz, N. and Bauer, R., 2023) [3] and increase the firms’ cost of equity capital (Du, J. et al., 2023) [4], and may even trigger a systemic financial crisis through a “green swan” event (Bolton et al., 2020) [5]. In view of this, developed countries and international organizations have been guiding and regulating in terms of macro policies, industry regulation and disclosure requirements for listed companies. In terms of standardization, in June 2023, the International Sustainability Standards Board (ISSB) issued two standards, “General Requirements for Disclosure of Sustainability-Related Financial Information (IFRS S1)” and “Climate-Related Disclosures” (IFRS S2) (hereinafter referred to as “S2”), aiming to establish a globally uniform disclosure bench-mark.

In November 2024, China’s Ministry of Finance, Ministry of Foreign Affairs, Securities Regulatory Commission and other 10 ministries and commissions jointly issued the Corporate Sustainability Disclosure Guidelines—Basic Guidelines (for Trial Implementation) (Caikuai [2024] No. 17), which guides enterprises to comprehensively examine sustainability impacts and actively practice the new development concept, and lays a solid foundation for the construction of a complete sustainability information framework institutional system has laid a solid foundation. Based on this, this paper takes Chinese A-share listed companies from 2004 to 2022 as a research sample to explore the impact of corporate climate risk information disclosure on green technology innovation.

The marginal contributions of this paper mainly include: first, enriching the theoretical linkage between climate risk disclosure and green technology innovation, focusing on and exploring the mediating roles played by media attention and agency costs, and providing a new perspective for understanding the complexity of climate action. Second, climate risk is subdivided into transition risk and physical risk, which clarifies the impact of different types of risk on green technological innovation, and provides a more refined analysis idea for climate risk management theory. Thirdly, this paper puts forward reform policy suggestions from the practical perspective, such as improving the information disclosure system, strengthening internal and external governance, so as to promote the benign operation of the “disclosure–innovation” closed loop in the low-carbon economy, which provides a reference for the optimization of the climate governance system in the sustainability guidelines of Chinese enterprises. In addition, this paper further explores the differential effects of enterprise size and financing constraints, and provides reference suggestions for enterprises, policy makers and financial institutions to reach a consensus on a low-carbon economy.

The subsequent structure of this paper is organized as follows: first, the second part combs through the relevant literature and puts forward hypotheses, analyzes the relevant studies and influence mechanisms of climate risk disclosure and green technology innovation, and introduces media attention and agency costs as mediating variables; the third part introduces the research methodology, data sources, and variable design; the fourth to the sixth part analyze the empirical results, conducts the robustness test, and investigates the transmission path; the seventh part further analyzes the heterogeneous effects of firm size and financing constraints; finally, we summarize the research conclusions and provides policy recommendations and practical insights.

2. Literature Review and Hypothesis Development

2.1. Literature Review

2.1.1. Relevant Research on Climate Risk Changes

In recent years, the frequent occurrence of climate disasters in many parts of the world has pushed scholars to explore its impact on the economy and society from multiple perspectives. At the macro and meso levels, they mainly cover economic growth and vulnerability, fluctuations in supply and demand in industries and the labor market, etc. For example, Botzen and Van (2009) [6] found that climate warming may increase the frequency and intensity of extreme weather events, worsening socio-economic vulnerabilities; and this it makes it possible for the whole industrial chain of the agri-food system, from primary production, processing, storage, and transportation to terminal consumption, to be suffer shocks (Godde, 2021) [7]. Chabot et al. (2023) [8] also confirmed that type III GHG emissions, long-term and short-term climate risks can undermine stability at both the level of individual institutions and the broader financial system. It has also been shown that climate tipping points increase global economic risks and even economic losses everywhere (Dietz et al., 2021) [9]. Groen et al. (2020) [10] and Kirchberger (2017) [11] focus on labor market impacts and find that climate risk reduces the supply of labor and increases affected industries and regions’ wage levels. At the micro level, scholars mainly study the economic consequences for market agents such as households and firms, e.g., Pankratz and Schiller (2024) [2] find that physical climate hazards, such as hot weather, significantly affect suppliers and increase the likelihood of termination of cooperation by downstream customers; Shun, W. et al. (2024) [1] demonstrated that climate policy uncertainty subjects firms to upgrading challenges, significantly inhibiting the improvement of total factor productivity; Huang et al. (2022) [12] found that firms with higher climate risk face higher bank lending costs; Li et al. (2024) [13] argued that climate risk is negatively correlated with firms’ innovation investments and innovation outputs, which suggests that climate risk may negatively affect firms’ innovation.

2.1.2. Environmental Disclosure and Firm Innovation

Research findings are divided on whether environmental disclosure contributes to corporate innovation: Alsaifi et al. (2020) [14] find that voluntary carbon emission reporting is positively correlated with improved financial outcomes for companies, not only in terms of better financial performance (Siddique et al., 2021) [15] but also in terms of prompting firms to adopt an “outside-in”-driven approach to achieve subsequent improvements and enhancements in carbon performance (Qian and Schaltegger, 2017) [16]. Feng et al. (2024) [17] revealed that climate risk disclosure positively impacts firms’ innovative capabilities by enhancing reputation, governance quality, and cost of capital efficiency. Tao et al. (2024) [18] find that climate-related disclosure promotes firms’ green innovation and that there is variability in industry attributes, ownership structure, and environmental regulations. Ilhan et al. (2023) [19] find that high-quality climate disclosure may reduce the cost of capital and improve market reputation. However, some studies have also found that environmental disclosure policies can crowd out capital and resources for firms’ green innovation, thus creating a disincentive effect on innovation (Wang, G. and Tang, B., 2024; Chen, H. et al., 2023) [20,21]. Shi et al. (2018) [22] find that environmental policies may unexpectedly inhibit firms’ innovative activities by increasing firms’ costs and altering location choices. In fact, from the perspective of research idea prevention, the differences in indicator construction and the transparency of environmental disclosures may also affect the reliability of the findings to a certain extent.

To summarize, the existing literature mainly focuses on the transmission mechanisms of reputational capital, internal governance and financing costs, or studies from the perspective of corporate “greenwashing” behavior and financial constraints. However, there is still a lack of research on the transmission mechanism of voluntary climate risk disclosure affecting green technology innovation. Based on this, this paper utilizes text analysis to construct climate risk disclosure indicators and systematically researches its impact on corporate green technological innovation, focusing on the transmission mechanism of media attention and agency cost in this process. Meanwhile, the heterogeneity analysis reveals the differences between enterprise size and financing constraints.

2.2. Hypothesis Formulation

2.2.1. Climate Risk Disclosure and Corporate Green Technology Innovation

Combining the results of previous research, this paper argues that the core logic of the impact of climate risk on corporate innovation can be summarized in the following three aspects: first, external pressure transmission. Climate risk disclosure requirements (e.g., the Task Force on Climate-related Financial Disclosure (TCFD) framework), ESG reports force companies to face environmental compliance pressure, to maintain capital market trust and reputation. According to stakeholder theory, companies need to balance all parties’ interests in their operations and proactively address environmental demands from external stakeholders—including society, governments, investors, and the public. They must demonstrate ecological accountability and a commitment to sustainability by reducing carbon emissions, mitigating environmental risks, and advancing targeted innovations.

