Does the Classified Reform of Chinese State-Owned Enterprises Alleviate Environmental, Social and Governance Decoupling?

Abstract

Accurate disclosure and proactive engagement in ESG practices are essential for achieving high-quality economic development, particularly as China addresses significant challenges during its reform journey. The Classified Reform of State-Owned Enterprises (CRSOE) is a strategic initiative by the Chinese government aimed at fostering this development. Our study leverages the implementation of the CRSOE as an exogenous shock, employing the difference-in-differences approach to assess the policy’s governance impact on ESG decoupling from the perspective of ownership heterogeneity. The policy was found to alleviate ESG decoupling, particularly pronounced among SOEs with special functions. The governance effect is achieved by reducing the aspiration–performance gap. Specifically, the policy effectively narrows the disparity between a company’s actual performance and the expected performance based on the industry average, thereby mitigating ESG decoupling. However, the policy’s impact can be weakened by factors such as political connections among executives and media attention. Furthermore, the CRSOE effectively addresses greenwashing practices within ESG decoupling, with a particularly strong effect on SOEs that fail to disclose ESG information in alignment with Global Reporting Initiative (GRI) standards. These findings highlight the importance of understanding the broader implications and underlying mechanisms of the policy. Therefore, building on the assessment of how the CRSOE policy impacts ESG decoupling, we also examine the mechanisms through which this policy operates and how its effectiveness varies under different conditions of heterogeneity. By extending the application of principal-agent theory and performance feedback theory, our research suggests that policymakers should prioritize market-driven reforms for fully competitive SOEs and promote a stronger emphasis on non-financial goals. Additionally, it is essential to mitigate the undue influence of political promotions on the management of all SOEs.

1. Introduction

Since the introduction of the ESG (Environmental, Social and Governance) concept, there has been a growing focus on sustainable development among global enterprises, with ESG research flourishing [1,2,3]. Specifically, ESG decoupling denotes the discrepancy between a company’s ESG disclosures and its actual performance, encompassing both the exaggeration termed as ESG greenwashing and the minimization known as ESG brownwashing [4]. Although the literature specifically addressing ESG decoupling is limited, a review reveals broad consensus among scholars. The deceptive nature of ESG decoupling poses significant risks to stakeholders’ rights and interests, including those of the broader public. Ultimately, its presence undermines efforts to achieve sustainable development [5,6,7]. In practice, these scandals emerge sporadically, exemplified by the April 2022 revelation that Mitsubishi Electric had engaged in four decades of transformer test falsification, leading to the revocation of its international certifications. Current research on ESG decoupling primarily addresses its influencing factors. However, only one study by Di and Li examines its consequences, indicating that this behavior impairs labor investment efficiency [8]. Most studies concentrate on micro-level corporate characteristics’ impact on ESG decoupling [4,6]. Only Aboud, Saleh and Eliwa [7] adopted a macro perspective, investigating the effects of EU Directive 2014/95’s implementation on ESG decoupling. However, there is a gap in exploring these factors from the policy perspectives of other countries or regions.

In China, since the Third Plenary Session of the Communist Party of China (CPC’s) 11th Central Committee, reforms have continuously advanced, leading to a rapid increase in the scale and a significant improvement in the operational efficiency of State-Owned Enterprises (SOEs) [9], and focused on exploring paths to high-quality economic development [10]. Greenness and openness, aligning with the concept of ESG, constitute important principles for high-quality economic development [10]. As China’s SOE reform advances into a new phase characterized by targeted reforms [11], the implementation of the CRSOE policy represents a renewed effort by China to foster high-quality economic development. Prior to the implementation of the CRSOE policy, SOEs navigated the dual yet often conflicting objectives of reaping economic benefits and fulfilling social responsibilities. The conflicts between these objectives resulted in notable confusion and inefficiency in decision-making. In order to resolve this issue, the CRSOE policy distinctly categorizes the functional positions of different SOEs into commercial and public sectors. Within the commercial sector, SOEs are further classified into fully competitive and special function entities, with distinct assessment criteria established for each category. Can this reform truly enable Chinese SOEs to refine their ESG strategies and contribute towards achieving the United Nations Sustainable Development Goals, as well as China’s vision for green, open and high-quality development?

Our study investigates the CRSOE as a potential determinant of ESG decoupling from a Chinese policy viewpoint. ESG has evolved into a crucial component of corporate strategy [12]. The concept of ESG strategy, including the disclosure and implementation of ESG information, plays a vital role in business management and is a key driver of ESG decoupling [6,13]. CRSOE has defined distinct functional positions for SOEs based on their classification, serving as a foundation for strategic planning and decision-making in business management [14], and is bound to influence corporate strategies and decisions. Eight years into the CRSOE, we examine whether it has addressed the principal-agent problem in SOEs and its impact on ESG decoupling governance.

The CRSOE, aimed specifically at SOEs, does not apply to non-SOEs. To carry out this examination, we began by considering ownership heterogeneity, distinguishing between SOEs and non-SOEs, and applied a difference-in-differences model. In this study, SOEs implementing the policy were designated as the experimental group, while non-state-owned enterprises served as the control group. We primarily examined the impact of the CRSOE on ESG decoupling from both overall and functional classification perspectives.

Our findings indicate that, following the CRSOE policy’s clarification of SOEs’ functional roles and primary objectives across industries, there was a significant reduction in overall ESG decoupling behavior among SOEs. Notably, SOEs in industries with specialized functional roles demonstrated a significant peer effect under a clearly unified goal orientation, aligning their business performance with the industry median-based performance expectations. For these SOEs, the CRSOE helped curb ESG decoupling behavior stemming from organizational inertia or risk-taking tendencies by reducing the aspiration–performance gap. However, the governance effect is diminished by senior management’s political ties and media attention. Additionally, the governance effect varies, primarily targeting greenwashing over brownwashing, and focusing on SOEs that have not disclosed ESG information according to Global Reporting Initiative (GRI) standards, as opposed to those that have.

Our study investigates the governance effects of the CRSOE on ESG decoupling and makes significant contributions in three dimensions: context, subjects and mechanisms. Firstly, from a research context perspective, studies on the influencing factors of ESG decoupling remain sparse. Particularly, research exploring governance effects from formal institutions, like policies, is rare [4,6,7,8]. Existing research primarily addresses mixed ownership reforms [15,16], while guiding policies like CRSOE have been largely overlooked [17]. Our study examines the CRSOE’s impact on ESG decoupling, thus addressing this research gap.

Secondly, concerning research subjects, ESG decoupling, derived from social responsibility decoupling, remains underexplored [4,6]. ESG decoupling, a deceptive strategic disclosure of non-financial information, exacerbates stakeholder information asymmetry and highlights the principal-agency problem between management and investors. Investigating CRSOE’s governance effects can help determine its effectiveness in reducing information asymmetry, providing a solution to this principal-agency dilemma.

Thirdly, from a mechanism perspective, we analyzed various CRSOE governance pathways on ESG decoupling, including aspiration–performance gaps, executive political connections and media attention, with emphasis on mechanism applicability across different SOE categories. This comprehensive analysis allows our research to offer policymakers sustainable development recommendations tailored to the realities of various enterprise types.

The remainder of this study is structured as follows. Section 2 delves into the background and key aspects of the CRSOE, integrating existing studies and theories on CRSOE and ESG decoupling to formulate a set of hypotheses. Section 3 outlines the sample selection methodology, data sources and the analytical model employed, alongside a brief descriptive statistical analysis. Section 4 presents the baseline regression results, provides an analysis and conducts a series of robustness checks. Section 5, offering additional insights, examines the mechanisms through which the CRSOE influences ESG decoupling and includes two heterogeneity analyses. The concluding section, Section 6, summarizes this study’s key findings, discusses theoretical and managerial implications, addresses this study’s limitations and outlines directions for future research.

2. Policy and Academic Backgrounds with Hypotheses

2.1. CRSOE

In the process of building socialism with Chinese characteristics, SOEs play an indispensable political and economic role. In the 2024 Fortune Global 500 list, China features 133 listed companies, 91 of which are SOEs, accounting for 68.4%. SOEs serve as an important instrument for the government to participate in the market economy, maintain social stability and control key factors [9]. Unlike private enterprises, SOEs pursue both economic and non-economic objectives, bearing governmental policy burdens [18,19]. The “dual system” functional positioning of Chinese SOEs has long reflected these burdens, leading to complex principal-agent issues [20]. This mixed functional positioning allows managers to act selectively in their own interests, compromising operational efficiency [21,22]. Between 1978 and 2013, SOE reform underwent three phases: decentralization and profit-sharing, institutional innovation and state asset development [17]. These reforms have enabled SOEs to gradually adapt to a market-oriented business environment, establish a modern enterprise system and enhance their capacity for asset value preservation and enhancement. However, due to the dual functional position of SOEs, the associated principal-agent problem remains unsolved [9,23]. The government, as the controlling shareholder, may require SOEs to forgo profits in favor of social responsibility goals [24], leading to agency conflicts with minority shareholders. This situation may also result in insufficient incentives for SOE management [25], fostering agency conflicts between management and controlling shareholders.

In November 2013, the Third Plenary Session of the CPC’s 18th Central Committee initially proposed “to accurately define the functions of various SOEs”. By August 2015, the strategic direction and guidelines for the CRSOE were first clarified by the Central Committee of the CPC and the State Council of China. The State-owned Assets Supervision and Administration Commission of the State Council and other ministries officially issued the Guiding Opinions on the Functional Definition and Classification of State-Owned Enterprises (Guiding Opinions) and implemented the CRSOE in December 2015, thus avoiding mission conflicts and contradictions in operational logic [17], marking the entry of SOEs reform into a new phase of classified reform [11]. This policy has since been continuously enriched and perfected. In August 2016, an implementation plan aimed at influencing SOE managers’ behavior with differentiated assessments was jointly issued by the SOASAC of the State Council and the Ministry of Finance, aligning with the functional positioning and development needs [14]. In June 2020, the Central Committee for Comprehensive Deepening Reform clarified that classified reform principles should steer the mixed-ownership reform, further promoting classified governance [17].

Guiding Opinions categorize SOEs into commercial and public welfare types. Commercial SOEs are subdivided into fully competitive and special function types, primarily aiming to preserve and enhance state-owned asset value. They seek to boost the state-owned economy’s vitality and expand state capital’s role via market-oriented operations. Fully competitive SOEs, operating in highly competitive sectors, closely align with SOE mixed-ownership reforms to prioritize economic gains. The state primarily assesses their operational indicators and market competitiveness. Special function SOEs, situated in sectors crucial to national security or natural monopolies, undertake significant national projects, and balance economic performance with social responsibilities. The state, while assessing their operational indicators, also focuses on assessing their special functions, special tasks and their ability to serve people’s livelihood. Public welfare SOEs focus on delivering public services and securing livelihoods and undertake some government service functions. They prioritize social responsibilities over economic gains. The state mainly assesses their service quality and operational efficiency, cost control, etc. Social evaluations also play a role in their assessment.

The recent literature on the reform effects in Chinese SOEs primarily investigates other policies like mixed-ownership reform [15,16,26,27]; however, with scant attention to classified reform, a key principle for implementing mixed-ownership reform [17]. Theoretically, while principal-agent theory has been explored with China as the context [28], the examination of specific principal-agent issues in SOEs reform has predominantly been from a macro perspective [29,30,31], with a notable lack of micro-level analysis for enterprises.

2.2. CRSOE and ESG Decoupling

ESG research, originating from corporate social responsibility studies, has swiftly emerged as a global focal point in recent years. Tracing its origins, Sheldon [32] introduced the concept of corporate social responsibility one century ago, viewing it as voluntary philanthropy. Subsequently, the scope of corporate social responsibility expanded, encompassing enterprises’ duties towards other stakeholders [33] and their ethical, economic, legal and philanthropic responsibilities [34]. Ultimately, Elkington [35] introduced the triple bottom line theory, categorizing corporate social responsibility performance into environmental, social and governance dimensions, marking the inception of ESG research. Since its formal introduction by the United Nations Global Compact in 2004, the comprehensive ESG concept has spurred scholarly focus on the interplay among its three dimensions, economic impacts of ESG practices, its role in risk mitigation and methods of ESG measurement [36,37], grounded in institutional and stakeholder theories.

