A Study on the Quality and Determinants of Climate Information Disclosure of A-Share-Listed Banks

Abstract

Against the backdrop of growing climate concerns, banks’ credit exposures as a financial services industry are extremely vulnerable to climate risks, and banks should make good disclosures to respond to stakeholder demands. This paper develops an evaluation system to evaluate the quality of climate information disclosure of A-share-listed banks with respect to the characteristics of the banking industry. It finds that the quality of climate information disclosure of A-share-listed banks is not high but is increasing year by year. A multiple regression model is also constructed to empirically investigate the factors influencing the quality of climate information disclosure. The results show that corporate size, dual listing and board size make a significant positive contribution to the quality of climate information disclosure of A-share-listed banks, while the shareholding ratio of institutional investors is negatively related to the quality of climate information disclosure.

1. Introduction

In recent years, climate change has had a significant impact on ecosystems, posing a serious challenge to human life, and has become a widespread concern worldwide. The greenhouse effect is one of the most talked about aspects of climate change, and CO₂ from human activities is a major cause of this phenomenon [1]. The Intergovernmental Panel on Climate Change (IPCC) released a report in April 2022 stating that carbon emissions are still on the rise [2]. Unlike greenhouse gases such as methane or nitrous oxide, the effects of carbon dioxide emissions will persist over time and these climate changes are irreversible [3]. Among these climate changes, changes in precipitation will have a dramatic impact on ecosystems in arid and semi-arid regions, with food insecurity and desertification of arable land in South Asia and South Africa [4]. Another important climate change is sea level rise, with carbon dioxide emissions being the main cause of sea level rise over the last decade [5]. This will lead to the gradual disappearance of ice sheets at the poles and the erosion of coastal areas [6]. The impact of climate change on ecosystems is also significant, with the composition of biomes on 5–20% of the US land area likely to change by 2100 [7]. Greenhouse gas emissions are largely generated by corporate activities, so corporate action is critical to curbing greenhouse gas emissions. Corporations can significantly improve energy efficiency through actions such as reducing emissions, developing new technologies and setting targets [8]. Over the period 2011–2020, Chinese companies could reduce their consumption of 9.1 billion tons of standard coal through green and low-carbon development compared to the baseline scenario [9].

The physical risks and transition risks of climate change are gaining increasing attention, with stakeholders demanding climate information disclosure from companies. Climate information is more than just carbon information. It focuses on information related to climate change. Climate information is defined by the International Sustainability Standards Board (ISSB) in International Financial Reporting Standard (IFRS) S2 [10] Climate-related Disclosures (draft S2) as “information about the climate-related risks and opportunities faced by an enterprise”. In order for stakeholders to possess a comprehensive understanding of the measures taken by an enterprise in response to climate risks and opportunities, the corresponding strategies, actions, performance and results of the enterprise are included, and the stakeholders are more demanding for climate information and are not satisfied only with the data on greenhouse gas emissions but attach equal importance to the corporate governance and strategic adjustments of enterprises under climate change.

Banks were once considered to be an environmentally insensitive sector as their business itself does not generate carbon emissions directly. In recent years, however, the banking sector’s important nexus role in climate risk response has begun to gain traction, and there is a growing expectation that banks will play an active role in the transition to net zero emissions. As a financial services industry, banks can actively address climate change by expanding green investments, limiting financing for high-emission activities and supporting carbon-intensive industries in transition to net zero emissions. Regulators in different regions are actively considering a portfolio of policies to encourage banks to coordinate all stakeholders to address climate change [11]. In the financial sector, the Chinese government has accelerated the pace of introducing relevant policies since 2020. The “Guidance on Promoting the Investment and Financing in Response to Climate Change” was issued in 2020, raising the subject of climate investment and financing to an unprecedented level. “Measures for the Administration of the Law-based Disclosure of Environmental Information by Enterprises”, issued in 2022, further clarifies and refines the targets and contents of disclosure, requiring the disclosure of information on carbon emissions, investment and financing projects in response to climate change.

In addition, the amount of climate-risk-related assets in banks’ loan exposures is significant, with the US, EU, China, Japan and Switzerland together accounting for USD 1.6 trillion [8]. Climate risk can lead to a sudden increase in risk premiums on various assets of banks, amplifying credit risk and liquidity risk. Globally, as the trend towards economic decarbonization intensifies, financial assets associated with carbon-intensive industries could suffer losses that could be amplified by financial leverage. Models by Dietz et al. suggest that, if no action is taken, the climate value at risk (the amount of losses due to climate change) of global financial assets would amount to USD 2.5 trillion [12]. Battiston and Stefano Mandel validate the results of this model, showing that a high proportion of banks’ loan portfolios are exposed to climate risk [13]. One of the most discussed risks is the climate policy risk of a ban on fossil fuels in order to meet the temperature control target. This could lead to most of the assets of fossil fuel companies being stranded, and companies in the transport sector, utilities, etc., would also be affected. This risk would also be transmitted to banks as banks are linked to these companies through loans, equity or bond holdings. In one estimate, the equity exposure of JP Morgan Chase is more than USD 30 billion. In addition to the loss of assets directly exposed to climate risk shocks, indirect losses in banks’ equity will occur due to the devaluation of counterparties’ debt obligations on the interbank credit market. In the case of Deutsche Bank, if the fossil fuel and utilities sectors were subject to a 100% policy shock, the resulting direct losses would lead to a shrinkage of equity of more than 25% and indirect losses would lead to another 5% loss [13].