Secondly, internal resources are reconfigured. S2 requires subjects to report climate risks and opportunities they face, where risks are categorized as physical risks and climate-related transformation risks. These requirements prompt companies to reassess the impact of environmental risks on long-term operations, push management to incorporate green technologies into strategic priorities, and shift more funds from traditional high-pollution businesses to green technology research and development. Meanwhile, according to the information asymmetry theory, in the context of increasingly stringent environmental policies and rising public awareness of environmental protection, climate risk disclosure can effectively reduce information asymmetry between enterprises and external stakeholders. It has been shown that enterprises’ awareness and disclosure of carbon risk improves information transparency, puts them under the supervision of the public and the capital market, and reduces the information asymmetry of investors or creditors (Jung et al., 2018) [23], and the significant improvement of information transparency can increase the R&D investment and patent output of enterprises (Wang, K., 2021) [24]. And both climate-related transition risk and physical risk affect the performance of global stock markets (Zhang, 2022) [25], as well as significantly affect the pricing and term structure of the credit default swap (CDS) market, especially the disclosure of transition risk significantly increases the CDS spreads of high-carbon-emitting firms (Kölbel et al., 2020) [26]. Conversely, when firms disclose the impact of climate risks (both physical and climate-related transition risks) on operational and financial performance, poor climate performance may not only lead to reputational damage, but also to higher audit fees (Yu and Si, 2023) [27] and higher financing costs (Huang et al., 2022) [12]. These negative consequences will objectively force enterprises to comprehensively identify and compensate for weaknesses in green transformation, further increase investment in green technology innovation, and optimize internal resource allocation, so as to better cope with climate change risks.

The third is competitive advantage building. Disclosure of climate risks and demonstration of green technology progress can strengthen the image of “environmental responsibility” of enterprises, make them more clearly recognize their own opportunities and challenges in the low-carbon transition, so as to accelerate the adjustment of technological innovation, business model and positioning of the value chain, enhance brand premium, help win policy subsidies or tax incentives, and further reduce the cost of innovation. Innovation costs. In addition, by voluntarily disclosing environment-related information, firms are able to enhance their reputational capital, strengthen their governance structure and reduce the transparency of information on financing costs (Feng et al., 2024) [17], which in turn helps to increase the number of green patents and R&D investment by firms. Compared with the climate transition risk, climate-related physical risks such as extreme weather and floods will form an immediate impact on enterprises by causing direct economic losses (e.g., asset damage and supply chain disruption), forcing enterprises to improve climate resilience through green technology innovation to cope with the risk; at the same time, the disclosure of physical risks may signal to the market that the enterprise is facing an existential crisis, triggering investor concerns about the long-term operating capability. At the same time, physical risk disclosure may signal to the market that the enterprise is facing an existential crisis, triggering investors’ concern about long-term operational capability, and then forcing the enterprise to accelerate the layout of green technology and reconstruct competitive advantages through innovation to stabilize investor expectations. Based on the above analysis, this paper proposes Hypothesis 1 and its sub-hypotheses:

H1. 

Climate risk disclosure has a positive effect on corporate green innovation, ceteris paribus;

H1a. 

Physical risk disclosure has a stronger positive impact on firms’ green technological innovation than climate-related transition risk disclosure.

2.2.2. Mediating Effects of Media Attention

As an important mechanism of corporate external governance, the role of media attention in corporate environmental behavior decision making has been increasingly emphasized by the academic community. The literature has been explored from different perspectives: in terms of governance effectiveness, Jin et al. (2024) [28] find that media reports can positively strengthen the role of institutional investors’ ESG activism in the promotion of corporate green innovation, and Lu, D. et al. (2015) [29] find that the traditional media significantly improves the quality of firms’ internal control through the mechanism of reputational pressure. Kong et al. (2020) [30] find that media attention motivates enterprises to increase their environmental protection efforts. At the level of behavioral driving mechanism, Jie and Jiahui (2023) [31] found that media attention positively promotes corporate green technology innovation and subsequently sustainable development. Zhang, Y. et al. (2021) [32] empirically found that media attention significantly promotes the green technology innovation behavior of heavily polluting enterprises, especially when there are more negative reports, enterprises are more inclined to adopt green technology to improve their public image, while Aerts and Cormier (2009) [33] demonstrated that negative news about the environment can significantly increase the level of environmental disclosure by enterprises, and McCombs and Shaw (1972) [34] illustrated that the media influences stakeholder perceptions through issue visualization.

In summary, by actively disclosing information on climate impacts, companies can demonstrate their attention to and commitment to environmental responsibility, and in particular, by revealing the climate risks they face, they enable the media to gain a more comprehensive understanding of their environmental performance and management measures. Increased transparency further encourages the media to be more willing to pay attention to and report on enterprise-related developments, as they often have access to more accurate and in-depth content, and the media often rank, compare and feature enterprises based on the disclosed data to analyze and monitor their performance in combating climate change, which pushes the environmental performance of enterprises to the forefront of public discussion.

On the contrary, when a firm’s environmental performance is reported by the media as backward or non-compliant, the firm will face great public pressure. In particular when there are more negative reports, firms are more inclined to adopt green technologies to improve their public image (Zhang, Y. et al., 2021) [32]. Meanwhile, for those enterprises with transparent disclosure and excellent performance, media reports can amplify their positive impacts and help them attract green investment, policy support, and government subsidies, thus obtaining sufficient financial support. Therefore, in order to avoid being at a disadvantage in public opinion and market competition, companies tend to continuously increase their investment in green technologies after climate risk disclosure, enhance their innovation capacity, and accelerate their green transformation. The monitoring role of the media can also make up for the lack of environmental regulation, increase public attention to the polluting behavior of enterprises, increase the reputational risk and regulatory pressure on enterprises, and motivate them to improve their behavior (Campa P. et al., 2018; Liu, Y. et al., 2023) [35,36].

Based on this, this paper proposes Hypothesis 2:

H2. 

Climate risk disclosure promotes corporate green technology innovation by increasing media attention.

2.2.3. Mediating Effects of Agency Costs

Principal-agent theory suggests that the separation of ownership and control leads to information asymmetry and conflicting objectives between shareholders and management, and the resulting agency costs constitute a central proposition in corporate governance research. The established literature centers on the dimensions of formal institutions and informal mechanisms: from the perspective of formal institutions, Kini and Williams (2012) [37] found that too large a gap in compensation within the executive team can make managers overly risky or short-sighted, and Watts and Zimmerman (1983) [38] found that audits, by independently verifying financial information can reduce management’s opportunistic behavior, thereby reducing monitoring costs and residual losses. Informally, research has focused on the fact that an increase in the proportion of women in the executive team (Jurkus et al., 2011) [39], private contracting (Henry, 2010) [40], and voluntary disclosure behavior (Chung et al., 2015) [41] can reduce principal-agent costs.