ESG measurement methods influence the content of corporate ESG disclosures and serve as a foundation for assessing ESG performance [36], propelling recent research advancements in ESG disclosures and performance. Research indicates that active ESG disclosure by companies can mitigate information asymmetry with investors, boost their reputation and consequently improve financial performance [38,39,40,41]. Regarding the impact of ESG disclosures, they enable corporations to communicate their commitment to social responsibility externally. Research indicates that disclosing greenhouse gas information can encourage corporations to reduce emissions [42]. Additionally, studies reveal a positive correlation between corporate environmental disclosure and performance [43], and ESG disclosures can further mitigate corporate earnings management [44]. Scholars have discovered that mandatory ESG disclosures improve price discovery efficiency [45]. Factors influencing ESG disclosures include national political, labor, cultural systems and corporate reputation characteristics [46]; a diverse board composition and strong ESG performance both encourage ESG information disclosure [47]. Moreover, enhancing ESG performance can lessen financing constraints, increase corporate value, recover social responsibility expenses and fulfill shareholder interests [39,48,49,50]. Good ESG performance is linked to reduced corporate risk, lower capital costs and enhanced corporate performance [48,51,52]; conversely, a company’s operational activities, financial decisions, governance and digital transformation can influence its ESG performance [53,54,55,56].

ESG disclosure refers to the statements made by enterprises about their efforts in environmental, social and governance (ESG) areas, while ESG performance reflects the measurable contributions they actually make in these fields. In response to external pressures, such as regulatory requirements, enterprises may strategically choose specific ESG disclosure content to mislead shareholders focused on sustainability. This is done to enhance their legitimacy and secure external approval. However, the actual ESG contributions may not align with these disclosures, resulting in a phenomenon known as ESG decoupling behavior [57,58,59,60]. Like the evolution from corporate social responsibility to ESG, early studies focused on corporate social responsibility decoupling. Meyer and Rowan [58] introduced the concept of decoupling, describing it as an organizational behavior that shows a mismatch between expectations and practices to obtain legitimacy. It is seen as a strategic response to institutional pressures [60]. Corporate social responsibility decoupling can “appease” stakeholders [59], but essentially deceives them, reducing transparency, increasing information asymmetry and diminishing market effectiveness. Research on ESG decoupling remains limited. Liu, Li and Meng [6] observed higher ESG decoupling levels in companies with diversified mutual fund investors; Eliwa, Aboud and Saleh [4] found a significant negative correlation between board gender diversity and ESG decoupling, with religion serving as a mediator.

Functional positioning serves as the foundation for strategic formulation and business decisions in enterprises, with ESG being integral to corporate strategy [12,14]. Therefore, the CRSOE’s approach to functional positioning for SOEs inevitably influences their ESG strategies, encompassing both disclosure and performance. Consequently, the CRSOE is bound to impact ESG decoupling. However, it remains to be seen whether this impact alleviates ESG decoupling and fosters high-quality SOEs development, or conversely, intensifies ESG decoupling, diverging from stakeholder expectations for SOEs’ ESG performance.

As discussed in the literature review of Section 2.1, due to SOEs’ dual functional position, they often face complex principal-agent issues, which include conflicts between controlling shareholders and management as well as conflicts with minority shareholders. According to principal-agency theory, there are conflicts between agents and principals over incentives and responsibilities [61]. In the context of SOEs, this manifests as two key conflicts. First, the government’s dual role as a controlling shareholder pursuing both financial and non-financial objectives complicates its ability to effectively incentivize and oversee management. Second, the government’s imposition of non-financial responsibilities on SOEs clashes with minority shareholders’ preference for financial returns. To address these issues, the CRSOE employs differentiated performance evaluation methods tailored to various categories of SOEs. This includes designing effective compensation contracts [62] to align management incentives, clarify responsibilities and balance the interests of minority shareholders. Specifically, this differentiated appraisal method enhances incentive mechanisms, clarifying SOE managers’ goals and resolving the contradictions from the dual functional positioning—balancing social responsibilities and economic benefits—thus reducing institutional pressure and enabling targeted decision-making. SOEs focusing on economic benefits no longer need to exaggerate ESG disclosures for social responsibility appraisals or investor expectations, reducing unnecessary ESG disclosures. Similarly, SOEs prioritizing public service can avoid compromising ESG performance for economic gains. These could facilitate effective ESG decoupling governance. In addition, this approach of differentiated assessment not only clarifies the primary responsibilities of different enterprises but also, through enhanced evaluation standards, restricts opportunistic behaviors like ESG decoupling. For SOEs with minority shareholders and considerable market competition, this approach ensures that minority shareholder interests are safeguarded from infringement due to potential collusion between controlling shareholders and management [63,64]. This ultimately enhances ESG decoupling governance.

Following the analysis, we propose the following hypothesis:

H1. 

Relative to non-SOEs, the CRSOE significantly alleviates ESG decoupling of SOEs.

2.3. ESG Decoupling Governance of SOEs with Different Functional Classifications by CRSOE

Although the CRSOE should generally play a significant role in governing SOEs’ ESG decoupling, the diverse industry characteristics and functional positions make it premature to determine the CRSOE’s governance impact on both special function and fully competitive SOEs’ ESG decoupling. For special function SOEs, their dual focus on economic and social responsibilities leads managers to prioritize projects crucial for societal well-being. Increased regulatory scrutiny further discourages opportunistic actions, aligning ESG disclosures more closely with actual achievements. This demonstrates the policy’s potentially pronounced governance effect on their ESG decoupling. In contrast, fully competitive SOEs, with their emphasis on economic gains and market competition, may find their managers keen on strategies that boost investment appeal and competitive edge. Enhanced ESG disclosures and performance, linked to corporate value enhancement and remuneration raise [65,66], suggest that policies should also exert a governance effect on their ESG decoupling. However, the comparative governance effectiveness of the CRSOE on ESG decoupling between these two SOE categories remains to be determined.

Drawing on the preceding analysis and comparing SOEs with non-SOEs—they are not impacted by the CRSOE—in equivalent industries and fields, we propose the following competing hypotheses to investigate the relative governance strength of the CRSOE on ESG decoupling between the two SOE types:

H2a. 

The CRSOE’s governance effect on ESG decoupling is stronger for special function SOEs than for fully competitive SOEs.

H2b. 

The CRSOE’s governance effect on ESG decoupling is stronger for fully competitive SOEs than for special function SOEs.

Figure 1 visually illustrates this study’s primary research hypotheses and effecting mechanisms.

3. Research Design

3.1. Sample Selection and Data Sources

We utilize the CRSOE’s implementation as an exogenous shock, constructing a DID model to empirically examine the policy’s effect on ESG decoupling.

Regarding sample selection, we identify SOEs that were impacted by policy shocks as the experimental group and non-SOEs, unaffected by the policy, as the control group, focusing on listed non-financial companies in China over 11 years from 2011 to 2021. The selection of this research period is driven by the necessity to validate the parallel trend assumption of the DID model, with 2016 marked as the initial year of policy enactment, encompassing five years on either side. The exclusion of 2022 from the sample period is due to the significant number of missing values in that year’s ESG decoupling data, leading to a substantial reduction in effective samples. Furthermore, to assess the policy’s influence on the ESG decoupling of SOEs with varied functional positions, also in alignment with the construction of experimental and control groups for the DID model, and drawing on the approach by Wei et al. [67], we categorize all sampled companies into SOEs with specific functions and their industry non-SOEs counterparts, and fully competitive SOEs alongside corresponding industry non-SOEs, based on broad industry classifications. SOEs with specific functions and their industry non-SOEs counterparts will be mentioned as Group A, and fully competitive SOEs alongside corresponding industry non-SOEs will be mentioned as Group B in the remaining parts of this study.

Using industry and operation fields as classification benchmarks, the CRSOE categorizes SOEs into commercial and public welfare types based on their functional positions. However, Chinese public welfare SOEs are unlikely to be publicly listed because they primarily focus on delivering public services and products, securing the well-being of the populace and performing certain governmental duties. Moreover, many public welfare SOEs are transformed from public institutions, so they do not primarily seek economic gains. Therefore, in line with the methodologies of prior Chinese scholars studying the CRSOE [67,68], we assume the absence of public welfare SOEs among listed firms, categorizing our sample’s SOEs into those with special functions and those that are fully competitive within the commercial sector.

The data on ESG decoupling used in this study are derived from ESG ratings provided by Sino-Securities and Bloomberg. Except for the data on newspaper media attention (papmedia) and online media attention (webmedia) provided in Section 5.1.3 which are derived from the Chinese Research Data Services Platform database, all other control variables and those used in further analysis originate from the China Securities Market and Accounting Research database. Companies were excluded from the sample for any of the following reasons:

• (1). Missing data essential for baseline regression;

• (2). Having an abnormal listing status (e.g., ST, *ST, PT), to prevent financial and management indicators being significantly deviant from normal values and potentially cause distortion in test results and research conclusions;

• (3). Changing their primary industry category, which might question their assignment to either the experimental or control group.

In total, we secured 3674 firm-year observations, with SOEs comprising 59.88% and non-SOEs 40.12% of the dataset. Group A represented 38.92%, while Group B made up 61.08%. All continuous variables underwent 5% and 95% winsorization, unless otherwise noted, to mitigate the effects of extreme outliers. The detailed composition of the observations can be found in Table 1.

3.2. Empirical Models and Variables Measurement

We construct a DID Model (1) to examine the effects of the CRSOE on ESG decoupling levels and assess the variance of these effects across different groups:

(1)${decoupling}_{i,t}={\beta }_0+{\beta }_1{SOE\_after}_{i,t}+{\beta }_2{Controls}_{i,t}+{Firm}_i+{Year}_t+{\epsilon }_{i,t}$

The subscripts i and t denote the company and the year, respectively. The dependent variable, decoupling, measures the degree of ESG decoupling, defined here as the inconsistency between a company’s ESG disclosures and its actual performance—specifically, the absolute difference between these two. ESG disclosure utilizes Bloomberg’s ESG disclosure score, which ranges from 0 to 100 points with equal weighting assigned to its three pillars: environmental, social and governance. Each pillar is divided into several specific themes and areas, with scores mainly based on the company’s own disclosure and publicly available ESG information (Accessing official data and documentation requires access via the Bloomberg Terminal: https://bba.bloomberg.net/?utm_source=bloomberg-menu&utm_medium=terminal (accessed on 29 November 2024). We also provide a third-party website for reference to the scoring system: https://www.esg-advising.com/insights/bloombergs-esg-disclosure-score (accessed on 29 November 2024).). ESG performance uses the ESG rating issued by Sino-Securities, which is categorized into nine levels from AAA to C and encompasses three pillars: environmental, social and governance. However, the key indicators for each pillar primarily reflect the actual efforts and achievements of companies in specific ESG domains and account for regulatory penalties against actions hindering sustainable development (For comprehensive details on the ESG ratings methodology, please visit the official website of Sino-Securities: https://www.chindices.com/files/Sino-Securities+Index+ESG+Ratings+Methodology+V2.1.pdf (accessed on 29 November 2024)). To ensure data comparability from two sources, the author standardized both before differentiation. The primary independent variable, SOE_after, is a DID interaction term of two dummy variables: state ownership and post-policy implementation period. It is assigned a value of 1 for SOEs impacted by the CRSOE from 2016 onwards, and 0 otherwise. The year 2016 is chosen as the starting point for policy impact, considering the policy’s late 2015 release and the subsequent corporate response period. SOE_after, the focal interaction term, has a regression coefficient indicating the policy’s observable effect on ESG decoupling. A significant negative coefficient suggests that the CRSOE effectively alleviates ESG decoupling, consistent with the hypothesis.

Controls encompass all relevant control variables. Drawing from existing studies on factors influencing ESG decoupling, we incorporate control variables that address both the financial characteristics and governance statuses of enterprises. Financial characteristics can serve as driving factors that influence an entity’s ESG decoupling behaviors. These characteristics include the company’s size (size) [4], reflecting asset richness; financial liquidity (liq) [69], indicating debt repayment capacity; operating performance (ROE) [6], highlighting profitability; asset turnover (turnover) [70], demonstrating asset usage efficiency; and corporate growth (growth) [4], showing variations in economic development status. Conversely, governance conditions such as management efficiency (mfee) [5], the proportion of independent directors (indep) [4] reflecting internal violations detection capabilities, duality (duality) [71] and company age (age) [71]—potentially affecting decision-making styles—constitute essential governance features influencing corporate ESG decoupling decisions. Additionally, Firm denotes individual fixed effects, mitigating omitted the variable bias constant over time at the firm level. Year signifies annual fixed effects, addressing omitted variable bias due to macroeconomic changes over time.

Moreover, within the mechanism testing in Section 5.1, we employed five variables, utilizing Models (2) and (3) to examine how the CRSOE influences ESG decoupling, while employing Model (4) to evaluate which factors might modify the strength of this influence.