However, the banking sector has been relatively unresponsive in terms of climate disclosure and has not pursued actions to address climate change with sufficient vigor. According to the Carbon Disclosure Project (CDP), the mean value of carbon disclosure score of 117 companies worldwide in 2018 was 2.79 points out of 7, while the standard deviation is 2.59 [14]. Banks were only able to achieve a moderate level of compliance in terms of the completeness of their climate information and still have significant room for improvement [15]. This is particularly true in developing countries, where the lack of a local mandatory disclosure regime and a disclosure framework for the financial sector provide little incentive for banks to voluntarily disclose climate information, and the lack of professional carbon auditors makes it difficult to account for quantitative indicators. These factors contribute to the low quality of climate information disclosure in developing countries.

In summary, banks are an important link in the transition of industries to a low-carbon economy, but their climate information disclosure is relatively lagging behind and the quality of disclosure still needs to be improved. Banks in developing countries lack local policy guidance and support from professional accounting institutions, and the road to climate information disclosure is fraught with difficulties. Due to methodological reasons and the fact that the banking industry often generates less greenhouse gas than heavy polluters, such as electricity, steel and chemical companies, banks are usually excluded from the research sample on environmental and climate risk issues. There are fewer studies on the quality of banks’ climate information disclosure and its influencing factors [15].

Since the 19th National Congress of the Communist Party of China, the Chinese government has actively promoted climate-related policies, continued to improve its environmental information disclosure system and established a collaborative management model across multiple sectors, but it has yet to introduce a mandatory disclosure system. Shanghai and Shenzhen Stock Exchange have a system for environmental information disclosure by listed companies dating back to 2006. However, it was not until 2021 that listed companies were encouraged to disclose their measures and results of carbon emission reduction in the “Standards for the Contents and Formats of Information Disclosure by Companies Offering Securities to the Public No. 2—Contents and Formats of Annual Reports (2021 Revision)”. A specific climate information disclosure system has not yet been established and there is a lack of institutional regulation.

Against this background, this study aims to address the following questions:

(1) How is the quality of climate information disclosure of A-share-listed banks?

(2) What are the influencing factors of climate disclosure quality of listed banks in the context that China has not yet established a mandatory disclosure system?

Therefore, this paper selects the banking industry as the research object, constructs an evaluation system and assesses the quality of climate information disclosure of listed companies in the banking industry based on ESG reports, social responsibility reports and environmental information disclosure reports of listed companies. This paper is aimed to present the progress of climate information disclosure of domestic banking enterprises in recent years. It also provides perspectives on climate information disclosure in the banking sector under different regulatory environments for comparison with studies in other regions. It also adopts an empirical research approach to study the factors influencing the quality of corporate climate information disclosure.

While China has not introduced a mandatory climate disclosure policy, the guidelines that have been issued covering SOEs and green finance are coming into play. China, as a market with both an A-share and a Hong Kong stock exchange, and with different disclosure guidelines for the two exchanges, can provide unique data to support the legitimacy and proactive disclosure theories. The sample in this paper takes place before the enactment of a mandatory disclosure regime in mainland China and is an interesting one. By quantitatively evaluating the quality of climate information disclosure of A-share-listed banks, this paper will provide targeted guidance for banks to subsequently improve the completeness, consistency and comparability aspects of climate information disclosure. The findings of this paper will also be useful to inform policymakers and regulators in China.

2. Literature Review

2.1. Methods for Evaluating the Quality of Climate Disclosure

Overseas researchers tend to measure the quality of corporate climate information disclosure in the following ways: the first is to refer to the corporate climate information disclosure scores published by CDP or other organizations. For example, in a study of 2298 companies in the United States from 2005 to 2019, the scores evaluated by CDP based on questionnaires filled out by companies were directly used to measure the quality of climate disclosure [16]. However, most institutions’ scores tend to evaluate the perspective of completeness only while lacking consideration of the reliability of disclosure.

The second method is to construct a disclosure quality index, which first classifies the climate information disclosed by enterprises, scores and assigns values according to qualitative or quantitative criteria and adds up the scores according to weights to obtain the enterprise climate disclosure quality score. This method can reflect the disclosure quality more accurately and completely, but most of the studies did not differentiate the evaluation index according to the industry characteristics [17]. For example, Kılıç and Kuzey constructed an evaluation system with 28 indicators, including climate-related policy, green credit, corporate governance, energy saving and emission reduction technology, when evaluating the quality of climate information disclosure of Turkish banks [18]. Haque and Deegan selected board size, carbon emission, research and development, potential liability, reporting benchmark, carbon pricing and external affairs in their study of Australian companies [17]. Demaria and Rigot, referring to the TCFD disclosure framework in their evaluation, selected 38 indicators under the four dimensions of governance, strategy, risk management and metrics and targets to construct a climate disclosure quality evaluation system [19].

The third method is content analysis. By analyzing the texts of annual reports or environmental disclosure reports, the quality of companies’ climate information disclosure is evaluated based on the length of the disclosure, or the frequency of the occurrence of specific keywords. In their study, Cosma et al. refer to the Task Force on Climate-Related Financial Disclosures (TCFD) framework in order to determine where keywords appear in the report [15]. Kouloukoui et al. analyze keywords such as “climate change”, “greenhouse gases”, “carbon dioxide” and “transition risks” for coding and measurement [20].

2.2. Quality of Climate Information Disclosure

The quality of climate disclosure in current studies is based on the evaluation of independent reports issued by companies (e.g., sustainability reports, ESG reports, etc.), but the point of contention is whether the quality of climate information disclosure is equivalent to the completeness of disclosure. In their study of Turkish automotive companies, Hoştut and Hof found that the companies with the fewest disclosure pages followed the Global Reporting Initiative (GRI) 305 classification to the letter and had higher-quality disclosures than other companies with more pages [21]. Comyns’ study of multinational oil and gas companies also found that the quality of reporting was not better when more GHG emissions data were published [22]. Therefore, researchers are considering incorporating the principles of comparability and reliability from GRI or IFRS disclosure principles into the evaluation system for the quality of climate disclosure.