Agency conflicts may lead management to avoid high-risk long-term investments and generate mismatch of innovation resources (Jensen et al., 2019) [42], and even reduce firms’ innovation efficiency through R&D manipulation (Wan, Y. and Xu, Y., 2019) [43]. Climate risk disclosure can effectively alleviate the information asymmetry between management and shareholders by improving corporate transparency. On the one hand, the act of disclosure makes shareholders more aware of management’s specific actions and commitments in green technology R&D and climate risk response, thus strengthening the supervision of management and avoiding its opportunistic behaviors, thus focusing more on long-term value creation. On the other hand, the act of disclosure reveals a company’s strategy for addressing climate change and the status of its resource allocation, helping shareholders to identify possible poor investment decisions by management (e.g., projects overly pursuing short-term profits) and to promote the tilting of resources toward medium- to long-term green technology R&D that is of sustainable development significance. In addition, climate risk disclosure further regulates management behavior and reduces shareholders’ monitoring costs by improving corporate internal governance (Feng L. et al., 2024) [17]. Therefore, this paper argues that climate risk disclosure prompts management to pay more attention to medium- and long-term value creation and to utilize more resources in the R&D and application of green technology innovation by reducing agency costs. Based on this, this paper proposes Hypothesis 3:

H3. 

Climate risk disclosure promotes firms’ green technology innovation by reducing the agency cost between shareholders and management.

3. Research Design

3.1. Sample Selection and Data Sources

To delve into the aforementioned issues, this paper takes as its sample the data of Chinese A-share listed companies spanning from 2004 to 2022. Subsequently, the following processing steps are implemented: First, financial listed companies are eliminated; second, delisted ST and PT companies are excluded; third, a 1% winsorization at both the upper and lower ends is conducted on all continuous variables. The climate risk disclosure information is sourced from the annual reports of listed companies. The data regarding companies’ green innovation are obtained from the CNRDS database, and the data on media attention are also retrieved from the CNRDS database. For the remaining variables, they are derived from the Cathay Pacific database.

3.2. Variable Definition and Measurement

3.2.1. Explained Variables

The dependent variable in this study is corporate green technology innovation (green_a), measured by the number of green patent applications submitted by listed companies, following the methodology of Xu, J. and Cui, J. (2020) [44]. Green patents are classified into three categories: total green patents (green_a), green invention patents (green_p, and green utility model patents (green_s), with all variables log-transformed to normalize the distribution.

3.2.2. Explanatory Variables

Currently, Chinese dictionaries do not provide a clear concept of climate change, so this paper draws on Guo and Huang’s (2024) [45] “seed word + Word2Vec similarity word expansion” approach to obtain the 100 most similar words reflecting corporate climate change risk disclosure covering a broad definition (see Appendix A 

Table A1. 100 Terms for Measuring Climate Change Risks.

Ozone	Energy Investment	Battery Electric	Water Resources
Large-scale Solar Energy	Extreme Climate	Renewable Resources	Circular Economy
Energy Consumption	Carbon Trading	Clean Air	Wind Farm
Catastrophe Insurance	Carbon Financing	Energy Storage Battery	Energy Consumption
Ice and Snow	Air Pollution	Electric Vehicle	Waste Discharge
Sponge City	Emission Reduction	Air Quality	Typhoon
Energy Regulation	Climate Change	Clean Fuel	Coastal Areas
Carbon Capture	Carbon Market	Recycling	Wind Power Generation
Plug-in Hybrid	Geothermal Energy	New Energy Vehicles	Degradation
Environmental Standards	Carbon Emission Reduction	Greenhouse Gases	Solar Energy
Green Mountains and Clear Waters	Energy Conservation and Emission Reduction	Carbon Sink	New Development Concepts
Carbon Peak	Climate Warming	Green Buildings	Environmental Disclosure
Low-Carbon Technology	Carbon Tax	Resource Conservation	Global Warming
Environmental Issues	Low Carbon	Pollution Control	Pollutant Emissions
Energy Demand	Energy Saving and Emission Reduction	Secondary Energy	Pollution Treatment
Carbon Sequestration	Climate Risks	Green Finance	Energy-Efficient Renovation
Tidal Energy	Carbon Neutrality	Forest Conservation	Sustainable Development
Environmentally Friendly	Green Transition	New Energy	Climate Improvement
Energy Crisis	Renewable	Carbon Dioxide	Clean Production
Carbon Tariff	Gas Emissions	Green Economy	Energy Storage
Charging Pile	Carbon Footprint	Global Energy	Green Economic Transition
Ecology	Low-Carbon Transition	New Energy Vehicles	World Energy Resources
Green Environmental Protection	Renewable Energy	Storm Surge	Sustainable Utilization
Carbon Emissions	Clean Power Sources	Energy Transformation	Low-Carbon City
Transmission Network	Clean Energy Supply	Energy Reduction	Circular Utilization

), the lexicon method was used to construct a measure of the degree of corporate climate change risk disclosure, and the degree of corporate-level climate change risk disclosure was measured by the percentage of the total number of climate risk-related words in the text of the CSR report.

3.2.3. Mediating Variables

In order to explore potential mechanisms, this study introduces two mediating variables. The first variable is media attention (ln_t), measured by aggregating positive and negative media coverage related to the firm and applying a logarithmic transformation to normalize the data. The second variable is the first type of agency cost (Ofee), which is represented by the operating expense ratio, which is calculated by dividing the sum of administrative and selling expenses by the total revenue from main operations.

3.2.4. Moderating Variables

Firm Size (Size): This study measures firm size using the number of employees.

Financing constraints (FC): This paper refers to Hadlock and Pierce (2009) [46] to establish a measure of corporate financing constraints FC index.

3.2.5. Control Variables

Control variables: Following the existing literature (Sun et al., 2024 [47]; Feng et al., 2024 [17]), this study controls for leverage ratio (Lev), return on total assets (ROA1), return on equity (ROE), revenue growth (Growth), board size (Board), Tobin’s Q (TobinQ), largest shareholder ownership (Top1), top three shareholders ownership (Top3), and management ownership (Mshare). The variables are defined as shown in

Table 1. Variable definitions.

Variable Type	Variable Name	Variable Symbol	Variable Definition
Dependent Variable	Corporate Green Technology Innovation	green_a	Logarithm of the sum of green invention and utility model patent applications, plus 1
		green_p	Logarithm of the number of green invention patent applications, plus 1
		green_s	Logarithm of the number of green utility model applications, plus 1
Independent Variable	Climate Risk Disclosure Index	pl	Measured as the percentage of climate risk-related terms in the corporate social responsibility report
Mediator Variable	Media Attention	ln_t	Logarithmic transformation of the total number of positive and negative media reports
	First-Category Agency Cost	Ofee	The sum of management and sales expenses divided by main business income
moderator variable	Enterprise size	Size	Number of employees employed by the enterprise
	Financing constraints	FC	FC index
Control Variable	Leverage Ratio	Lev	Total liabilities/total assets
	Return on Assets	ROA1	Net profit/average total assets
	Return on Equity	ROE	Net profit/average owner’s equity
	Revenue Growth Rate	Growth	Revenue growth rate
	Board Size	Board	It is the natural logarithm of the count of board members
	Tobin’s Q	TobinQ	(Market value of equity + liabilities)/total assets
	Largest Shareholder Ownership	Top1	Found by dividing the shares held by the largest shareholder by the total number of shares
	Top Three Shareholders Ownership	Top3	The proportion of shares owned by the three largest shareholders compared to the overall number of shares
	Management Shareholding Ratio	Mshare	The percentage of shares held by directors, supervisors, and senior management in relation to the total capital of the company

.