(2)${\mathit{absapgap}}_{i,t}={\beta }_0+{\beta }_1{\mathit{SOE}\_\mathit{after}}_{i,t}+{\beta }_2{\mathit{Controls}}_{i,t}+{\mathit{Firm}}_i+{\mathit{Year}}_t+{\epsilon }_{i,t}$

(3)${\mathit{decoupling}}_{i,t}={\beta }_0+{\beta }_1{\mathit{SOE}\_\mathit{after}}_{i,t}+{\beta }_2{\mathit{absapgap}}_{i,t}+{\beta }_3{\mathit{Controls}}_{i,t}+{\mathit{Firm}}_i+{\mathit{Year}}_t+{\epsilon }_{i,t}$

(4)${\mathit{decoupling}}_{i,t}={\beta }_0+{\beta }_1{\mathit{SOE}\_\mathit{after}}_{i,t}+{\beta }_2{N}_{i,t}+{\beta }_3{\mathit{SOE}\_\mathit{after}\_N}_{i,t}+{\beta }_4{\mathit{Controls}}_{i,t}+{\mathit{Firm}}_i+{\mathit{Year}}_t+{\epsilon }_{i,t}$

In the aforementioned models, N represents the specific moderator variables employed. Next, we outline the measurement methods of all mechanism variables. Firstly, we use the absolute value of the aspiration–performance gap (absapgap) to test the CRSOE’s ability to inhibit ESG decoupling through gap reduction. Drawing from existing research [72], this variable captures the difference between an enterprise’s actual and expected performance; the former is determined by the firm’s annual return on assets (ROA). The expected performance is computed by averaging the previous year’s industry ROA, excluding the firm itself, and combining 50% of this average with 50% of the previous year’s expected performance. Notably, the expected performance for 2010, the year preceding the sample period, is replaced by that year’s average industry ROA, excluding the enterprise itself. Secondly, we assess the presence of political connections (pc) and their level (pclevel) to examine potential obstruction to the CRSOE’s governance effectiveness on ESG decoupling. Drawing from the existing literature [73], pc is a dummy variable showing whether the chairman or general manager holds or previously held government office or CPC affiliation, People’s Congress or the Chinese People’s Political Consultative Conference representation, while pclevel ranks these connections: township-level or county reps = 1; division-level or municipal reps = 2; department-level or provincial reps = 3; ministry-level or national reps = 4. Thirdly, we use newspaper media attention (papmedia) and online media attention (webmedia) to explore their effects on the governance of ESG decoupling by the CRSOE. The news data are sourced from the Chinese Research Data Services Platform, which aggregates information from over 600 major newspapers and more than 400 leading online media outlets (Access to detailed information requires logging into their official website to view the description of the China Financial News Database (CFND): https://www.cnrds.com/Home/Login (accessed on 29 November 2024)). papmedia is valued by the logarithm of a specific company’s name mentions in targeted newspaper content over the year, plus one. The measurement method for webmedia follows that of papmedia.

Definitions of all variables can be found in Table 2.

3.3. Descriptive Statistical Features

Descriptive statistics for the variables are presented in Table 3. The mean value for ESG decoupling (decoupling) is 0.96, indicating its widespread presence among Chinese enterprises. The average deviation between ESG disclosure and performance is close to one standard deviation. The minimum value is 0.07, the standard deviation is 0.66 and the maximum value is 2.4, indicating significant variation in ESG decoupling across enterprises. Some companies exhibit severe ESG decoupling, highlighting a noteworthy issue. The DID interaction term (SOE_after) has a mean of 0.33, showing that post-policy implementation for state-owned enterprise observations makes up about 33% of the study period’s total. With a balanced panel and considering the policy implementation timing, it is inferred that state-owned enterprises account for about 60% of the total sample, corroborating earlier data. Overall, the descriptive statistics of each variable are largely consistent with those found in existing research.

4. Empirical Results and Analysis

4.1. Baseline Regression Results and Analysis

The regression results are presented in Table 4. To validate the previously proposed hypotheses, we initially investigate the univariate impact of the CRSOE on ESG decoupling levels. Subsequently, we employ Model (1) to assess the policy’s influence on ESG decoupling from both overall and functional classification viewpoints. Controlling for both individual and annual fixed effects, and using ESG decoupling as the dependent variable, column (1) examines the univariate effect of the CRSOE DID interaction term (SOE_after) on ESG decoupling levels. Subsequently, column (2) incorporates control variables to thoroughly evaluate the CRSOE’s impact on ESG decoupling levels. Lastly, enterprises are categorized into Group A and Group B. Columns (3) and (4) explore the policy’s governance effect on ESG decoupling in both groups, facilitating a comparison of governance strength between them.

The regression results from column (1) reveal that the CRSOE interaction term’s coefficient is −0.24, significantly negative at the 1% level. This suggests a negative relationship between the CRSOE and ESG decoupling. Upon incorporating control variables, the robustness of column (2)’s regression result suggests an overall governance effect of the CRSOE on ESG decoupling. This indication confirms H1: the CRSOE significantly alleviates ESG decoupling of SOEs. This allows the enterprises to align with their assigned clear functional positions as goals, ultimately achieving the effectiveness of governance ESG decoupling. This result also demonstrates that the principal-agency problem in SOEs, as previously proposed by us, is alleviated. At the same time, this result lays the foundation for further testing of the following hypotheses.

Upon jointly analyzing the results from columns (3) and (4), we find the CRSOE interaction term’s coefficients to be −0.22 and −0.16, respectively, both significantly negative at the 1% level. Additionally, the difference between groups is statistically significant at the 1% level. This indicates that, whether for special functional SOEs or fully competitive ones, the policy significantly enhances the governance of their ESG decoupling. This aligns with our findings in Section 2 of the analysis. More importantly, column (3)’s larger regression coefficient suggests that the governance effects of the CRSOE on ESG decoupling are more pronounced in special functional SOEs than in fully competitive ones, supporting H2a while refuting H2b. From the standpoint of the principal-agency conflict involving incentives and responsibilities [61], this likely stems from the following: Managers in fully competitive SOEs are driven chiefly by economic performance incentives. In contrast, managers in special functional SOEs face greater pressure from social responsibility assessments. As a result, responsibility conflicts in these enterprises are more effectively addressed compared to incentive conflicts in fully competitive enterprises.

4.2. Robustness Checks

4.2.1. Parallel Trend Test

To qualify the CRSOE in China as a quasi-natural experiment, it is necessary for both the experimental group (SOEs) and the control group (non-SOEs) to exhibit consistent ESG decoupling trends prior to the policy’s enactment. Using the event study method, this study constructs Model (5) to verify the parallel trend assumption and to examine the CRSOE’s dynamic effects:

(5)${\mathit{decoupling}}_{i,t}={\beta }_0+\sum {\beta }_1{\mathit{SOE}\_\mathit{year}}_{i,l}+{\beta }_2{\mathit{Controls}}_{i,t}+{\mathit{Firm}}_i+{\mathit{Year}}_t+{\epsilon }_{i,t}$

In Model (5), the subscript $l$ denotes the lth year relative to 2015 (using subscript number 0), the implementation year of the CRSOE, with values ranging from −4 to 6. Essentially, the model integrates annual dummy variables with cross-products of enterprise ownership-type dummies. Other variables are defined as in Model (1). We conduct estimations using four and one year(s) prior to the CRSOE’s implementation as the base year, respectively, presenting the findings in Table 5.

As the table illustrates, columns (1) and (2) detail the full sample’s parallel trend test and dynamic effect observations. Before the CRSOE implementation, SOEs and non-SOEs showed no significant difference in ESG decoupling. Post-implementation, a significant and generally negative difference emerged, aligning with the baseline regression. This confirms the parallel trend and indicates a strong, sustained policy effect. Parallel trend tests and dynamic effect observations for Group A were conducted in columns (3) and (4), with Group B undergoing the same tests in columns (5) and (6). Before policy implementation, columns (3) and (4) show no significant ESG decoupling difference between the groups, passing the parallel trend test. Post-implementation, significant differences at the 5% level emerged in the first and fifth years, with a notable increase in absolute t-values and a consistent negative regression coefficient direction, indicating a sustained policy effect. The findings in columns (5) and (6) mirror those in columns (1) and (2), confirming the parallel trend and indicating a strong, sustained policy effect.

4.2.2. DID Test Based on Propensity Score Matching (PSM)

We construct a DID interaction term in the baseline regression to verify the hypothesis that the CRSOE alleviates ESG decoupling. However, as mentioned in Section 3, the DID interaction term is derived from multiplying two dummy variables: whether the property rights are state-owned, and whether the time is post-policy implementation, the variation in property rights also influences the interaction term’s value. Is the observed difference in ESG decoupling levels between the experimental and control groups attributable to CRSOE implementation, or to inherent differences stemming from varying property rights? To address this issue, we employ a PSM-DID test for re-estimating the baseline regression’s findings.

During the PSM phase, all control variables in Model (1) were used as matching variables, employing three matching principles: 1:1 nearest neighbor, 0.05 caliper radius and kernel matching. The matching effectively eliminated most inter-group differences in control variables between SOEs and non-SOEs, demonstrating the applicability of this robustness testing method. Nearest-neighbor matching yielded 1657 company–year observations, whereas radius and kernel matching both resulted in 3653 observations. Regression results post-PSM are presented in Table 6. Columns (1)–(3) show results for the full sample, Group A and Group B using 1:1 nearest-neighbor matching. Columns (4)–(6) and (7)–(9) follow the same order, showcasing results from 0.05 caliper radius and kernel matching, respectively. The regression coefficient of the interaction term SOE_after is significantly positive at the 1% level across Columns (1)–(9), aligning in direction and magnitude with prior baseline regression results. This suggests that our findings are robust, unaffected by the inherent differences in enterprise characteristics due to varying property rights.

4.2.3. Placebo Test

In this section, we perform a placebo test on the baseline regression by randomly generating the SOE_after interaction term. This aims to further eliminate noise interference from other random factors or concurrent policies and events affecting the CRSOE. Figure 2, Figure 3 and Figure 4 display the probability density distributions of the interaction term coefficients for the full sample, Group A and Group B, respectively. These were obtained through 500 regression estimates using Model (1) and the Monte Carlo simulation method, with companies affected by the CRSOE randomly designated. The coefficients of the interaction terms from the simulation regression are near 0. The vertical dashed lines in the figures show the coefficients of the interaction terms in the baseline regression, which are outside the distribution of the 500 simulation regressions. We also obtained the t-statistics for the interaction terms from the regression estimates and created their probability density distribution diagrams, which are not shown due to space limitations. The placebo test results suggest that the findings supported by the baseline regression are not due to conventional random factors or unobservable elements.

4.2.4. Other Robustness Checks

In addition to the tests above, we employed three methods for robustness checks to ensure the stability of our baseline regression results: employing all control variables lagged by one period, incorporating potentially omitted control variables and winsorizing at the 1% and 99% quantile levels. The regression results are presented in Table 7.

Columns (1)–(3), respectively, display the regression estimates for the full sample, Group A and Group B, after all control variables were lagged by one period. Analysis reveals that the regression coefficients are significantly negative at the 1%, 5% and 10% statistical levels, respectively, with Group A’s coefficient absolute value remaining larger than that of Group B.

Columns (4)–(6) present the regression estimates for the full sample, Group A and Group B, following the inclusion of three common control variables: leverage ratio (lev), audit firm authority (big4) and board size (board) based on Model (1). The results show that, post-inclusion of the specified control variables, the regression coefficients align closely with those of the baseline regression. The additional control variables exhibit no significant impact, suggesting that they are neither suitable as influencers of ESG decoupling in this study nor constitute omitted variables. Possible explanations include the leverage ratio’s dual impact on a company’s ESG disclosure and performance, potentially leaving the ESG decoupling level unchanged. Additionally, audit institutions primarily focus on financial conditions, with audits and attestation services for non-financial aspects like ESG not yet widespread. Similarly, board size may influence a company’s ESG disclosure and performance to a comparable extent.

Furthermore, this section relaxes the criteria for excluding extreme observations in continuous variables, aiming to preserve more observational data. Columns (7)–(9) display the regression estimates for the full sample, Group A and Group B, following 1% and 99% quantile winsorizing of all continuous variables. The results indicate that, post-winsorizing, the regression coefficients and significance levels closely align with the baseline regression. In summary, our study’s conclusions are held under the robustness checks described above, affirming the strong robustness of Model (1).

5. Further Analysis

5.1. Influencing Mechanisms

5.1.1. Aspiration–Performance Gap

According to corporate behavior theory, corporate managers, viewed as bounded rational decision-makers within organizations [74], strategically decide based on past experiences and established pathways [75]. With the ongoing evolution of corporate behavior theory, performance feedback theory has gradually been developed. Performance feedback theory illustrates how business conditions influence decision-making. The aspiration–performance gap serves as a key tool within this theory for characterizing such conditions. Corporate managers make decisions based on feedback derived from the aspiration–performance gap [74,76]. Analyzing based on this theory, a positive aspiration–performance gap demotivates a company from seeking resources, causing organizational inertia and reduced investment behaviors [77]; then, the company might cut investments and covertly decrease its ESG performance. Conversely, a negative gap prompts increased resource searching and risk-taking [78]. This indicates that any aspiration–performance gap, positive or negative, deepens ESG decoupling. The CRSOE redefines SOEs’ functional positions, granting them clearer, unified business goals and stronger social ties within the same industry. Its operations align more closely with market mechanisms, creating a same-group effect [79] that aligns same-group companies’ performance with industry expectations, thereby narrowing the aspiration–performance gap. In summary, the CRSOE is likely to alleviate ESG decoupling by reducing the aspiration–performance gap.