Regardless of the assessment methodology, the results show that the quality of corporate climate disclosure in these samples is unsatisfactory, in both developing and developed countries. This is supported by Kılıç and Kuzey’s study in Turkey [18], Kouloukoui et al.’s study in Brazil [20] and Giannarakis’ study in Europe [23]. In their study of Malaysian companies, Nik Ahmad and Hossain found that climate disclosures are rhetorical, attempting to show concern about climate change, and lacking in substance [24].

While there is not much evidence from the banking sector, the quality of climate information disclosure by banks in the only studies available is also poor. Kılıç and Kuzey found that, while climate-change-related disclosure by Turkish banks increased over the period 2010–2016, few banks had dedicated board committees to manage environmental or climate matters, and over 50% of banks did not include climate-change-related issues in their corporate policies or lending decisions [18]. Cosma et al. found that, after the release of the TCFD framework, climate disclosure by European banks remained immature, with 27%, 62%, 60% and 57% disclosure for governance, strategy, risk management and indicator targets, respectively, and many companies did not even make climate disclosures at all [15].

2.3. Factors Influencing Climate Information Disclosure

Existing studies have focused on the link between intrinsic company characteristics and external factors on climate information disclosure. Intrinsic factors include firm size, financial performance, industry classification and government ownership [25,26,27,28]. These factors of influence correspond to different theoretical foundations. As with environmental information disclosure, the main theoretical foundations of climate information disclosure include stakeholder theory, legitimacy theory, voluntary disclosure theory and agency theory.

Firm size has been repeatedly mentioned in previous studies, with researchers suggesting that larger firms tend to be involved with more stakeholders. Based on legitimacy theory, while large firms receive more attention from regulators and stakeholders, they tend to act legitimately. A study by Freedman and Jaggi using a sample of firms in heavily polluting industries after Kyoto Protocol [25], Prado-Lorenzo et al. on the top developed countries (US, Australia, Canada and EU) in terms of market capitalization of 101 firms in the top 500 of developed countries (US, Australia, Canada and EU) as a sample [26], a study by Stanny and Ely on constituents of the S&P 500 [29] and a study by Hackston and Milne on New Zealand all confirm this [30]. Some of the studies did not find an association between firm size and disclosure, possibly because the majority of the study sample was categorized as larger firms [31].

In terms of financial performance, researchers have focused on firm profitability, but empirical studies have not yet yielded clear and conclusive results. There are two different views based on agency theory. One view is that management, as agents, is more concerned with the financial performance of the company in relation to its own interests. The other view is that agency costs can be reduced through climate information disclosure [23]. Giannarakis found a positive effect of profitability on the quality of climate disclosure at the 10% level in a study of 215 firms in Europe [23]. However, the results from another study of New Zealand firms that examined current and lagged profitability found no positive relation between profitability and climate disclosure quality [30].

In terms of corporate ownership, state-owned enterprises tend to receive more attention and are required to be socially and environmentally friendly, so state-owned holdings are associated with more voluntary disclosure. Investors’ demand for such disclosure tends to be higher in developed capital markets [32]. A study of oil and gas companies in Saudi Arabia showed that government ownership was positively associated with corporate social responsibility disclosure [27]. Studies of Malaysian listed companies found that government ownership was significantly associated with corporate efforts in climate mitigation and tended to disclose more climate information [33,34], and a study in India confirmed this view [35].

Corporate governance has also received a great deal of attention [36,37,38]. Agency theory suggests that the board acts as a representative of shareholders and acts as a watchdog over management [28,39]. Fama and Jensen argue that the board’s effectiveness in monitoring depends on the extent to which it is independent of management [28]. Motivated by a desire to maintain their reputation, an independent board of directors will effectively monitor management and keep an eye on the interests of shareholders. However, the results of empirical studies have not yet led to consistent conclusions: a meta-analysis showed that the positive association between board independence and voluntary disclosure only occurred in countries with high investor protection rights [40]. Forker, Michelon et al. showed that the proportion of independent directors was not associated with an active sustainability disclosure strategy [38,41]. Leung, Amran verified the positive correlation between the proportion of independent executives and climate disclosure [31,42]. However, some studies even showed a negative correlation between the proportion of independent directors and disclosure [43]. A domestic study showed that improving the independent director system in the context of multiple agency relationships could help enhance the risk control of bank executives [44]. In addition to the proportion of independent directors, board size and the establishment of an environmental protection department have also been used in studies to characterize this factor of corporate governance, with Yang Tui and Li Yu Xiaolu’s study of 502 Chinese heavily polluting industries finding that the establishment of an environmental protection department had a more significant positive effect on environmental information disclosure compared to board size [45].

External pressures include those from (institutional) investors [46,47], public opinion [48] and regulation [49]. For institutional investors, shareholder activism suggests that this is largely due to the fact that investors are becoming increasingly concerned about the climate risk exposure of their underlying investments and that investors can pressure management to disclose climate information by putting pressure on them. A study of 265 US-listed companies suggests that institutional investors, particularly long-term institutional investors, increase voluntary disclosure of climate change risks [50]. In a resource constraint perspective, Luo et al. find that carbon disclosure is positively related to resource availability and that this relationship is more pronounced in developing countries [51].

However, there is no consensus in the academic community regarding the impact of some of the influencing factors, such as profitability and industry attributes [52], which may be due to the differences in the definition of carbon disclosure quality used in different studies.