3.3. Model Design

The Baseline Regression Model

$${green\_a}_{i,t}={\beta }_0+{\beta }_1{pl}_{i,t}+{\beta }_2{controls}_{i,t}+{YEAR}_t+{\gamma }_i+{\epsilon }_{i,t}$$

(1)

Model (1) is the baseline regression model of this paper. ${\beta }_1$ is the focus of this paper, and its economic meaning is the impact of firms’ climate risk disclosure on their green technology innovation. YEAR and γ are year fixed effects and individual fixed effects, and this paper adopts the standard error of clustering at the firm level.

4. Empirical Results and Analysis

4.1. Descriptive Statistics

The results of the descriptive statistics of the main variables are shown in

Table 2. Descriptive statistics.

Variable	(1)	(2)	(3)	(4)	(5)
	N	Mean	SD	Min	Max
pl	33,782	0.120	0.271	0	1.308
green_a	33,782	0.323	0.723	0	3.466
green_p	33,782	0.215	0.569	0	2.996
green_s	33,782	0.186	0.507	0	2.639
Lev	33,782	0.414	0.194	0.027	0.908
ROA1	33,782	0.046	0.061	−0.373	0.257
ROE	33,782	0.075	0.114	−0.926	0.470
Growth	33,782	0.175	0.372	−0.658	4.024
Board	33,782	2.137	0.203	1.099	2.773
Top1	33,782	0.348	0.150	0.080	0.758
Top3	33,782	0.492	0.153	0.151	0.887
TobinQ	33,782	1.945	1.206	0.802	15.610
Mshare	33,782	0.134	0.195	0	0.706

, which shows that the standard deviation of the total number of applications for green patents of the dependent variable enterprises is large, and the green innovation gap between enterprises is more obvious, and similarly, the reporting of climate risk information exhibits a wide range of variation.

4.2. The Impact of Climate Risk Information Disclosure on Corporate Green Technology Innovation

The empirical results are shown in

Table 3. Baseline regression results.

	(1)	(2)	(3)	(4)	(5)
	green_a	green_p	green_s	green_a	green_a
pl	0.0972 ***	0.0904 ***	0.0430 **
	(3.7988)	(4.2880)	(2.1498)
pl_phy				0.2434 **
				(2.0997)
pl_tranf					0.0692 **
					(2.0193)
Lev	0.1007 **	0.0857 ***	0.0529 *	−0.1788	−0.1933 *
	(2.5642)	(2.6263)	(1.9140)	(−1.5696)	(−1.7057)
ROA1	−0.2145	−0.0943	−0.1109	−0.0133	−0.0083
	(−1.2580)	(−0.7557)	(−0.8527)	(−0.0413)	(−0.0258)
ROE	0.1619 *	0.0952	0.0813	0.0911	0.0938
	(1.8467)	(1.5390)	(1.1814)	(0.5526)	(0.5715)
Growth	−0.0259 ***	−0.0219 ***	−0.0177 ***	−0.0189	−0.0176
	(−3.8782)	(−4.1603)	(−3.3668)	(−1.1157)	(−1.0406)
Board	−0.0158	−0.0312	0.0147	−0.1246	−0.1259
	(−0.4608)	(−1.0969)	(0.5945)	(−1.5742)	(−1.5927)
TobinQ	−0.0019	0.0010	−0.0022	−0.0051	−0.0051
	(−0.4601)	(0.2964)	(−0.7949)	(−0.5599)	(−0.5720)
Top1	−0.1097	−0.0745	−0.0988	0.0011	−0.0020
	(−1.2282)	(−0.9747)	(−1.6085)	(0.0050)	(−0.0090)
Top3	−0.1228	−0.0563	−0.0460	−0.5813 **	−0.5704 **
	(−1.4814)	(−0.8462)	(−0.8081)	(−2.5373)	(−2.5028)
Mshare	0.0513	0.0010	0.0326	0.0458	0.0662
	(1.0349)	(0.0261)	(0.9226)	(0.2799)	(0.4079)
_cons	0.4005 ***	0.2880 ***	0.1873 ***	1.1125 ***	1.0920 ***
	(4.8684)	(4.1844)	(3.2446)	(5.3785)	(5.2891)
Company FE	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES
N	33,268	33,268	33,268	7252	7252
r2_a	0.6114	0.5800	0.5523	0.7292	0.7293

Note: * p < 0.1, ** p < 0.05, and *** p < 0.01, with t-values in parentheses.

. In column (1), the coefficient of climate risk disclosure is positive at the 1% significance level, indicating that climate risk disclosure significantly promotes enterprises’ green technological innovation. Further analysis reveals that climate risk disclosure has a significantly positive impact on different types of patents, with a slightly stronger effect on green invention patents. This suggests that climate risk disclosure can, to some extent, encourage enterprises’ green technological innovation behavior, thereby validating Hypothesis 1. This study further divides climate disclosure into climate risk disclosure and physical risk disclosure based on the S2 standard (see Appendix A 

Table A2. Climate risk classification.

Climate Transition Risk Categories	Glossary of Core Concepts
Policy and Regulatory Risks	Carbon Tax, Carbon Tariff, Environmental Standards, Energy Regulation, Carbon Emissions, Peak Carbon, Carbon Emission Reduction, Carbon Neutrality
Technology Substitution Risk	Renewable Energy, New Energy Vehicles, Electric Vehicles, Charging Piles, Carbon Capture Carbon Sequestration
Market and Demand Risk	Carbon Market, Carbon Trading, Financing, Green Finance, Energy Consumption, Energy Demand, Circular Economy
Stranded Asset Risk	Fossil Energy, Energy Crisis
Legal and Litigation Risk	Environmental Disclosure, Pollution Control, Environmental Litigation
Reputation Risk	Environmentally friendly, green water and mountains, sustainable development, ecology, resource conservation
Physical Risks	Core Vocabulary
Acute Risks	Typhoon, Storm Surge, Extreme Weather, Sea Level Rise
Chronic Risks	Climate Warming, Climate Change, Climate Risk, Air Pollution, Carbon Dioxide, Coastal Areas, Environmental Issues, Pollutant Emissions, Global Warming, Ozone, Pollutant Discharge

) to explore the impact of different types of climate risks. The regression results, as shown in columns (4) and (5) of Table 3, indicate that both physical risks and climate-related transition risks promote enterprises’ green innovation. However, the impact of physical risks on enterprises’ green technological innovation is stronger.