Table 8 presents the regression results. We measure the bidirectional gap between performance and expectations using absapgap and assess its mediating effect on ESG decoupling across the full sample, Group A and Group B. After excluding samples that lack the absolute values of the aspiration–performance gap variable, the regression results of Model 1, Model 2 and Model 3 are reported in columns (1), (2) and (3) for the full sample. Among them, the interaction regression coefficients in column (1) and column (2) are significant at the 1% and 10% statistical levels, respectively, and both are negative. The regression coefficient of absapgap in column (3) is significantly positive at the 10% statistical level, indicating that, overall, the CRSOE reduces ESG decoupling by lowering the aspiration–performance gap. This is consistent with our previous predictions. Model 1, Model 2 and Model 3 regression results for Group A samples are presented in columns (4), (6) and (8), respectively. Among them, the interaction regression coefficients in column (4) and column (6) are significantly negative at the 5% and 1% statistical levels, respectively. The regression coefficient of absapgap in column (8) is significantly positive at the 10% statistical level. This suggests that the policy can reduce ESG decoupling in SOEs with special functional positions by lowering the aspiration–performance gap. Model 1, Model 2 and Model 3 regression results for Group B samples are presented in columns (5), (7), and (9), respectively. However, their interaction regression coefficients are not significant. The interaction regression coefficient in column (5) is negative, and the t-value’s absolute value nearly reaches significance, potentially due to a smaller sample size. This does not affect the baseline regression’s conclusions. However, the lack of significance in columns (7) and (9) indicates that the policy cannot reduce the ESG decoupling of fully competitive SOEs through this method. This lack of impact may stem from fully competitive SOEs facing fewer financing constraints and gaining higher market shares than non-SOEs post-CRSOE, diminishing the same-group effect [80] and its influence on the aspiration–performance gap.

5.1.2. Political Ties of Executives

In emerging markets, including China, the growing influence of political connections on corporate economy and behavior has been gaining increasing attention [81]. The existing literature shows that scholars have extensively studied the financial impacts of corporate political connections, generally divided into positive and negative views [82,83]. Research on non-financial impacts of political connections primarily focuses on innovation performance, including patent applications and R&D expenditures [84,85]. A limited number of studies focus on sustainable development, finding that corporate political connections can promote both radical and incremental environmental innovations in specific contexts [86]. Moreover, executives’ political connections often lead to more proactive environmental strategies [87]. However, there is currently insufficient literature on the role of political connections with a focus on SOEs, and even less on ESG decoupling as inappropriate corporate behavior. We still need to investigate how political connections influence the effectiveness of formal institutions in managing ESG decoupling.

Chinese SOEs represent the concept of public ownership by all citizens. Governments at all levels and the State-owned Assets Supervision and Administration Commissions exercise agency property rights over SOEs, representing property rights on behalf of all state citizens. Consequently, the government effectively controls the productive resources and residual claims of SOEs, becoming the controlling or even sole shareholder. This inevitably results in not only regulatory but also investment relationships between government departments and SOEs, leading to complex connections distinct from private enterprises: the government acts as both investor and regulator, and may also become a creditor by lending to SOEs. State control over SOEs is ultimately exerted through appointing management personnel [88]. Consequently, SOEs’ executives, often with political backgrounds or government positions, may possess strong political motives. For these executives, the incentives of political promotion, prestige and power far outweigh economic incentives [89]. When SOEs’ CEOs are politically promoted, the value of cash holdings significantly decreases [90]. The disparity in incentive utility leads to varied behavioral decisions; SOEs’ executives may exaggerate ESG information disclosure to fulfill political objectives and enhance political performance, thereby weakening the CRSOE’s alleviation in ESG decoupling.

The results from the regression analysis using Model (4) are presented in Table 9. Columns (1)–(3) present the regression results for the full sample, Group A and Group B, respectively, using the presence of political ties as a moderating variable. In column (1), the regression coefficient for the interaction term SOE_after_pc is significantly positive at the 5% level, contrasting with the DID term SOE_after, albeit with a smaller absolute value. This suggests that the presence of political ties among SOE chairmen and general managers diminishes the governance effect of the CRSOE on ESG decoupling in the full sample. In columns (2) and (3), the regression coefficients for SOE_after_pc are significantly positive at the 5% level and insignificant, respectively, with both coefficients having smaller absolute values compared to SOE_after. This indicates that political ties among SOE chairmen and general managers notably weaken the governance effect of the CRSOE on ESG decoupling in special function SOEs, but not in fully competitive SOEs. This implies that post-reform, fully competitive SOEs prioritize economic benefits and market competitiveness, severing political incentives for executives with political ties. Consequently, their governance of ESG decoupling is unaffected by political ties, reflecting a measure of the CRSOE’s success.

Columns (4)–(6) further outline the regression outcomes for the full sample, Group A and Group B, considering the degree of political connections as the moderating variable. Notably, in columns (4) and (5), the regression coefficients for SOE_after_pc are significantly positive at the 1% level, whereas in column (6), they are not significant. This finding more distinctly corroborates the conclusions from the analysis in columns (1)–(3), offering a more detailed and accurate assessment of the moderating variable.

5.1.3. Media Attention

The rapid growth of the media industry has placed it in a crucial role in modern society, leading to increased research on its influence on businesses and the financial sector [91]. Specifically, concerning media attention’s impact on corporate governance and decision-making, existing studies primarily explore four areas: mergers and acquisitions, information disclosure, investor relations and financial fraud. For instance, media attention can enhance the success probability of mergers and predict future mergers [92,93]. It compels companies to disclose information as shareholders demand, shaping investor beliefs to reinforce corporate governance [94,95]. Media attention effectively identifies corporate fraud and misconduct, prompting behavior change or government sanctions [96,97]. Although research has considered media attention’s impact on information disclosure and financial fraud, there is a lack of studies on its role in external policy governance of ESG decoupling as a non-financial information fraud in sustainable development.

As a classic external governance mechanism for companies, media attention serves both as a source of corporate legitimacy and as an origin of crises [98]. Media attention not only enhances the quality of corporate ESG disclosures [99] but also enables companies to reduce agency costs and improve their ESG performance under increased analyst scrutiny [100]. By exerting external market pressure, media attention exposes environmental and other regulatory violations, which are not easily detected [101], but may also prompt short-sighted behavior in corporate management, reluctant to invest in ESG [102,103]. By applying external pressure, media attention simultaneously enhances ESG disclosure and performance, and thus may either strengthen or weaken the governance effect of the CRSOE on ESG decoupling.

Similar to the examination of political ties’ moderating effects, we continue to employ Model (4) for regression analysis, with the findings presented in Table 10. Columns (1)–(3) present the regression results for the whole sample, Group A and Group B, using print media attention as the moderating variable. Column (1) shows the regression coefficient of the interaction term SOE_after_papmedia to be significantly positive at the 1% level, with a sign opposite to that of the DID term SOE_after, but with a relatively smaller absolute value, suggesting that print media attention received by SOEs diminishes the governance impact of the CRSOE on ESG decoupling for the whole sample. The findings in column (2) mirror those in column (1), yet the regression coefficient for the interaction term SOE_after_papmedia in column (3) is not significant, highlighting that the diminishing effect of print media attention on the CRSOE’s alleviation of ESG decoupling is pronounced only in special function SOEs, not in fully competitive SOEs.

Columns (4)–(6) outline the regression outcomes for the whole sample, Group A and Group B, with online media attention serving as the moderating variable. The regression coefficients for the interaction term SOE_after_webmedia in all three columns are significantly positive at the 1% level, with signs opposite to the DID term SOE_after, yet with smaller absolute values, indicating that online media attention, for both special function and fully competitive SOEs, reduces the governance effectiveness of the CRSOE on ESG decoupling.

Overall, media attention on SOEs may exert excessive pressure, potentially triggering management’s near-sighted behavior, thereby increasing the motivation for ESG decoupling and diminishing the governance impact of the CRSOE on such decoupling. Furthermore, the moderating effects of media attention vary across different media types. The impact of online media attention in weakening the CRSOE’s alleviation of ESG decoupling is notably more pronounced and intense than that of print media. This could be attributed to the internet era [104], wherein the public and stakeholders have easier access to online media content and can respond more readily. Additionally, the influence of political review on Chinese newspapers [105] contrasts with the relative freedom of online media, which enjoys higher public trust and exerts greater pressure on SOEs than print media.

Moreover, the functional positioning of enterprises influences the moderating effect of media attention on the CRSOE’s alleviation of ESG decoupling. Compared to special function SOEs, the weakening effect of media attention on the alleviation of ESG decoupling by the CRSOE in fully competitive SOEs is less significant and intense. Given their focus on maintaining reputation and enhancing market competitiveness, these enterprises prioritize analyst tracking and are more proactive in managing the external pressures from media attention, thereby mitigating short-sighted behavior to some extent.

5.2. Heterogeneity Analysis

5.2.1. Greenwashing and Brownwashing: Two Types of ESG Decoupling

ESG decoupling refers to the inconsistency between ESG disclosures and actual performance, encompassing two phenomena: greenwashing, where disclosures surpass actual performance, and brownwashing, where actual performance exceeds disclosed information [4]. Factors such as strict external regulations, incentives from consumers and investors, company-level corporate characteristics and managers’ psychological motivations can all contribute to greenwashing [106]. Meanwhile, pressures from the external governance environment and enhanced stakeholder oversight through internal governance mechanisms can diminish the likelihood of companies engaging in greenwashing [107]. The aim of brownwashing is to avert undue scrutiny from stakeholders and competitive disadvantages [108]. For Chinese firms, this also reflects the doctrine of the mean, indicating a reluctance to face the evaluative pressures associated with high levels of ESG practices. Regardless of whether it is greenwashing or brownwashing, both lead to significant issues like information asymmetry and decreased transparency. A more detailed analysis into whether the CRSOE exerts a governance role on greenwashing or brownwashing can undoubtedly offer a more precise evaluation of policy impacts.

We define ESG greenwashing (greenwash) as the discrepancy between the standardized scores of ESG disclosures and ESG performance, while ESG brownwashing (brownwash) is defined as the difference in the opposite direction. Companies are categorized based on whether they have at least six years (over half) of greenwash greater than 0 between 2011 and 2021, with those exceeding this threshold labeled as greenwashing companies and the others as brownwashing companies. Contrary to the baseline regression and in accordance with Model (1), this analysis treats ESG greenwashing and brownwashing as dependent variables, conducting regression tests separately on the entire sample, as well as on greenwashing and brownwashing companies within Group A and Group B. Due to the reduced sample size per group and the scarcity of extreme values in the data for ESG greenwashing and brownwashing, all continuous variables undergo winsorizing at the 1% and 99% quantiles prior to regression.

Table 11 presents the regression results for the entire sample, categorized by ESG decoupling type and employing various dependent variables across columns (1)–(3). Column (1) serves as the baseline regression reference for the entire sample. In column (2), the regression on greenwashing companies reveals a significantly negative interaction term at the 10% statistical level, suggesting that the CRSOE markedly alleviates the ESG greenwashing among these companies. Column (3)’s regression on brownwashing companies shows an insignificant interaction term, indicating that the CRSOE does not significantly influence the ESG brownwashing in these companies. These findings suggest that the CRSOE, through assigning SOEs’ specialized functional roles and abandoning the “one-size-fits-all” regulatory assessment, not only mitigates external environmental pressures on SOEs but also encourages SOE managers to enhance internal governance mechanisms aligned with corporate functional objectives. Therefore, it can be argued that the motivation for SOEs to greenwash has weakened to some extent, and that the CRSOE has achieved certain effects in the governance of greenwashing. However, the reforms cannot influence factors leading to ESG brownwashing, such as peer pressure and excessive attention from investors. SOEs remain concerned that honestly disclosing their excellent ESG performance might create pressure to maintain high expenditures in this area to uphold their reputation. As a result, they continue to engage in ESG brownwashing behavior. This may explain why the CRSOE has no significant impact on ESG brownwashing behavior.

Columns (4)–(9) present the regression outcomes for Group A and Group B, further segmented by ESG decoupling type and employing varied dependent variables. Columns (4) and (5) act as baseline regressions for Group A and Group B, respectively, serving as references. In columns (6) and (7), regressions on greenwashing companies in both Group A and Group B reveal no significant inter-group differences, suggesting a uniform governance effect of the CRSOE on ESG greenwashing across SOEs with special functions and those in fully competitive sectors. Columns (8) and (9)’s regressions on brownwashing companies in both groups also show no significant inter-group differences, indicating a consistent lack of noticeable CRSOE governance effect on ESG brownwashing across SOEs, regardless of their functional position or competitive environment. Consequently, further analysis of these results is deemed unnecessary.