3. Hypotheses

In the context of a mandatory disclosure regime not yet in place in China, we primarily want to test the impact of listed banks’ own characteristics on the quality of climate information disclosure. Based on stakeholder theory, we, therefore, start with firm size and ownership. Based on voluntary disclosure theory, we selected the proportion of independent directors and the board size as proxies for the efficiency of the company’s board of directors. For the financial performance of the company, we selected the return on total assets (ROA) and revenue growth rate. We also selected the indicator of institutional investors’ shareholding ratio to test the performance of shareholder activism among A-share-listed banks. Finally, given that some banks are also listed on Hong Kong’s stock exchange, which has higher climate disclosure requirements, we also selected whether the companies are dual-listed to test the impact of exchange disclosure guidelines on the quality of climate disclosure.

3.1. Company Size

The size of a company refers to the volume of its operations. A company’s business activities often have a direct or indirect impact on climate change, and the company must arrange for additional operational activities to manage the climate impacts of the company’s operations. According to stakeholder theory, management needs to reconcile the interests of all parties. The larger a company is, the greater its environmental and climate responsibilities, and the more likely it is to be noticed by government regulators and the press, etc. The company will need to make more extensive climate information disclosures in order to meet the needs of its stakeholders and maintain its image and competitiveness in all areas. The following hypotheses are, therefore, proposed:

Hypothesis 1 (H1).

Firm size is positively related to corporate climate information disclosure.

Most previous studies have used total assets, sales and market capitalization or the logarithm of these indicators to measure firm size. Considering the high proportion of liabilities in the total assets of banks and the large fluctuations in market capitalization of A-share companies within an annual range, this paper selects the annual revenue of a company as an indicator of firm size to better characterize the size of the company within a year.

3.2. Ownership

By the end of May 2022, eight mainland Chinese banks had become TCFD-supporting institutions, including all six large state-owned commercial banks, making SOEs more advanced in taking the initiative to benchmark against international disclosure standards. Given that SOEs have always been the mainstay of the national economy under China’s economic system, and that SOEs also have a social mission and a huge responsibility in climate change and energy transition, the following hypothesis is proposed:

Hypothesis 2 (H2).

The quality of climate information disclosure by SOEs is significantly better than that of non-SOEs.

3.3. Dual Listing

In 2019, the SEHK released a new version of its Environmental, Social and Governance Reporting Guidelines, which, for the first time, includes the topic of “climate change”, aligning with the TCFD framework and making it mandatory for companies to disclose “mitigation measures to identify and address significant climate-related issues that have and may have an impact on the issuer”. The 2022 Review of Environmental, Social and Governance Disclosure Practices, published in 2022, also focused on climate disclosure, reporting that over 85% of issuers disclosed the new climate-related requirements. In contrast, no mandatory disclosure requirements have been announced for A-shares to date, and, in terms of disclosure requirements, the SEHK requirements are more stringent, so it is inferred that the quality of information disclosure for H-share-listed enterprises is higher than that for A-shares [53]. Therefore, the following hypothesis is proposed:

Hypothesis 3 (H3).

The quality of climate information disclosure is significantly better for companies listed on both stock exchanges than for companies with a single domestic listing.

3.4. Profitability

Profitability is one of the key financial indicators of a firm, reflecting its ability to earn a profit per unit of capital/equity. The more profitable a company is, the more willing its management will be in pursuing social performance beyond financial indicators. In addition, companies incur costs in making climate disclosures and addressing climate risks, which have an impact on their profitability. If a firm is more profitable, the impact of disclosure costs is smaller and firms will be more inclined to disclose climate information and take action in response. The following hypothesis is proposed:

Hypothesis 4 (H4).

The quality of climate information disclosure is better for more profitable firms than for less profitable firms.

Previous studies have often chosen return on total assets ROA or return on net assets ROE as a symbol of corporate profitability, but ROA tends to focus on corporate profitability efficiency, while ROE is more from the perspective of financial and shareholders’ equity, so total assets ROA is chosen in this paper.

3.5. Revenue Growth Rate

The growth rate of corporate revenue is an indicator of the growth capability in financial performance. Climate disclosure entails significant costs in the current period, particularly in relation to the accounting of carbon emissions, stress testing based on different climate scenarios, etc. However, the benefits, such as reducing financing costs, do not accrue in the current period. Companies in an expansionary phase, therefore, tend to invest more resources and energy in their own growth than in climate- or environment-related investments, and the quality of climate information disclosure tends to be poor. The following hypothesis is proposed:

Hypothesis 5 (H5).

The quality of climate information disclosure is lower for companies with faster revenue growth than for companies with lower revenue growth.

3.6. Proportion of Independent Directors

In principal–agent theory, the duty of independent directors is to supervise the conduct of management team to disclose more comprehensive climate information. In the face of an increasingly complex and volatile financial environment, it is necessary for commercial banks to enhance the professionalism and diversity of independent directors and strengthen the ability of independent directors to perform their duties. Independent directors should shoulder the responsibility of identifying risks and proposing countermeasures and supervising the board of directors and management for climate information disclosure. The following hypothesis is proposed:

Hypothesis 6 (H6).

The proportion of independent directors is positively related to corporate climate information disclosure.

3.7. Board Size

The larger the board of directors, the more likely it is that a dedicated person or a special committee will be assigned to be responsible for operations related to environmental or climate information disclosure, which will help the enterprise to carry out climate risk response work. The following hypothesis is, therefore, proposed:

Hypothesis 7 (H7).

Board size is positively related to corporate climate information disclosure.

3.8. Shareholding of Institutional Investors

Institutional investors who hold a large number of shares in a company can even mobilize other shareholders to make requests to management to disclose climate information. In terms of expertise, institutional investors tend to have a deeper understanding of climate investment. In terms of tolerance for climate risks in their portfolios, institutional investors are much less tolerant than ordinary investors. In terms of motivation, institutional shareholders are generally significantly higher than ordinary investors, who are more interested in climate risk disclosure and subsequent risk response by companies. The following hypothesis is, therefore, proposed:

Hypothesis 8 (H8).