5. Robustness Tests

5.1. Handling Endogeneity Issues: The Instrumental Variables Method

There may be a bidirectional causal relationship between corporate climate risk information disclosure and green technology innovation, for this reason, taking into account, local regulation, facing climate risk and other factors, the disclosure level of other enterprises in the same city has a certain correlation with the disclosure behavior of the target enterprise, but its direct impact on the green innovation activities of a single enterprise is small, this paper selects the enterprise’s city in the same year the other enterprises climate risk information The mean value (IV) of the disclosure level of other enterprises in the same city in the same year is chosen as the instrumental variable.

Table 4. The instrumental variable test.

	First Stage		Second Stage
	pl	green_a	green_p	green_s
IV	0.8424 ***
	(23.8068)
pl		0.1375 ***	0.0784 **	0.0966 **
		(2.7860)	(2.0555)	(2.5490)
Lev	0.0509 ***	0.0981 **	0.0865 ***	0.0494 *
	(2.7307)	(2.4961)	(2.6375)	(1.7963)
ROA1	−0.1019	−0.2104	−0.0955	−0.1055
	(−1.3355)	(−1.2319)	(−0.7635)	(−0.8098)
ROE	0.0698 *	0.1594 *	0.0959	0.0780
	(1.7970)	(1.8129)	(1.5433)	(1.1315)
Growth	−0.0123 ***	−0.0253 ***	−0.0221 ***	−0.0170 ***
	(−3.9595)	(−3.7609)	(−4.1491)	(−3.2014)
Board	−0.0041	−0.0164	−0.0310	0.0139
	(−0.2817)	(−0.4787)	(−1.0879)	(0.5654)
TobinQ	−0.0007	−0.0018	0.0010	−0.0022
	(−0.4600)	(−0.4434)	(0.2902)	(−0.7605)
Top1	−0.0291	−0.1077	−0.0751	−0.0961
	(−0.6612)	(−1.2065)	(−0.9811)	(−1.5591)
Top3	0.0528	−0.1254	−0.0555	−0.0495
	(1.2668)	(−1.5138)	(−0.8345)	(−0.8650)
Mshare	0.0142	0.0507	0.0012	0.0318
	(0.7077)	(1.0231)	(0.0306)	(0.9018)
Kleibergen–Paap rk LM statistic			162.360 [0.000]
Kleibergen–Paap rk Wald F statistic			566.783
_cons	−0.0110
	(−0.3027)
Company FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
N	33,268	33,268	33,268	33,268
r2_a	0.5746	0.0027	0.0029	0.0002

Note: * p < 0.1, ** p < 0.05, and *** p < 0.01, with t-values in parentheses.

demonstrates the two-stage least squares estimation results of the instrumental variable analysis. From the results, it can be seen that the instrumental variables are valid and the promotion effect of climate risk information disclosure on firms’ green technology innovation is significantly positive under IV-2SLS estimation. Thus, the results of the instrumental variable analysis indicate that the findings remain robust after controlling for endogeneity issues.

5.2. One-Period Lag

The lagged effect of climate risk information disclosure on corporate green technology innovation is considered in this study. To address this, we employ a one-period lagged disclosure variable (L1_pl) as the core explanatory variable and estimate its relationship with firms’ sustainable innovation outcomes. As shown in

Table 5. The one-period lag robustness test.

	(1)	(2)	(3)
	green_a	green_p	green_s
L1_pl	0.0907 ***	0.0758 ***	0.0495 **
	(3.5448)	(3.5718)	(2.4481)
Lev	0.0956 **	0.0855 ***	0.0475 *
	(2.4092)	(2.5806)	(1.7094)
ROA1	−0.2078	−0.1055	−0.0907
	(−1.2057)	(−0.8329)	(−0.6919)
ROE	0.1592 *	0.1001	0.0728
	(1.7979)	(1.5987)	(1.0496)
Growth	−0.0253 ***	−0.0218 ***	−0.0171 ***
	(−3.7723)	(−4.0979)	(−3.2476)
Board	−0.0149	−0.0313	0.0161
	(−0.4334)	(−1.0935)	(0.6539)
TobinQ	−0.0018	0.0012	−0.0024
	(−0.4431)	(0.3346)	(−0.8443)
Top1	−0.1140	−0.0743	−0.1021 *
	(−1.2703)	(−0.9664)	(−1.6608)
Top3	−0.1241	−0.0613	−0.0452
	(−1.4873)	(−0.9160)	(−0.7914)
Mshare	0.0706	0.0172	0.0423
	(1.4455)	(0.4393)	(1.2143)
_cons	0.4011 ***	0.2908 ***	0.1850 ***
	(4.8629)	(4.1964)	(3.2087)
Company FE	YES	YES	YES
Year FE	YES	YES	YES
N	32,578	32,578	32,578
r2_a	0.6131	0.5813	0.5535

Note: * p < 0.1, ** p < 0.05, and *** p < 0.01, with t-values in parentheses.

, the coefficients remain statistically significant at the 1% level, indicating that lagged climate risk disclosure significantly increases the total volume of environmentally focused patents. Further disaggregation by patent type reveals consistently positive coefficients across all categories, confirming the robustness of the results.

5.3. Replacing the Explanatory Variable

This study replaces the original explanatory variable with the total word frequency of climate risk disclosure in the text and applies a logarithmic transformation after adding 1. As shown in

Table 6. The replaced explanatory variable robustness test.

	(1)	(2)	(3)
	green_a	green_p	green_s
ln_num	0.0249 ***	0.0227 ***	0.0106 ***
	(4.5771)	(5.0114)	(2.5943)
Lev	0.1012 ***	0.0862 ***	0.0532 *
	(2.5794)	(2.6461)	(1.9261)
ROA1	−0.2146	−0.0946	−0.1112
	(−1.2596)	(−0.7590)	(−0.8542)
ROE	0.1610 *	0.0944	0.0810
	(1.8384)	(1.5305)	(1.1770)
Growth	−0.0255 ***	−0.0216 ***	−0.0176 ***
	(−3.8293)	(−4.1124)	(−3.3458)
Board	−0.0166	−0.0319	0.0144
	(−0.4841)	(−1.1225)	(0.5814)
TobinQ	−0.0019	0.0010	−0.0023
	(−0.4728)	(0.2828)	(−0.8028)
Top1	−0.1063	−0.0714	−0.0974
	(−1.1950)	(−0.9397)	(−1.5898)
Top3	−0.1289	−0.0617	−0.0485
	(−1.5575)	(−0.9301)	(−0.8528)
Mshare	0.0506	0.0005	0.0323
	(1.0220)	(0.0118)	(0.9154)
_cons	0.3993 ***	0.2869 ***	0.1868 ***
	(4.8585)	(4.1805)	(3.2349)
Company FE	YES	YES	YES
Year FE	YES	YES	YES
N	33,268	33,268	33,268
r2_a	0.6117	0.5803	0.5523

Note: * p < 0.1 and *** p < 0.01, with t-values in parentheses.

, climate risk disclosure significantly boosts green patent filings at the 1% significance level, including both invention patents and utility model patents.

5.4. The Placebo Test

In the above tests, the relationship between climate risk information disclosure and corporate green technology innovation may reflect natural time trends or be influenced by other random factors. To address this, this paper conducts a placebo test by randomly constructing a climate risk information disclosure index. Specifically, random values are assigned to generate a pseudo climate risk information disclosure dummy variable. This pseudo variable is used to replace the actual climate risk information disclosure variable, and the sample is then substituted into model (1) for regression analysis. The process is repeated 1000 times.