5.2.2. GRI Standards Adherence

In 1997, the United Nations Environment Program and the Coalition for Environmentally Responsible Economies in the United States launched the GRI, the first to set independent sustainability reporting standards. The GRI standards system, evolving over time, is now a globally adopted ESG reporting framework among listed companies [109]. Research shows that following GRI standards enhances corporate social responsibility reporting [110] and positively affects ESG practice disclosures [111]. Companies adhering to GRI standards for ESG disclosure achieve greater carbon emission reductions [112]. Thus, ESG disclosures following GRI standards not only elevate disclosure levels but also genuinely enhance ESG performance. Analyzing the variability in GRI standard adherence can reveal its potential to substitute the CRSOE’s governance in ESG decoupling.

To examine the heterogeneity arising, we subdivided the entire sample, Group A and Group B based on their adherence to GRI standards for social responsibility reporting, and performed regression analyses as per Model (1), with findings presented in Table 12. Columns (1)–(3) show the regression outcomes for the whole sample, categorized by GRI standard adherence, with column (1) acting as the baseline regression for reference. Column (2) reveals an insignificant interaction term, suggesting that the CRSOE’s governance effect on ESG decoupling is not pronounced in SOEs following GRI standards for their reports. Column (3) indicates a significantly negative interaction term at the 1% level, with a larger absolute value than in column (1), suggesting that the CRSOE significantly lowers the ESG decoupling level in SOEs not adhering to GRI standards, with a stronger policy impact than on the whole sample. This outcome could stem from SOEs adhering to GRI standards for ESG disclosure already exhibiting higher disclosure levels and ESG performance focus, resulting in lower ESG decoupling, thus diminishing the CRSOE’s incremental impact. For SOEs not following GRI standards for ESG disclosure, the evident governance effect of the CRSOE on ESG decoupling confirms the earlier hypothesis.

Columns (4)–(9) show the categorized regression outcomes for Group A and Group B, based on GRI standard adherence, with columns (4) and (5) as baseline regressions for reference. Regression analysis in columns (6) and (7) for companies adhering to GRI standards shows a significant positive interaction term in column (6) at the 5% level, but not in column (7). This suggests that the CRSOE intensified ESG decoupling in special function SOEs adhering to GRI standards, a finding that contradicts earlier analysis and, as seen with column (2), stems from a small sample size, reducing its reliability. In columns (8) and (9), regression for companies not following GRI standards revealed significantly negative interaction terms at the 1% level. The larger absolute value in column (8) compared to column (9) indicates that the CRSOE’s governance on ESG decoupling is more pronounced in special function SOEs not adhering to GRI standards, aligning with baseline findings.

6. Conclusions

6.1. Major Findings of the Study

First, from the perspective of ownership heterogeneity, we apply the principal-agency theory to examine the impact of China’s CRSOE policy, a strategy designed to enhance the functional positioning of SOEs, on the governance of ESG decoupling. This investigation aims to explore the reasons why Chinese SOEs have consistently maintained rapid development. Our findings indicate that the CRSOE policy, by establishing a clear functional position and implementing differentiated performance evaluations, has enhanced SOE operations and governance through improved incentives and accountability supervision. It alleviates principal-agency conflicts within SOEs and mitigates the ESG decoupling from an overall and categorical perspective, including special function and fully competitive categories. By comparing the governance effects of these two categories, we observe that the CRSOE has a stronger ESG decoupling governance effect on special function SOEs than on fully competitive ones. This discrepancy may stem from the higher accountability and supervision pressures faced by managers of special function SOEs, as opposed to the economic incentives that primarily drive managers of fully competitive SOEs. These factors contribute to a more significant reduction in principal-agent conflicts within special function SOEs.

Next, we analyze the mechanism through which the CRSOE governs ESG decoupling. According to corporate behavior and performance feedback theories, the CRSOE leverages peer effects by reducing the aspiration–performance gap in both overall and special function SOEs. This adjustment influences corporate decision-making and achieves governance by addressing ESG decoupling. However, this mechanism does not apply to fully competitive SOEs. This may be because, despite operating in highly competitive markets, they maintain competitive financing advantages. Consequently, the peer effect is minimal, and the aspiration–performance gap remains largely unaltered. Political promotions may provide strong incentives that surpass economic ones, political ties of executives turn out to diminish the ESG decoupling governance in general and special function SOEs. This does not affect fully competitive SOEs, indicating the policy’s limited success in incentive enhancement. Press media attention exerts undue pressure, undermining ESG decoupling governance in general and special function SOEs. However, this moderation effect is lessened, maybe because of government controls and public skepticism. Conversely, online media’s broader reach significantly impairs ESG decoupling governance across all SOE types.

Finally, heterogeneity tests reveal that the CRSOE policy only reduced instances of greenwashing within ESG decoupling. Enterprises appear hesitant to manage the excessive attention and pressure associated with maintaining high ESG performance levels. As a result, the CRSOE policy fails to address the brownwashing arising from this reluctance. Furthermore, the CRSOE primarily impacts SOEs lacking GRI standards compliance, showing little effect on those adhering to GRI standards with high ESG disclosure and performance. This suggests that GRI standards implementation may serve as a substitute for certain aspects of the CRSOE.

6.2. Theoretical Insights

We explore the governance path of ESG decoupling against the backdrop of the CRSOE, offering numerous insights for the theory’s development and practical application.

First, we delve into the underexplored area of Chinese SOEs’ functional classification, examining the CRSOE as a potential factor influencing ESG decoupling. Simultaneously, we categorize ESG decoupling into greenwashing and brownwashing. This approach not only enhances our understanding of ESG decoupling’s nuances but also uncovers the CRSOE as a novel factor in this context, thereby augmenting the empirical evidence from China for future studies.

Second, we assess the influence of the CRSOE as an indicator of decision-making behavior that reflect enterprises’ high-quality development on ESG decoupling. We further conduct analyses based on functional classifications to thoroughly evaluate the CRSOE’s effects. On one hand, our study validates the CRSOE’s appropriateness of categorizing SOEs by type; on the other hand, it identifies that the CRSOE has areas for enhancement in the specific oversight and market-driven incentives of fully competitive SOEs. Our study explores methods for SOEs in China to achieve sustainable, high-quality development through ESG decoupling governance. Additionally, it showcases the achievements and areas for improvement in Chinese SOEs’ high-quality development in dimensions like greenness and openness to the world. This research provides significant support for the development of economic theories in China and other emerging countries globally.

Third, based on prior studies, we select the CRSOE—a policy altering the goals of SOEs—as a focal point, aiming to elucidate the sequence whereby corporate goals dictate governance actions like incentives and oversight, subsequently influencing decision-making and actions. This enriches the micro-level discourse on the principal-agent dilemma within the CRSOE, reinforcing and expanding principal-agent and optimal contract theories within China’s milieu. Furthermore, validating the CRSOE’s effectiveness in narrowing the aspiration–performance gap for alleviating ESG decoupling enriches the practical applications of corporate behavior and performance feedback theories.

6.3. Managerial Insights

Our research conclusions offer several managerial insights, which we hope will be of practical significance to the Chinese government in further supporting enterprises and fostering a market economy characterized by high-quality development and fair competition.

First, to further refine the CRSOE, it is advisable to enhance the marketization of fully competitive SOEs. This involves maximizing the market economy’s role, reducing or eliminating the political oversight of these enterprises. By setting SOEs as models, China could aim to foster an open, transparent and efficient competitive environment in related sectors, aligning with international business standards. This external shift encourages SOEs to actively engage with the market, leading to same-group effects, an alleviation in ESG decoupling, and improved transparency.

Second, aiming to serve enterprises and enhance the market economy’s quality, we suggest reducing or eliminating political promotion incentives for SOEs. These incentives not only diminish their awareness of market competition, focusing solely on career advancement but also unfairly advantage them over non-SOEs, undermining the free market’s fairness and transparency. Specifically, the government could act as a non-SOEs shareholder, ensuring a clear separation of management and control rights, and appoint professional managers for enterprise operations.

Third, alongside driving fully competitive SOEs towards economic gains and market competitiveness, it is crucial to motivate them not to overlook non-financial objectives. Truthful ESG disclosure and proactive performance can ultimately align with economic goals. The relationship between these goals is not oppositional but complementary. In assessing these SOEs’ economic objectives, indicators should be innovatively established to evaluate financial performance derived from fulfilling social responsibilities, appealing to investors with a preference for ESG responsibilities.

6.4. Research Limitations and Prospects

Our study acknowledges several limitations, which future research could address to deepen understanding and foster innovation.

First, this paper focuses exclusively on the governance role of the CRSOE in ESG decoupling within China, without tackling policies relevant to ESG decoupling governance in SOEs globally, limiting its international applicability. Future research could investigate how related policies in other countries impact ESG decoupling and their underlying mechanisms. Additionally, research with a cross-national perspective could examine the effects of institutional differences on policy governance of ESG decoupling.

Second, this paper categorizes ESG decoupling into greenwashing and brownwashing, providing a relatively straightforward framework. This approach may oversimplify the complex interrelationships between the two phenomena. Future research could use greenwashing and brownwashing as entry points to deeply probe into the mechanisms behind ESG decoupling’s occurrence and governance. Additionally, further exploration could investigate the potential conversion relationship between greenwashing and brownwashing, and to identify factors influencing enterprise behavior towards these activities.

Third, while this paper identifies factors influencing ESG decoupling, it does not extend the logical examination to understand how ESG decoupling, as an enterprise behavior, influences financial performance and operational objectives. Future research could address this gap by employing advanced methodologies, such as fuzzy set qualitative comparative analysis, to construct a comprehensive logical framework centered on ESG decoupling, encompassing its causes and effects.

Author Contributions

Conceptualization, H.Z., D.W. and Z.Z.; Data curation, H.Z.; Formal analysis, H.Z.; Funding acquisition, H.Z., D.W. and Z.Z.; Investigation, H.Z.; Methodology, H.Z.; Project administration, X.H.; Software, H.Z.; Supervision, Dongmei Wang, Z.Z. and X.H.; Validation, Dongmei Wang; Visualization, H.Z.; Writing—original draft, H.Z.; Writing—review and editing, Dongmei Wang, Z.Z. and X.H. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Restrictions apply to the availability of these data. Data were obtained from Bloomberg Terminal, Sino-Securities, China Stock Market & Accounting Research Database, Chinese Research Data Services Platform and are available from the authors with the permission of those third parties, respectively.

Conflicts of Interest

The authors declare no conflicts of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures and Tables

Figure 1. Primary research hypotheses and effecting mechanisms. The content within the dotted box can be regarded as a whole to participate in the logical deduction process.

Figure 2. Placebo test for the full sample. The vertical dotted line shows the coefficient of the interaction term in the baseline regression.

Figure 3. Placebo test for Group A. The vertical dotted line shows the coefficient of the interaction term in the baseline regression.

Figure 4. Placebo test for Group B. The vertical dotted line shows the coefficient of the interaction term in the baseline regression.

Table 1

Observations composition.

	Group A	Group B	Total
SOEs (Experimental group)	1078	1122	2200
Non-SOEs (Control group)	352	1122	1474
Total	1430	2244	3674

Due to missing values in certain variables, there is a slight difference between the observations presented in this table and the tests in Section 5.1.1 and Section 5.1.3.

Table 2

Variables’ definitions.

Variable Name	Symbol	Definition
ESG decoupling degree	decoupling	As explained in the text.
DID interaction term	SOE_after	As explained in the text.
company size	size	The natural logarithm of total assets.
financial liquidity	liq	Ratio of current assets to current liabilities.
company operation performance	ROE	Ratio of net profit to shareholder equity balance.
asset turnover	turnover	Ratio of operating revenue to total assets.
company growth	growth	Ratio of the difference between the current year’s total operating income and the previous year’s total operating income to the previous year’s total operating income.
company management efficiency	mfee	Ratio of administrative expenses to operating income.
the proportion of independent directors	indep	Ratio of number of independent directors to number of board members.
CEO duality	duality	If the chairman and general manager are the same person, the value is 1, otherwise it is 0.
company age	age	Add 1 to the difference between the data year and the establish year, and then take the natural logarithm.
absolute value of the aspiration–performance gap	absapgap	As explained in the text.
political connections	pc	As explained in the text.
political connections level	pclevel	As explained in the text.
newspaper media attention	papmedia	As explained in the text.
online media attention	webmedia	As explained in the text.

Table 3

Descriptive statistics for the primary variables.