The quality of climate information disclosure of companies with a higher proportion of institutional shareholding is higher than that of companies with a lower proportion of institutional shareholding.

4. Methods

4.1. Sample

The evaluation object of this paper is A-share-listed banks. In terms of time frame, as the climate disclosure of most A-share-listed banks before 2018 is almost zero and has no evaluation significance, the sample period is chosen from 2018 to 2021 to ensure both the variability among companies and the validity of the disclosure. To ensure the variability of the sample, 6 companies listed after 1 January 2020 were further excluded, leaving a total of 36 A-share-listed companies for evaluation. The content of the evaluation is the published social responsibility reports (CSR), environmental, social and governance (ESG) reports, green finance topic (TCFD) reports, sustainability reports and environmental information disclosure reports downloaded from the investor relations/sustainability section of each company’s website.

4.2. Measurement of Variables

4.2.1. Dependent Variables—Climate Disclosure Quality

Common methods of evaluating the quality of climate information disclosure include index methods, content analysis and direct reference to third-party evaluation data [3]. In this paper, we combine the content analysis method and the index method to construct an evaluation system for climate information disclosure quality, select indicators to construct the evaluation system through the index method and assign values to the evaluation indicators through the content analysis method.

According to the TCFD principles for effective disclosure (relevant, specific and complete, clear and understandable, consistent, comparable, reliable, verifiable and objective, timely), the GRI Principles (accuracy, balance, clarity, comparability, completeness, sustainability context, timeliness, verifiability) [54], the Guidelines for Financial Institutions Environmental Information Disclosure (accuracy timely, consistent and coherent) and the Guidelines on Green Finance for Banking and Insurance Institutions, an evaluation system containing six dimensions of completeness, consistency, comparability, reliability, timeliness and green finance, were eventually constructed. Based on Demaria and Rigot’s [35] climate information disclosure quality evaluation system, combined with some indicators of Haque [17], Moroney [55], Park [56], Kılıç [18] and Furrer [57], we select 9 primary indicators and 40 secondary indicators, as shown in

Table 1. Climate disclosure evaluation system.

Dimension	Primary Indicators	Secondary Indicators
Completeness	Governance (C1)	Oversight Responsibilities of the Board (C1,1)
		Specialized Positions of the board (C1,2)
		Periodic review of the board (C1,3)
		Management Views on Climate Issues (C1,4)
		Setting up a climate risk management unit (C1,5)
		Set up a performance evaluation system for Greenhouse gas emissions (C1,6)
		Pay tied to climate goals (C1,7)
		Participate in climate-related activities(C1,8)
	Strategy (C2)	Risk description (C2,1)
		Opportunity Description(C2,2)
		The impact of risk (C2,3)
		The impact of opportunities (C2,4)
		Responses to Climate Risks (C2,5)
		Responses to Climate Opportunities (C2,6)
		Stress tests based on different climate scenarios (C2,7)
	Risk Management (C3)	Process for identifying and assessing climate risks (C3,1)
		Processes for managing climate risk (C3,2)
		Integrating climate risk into overall risk management (C3,3)
	Metrics and Targets (C4)	The amount of greenhouse gas emission reduction (C4,1)
		Buy and/or develop renewable energy (C4,2)
		Require suppliers to reduce greenhouse gas emissions (C4,3)
		Greenhouse gas emission per person/revenue (C4,4)
		Neutral carbon dioxide emission target (C4,5)
		Peak carbon dioxide emission target (C4,6)
		Emission reduction target (C4,7)
		Scope 1, 2 emission of greenhouse gases (C4,8)
		Scope 3 emission of greenhouse gases (C4,9)
		Appraisal and rewards for climate goals (C4,10)
Consistency	Consistency (C5)	Consistent with previous disclosures (C5,1)
		The same criteria (C5,2)
Comparability	Comparability (C6)	Disclosure of multi-year data (C6,1)
		Comply with the international disclosure framework (C6,2)
Reliability	Third-Party Certification (C7)	Third-party certification (C7,1)
Timeliness	Release Time (C8)	Release in 4 months after the end of the year (C8,1)
Green finance	Green Banking (C9)	Green credit scale (C9,1)
		Green credit category (C9,2)
		Green bond issuance scale (C9,3)
		Green wealth management issuance scale (C9,4)
		Green finance projects convert carbon dioxide emission reductions (C9,5)
		Credit balance of two high and one surplus industries (C9,6)

. Specifically, the primary indicators under the integrity dimension follow the TCFD frame. Based on the indicators in the above research and guidelines of the China government, indicators for stress test, neutral carbon dioxide emission target, peak carbon dioxide emission target, green bonds, green wealth management and loans to the polluting industries are added.

We have assigned to each indicator a value according to the following criteria: assigned the value of zero if the analyzed company does not disclose related information, assigned the value of half if the analyzed company discloses related information partly or in cases and assigned the value of one if the analyzed company discloses quantitatively or in detail. Thus, the CDI was measured as follows:

$$CDI={\sum }_i{\sum }_j{C}_{i,j}/j$$

(1)

i: the number of primary indicators

j: the number of secondary indicators of each primary indicator

4.2.2. Independent Variable

All the companies in this study are from a single industry sector. Size, ownership, dual-listed, ROA, revenue growth, independent director ratio, board size, investor shareholder ratio are used as a control variable. Definition of each variable is shown in

Table 2. Independent variables.