Figure 1 presents the results of the placebo test. The coefficient distribution is symmetric around zero, with the placebo estimates significantly differing from the baseline regression coefficient (0.0972; see Table 3, column 1). Additionally, the p-values for most placebo coefficients are statistically insignificant, as 90% of them exceed 0.1. These findings validate the benchmark results, thereby supporting a robust association between climate risk disclosure and corporate green innovation.

5.5. The Replacement Model

Given that the explanatory variable green patent applications is a count-type ordinal data and suffers from the problem of being significantly over-discrete, this paper further tests it using a fixed-effects negative binomial regression model. The results reported in

Table 7. The negative binomial regression test.

	(1)	(2)	(3)
	GREEN_A	GREEN_P	GREEN_S
pl	0.1466 **	0.1622 **	0.0440
	(2.3239)	(2.2838)	(0.4544)
Lev	0.3501 **	0.3640 *	0.3000
	(2.1054)	(1.7331)	(1.4046)
ROA1	−1.0359	−0.5419	−1.1319
	(−1.2692)	(−0.5594)	(−1.0502)
ROE	0.9179 **	0.7225	0.9143 *
	(2.3176)	(1.5385)	(1.8862)
Growth	−0.1101 **	−0.1289 ***	−0.1253 ***
	(−2.4032)	(−2.9738)	(−2.7233)
Board	0.1249	0.1199	0.2497
	(0.8006)	(0.6447)	(1.5748)
TobinQ	0.0014	0.0184	−0.0086
	(0.0840)	(0.9011)	(−0.3911)
Top1	0.1693	0.3012	−0.1899
	(0.4705)	(0.6545)	(−0.4149)
Top3	−0.1940	−0.2096	0.1721
	(−0.6030)	(−0.4726)	(0.4286)
Mshare	0.2005	−0.0582	0.2517
	(1.2453)	(−0.2348)	(1.1972)
_cons	−2.5263 ***	−2.6483 ***	−2.5813 ***
	(−6.3471)	(−5.1135)	(−5.6149)
Company FE	YES	YES	YES
Year FE	YES	YES	YES
N	18,795	15,826	14,758

Note: * p < 0.1, ** p < 0.05, and *** p < 0.01, with t-values in parentheses.

show that after controlling for firm-individual and time effects, climate risk disclosure (pl) still exhibits a significant facilitating effect on green technology innovation (GREEN_A). Distinguishing between green patent categories reveals a large and significant impact coefficient on invention patents (GREEN_P), which is consistent with the quantitative comparison pattern of the benchmark regression. Overall, the impact of climate risk disclosure on corporate green technology innovation is still significantly positive, further supporting the empirical findings of this paper.

6. Mechanism Analysis

6.1. Media Attention

$${green\_a}_{i,t}={\beta }_0+{\beta }_1{pl}_{i,t}+{\beta }_2{controls}_{i,t}+{YEAR}_t+{\gamma }_i+{\epsilon }_{i,t}$$

(2)

$${ln\_t}_{i,t}={\alpha }_0+{\alpha }_1{pl}_{i,t}+{\alpha }_2{controls}_{i,t}+{YEAR}_t+{\gamma }_i+{\epsilon }_{i,t}$$

(3)

$${green\_a}_{i,t}={\gamma }_0+{\gamma }_1{pl}_{i,t}+{\gamma }_2{ln\_t}_{i,t}+{\gamma }_3{controls}_{i,t}+{YEAR}_t+{\gamma }_i+{\epsilon }_{i,t}$$

(4)

Table 8. The media attention mediation effect test.

	(1)	(2)	(3)
	green_a	ln_t	green_a
pl	0.0972 ***	0.1440 ***	0.0915 ***
	(3.7988)	(4.2832)	(3.5394)
ln_t			0.0281 ***
			(5.9121)
Lev	0.1007 **	0.4275 ***	0.0938 **
	(2.5642)	(5.2936)	(2.3679)
ROA1	−0.2145	0.5940 *	−0.1798
	(−1.2580)	(1.6851)	(−1.0775)
ROE	0.1619 *	0.3547 **	0.1124
	(1.8467)	(1.9860)	(1.3341)
Growth	−0.0259 ***	0.0214	−0.0267 ***
	(−3.8782)	(1.5033)	(−3.9393)
Board	−0.0158	0.1203 **	−0.0210
	(−0.4608)	(1.9748)	(−0.5988)
TobinQ	−0.0019	0.0712 ***	−0.0035
	(−0.4601)	(9.5080)	(−0.8324)
Top1	−0.1097	−0.4857 ***	−0.1023
	(−1.2282)	(−2.6766)	(−1.1365)
Top3	−0.1228	0.3069 *	−0.1263
	(−1.4814)	(1.8864)	(−1.5032)
Mshare	0.0513	0.2620 **	0.0419
	(1.0349)	(2.5627)	(0.8289)
_cons	0.4005 ***	2.5104 ***	0.3347 ***
	(4.8684)	(17.2156)	(3.9590)
Company FE	YES	YES	YES
Year FE	YES	YES	YES
N	33,268	32,321	32,321
r2_a	0.6114	0.8353	0.6148

Note: * p < 0.1, ** p < 0.05, and *** p < 0.01, with t-values in parentheses.

neatly sums up the estimation outcomes for models (2) through (4). The regression coefficients of both climate risk disclosure and media attention in relation to corporate green technology innovation are statistically significant at the 1% mark. This clearly shows that media attention has a partial mediating function. These results offer empirical backing for Hypothesis 2.

Further validation through 1000 bootstrapping iterations confirms this mechanism: as shown in

Table 9. The bootstrap method mediation effect test.

Dependent Variable	Green Technology Innovation
Mediating Variable	Media Attention
Indirect Effect	0.004 ***
Direct Effect	0.092 ***

Note: *** p < 0.01.

, the estimated indirect effect of media attention is 0.004, with a confidence interval of [0.002, 0.006]. These results suggest that media attention serves as a significant mediator in transmitting the impact of climate risk disclosure on green technology innovation outcomes.

6.2. Agency Costs

$${green\_a}_{i,t}={\beta }_0+{\beta }_1{pl}_{i,t}+{\beta }_2{controls}_{i,t}+{YEAR}_t+{\gamma }_i+{\epsilon }_{i,t}$$

(5)

$${Ofee}_{i,t}={\alpha }_0+{\alpha }_1{pl}_{i,t}+{\alpha }_2{controls}_{i,t}+{YEAR}_t+{\gamma }_i+{\epsilon }_{i,t}$$

(6)

$${green\_a}_{i,t}={\gamma }_0+{\gamma }_1{pl}_{i,t}+{\gamma }_2{Ofee}_{i,t}+{\gamma }_3{controls}_{i,t}+{YEAR}_t+{\gamma }_i+{\epsilon }_{i,t}$$

(7)

Models (5)–(7) conduct a mediating test for agency costs. The results are presented in

Table 10. The agency cost mediation effect test.