Variable	Observations	Mean	Standard Deviation	Min	P50	Max
decoupling	3674	0.958	0.664	0.0700	0.815	2.400
SOE_after	3674	0.327	0.469	0.000	0.000	1.000
size	3674	23.462	1.229	21.543	23.354	26.051
liq	3674	1.709	1.077	0.439	1.397	4.612
ROE	3674	0.098	0.068	−0.006	0.090	0.242
turnover	3674	0.669	0.365	0.169	0.606	1.529
growth	3674	0.135	0.207	−0.207	0.108	0.627
mfee	3674	0.066	0.040	0.015	0.057	0.157
indep	3674	0.369	0.042	0.333	0.364	0.455
duality	3674	0.154	0.361	0.000	0.000	1.000
age	3674	2.938	0.293	2.303	2.996	3.367
absapgap	3267	0.023	0.023	0.001	0.015	0.085
pc	3674	0.371	0.483	0.000	0.000	1.000
pclevel	3674	1.215	1.671	0.000	0.000	4.000
papmedia	3509	4.625	1.873	0.693	4.290	12.992
webmedia	3674	5.723	1.124	1.386	5.694	10.814

Table 4

Impacts of CRSOE on ESG decoupling.

	(1)	(2)	(3)	(4)
	Full Sample	Full Sample	Group A	Group B
	Decoupling	Decoupling	Decoupling	Decoupling
SOE_after	−0.237 ***	−0.188 ***	−0.223 ***	−0.156 ***
	(−5.74)	(−4.36)	(−2.58)	(−2.98)
size		0.094 **	−0.021	0.153 ***
		(2.48)	(−0.31)	(3.32)
liq		−0.060 ***	0.028	−0.077 ***
		(−2.95)	(0.69)	(−3.34)
ROE		−0.253	−0.226	−0.420
		(−1.05)	(−0.63)	(−1.28)
turnover		0.131 *	0.115	0.176 *
		(1.70)	(0.99)	(1.70)
growth		−0.116 *	−0.083	−0.101
		(−1.94)	(−0.96)	(−1.24)
mfee		0.452	1.348	0.871
		(0.81)	(1.36)	(1.21)
indep		1.033 **	2.350 ***	0.303
		(2.40)	(3.25)	(0.56)
duality		0.027	0.198 **	−0.049
		(0.61)	(2.52)	(−0.92)
age		0.427 *	1.622 ***	−0.248
		(1.85)	(4.03)	(−0.90)
constant	0.639 ***	−3.336 ***	−4.669 **	−2.336 *
	(5.28)	(−2.75)	(−2.15)	(−1.74)
diff.			p = 0.0000 ***
Firm FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
N	3674	3674	1430	2244
adj. R2	0.2058	0.2120	0.2798	0.1754

The t statistics in parentheses use robust standard errors under double fixed effects of company and year; ***, ** and * indicate that the coefficients are significant at the 1%, 5% and 10% statistical levels, respectively; diff. display the Chow test is used to test the differences between groups and the result is the p value of the product of the group dummy variable and the main independent variable. The following tables are the same as this table.

Table 5

Parallel trend test.

	(1)	(2)	(3)	(4)	(5)	(6)
	Full Sample	Full Sample	Group A	Group A	Group B	Group B
	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling
pre4		0.121		0.125		0.155
		(1.17)		(0.62)		(1.21)
pre3	−0.082	0.039	−0.064	0.060	−0.123	0.031
	(−0.79)	(0.39)	(−0.34)	(0.31)	(−0.98)	(0.26)
pre2	−0.104	0.017	−0.076	0.049	−0.147	0.007
	(−0.99)	(0.16)	(−0.37)	(0.23)	(−1.16)	(0.06)
pre1	−0.121		−0.125		−0.155
	(−1.17)		(−0.62)		(−1.21)
current	−0.357 ***	−0.237 **	−0.314 *	−0.189	−0.357 ***	−0.203 *
	(−3.73)	(−2.57)	(−1.66)	(−0.98)	(−3.07)	(−1.81)
post1	−0.377 ***	−0.257 ***	−0.417 **	−0.292	−0.363 ***	−0.209 *
	(−3.98)	(−2.84)	(−2.31)	(−1.60)	(−3.14)	(−1.88)
post2	−0.297 ***	−0.176 **	−0.285 *	−0.160	−0.271 **	−0.116
	(−3.14)	(−1.96)	(−1.69)	(−0.94)	(−2.29)	(−1.02)
post3	−0.353 ***	−0.232 **	−0.327 *	−0.202	−0.348 ***	−0.194 *
	(−3.61)	(−2.49)	(−1.77)	(−1.09)	(−2.89)	(−1.68)
post4	−0.334 ***	−0.214 **	−0.310	−0.186	−0.367 ***	−0.213 *
	(−3.26)	(−2.17)	(−1.54)	(−0.92)	(−2.98)	(−1.79)
post5	−0.284 ***	−0.164 *	−0.395 **	−0.270	−0.247 *	−0.093
	(−2.77)	(−1.66)	(−2.01)	(−1.36)	(−1.96)	(−0.76)
post6	−0.298 ***	−0.177	−0.328	−0.203	−0.294 **	−0.140
	(−2.65)	(−1.63)	(−1.61)	(−0.98)	(−2.12)	(−1.03)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
N	3674	3674	1430	1430	2244	2244
adj. R2	0.2142	0.2142	0.2775	0.2775	0.1765	0.1765

The t statistics in parentheses use robust standard errors under double fixed effects of company and year; ***, ** and * indicate that the coefficients are significant at the 1%, 5% and 10% statistical levels, respectively.

Table 6

Regression results post-PSM.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Full Sample	Group A	Group B	Full Sample	Group A	Group B	Full Sample	Group A	Group B
	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling
SOE_after	−0.370 ***	−0.522 ***	−0.290 ***	−0.193 ***	−0.233 ***	−0.160 ***	−0.193 ***	−0.233 ***	−0.160 ***
	(−5.06)	(−3.85)	(−3.24)	(−4.45)	(−2.67)	(−3.04)	(−4.45)	(−2.67)	(−3.04)
diff.		p = 0.0047 ***			p = 0.0000 ***			p = 0.0000 ***
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	1657	572	1085	3653	1425	2228	3653	1425	2228
adj. R2	0.2039	0.2885	0.1706	0.2142	0.2804	0.1777	0.2142	0.2804	0.1777

The t statistics in parentheses use robust standard errors under double fixed effects of company and year; *** indicate that the coefficients are significant at the 1%, 5% and 10% statistical levels, respectively; diff. display the Chow test is used to test the differences between groups and the result is the p value of the product of the group dummy variable and the main independent variable.

Table 7

Other robustness checks.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Full Sample	Group A	Group B	Full Sample	Group A	Group B	Full Sample	Group A	Group B
	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling
SOE_after	−0.158 ***	−0.239 **	−0.106 *	−0.187 ***	−0.229 ***	−0.152 ***	−0.198 ***	−0.243 **	−0.158 ***
	(−3.44)	(−2.55)	(−1.94)	(−4.31)	(−2.63)	(−2.88)	(−4.10)	(−2.51)	(−2.71)
lev				0.080	0.005	0.162
				(0.45)	(0.02)	(0.69)
big4				0.002	−0.435 *	0.103
				(0.02)	(−1.76)	(0.98)
board				−0.170	−0.080	−0.294
				(−1.16)	(−0.36)	(−1.49)
diff.		p = 0.0000 ***			p = 0.0001 ***			p = 0.0001 ***
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	3340	1300	2040	3674	1430	2244	3674	1430	2244
adj. R2	0.2414	0.3076	0.2090	0.2117	0.2812	0.1758	0.2173	0.2694	0.1900

The t statistics in parentheses use robust standard errors under double fixed effects of company and year; ***, ** and * indicate that the coefficients are significant at the 1%, 5% and 10% statistical levels, respectively; diff. display the Chow test is used to test the differences between groups and the result is the p value of the product of the group dummy variable and the main independent variable.

Table 8

Influencing mechanisms: aspiration–performance gap.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Full Sample	Full Sample	Full Sample	Group A	Group B	Group A	Group B	Group A	Group B
	Decoupling	Absapgap	Decoupling	Decoupling	Decoupling	Absapgap	Absapgap	Decoupling	Decoupling
SOE_after	−0.119 ***	−0.002 *	−0.117 **	−0.210 **	−0.079	−0.009 ***	0.000	−0.198 **	−0.079
	(−2.61)	(−1.72)	(−2.56)	(−2.37)	(−1.41)	(−3.57)	(0.15)	(−2.23)	(−1.42)
absapgap			1.008 *					1.375 *	0.858
			(1.68)					(1.69)	(1.14)
diff.				p = 0.0000 ***		p = 0.0433 **		p = 0.0000 ***
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	3267	3267	3267	1375	1892	1375	1892	1375	1892
adj. R2	0.2049	0.3096	0.2055	0.2813	0.1553	0.2475	0.3427	0.2819	0.1555

The t statistics in parentheses use robust standard errors under double fixed effects of company and year; ***, ** and * indicate that the coefficients are significant at the 1%, 5% and 10% statistical levels, respectively; diff. display the Chow test is used to test the differences between groups and the result is the p value of the product of the group dummy variable and the main independent variable.

Table 9

Influencing mechanisms: political ties of executives.

	(1)	(2)	(3)	(4)	(5)	(6)
	Full Sample	Group A	Group B	Full Sample	Group A	Group B
	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling
SOE_after	−0.213 ***	−0.272 ***	−0.148 ***	−0.223 ***	−0.287 ***	−0.155 ***
	(−4.74)	(−3.09)	(−2.70)	(−4.98)	(−3.26)	(−2.84)
pc	−0.046	−0.007	−0.060
	(−1.28)	(−0.11)	(−1.36)
SOE_after_pc	0.115 **	0.215 **	−0.015
	(2.05)	(2.52)	(−0.20)
pclevel				−0.022 **	0.000	−0.028 **
				(−2.06)	(0.00)	(−2.13)
SOE_after_pclevel				0.053 ***	0.086 ***	0.009
				(3.04)	(3.34)	(0.36)
diff.		p = 0.0194 **			p = 0.0537 *
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
N	3674	1430	2244	3674	1430	2244
adj. R2	0.2125	0.2829	0.1756	0.2139	0.2862	0.1767

The t statistics in parentheses use robust standard errors under double fixed effects of company and year; ***, ** and * indicate that the coefficients are significant at the 1%, 5% and 10% statistical levels, respectively; diff. display the Chow test is used to test the differences between groups and the result is the p value of the product of the group dummy variable and the main independent variable.

Table 10

Influencing mechanisms: media attention.

	(1)	(2)	(3)	(4)	(5)	(6)
	Full Sample	Group A	Group B	Full Sample	Group A	Group B
	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling
SOE_after	−0.363 ***	−0.464 ***	−0.236 **	−0.852 ***	−1.007 ***	−0.606 ***
	(−5.21)	(−3.96)	(−2.54)	(−6.84)	(−5.40)	(−3.37)
papmedia	−0.053 ***	−0.008	−0.089 ***
	(−2.69)	(−0.27)	(−3.37)
SOE_after_papmedia	0.040 ***	0.055 ***	0.020
	(3.21)	(3.00)	(1.18)
webmedia				−0.009	−0.009	0.005
				(−0.32)	(−0.23)	(0.15)
SOE_after_webmedia				0.122 ***	0.143 ***	0.084 ***
				(5.80)	(4.87)	(2.67)
diff.		p = 0.0122 **			p = 0.0004 ***
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
N	3509	1408	2101	3509	1408	2101
adj. R2	0.2123	0.2737	0.1802	0.2168	0.2832	0.1779

The t statistics in parentheses use robust standard errors under double fixed effects of company and year; *** and ** indicate that the coefficients are significant at the 1% and 5% statistical levels, respectively; diff. display the Chow test is used to test the differences between groups and the result is the p value of the product of the group dummy variable and the main independent variable.

Table 11

Heterogeneity analysis: greenwashing and brownwashing.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Full Sample	Full Sample	Full Sample	Group A	Group B	Group A	Group B	Group A	Group B
	Decoupling	Greenwash	Brownwash	Decoupling	Decoupling	Greenwash	Greenwash	Brownwash	Brownwash
SOE_after	−0.198 ***	−0.152 *	−0.068	−0.223 ***	−0.156 ***	−0.348 **	−0.041	−0.183	0.039
	(−4.10)	(−1.85)	(−1.03)	(−2.58)	(−2.98)	(−2.02)	(−0.42)	(−1.62)	(0.44)
diff.				p = 0.0001 ***		p = 0.5800		p = 0.1906
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	3674	1771	1903	1430	2244	583	1188	847	1056
adj. R2	0.2173	0.4853	0.5390	0.2798	0.1754	0.5273	0.4708	0.6053	0.4791

The t statistics in parentheses use robust standard errors under double fixed effects of company and year; ***, ** and * indicate that the coefficients are significant at the 1%, 5% and 10% statistical levels, respectively; diff. display the Chow test is used to test the differences between groups and the result is the p value of the product of the group dummy variable and the main independent variable.