Variable	Definition
Size	Logarithm of annual revenue
SOE	1 for SOEs, 0 for non-SOEs
Abr	1 for both China Mainland and Hong Kong listings, 0 for China Mainland only
ROA	(Total profit/Average total assets) × 100%
Growth	Current year revenue/Prior year revenue × 100% − 1
IDR	Number of independent directors/number of board members × 100%
BS	Number of board members
Inv	Number of institutional holdings/number of shares outstanding × 100%

:

4.3. Model

To identify which factors have a significant influence on the A-listed banks’ climate information disclosure quality, we construct a multiple linear regression model:

$$CDI={{\beta }_0+\beta }_{1}Size+{\beta }_{2}SOE+{\beta }_{3}Abr+{\beta }_{4}ROA+{\beta }_{5}Growth+{\beta }_{6}ID+{\beta }_{7}Inv+\epsilon$$

(2)

5. Results

5.1. Descriptive Statistics

Based on Equation (1) and ESG reports, social responsibility reports and environmental information disclosure reports of the sample, we calculated the quality of climate disclosure scores for A-share-listed banks in 2018–2021, and the detailed data are shown in Appendix A. The mean of the climate disclosure quality score for A-share-listed banks was 2.818 points (out of a possible maximum 9 points), with a standard deviation of 1.613. The minimum score was 1 point, and a maximum value of 7.850 points. The results indicate that the quality of climate disclosure of A-listed banks is generally low and varies significantly among enterprises.

Descriptive statistics for the independent continuous variables are shown in

Table 3. Descriptive statistics for continuous independent variables.

Variables	N	Mean	Std. Deviation	Min	Max
CDI	144	2.818	1.613	1	7.850
Size	144	6.037	1.698	3.305	9.137
ROA	144	0.847	0.163	0.501	1.372
Growth	144	0.106	0.091	−0.096	0.578
IDR	144	0.384	0.043	0.308	0.545
BS	144	13.65	2.016	9	19
Inv	136	50.84	28.16	0.098	118.2

:

The number of dual-listed companies in the sample was 52, accounting for 36.11%. The number of SOEs in the sample was 24, accounting for 16.67%, which was significantly less than that of private enterprises.

Descriptive statistics for the binary independent variables are shown in

Table 4. Descriptive statistics for binary independent variables.

Variables	Value	Frequency	Percent
Dual-listed	0	92	63.89%
	1	52	36.11%
SOE	0	120	83.33%
	1	24	16.67%

:

Table 5. Correlation matrix of all variables.

	CDI	Size	SOE	Abr	ROA	Growth	IDR	BS	Inv
CDI	1
Size	0.690 ***	1
SOE	0.505 ***	0.651 ***	1
Abr	0.584 ***	0.564 ***	0.595 ***	1
ROA	0.095	0.184 **	0.104	−0.043	1
Growth	−0.159 *	−0.166 **	−0.180 **	−0.185 **	0.048	1
IDR	0.188 **	0.244 ***	0.280 ***	0.193 **	0.031	−0.099	1
BS	0.288 ***	0.250 ***	0.042	0.075	0.125	0.034	−0.296 ***	1
Inv	0.385 ***	0.771 ***	0.456 ***	0.301 ***	0.153 *	−0.218 **	0.171 **	0.099	1

* p < 0.1, ** p < 0.05, *** p < 0.01.

shows the correlation matrix for the variables in our analysis. It is evident that CDI is significant and positively related to size, SOE, Abr, IDR, BS and Inv. Table 5 also shows growth with a negative correlation between CDI. Table 5 further shows low correlations between the independent variables, implying that multicollinearity is unlikely to be a concern for this dataset.

5.2. Regression Results

Based on Equation (2), we use STATA 17.0 for multiple linear regression. The regression results are shown in

Table 6. Regression Results.

	Coef.	St. Err.	t-Value	p-Value	[95% Conf]	[Interval]	Sig.
Size	0.673	0.124	5.43	0	0.428	0.918	***
SOE	0.054	0.367	0.15	0.883	−0.672	0.781
Abr	0.711	0.281	2.53	0.012	0.156	1.266	**
ROA	−0.032	0.612	−0.05	0.958	−1.243	1.178
Growth	−0.77	1.071	−0.72	0.473	−2.889	1.349
IDR	2.108	2.47	0.85	0.395	−2.78	6.995
BS	0.105	0.055	1.91	0.059	−0.004	0.214	*
Inv	−0.015	0.006	−2.56	0.012	−0.026	−0.003	**
Constant	−2.909	1.364	−2.13	0.035	−5.608	−0.211	**

* p < 0.1, ** p < 0.05, *** p < 0.01.

.

The F-test value was 20.682 and Prob = 0.0000 for the dependent variable, the model was significant at the 1% level and the fit was good and statistically significant. The R2 was 0.566, which had good explanatory power, and the above variables had an impact on the CDI for the explanatory variable climate disclosure quality. The variance inflation factor VIF was calculated using STATA at 2.08, which is close to 1 and less than 10, and it can be assumed that there is no multicollinearity problem in the multiple regression model.

6. Discussion

6.1. The Quality of Carbon Information Disclosure by A-Listed Banks

According to our evaluation system, the overall climate information disclosure quality of A-share-listed banks is uneven and low, which is consistent with previous research [15,18]. The level of disclosure on governance, strategy and risk management lags behind the international average by a large margin [58], reflecting the fact that A-share-listed banks lack corresponding management mechanisms. Most listed banks have not yet established a dedicated climate management committee or a corresponding risk management department, and most companies’ climate business is still under the responsibility of the original environmental protection department, lacking special training on climate information disclosure. Company boards and management have inadequate knowledge of transition and physical risks under climate change and lack a top-down risk management process. There is an urgent need to prioritize raising the risk awareness of listed banks in order to effectively carry out the subsequent challenging responses, such as carbon asset statistics and stress testing. This is consistent with Buranatrakul and Swierczek’s finding that climate change strategic actions in Asia are not satisfying [59].