	(1)	(2)	(3)
	green_a	Ofee	green_a
pl	0.0972 ***	−0.0054 **	0.0952 ***
	(3.7988)	(−2.2210)	(3.6527)
Ofee			−0.2569 ***
			(−4.3337)
Lev	0.1007 **	−0.0838 ***	0.0781 *
	(2.5642)	(−9.2054)	(1.9540)
ROA1	−0.2145	−0.2203 ***	−0.2519
	(−1.2580)	(−5.0030)	(−1.4783)
ROE	0.1619 *	−0.0579 ***	0.1348
	(1.8467)	(−2.7745)	(1.5865)
Growth	−0.0259 ***	−0.0198 ***	−0.0313 ***
	(−3.8782)	(−12.6715)	(−4.5188)
Board	−0.0158	−0.0093	−0.0146
	(−0.4608)	(−1.3256)	(−0.4213)
TobinQ	−0.0019	0.0040 ***	−0.0007
	(−0.4601)	(4.9237)	(−0.1809)
Top1	−0.1097	−0.0183	−0.1124
	(−1.2282)	(−0.9334)	(−1.2597)
Top3	−0.1228	0.0436 **	−0.1069
	(−1.4814)	(2.5035)	(−1.2814)
Mshare	0.0513	0.0153	0.0465
	(1.0349)	(1.4471)	(0.9373)
_cons	0.4005 ***	0.2074 ***	0.4464 ***
	(4.8684)	(12.7832)	(5.2834)
Company FE	YES	YES	YES
Year FE	YES	YES	YES
N	33,268	32,858	32,858
r2_a	0.6114	0.7870	0.6121

Note: * p < 0.1, ** p < 0.05, and *** p < 0.01, with t-values in parentheses.

. In column (3), the regression coefficients for climate risk information disclosure (pl) and first-category agency costs (Ofee) on corporate green technology innovation are 0.0952 and −0.2569, respectively, both statistically significant at the 1% level. This suggests that climate risk information disclosure enhances corporate green technology innovation by reducing first-category agency costs. This study employs the Bootstrap method with 1000 resamples.

Table 11. The bootstrap method mediation effect test.

Dependent Variable	Green Technology Innovation
Mediating Variable	First-Category Agency Cost
Indirect Effect	0.001 ***
Direct Effect	0.095 ***

Note: *** p < 0.01.

shows that the indirect effect of first-category agency costs is significant at the 1% level, with a confidence interval of [0.000471, 0.002353], which does not include 0. Therefore, Hypothesis 3 is supported.

7. Heterogeneity Analysis

In this paper, we add model (8) and model (9) to the benchmark model, incorporating the interaction terms (pl × Size, pl × FC) of climate risk disclosure (pl) with firm size (size) and financing constraints (FC), respectively, to explore the differentiated characteristics between climate risk disclosure and green technology innovation. Among them, we focus on the coefficients of the mutual cross-multiplication terms, which, if the direction of the coefficients is consistent with the baseline regression, indicate that the moderating variables have a facilitating effect on the main effect; conversely, they reflect an inhibitory effect.

7.1. Heterogeneity in Firm Size

$${green\_a}_{i,t}={\omega }_0+{\omega }_1{pl}_{i,t}+{\omega }_2{Size}_{i,t}+{\omega }_3{pl}_{i,t}\times {Size}_{i,t}+{\omega }_4{controls}_{i,t}+{YEAR}_t+{\gamma }_i+{\epsilon }_{i,t}$$

(8)

Table 12. Analysis of firm size heterogeneity.

	(1)	(2)	(3)
	green_a	green_p	green_s
pl	−0.4664 ***	−0.3650 ***	−0.2678 **
	(−3.2134)	(−2.8441)	(−2.2334)
Pl×Size	0.0641 ***	0.0518 ***	0.0354 **
	(3.5847)	(3.2848)	(2.3568)
Size	0.0444 ***	0.0350 ***	0.0237 ***
	(4.7092)	(4.3640)	(3.8089)
Lev	0.0441	0.0412	0.0218
	(1.1150)	(1.2493)	(0.7905)
ROA1	−0.2078	−0.0904	−0.1053
	(−1.2229)	(−0.7286)	(−0.8090)
ROE	0.1385	0.0773	0.0679
	(1.5907)	(1.2635)	(0.9879)
Growth	−0.0274 ***	−0.0231 ***	−0.0185 ***
	(−4.0640)	(−4.3481)	(−3.4759)
Board	−0.0305	−0.0428	0.0064
	(−0.9022)	(−1.5195)	(0.2618)
TobinQ	0.0004	0.0028	−0.0011
	(0.0887)	(0.8259)	(−0.3852)
Top1	−0.0951	−0.0627	−0.0909
	(−1.0755)	(−0.8322)	(−1.4935)
Top3	−0.1575 *	−0.0840	−0.0649
	(−1.9072)	(−1.2698)	(−1.1502)
Mshare	0.0670	0.0130	0.0412
	(1.3504)	(0.3302)	(1.1645)
_cons	0.1239	0.0702	0.0408
	(1.1571)	(0.7859)	(0.5598)
Company FE	YES	YES	YES
Year FE	YES	YES	YES
N	33,240	33,240	33,240
r2_a	0.6133	0.5821	0.5532

Note: * p < 0.1, ** p < 0.05, and *** p < 0.01, with t-values in parentheses.

shows the results of the regression on firm size heterogeneity. The data show that the coefficient on the interaction term pl × Size is significantly positive at least at the 5% level, indicating a significant positive moderating effect of firm size. Specifically, the increase in firm size increases the difficulty for shareholders to monitor managers (Lv, C. et al., 2011) [48], making the information asymmetry problem more prominent. In larger firms, due to the difficulty of monitoring, managers may be more inclined to focus on short-term performance and personal interests while neglecting long-term value creation. As mentioned earlier, climate risk disclosure can reduce agency costs between shareholders and managers, limit managers’ opportunistic behavior, and ensure that firms’ strategic decisions are more consistent with long-term interests. Large corporations, which are more likely to receive public and media attention, tend to be accompanied by higher social expectations and public opinion pressure after their climate risk disclosures. If disclosure behavior is inconsistent with actual actions, it may lead to reputational damage and even trigger distrust among shareholders and investors. This external pressure forces large corporations to take more proactive actions, such as demonstrating their commitment to sustainable development through green technological innovations, in order to defuse potential public opinion risks and maintain market trust. With their strong technological and financing capabilities, large companies can quickly translate their disclosure commitments into concrete green innovation practices. In addition, their strengths in supply chain management and production process optimization, as well as their ability to integrate resources, provide strong support for the systematic implementation of green technological innovations, thus significantly enhancing their ability to cope with climate risks.

7.2. Heterogeneity Analysis of Financing Constraints

$${green\_a}_{i,t}={\theta }_0+{\theta }_1{pl}_{i,t}+{\theta }_2{FC}_{i,t}+{\theta }_3{pl}_{i,t}\times {FC}_{i,t}+{\theta }_4{controls}_{i,t}+{YEAR}_t+{\gamma }_i+{\epsilon }_{i,t}$$

(9)

Table 13. Heterogeneity analysis of financing constraints.