Table 12

Heterogeneity analysis: GRI standards adherence.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Full Sample	Full Sample	Full Sample	Group A	Group B	Group A	Group B	Group A	Group B
	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling	Decoupling
SOE_after	−0.188 ***	0.020	−0.296 ***	−0.223 ***	−0.156 ***	0.594 **	−0.179	−0.380 ***	−0.223 ***
	(−4.36)	(0.12)	(−6.36)	(−2.58)	(−2.98)	(2.22)	(−0.93)	(−3.98)	(−3.96)
diff.		p = 0.0000 ***		p = 0.0001 ***		p = 0.0243 **		p = 0.0006 ***
(6) = (8)						p = 0.0076 ***			-
(7) = (9)							p = 0.0009 ***
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Firm FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	3674	663	3011	1430	2244	325	338	1105	1906
adj. R2	0.2120	0.3737	0.2325	0.2798	0.1754	0.3790	0.3891	0.3384	0.1788

The t statistics in parentheses use robust standard errors under double fixed effects of company and year; *** and ** indicate that the coefficients are significant at the 1% and 5% statistical levels, respectively; diff. display the Chow test is used to test the differences between groups and the result is the p value of the product of the group dummy variable and the main independent variable.

References

1. United Nations. Who Cares Wins: Connecting Financial Markets to a Changing World; United Nations: New York, NY, USA, 2004.

2. Jain, K.; Tripathi, P. Mapping the environmental, social and governance literature: A bibliometric and content analysis. J. Strateg. Manag.; 2023; 16, pp. 397-428. [DOI: https://dx.doi.org/10.1108/JSMA-05-2022-0092]

3. Liu, M.; Luo, X.; Lu, W.-Z. Public perceptions of environmental, social, and governance (ESG) based on social media data: Evidence from China. J. Clean. Prod.; 2023; 387, 135840. [DOI: https://dx.doi.org/10.1016/j.jclepro.2022.135840]

4. Eliwa, Y.; Aboud, A.; Saleh, A. Board gender diversity and ESG decoupling: Does religiosity matter?. Bus. Strateg. Environ.; 2023; 32, pp. 4046-4067. [DOI: https://dx.doi.org/10.1002/bse.3353]

5. García-Sánchez, I.-M.; Hussain, N.; Aibar-Guzmán, C.; Aibar-Guzmán, B. Assurance of corporate social responsibility reports: Does it reduce decoupling practices?. Bus. Ethics Environ. Responsib.; 2022; 31, pp. 118-138. [DOI: https://dx.doi.org/10.1111/beer.12394]

6. Liu, Y.; Li, W.; Meng, Q. Influence of distracted mutual fund investors on corporate ESG decoupling: Evidence from China. Sustain. Account. Manag. Policy J.; 2023; 14, pp. 184-215. [DOI: https://dx.doi.org/10.1108/SAMPJ-10-2021-0401]

7. Aboud, A.; Saleh, A.; Eliwa, Y. Does mandating ESG reporting reduce ESG decoupling? Evidence from the European Union’s Directive 2014/95. Bus. Strateg. Environ.; 2023; 33, pp. 1305-1320. [DOI: https://dx.doi.org/10.1002/bse.3543]

8. Di, R.; Li, C. The cost of hypocrisy: Does corporate ESG decoupling reduce labor investment efficiency?. Econ. Lett.; 2023; 232, 111355. [DOI: https://dx.doi.org/10.1016/j.econlet.2023.111355]

9. Lin, K.J.; Lu, X.; Zhang, J.; Zheng, Y. State-owned enterprises in China: A review of 40 years of research and practice. China J. Account. Res.; 2020; 13, pp. 31-55. [DOI: https://dx.doi.org/10.1016/j.cjar.2019.12.001]

10. Bei, J. Study on the “high-quality development” economics. China Political Econ.; 2018; 1, pp. 163-180. [DOI: https://dx.doi.org/10.1108/CPE-10-2018-016]

11. Huang, Q. How “New SOEs” Come of Age: Four Decades of China’s SOE Reform. China Econ.; 2018; 13, pp. 58-83. [DOI: https://dx.doi.org/10.19602/j.chinaeconomist.2018.01.03]

12. Lokuwaduge, C.S.D.; Heenetigala, K. Integrating Environmental, Social and Governance (ESG) Disclosure for a Sustainable Development: An Australian Study. Bus. Strateg. Environ.; 2017; 26, pp. 438-450. [DOI: https://dx.doi.org/10.1002/bse.1927]

13. Rahman, H.U.; Zahid, M.; Al-Faryan, M.A.S. ESG and firm performance: The rarely explored moderation of sustainability strategy and top management commitment. J. Clean. Prod.; 2023; 404, 136859. [DOI: https://dx.doi.org/10.1016/j.jclepro.2023.136859]

14. Zhu, T.; Tu, Y. Specialized Functional Positioning and Investment Herd Behavior of Competitive State-owned Enterprises: A Quasi-natural Experiment Based on the Classification Reform of State-owned Enterprises. J. Financ. Econ.; 2023; 49, pp. 20-34. (In Chinese) [DOI: https://dx.doi.org/10.16538/j.cnki.jfe.20220519.202]

15. Han, S.; Pan, Y.; Mygrant, M.; Li, M. Differentiated environmental regulations and corporate environmental responsibility: The moderating role of institutional environment. J. Clean. Prod.; 2021; 313, 127870. [DOI: https://dx.doi.org/10.1016/j.jclepro.2021.127870]

16. Li, B.; Xu, L.; McIver, R.P.; Liu, X.; Pan, A. Mixed-Ownership Reform and Private Firms’ Corporate Social Responsibility Practices: Evidence from China. Bus. Soc.; 2022; 61, pp. 389-418. [DOI: https://dx.doi.org/10.1177/0007650320958762]

17. Huang, Q. Classified Reform of State-owned Enterprises. Econ. Res. J.; 2022; 57, pp. 4-12. (In Chinese)

18. Ramanadham, V. The Economics of Public Enterprise; Routledge: London, UK, 1991.

19. Lin, J.Y.; Tan, G. Policy burdens, accountability, and the soft budget constraint. Am. Econ. Rev.; 1999; 89, pp. 426-431. [DOI: https://dx.doi.org/10.1257/aer.89.2.426]

20. Shleifer, A. State versus Private Ownership. J. Econ. Perspect.; 1998; 12, pp. 133-150. [DOI: https://dx.doi.org/10.1257/jep.12.4.133]

21. Holmstrom, B.; Milgrom, P. Multitask Principal–Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design. J. Law Econ. Organ.; 1991; 7, pp. 24-52. [DOI: https://dx.doi.org/10.1093/jleo/7.special_issue.24]

22. Ocasio, W. Towards an attention-based view of the firm. Strateg. Manag. J.; 1997; 18, pp. 187-206. [DOI: https://dx.doi.org/10.1002/(SICI)1097-0266(199707)18:1+<187::AID-SMJ936>3.0.CO;2-K]

23. Scharfstein, D.S.; Stein, J.C. Herd behavior and investment. Am. Econ. Rev.; 1990; 80, pp. 465-479.

24. Dong, X.-Y.; Putterman, L. Soft budget constraints, social burdens, and labor redundancy in China’s state industry. J. Comp. Econ.; 2003; 31, pp. 110-133. [DOI: https://dx.doi.org/10.1016/S0147-5967(02)00012-4]

25. Chang, E.C.; Wong, S.M.L. Governance with multiple objectives: Evidence from top executive turnover in China. J. Corp. Financ.; 2009; 15, pp. 230-244. [DOI: https://dx.doi.org/10.1016/j.jcorpfin.2008.10.003]

26. Zhang, X.; Yu, M.; Chen, G. Does mixed-ownership reform improve SOEs’ innovation? Evidence from state ownership. China Econ. Rev.; 2020; 61, 101450. [DOI: https://dx.doi.org/10.1016/j.chieco.2020.101450]

27. Lo, D.; Gao, L.; Lin, Y. State ownership and innovations: Lessons from the mixed-ownership reforms of China’s listed companies. Struct. Change Econ. Dyn.; 2022; 60, pp. 302-314. [DOI: https://dx.doi.org/10.1016/j.strueco.2021.12.002]

28. Qian, Y. Enterprise reform in China: Agency problems and political control. Econ. Transit.; 1996; 4, pp. 427-447. [DOI: https://dx.doi.org/10.1111/j.1468-0351.1996.tb00181.x]

29. Siqueira, K.; Sandler, T.; Cauley, J. Common agency and state-owned enterprise reform. China Econ. Rev.; 2009; 20, pp. 208-217. [DOI: https://dx.doi.org/10.1016/j.chieco.2009.02.001]

30. Liu, X.; Shen, J.H.; Deng, K. Endowment Structure, property rights and reforms of large state-owned enterprises (SOEs) in China: Past, present and future. Struct. Change Econ. Dyn.; 2022; 62, pp. 675-692. [DOI: https://dx.doi.org/10.1016/j.strueco.2022.05.006]

31. Huang, X.; Bao, Y.; Yu, X.; Cai, B.; Wang, Q. Exploration of the long-term performance-oriented incentive mechanism for balancing environmental and economic policies. J. Clean. Prod.; 2023; 429, 138870. [DOI: https://dx.doi.org/10.1016/j.jclepro.2023.138870]

32. Sheldon, O. The Philosophy of Management; Pitman: New York, NY, USA, 1924.

33. Freeman, R.E. Strategic Management: A Stakeholder Approach; Pitman: Boston, MA, USA, 1984.

34. Carroll, A.B. The pyramid of corporate social responsibility: Toward the moral management of organizational stakeholders. Bus. Horiz.; 1991; 34, pp. 39-48. [DOI: https://dx.doi.org/10.1016/0007-6813(91)90005-G]

35. Elkington, J. Cannibals with Forks: The Triple Bottom Line of 21st Century Business; New Society Pub.: Gabriola Island, BC, Canada, 1997.

36. Li, T.-T.; Wang, K.; Sueyoshi, T.; Wang, D.D. ESG: Research Progress and Future Prospects. Sustainability; 2021; 13, 11663. [DOI: https://dx.doi.org/10.3390/su132111663]

37. Truant, E.; Borlatto, E.; Crocco, E.; Sahore, N. Environmental, social and governance issues in supply chains. A systematic review for strategic performance. J. Clean. Prod.; 2024; 434, 140024. [DOI: https://dx.doi.org/10.1016/j.jclepro.2023.140024]

38. Yuan, X.; Li, Z.; Xu, J.; Shang, L. ESG disclosure and corporate financial irregularities–Evidence from Chinese listed firms. J. Clean. Prod.; 2022; 332, 129992. [DOI: https://dx.doi.org/10.1016/j.jclepro.2021.129992]

39. Cheng, B.; Ioannou, I.; Serafeim, G. Corporate social responsibility and access to finance. Strateg. Manag. J.; 2014; 35, pp. 1-23. [DOI: https://dx.doi.org/10.1002/smj.2131]

40. Weber, O. Environmental, Social and Governance Reporting in China. Bus. Strateg. Environ.; 2014; 23, pp. 303-317. [DOI: https://dx.doi.org/10.1002/bse.1785]

41. Odriozola, M.D.; Baraibar-Diez, E. Is Corporate Reputation Associated with Quality of CSR Reporting? Evidence from Spain. Corp. Soc. Responsib. Environ. Manag.; 2017; 24, pp. 121-132. [DOI: https://dx.doi.org/10.1002/csr.1399]

42. Tomar, S. Greenhouse Gas Disclosure and Emissions Benchmarking. J. Account. Res.; 2023; 61, pp. 451-492. [DOI: https://dx.doi.org/10.1111/1475-679X.12473]

43. Brammer, S.J.; Pavelin, S. Corporate reputation and social performance: The importance of fit. J. Manag. Stud.; 2006; 43, pp. 435-455. [DOI: https://dx.doi.org/10.1111/j.1467-6486.2006.00597.x]

44. Borralho, J.M.; Hernandez-Linares, R.; Gallardo-Vazquez, D.; Paiva, I.C.d.S. Environmental, social and governance disclosure’s impacts on earnings management: Family versus non-family firms. J. Clean. Prod.; 2022; 379, 134603. [DOI: https://dx.doi.org/10.1016/j.jclepro.2022.134603]

45. Zhang, Q.; Ding, R.; Chen, D.; Zhang, X. The effects of mandatory ESG disclosure on price discovery efficiency around the world. Int. Rev. Financ. Anal.; 2023; 89, 102811. [DOI: https://dx.doi.org/10.1016/j.irfa.2023.102811]

46. Baldini, M.; Dal Maso, L.; Liberatore, G.; Mazzi, F.; Terzani, S. Role of Country- and Firm-Level Determinants in Environmental, Social, and Governance Disclosure. J. Bus. Ethics; 2018; 150, pp. 79-98. [DOI: https://dx.doi.org/10.1007/s10551-016-3139-1]