In our evaluation system, quantitative disclosure includes stress tests, CO₂ emissions of Scope 1, 2, 3, green credit, green bonds, green finance, credit in polluting industries and green finance carbon efficiency. With the exception of green credit, the disclosure rate of all other indicators is low. Unlike previous environmental information disclosure, climate information disclosure is more challenging and involves more complex calculation of quantitative indicators. This requires the government to introduce unified accounting standards and enterprises to set up relevant departments to take the initiative to dovetail and undertake high-quality data accounting and disclosure. Since 2022, “China Green Bond Principles” and “Accelerating the Establishment of a Unified and Standardized Statistical and Accounting System for Carbon Emissions” have been issued. We are likely to observe vigorous promotion of the establishment of relevant standards in the climate sector in 2022.

In terms of the caliber of carbon dioxide emissions, most commercial banks only choose the headquarters or the main business branches for disclosure. The carbon emissions per person or the carbon intensity per unit of revenue are calculated by dividing the carbon emissions of incomplete caliber by the whole number of employees or the amount of revenue, thus leading to the distortion of such data. Moreover, the caliber chosen for climate information disclosure is not consistent from year to year. With the proficient application of accounting standards and methods, the caliber of carbon emissions is often expanding. Therefore, the inconsistent caliber will lead to a lack of comparability of carbon dioxide emissions data between the preceding and following years. This is also true for the carbon emission reduction benefits generated by green finance. Some banks only choose to account for the carbon emission reduction benefits generated by green credits, while others account for green finance as a whole, including green credits, bonds and various investment and financing projects.

According to this paper, only 46.53% of the data disclosed by 36 listed banking companies during 2018–2021 have been verified by auditors, generally scoring moderate in the evaluation indicator of reliability. It should be noted, however, that some of these listed banks do not disclose CO₂ emissions data, which is not included in the third-party audit certification. If both carbon emissions data are disclosed and third-party audit certificates are provided, only 55 observations out of 144 qualify, which means the scoring rate of valid carbon emissions reliability is only 38.19%. Moreover, listed banks rarely disclose their accounting methods for the carbon benefits generated by green financial products, and there is no way to verify the correctness of their accounting methods and the authenticity of their data.

6.2. The Determinants of Carbon Information Disclosure by A-Listed Banks

This paper shows that corporation size is highly positively correlated with climate disclosure. This finding is consistent with most previous research [23,26,37,51] and proves stakeholder theory, with larger companies tending to do better on disclosure due to the large number of stakeholders and the greater public pressure they face. Additionally, unlike previous studies that tend to select the top-ranked companies in terms of size as the subject of study [31], the sample chosen for this paper is listed companies in the banking sector, covering both state-owned banks and city merchant banks, with large differences in size to effectively prove the point.

Firms listed both locally and abroad perform better on the quality of climate information disclosure than those listed only in mainland China, consistent with studies by Hackston and Milne, Cui et al. [30,60]. The offshore issues of banks in the sample all occurred in Hong Kong shares, and the earlier origin of the climate information disclosure regime on the SEHK, with better and more stringent disclosure information content, made the disclosure experience of companies listed in both places richer. Moreover, the banks listed in Hong Kong need to consider global investors’ requirements. Overseas investors are more active in terms of shareholder activism than domestic ones, which also contributes to the improvement in the quality of climate disclosure of Hong-Kong-listed companies.

The quality of climate information disclosure of state-owned enterprises is not significantly better than that of private enterprises, which is inconsistent with the hypothesis and a former article [23]. Agency theory suggests that ownership is a significant factor in climate change disclosure, but, in our research, it is not compared to factors such as size and listed abroad. SOEs are not set up only for economic outcomes but also to achieve some of the political and social issues. Disclosure of a company’s social responsibility actions can demonstrate that the company is operating in accordance with the expectations of the state, ultimately legitimizing its existence. However, there is no significant difference between the two in terms of the findings of this paper’s research study. All six state-owned banks have now joined the TCFD, but it will take time to improve the completeness of disclosure in line with the TCFD framework.

This paper shows that ROA and the growth rate of corporate revenue, as proxies for financial performance, have a significant impact on corporate climate information disclosure. This is consistent with Hackston and Milne [30] but inconsistent with Giannarakis [23]. One possible reason is that companies do not disclose climate information based on cost and financial performance considerations, and that companies with better financial performance do not have a stronger incentive to disclose climate information. The results of this paper even suggest that financial performance has a negative impact on the quality of climate information disclosure, indicating that companies with limited management energy are unable to address climate risks while achieving financial performance and perform poorly in terms of information disclosure.

In terms of corporate governance, this paper shows that the proportion of independent directors does not have a positive impact on the quality of climate information disclosure, but board size has a positive impact on climate information disclosure at the 5% level of significance. The former conclusion is consistent with Forker [41] and Michelon [38]. The latter conclusion is consistent with Giannarakis [23] and Yang [45]. Currently, a relatively high proportion of independent directors of banking companies have academic backgrounds, but a relatively low proportion have risk management experience and therefore do not have sound climate risk management capabilities in terms of professional competence, leading to the reason why the proportion of independent directors has no significant positive effect on climate information disclosure. Further, perhaps this is a possible reason that agency theory fails to explain this sample. The larger the board of directors, the more comprehensive the examination of risks in all aspects of company development, and, the more attention paid to the legitimacy of the company’s operations and the protection of stakeholders’ interests, the more active the action on climate information disclosure and the higher the quality of disclosure.