	(1)	(2)	(3)
	green_a	green_p	green_s
pl	0.1519 ***	0.1375 ***	0.0801 **
	(3.7112)	(4.0629)	(2.4410)
Pl×FC	−0.2331 ***	−0.2007 ***	−0.1570 **
	(−2.7233)	(−2.8394)	(−2.3412)
FC	−0.0322	−0.0277	−0.0137
	(−1.0491)	(−1.1179)	(−0.6305)
Lev	0.0660	0.0558	0.0343
	(1.5167)	(1.5434)	(1.1280)
ROA1	−0.2059	−0.0869	−0.1067
	(−1.2026)	(−0.6947)	(−0.8167)
ROE	0.1629 *	0.0960	0.0819
	(1.8555)	(1.5528)	(1.1885)
Growth	−0.0257 ***	−0.0217 ***	−0.0174 ***
	(−3.8507)	(−4.1298)	(−3.3169)
Board	−0.0178	−0.0329	0.0137
	(−0.5218)	(−1.1649)	(0.5601)
TobinQ	−0.0024	0.0006	−0.0024
	(−0.5817)	(0.1688)	(−0.8640)
Top1	−0.1001	−0.0662	−0.0928
	(−1.1213)	(−0.8692)	(−1.5127)
Top3	−0.1407 *	−0.0716	−0.0577
	(−1.7081)	(−1.0864)	(−1.0246)
Mshare	0.0537	0.0031	0.0331
	(1.0774)	(0.0778)	(0.9336)
_cons	0.4414 ***	0.3232 ***	0.2080 ***
	(5.2368)	(4.5991)	(3.5051)
Company FE	YES	YES	YES
Year FE	YES	YES	YES
N	33,268	33,268	33,268
r2_a	0.6117	0.5804	0.5525

Note: * p < 0.1, ** p < 0.05, and *** p < 0.01, with t-values in parentheses.

shows the results of the financing constraint heterogeneity regression. The data show that the coefficient on the interaction term pl × FC is significantly negative, at least at the 5% level, suggesting a significant negative moderating effect of financing constraints. Specifically, financing constraints impair firms’ green innovation capabilities (Yu et al., 2021; Madrid-Guijarro et al., 2016) [49,50]. At the same time, climate risk can significantly increase firms’ financing costs (Huang et al., 2022) [12], thus having an important impact on the financing environment for firms. In this context, enterprises with stronger financing capacity may be more inclined to cope with the potential impacts of climate risk through green technology innovation, mainly because: enterprises with stronger financing capacity usually have more complex business models and wider market layouts, and have more resources to support green technology research and development. After the disclosure of climate risk, enterprises are more capable of taking innovative measures quickly. Green technology innovation not only reduces the potential negative impact of climate risk on financing, but also demonstrates to investors their commitment and action in sustainable development, strengthens their resilience to climate change, further attracts ESG investment, and strengthens their financing ability, forming a “disclosure–green innovation–financing” cycle. This will further attract ESG investment and enhance financing capacity, forming a positive cycle of “disclosure–green innovation–financing”. On the other hand, if such enterprises do not take corresponding green technology innovation actions after disclosing climate risk information, their reputations may be negatively affected. Therefore, enterprises with higher financing capacity tend to promote green technology innovation after climate risk disclosure, so as to consolidate their reputation and investor trust through practical actions, and thus maintain their competitive advantages in the capital market.

8. Conclusions and Recommendations

8.1. Conclusions of This Study

This paper conducts an empirical investigation into how the disclosure of climate—risk information affects corporate green technology innovation. It makes use of the relevant data of Chinese A-share listed companies spanning from 2004 to 2022. The findings suggest that disclosing climate—risk information can give a boost to corporate green technology innovation. It does so by ramping up media attention and cutting down agency costs. Specifically, media attention, as an external governance mechanism, amplifies the signal of corporate environmental responsibility fulfillment, creating a dual driving force of public monitoring pressure and access to policy resources. Reduced agency costs, on the other hand, promote long-term green R&D investment by enhancing shareholders’ constraints on management and reducing resource mismatch. Heterogeneity effect analysis shows that the above effects are more significant in large enterprises and enterprises with lower financing constraints, reflecting the necessity of synergistic optimization of climate risk management strategies with corporate resource endowments.

This paper reveals the transmission logic of climate risk disclosure on green innovation through enhancing the effectiveness of media monitoring and mitigating agency conflicts: on the one hand, the active disclosure of climate risk information can attract media focus, and force enterprises to accelerate the layout of green technology through reputational incentives and accountability pressure, forming the external driving path of “disclosure–media focus–green technology investment”. On the other hand, disclosure enhances the efficiency of shareholders’ monitoring of management’s climate action, suppresses short-term opportunistic behavior, and promotes the tilting of resources to sustainable technology exploration, thus verifying the role of non-financial information disclosure in optimizing corporate governance. This finding offers a novel vantage point for comprehending the decision-making mechanism of corporate climate action. In addition, this paper finds that there is significant heterogeneity in the impact of climate risk disclosure on corporate green technology innovation in terms of firm size and financing constraints, which provides empirical references for government regulators to formulate differentiated climate risk management strategies.

8.2. Practical Implications

First, for government regulators, differentiated disclosure standards should be constructed, with mandatory disclosure of quantitative transformation risk data for high-carbon-emitting industries, while disclosure of climate adaptation measures should be emphasized for physical risk-sensitive industries such as agriculture and tourism. Regulators can specify specific disclosure requirements and standards by tier, industry and stage according to the scale of the enterprise, the motivation for financing and its position in the industrial chain, so as to ensure that the information disclosed by enterprises is true, complete and comparable.

Second, for enterprises, first, annual targets for green technology patents should be set and publicly displayed in climate reports to consolidate customer trust and market competitiveness. Second, indicators related to green technology innovation should be introduced into management appraisals and linked to management remuneration and performance incentives to enhance their initiative and motivation in addressing climate change.

Third, for the media and investors, it is recommended to develop an early warning system for climate risk public opinion, focusing on monitoring the degree of match between the transition risk commitments and actions of new energy and manufacturing enterprises, and strengthening the supervision of enterprises that “deviate from their words and actions” through negative news alerts, financing constraints and other measures. At the same time, a white paper on climate innovation index should be released jointly with industry associations to assess the disclosure quality and innovation effectiveness of enterprises by industry, and to guide the flow of capital to high-quality enterprises that genuinely promote green transformation.

Finally, there are some shortcomings in this paper: due to the limitations of the sample, this paper only discusses Chinese listed enterprises, and the observation of non-listed enterprises needs to be further expanded; moreover, since China has not yet established a unified climate risk disclosure framework, enterprises have greater autonomy in the disclosure content, indicator selection and quantitative caliber, and this paper is somewhat subjective in the selection and construction of climate risk disclosure indicators. In the future, non-listed enterprises and multinational corporations can be included in the scope of comparative analysis, and the quality of this study can be further improved with the help of the requirements of corporate sustainability disclosure standards.