This study shows that the shareholding ratio of institutional investors has a significant negative impact on climate information disclosure, indicating that investment institutions are not fulfilling their obligations as professional investors to monitor the climate information disclosure of investee companies. This conclusion leads to an opposite view relative to Cotter and Najah [47] and Reid and Toffel [49]. This shows that the concept of socially responsible investment is not well recognized among local investment institutions in China, that investment institutions have not established a sound concept of socially responsible investment and that the heads of investment institutions do not have strong awareness of climate risk. As a result, investors do not take an active role in the management of companies in which they invest. Combined with the previous data, which show that the shareholding ratio of institutional investors is positively correlated with financial performance, such as total gearing ratio ROA and corporate revenue growth rate, institutional investors attach much more importance to financial performance than to the fulfillment of corporate social responsibility.

7. Conclusions

The climate information disclosure of A-share-listed banks started late and the overall level is low. All four primary indicators under the integrity dimension have scores below the international average. However, it shows a rising trend year by year, and the quality of disclosure improved significantly in 2021. Climate information disclosure performs poorly in completeness and consistency, moderately well in the three dimensions of comparability, reliability and green finance and excels in timeliness. Among the evaluation indicators, the disclosure rate of quantitative indicators significantly lags behind that of qualitative indicators. The disclosure caliber and selection of indicators are not uniform, so the reliability still needs to be improved. The main reason behind the result is that China does not yet have a mandatory climate information disclosure policy.

The empirical results on the factors influencing the quality of climate information disclosure found that enterprise size, dual listing and board size had a significant positive contribution to the quality of climate information disclosure of A-share-listed banks. The shareholding ratio of institutional investors was negatively related to the quality of climate information disclosure. ROA, annual revenue growth rate and proportion of independent directors had no significant effect on climate information disclosure.

This suggests that the size of the firm, the disclosure policy and corporate governance all contribute to the quality of the firm’s climate information disclosure to varying degrees. Institutional investors, however, are not motivated to participate in climate information disclosure and do not sufficiently consider climate information disclosure in their investment decisions.

Limited by the late start in climate information disclosure in China, fewer companies conduct climate information disclosure, not to mention continuous disclosure for years. In this paper, we have selected listed banks with good quality of sample data but limited in terms of time series length. This paper’s climate information disclosure quality evaluation system only targets A-share-listed banks. Although the internal validity of the multiple regression model is ensured, the sample diversity is limited and can be subsequently extended to other industries.

In terms of the selection of determining factors, we mainly consider important firm characteristics, covering financial attributes, ownership, governance structure, dual listing and institutional investors. We do not consider external variables, such as policy and public monitoring. We do not test carbon performance as an influencing factor because data on CO₂ emissions in Scope 1 and 2 are insufficient and the caliber of disclosure is inconsistent.

8. Suggestions

8.1. For Policymakers and Regulators

Benchmark with international mainstream climate disclosure standards and improve disclosure completeness as soon as possible. Considering that there is currently no localized disclosure framework for climate information disclosure, it is recommended that climate authorities, such as the Ministry of Ecology and Environment, in conjunction with the Securities and Futures Commission, should learn from mainstream information disclosure frameworks, such as TCFD and CDP. To unify statistical caliber, accounting standards and disclosure frameworks as far as possible is the first priority so that the comparability and consistency of climate information can be improved. Further, supplement the existing climate information disclosure framework and achieve compliance with industry characteristics. Promote pilot climate information disclosure by industries and enterprises, and explore local practices of the climate information disclosure framework in practice.

Improve climate information disclosure requirements. Considering the banking industry’s dual role in climate information disclosure, indicators such as carbon emissions, emission reduction targets and carbon performance of green finance should gradually transition from encouraging disclosure to mandating disclosure. As China’s industry is currently in the process of transitioning to a low-carbon economy and the share of clean technology revenue has not yet taken off, the government should encourage banks to make disclosures on the share of fossil fuel revenue and the share of clean energy revenue. Encourage companies to make Scope 3 carbon disclosure and accounting for the full life cycle carbon footprint of their products, on top of Scope 1 and 2 carbon emissions.

Establish uniform accounting standards. Considering that the reliability of disclosure and the disclosure rate of quantitative indicators by listed banks in China still need to be improved, it is recommended that the construction of a greenhouse gas statistics and accounting system be accelerated. Unifying and standardizing carbon accounting methods and providing technical support and relevant training for quantitative disclosure are also urgent.

8.2. For Banks

Firstly, bank boards and management should be fully aware of the significant impact and opportunities of physical and transformation risks under climate change on the company’s credit business and develop strategies accordingly, integrating consideration of climate factors into daily business processes. In response to the needs of relevant stakeholders, the board of directors should establish a green finance committee and a climate risk committee to regularly review corporate climate information disclosure and efforts related to risk and opportunity. What is more, banks should proactively raise awareness of social responsibility, enhance the board’s climate governance capabilities, actively participate in climate-governance-related activities and training and make a statement in the annual report or sustainability report. Bank boards should select and recruit independent directors with climate risk management capabilities.

Secondly, the bank board must take the initiative to benchmark the TCFD disclosure framework and further improve the quality of climate information disclosure. Banks should actively examine climate risk exposures and loans to the polluting industries and conduct stress tests under different climate scenarios to ensure that the remaining liquidity can cope with losses and defaults arising from climate risks. Efforts are needed to promote the development of green credit and green bonds for climate-friendly business and ensure that financial instruments for carbon emissions reduction are put into practice. The management should clarify and unify the statistical caliber of climate information disclosure and promote the statistical accounting of Scope 1/2’s CO₂ emissions, clean energy share and carbon emissions intensity, as well as the accounting of carbon performance generated by green finance. On top of this, undertake efforts to learn how to account Scope 3’s CO₂ emissions.

Before banks are ready to set carbon targets, their own credit structure and climate risk management capabilities should be well evaluated. National and local policies should be considered as it pertains to a scientific and reasonable target. After this, carry out target decomposition and set phased assessment. Banks need to engage a professional carbon accounting agency to carry out verification of disclosure.