47. Kizys, R.; Tzouvanas, P.; Donadelli, M. From COVID-19 herd immunity to investor herding in international stock markets: The role of government and regulatory restrictions. Int. Rev. Financ. Anal.; 2021; 74, 101663. [DOI: https://dx.doi.org/10.1016/j.irfa.2021.101663] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38620612]

48. Friede, G.; Busch, T.; Bassen, A. ESG and financial performance: Aggregated evidence from more than 2000 empirical studies. J. Sustain. Financ. Investig.; 2015; 5, pp. 210-233. [DOI: https://dx.doi.org/10.1080/20430795.2015.1118917]

49. Lee, K.-H.; Cin, B.C.; Lee, E.Y. Environmental Responsibility and Firm Performance: The Application of an Environmental, Social and Governance Model. Bus. Strateg. Environ.; 2016; 25, pp. 40-53. [DOI: https://dx.doi.org/10.1002/bse.1855]

50. Nollet, J.; Filis, G.; Mitrokostas, E. Corporate social responsibility and financial performance: A non-linear and disaggregated approach. Econ. Model.; 2016; 52, pp. 400-407. [DOI: https://dx.doi.org/10.1016/j.econmod.2015.09.019]

51. Verheyden, T.; Eccles, R.G.; Feiner, A. ESG for all? The impact of ESG screening on return, risk, and diversification. J. Appl. Corp. Financ.; 2016; 28, pp. 47-55. [DOI: https://dx.doi.org/10.1111/jacf.12174]

52. Miyai, H.; Sugiura, Y. Environmental, Social, and Governance Investment and Material Disclosure by Companies. SSRN; 2018; [DOI: https://dx.doi.org/10.2139/ssrn.3109050]

53. Wang, S.; Esperanca, J.P. Can digital transformation improve market and ESG performance? Evidence from Chinese SMEs. J. Clean. Prod.; 2023; 419, 137980. [DOI: https://dx.doi.org/10.1016/j.jclepro.2023.137980]

54. Alam, M.J.; Ahmed, K.S.; Hossen, B.; Mozammel, H.; Hoque, A. Storage pests of maize and their status in Bangladesh. J. Biosci. Agric. Res.; 2019; 20, pp. 1724-1730. [DOI: https://dx.doi.org/10.18801/jbar.200219.210]

55. Xu, M.; Yu, N.; Kim, J. Superhydrophobic drag reduction in high-speed towing tank. J. Fluid Mech.; 2021; 908, A6. [DOI: https://dx.doi.org/10.1017/jfm.2020.872]

56. Gu, Y.; Ho, K.-C.; Yan, C.; Gozgor, G. Public environmental concern, CEO turnover, and green investment: Evidence from a quasi-natural experiment in China. Energy Econ.; 2021; 100, 105379. [DOI: https://dx.doi.org/10.1016/j.eneco.2021.105379]

57. Ferreira-Quilice, T.; Hernández-Maestro, R.M.; Duarte, R.G. Corporate sustainability transitions: Are there differences between what companies say and do and what ESG ratings say companies do?. J. Clean. Prod.; 2023; 414, 137520. [DOI: https://dx.doi.org/10.1016/j.jclepro.2023.137520]

58. Meyer, J.W.; Rowan, B. Institutionalized organizations: Formal structure as myth and ceremony. Am. J. Sociol.; 1977; 83, pp. 340-363. [DOI: https://dx.doi.org/10.1086/226550]

59. Aguinis, H.; Glavas, A. What We Know and Don’t Know About Corporate Social Responsibility: A Review and Research Agenda. J. Manag.; 2012; 38, pp. 932-968. [DOI: https://dx.doi.org/10.1177/0149206311436079]

60. Khanchel, I.; Lassoued, N.; Gargouri, R. Have corporate social responsibility strategies mattered during the pandemic: Symbolic CSR versus substantive CSR. Corp. Soc. Responsib. Environ. Manag.; 2023; 31, pp. 1380-1398. [DOI: https://dx.doi.org/10.1002/csr.2632]

61. Stigler, G.J.; Friedland, C. The literature of economics: The case of Berle and Means. J. Law Econ.; 1983; 26, pp. 237-268. [DOI: https://dx.doi.org/10.1086/467032]

62. Jensen, M.; Meckling, W.H. Theory of the firm: Managerial behavior, agency costs and ownership structure. J. Financ. Econ.; 1976; 3, pp. 305-360. [DOI: https://dx.doi.org/10.1016/0304-405X(76)90026-X]

63. Zhang, M.; Gao, S.; Guan, X.; Jiang, F. Controlling Shareholder-Manager Collusion and Tunneling: Evidence from China. Corp. Gov. Int. Rev.; 2014; 22, pp. 440-459. [DOI: https://dx.doi.org/10.1111/corg.12081]

64. Wang, K.; Xiao, X. Controlling shareholders’ tunneling and executive compensation: Evidence from China. J. Account. Public Policy; 2011; 30, pp. 89-100. [DOI: https://dx.doi.org/10.1016/j.jaccpubpol.2010.09.014]

65. Yu, E.P.-Y.; Guo, C.Q.; Luu, B.V. Environmental, social and governance transparency and firm value. Bus. Strateg. Environ.; 2018; 27, pp. 987-1004. [DOI: https://dx.doi.org/10.1002/bse.2047]

66. Wu, S.; Li, X.; Du, X.; Li, Z. The Impact of ESG Performance on Firm Value: The Moderating Role of Ownership Structure. Sustainability; 2022; 14, 14507. [DOI: https://dx.doi.org/10.3390/su142114507]

67. Wei, M.; Cai, G.; Liu, J. Research on Classified Governance of Chinese State Owned Listed Companies. J. Sun Yat-sen Univ.; 2017; 57, pp. 175-192. (In Chinese) [DOI: https://dx.doi.org/10.13471/j.cnki.jsysusse.2017.04.019]

68. Xu, W.; Zhang, R.; Liu, Y.; Liu, P. Classification Governance, Controling Party Governance Mechanism and Innovation Bonus Release: A Study Based on the Analysis of State-controled Listed companies. Nankai Bus. Rev.; 2018; 21, pp. 90-102. (In Chinese) [DOI: https://dx.doi.org/10.3969/j.issn.1008-3448.2018.03.010]

69. Arouri, M.; El Ghoul, S.; Gomes, M. Greenwashing and product market competition. Financ. Res. Lett.; 2021; 42, 101927. [DOI: https://dx.doi.org/10.1016/j.frl.2021.101927]

70. Gull, A.A.; Hussain, N.; Khan, S.A.; Mushtaq, R.; Orij, R. The power of the CEO and environmental decoupling. Bus. Strateg. Environ.; 2023; 32, pp. 3951-3964. [DOI: https://dx.doi.org/10.1002/bse.3347]

71. Yang, L.; Ye, M.; Wang, H.; Lu, W. Female power, ownership and ESG decoupling: Evidence from China. Int. J. Gend. Entrep.; 2024; 16, pp. 341-366. [DOI: https://dx.doi.org/10.1108/IJGE-12-2023-0303]

72. Chen, W.-R. Determinants of firms’ backward-and forward-looking R&D search behavior. Organ. Sci.; 2008; 19, pp. 609-622. [DOI: https://dx.doi.org/10.1287/orsc.1070.0320]

73. Fan, J.P.; Wong, T.J.; Zhang, T. Politically connected CEOs, corporate governance, and Post-IPO performance of China’s newly partially privatized firms. J. Financ. Econ.; 2007; 84, pp. 330-357. [DOI: https://dx.doi.org/10.1016/j.jfineco.2006.03.008]

74. Cyert, R.M.; March, J.G. A Behavioral Theory of the Firm; Prentice Hall/Pearson Education: Upper Saddle River, NJ, USA, 1963; 322.

75. Argote, L.; Greve, H.R. A behavioral theory of the firm—40 years and counting: Introduction and impact. Organ. Sci.; 2007; 18, pp. 337-349. [DOI: https://dx.doi.org/10.1287/orsc.1070.0280]

76. Lant, T.K. Aspiration level adaptation: An empirical exploration. Manag. Sci.; 1992; 38, pp. 623-644. [DOI: https://dx.doi.org/10.1287/mnsc.38.5.623]

77. Kotiloglu, S.; Chen, Y.; Lechler, T. Organizational responses to performance feedback: A meta-analytic review. Strateg. Organ.; 2021; 19, pp. 285-311. [DOI: https://dx.doi.org/10.1177/1476127019883361]

78. Bromiley, P. Testing a causal model of corporate risk taking and performance. Acad. Manag. J.; 1991; 34, pp. 37-59. [DOI: https://dx.doi.org/10.2307/256301]

79. Manski, C.F. Economic analysis of social interactions. J. Econ. Perspect.; 2000; 14, pp. 115-136. [DOI: https://dx.doi.org/10.1257/jep.14.3.115]

80. He, W.; Wang, Q. The peer effect of corporate financial decisions around split share structure reform in China. Rev. Financ. Econ.; 2020; 38, pp. 474-493. [DOI: https://dx.doi.org/10.1002/rfe.1088]

81. Adel Tantawy, A.; Amankwah-Amoah, J.; Puthusserry, P. Political ties in emerging markets: A systematic review and research agenda. Int. Mark. Rev.; 2023; 40, pp. 1344-1378. [DOI: https://dx.doi.org/10.1108/IMR-09-2022-0197]

82. Wang, T.; Zhang, T.; Shou, Z. The double-edged sword effect of political ties on performance in emerging markets: The mediation of innovation capability and legitimacy. Asia Pac. J. Manag.; 2021; 38, pp. 1003-1030. [DOI: https://dx.doi.org/10.1007/s10490-019-09686-w]

83. Shen, L.; Zhou, K.Z.; Wang, K.; Zhang, C. Do political ties facilitate operational efficiency? A contingent political embeddedness perspective. J. Oper. Manag.; 2023; 69, pp. 159-184. [DOI: https://dx.doi.org/10.1002/joom.1203]

84. Kwak, J.; Chang, S.-Y.; Jin, M. The effects of political ties on innovation performance in China: Differences between central and local governments. Asian Bus. Manag.; 2023; 22, pp. 300-329. [DOI: https://dx.doi.org/10.1057/s41291-021-00167-x]

85. Ji, J.; Yi, Y. Value variations of political ties in the acquisition of R&D capabilities: The roles of non-SOEs and dysfunctional competition. Ind. Mark. Manag.; 2022; 101, pp. 1-11. [DOI: https://dx.doi.org/10.1016/j.indmarman.2021.11.012]

86. Liao, Z.; Weng, C.; Long, S.; Xiao, Z. Do social ties foster firms’ environmental innovation? The moderating effect of resource bricolage. Technol. Anal. Strateg. Manag.; 2021; 33, pp. 476-490. [DOI: https://dx.doi.org/10.1080/09537325.2020.1821876]

87. Yan, Z.; Peng, X.; Lee, S.; Fang, P. Chasing the light or chasing the dark? top managers’ political ties and corporate proactive environmental strategy. Technol. Anal. Strateg. Manag.; 2023; 35, pp. 1341-1354. [DOI: https://dx.doi.org/10.1080/09537325.2021.2003774]

88. Lin, L.-W. A Network Anatomy of Chinese State-Owned Enterprises. World Trade Rev.; 2017; 16, pp. 583-600. [DOI: https://dx.doi.org/10.1017/S1474745617000210]

89. Niskanen, W.A. Bureaucracy and Representative Government; Aldine Atherton: Chicago, IL, USA, 1971.

90. Lou, X.; Qian, A.; Zhang, C. Do CEO’s political promotion incentives influence the value of cash holdings: Evidence from state-owned enterprises in China. Pac. Basin Financ. J.; 2021; 68, 101617. [DOI: https://dx.doi.org/10.1016/j.pacfin.2021.101617]

91. Raimondo, C. The Media and the Financial Markets: A Review. Asia-Pac. J. Financ. Stud.; 2019; 48, pp. 155-184. [DOI: https://dx.doi.org/10.1111/ajfs.12250]

92. Ahern, K.R.; Sosyura, D. Who Writes the News? Corporate Press Releases during Merger Negotiations. J. Financ.; 2014; 69, pp. 241-291. [DOI: https://dx.doi.org/10.1111/jofi.12109]

93. Ahern, K.R.; Sosyura, D. Rumor Has It: Sensationalism in Financial Media. Rev. Financ. Stud.; 2015; 28, pp. 2050-2093. [DOI: https://dx.doi.org/10.1093/rfs/hhv006]

94. Dyck, A.; Zingales, L.; Islam, R. The Right to Tell: The Role of Mass Media in Economic Development; World Bank Publications: Chicago, IL, USA, 2002; pp. 107-135.

95. Boulland, R.; Degeorge, F.; Ginglinger, E. News Dissemination and Investor Attention. Rev. Financ.; 2016; 21, pp. 761-791. [DOI: https://dx.doi.org/10.1093/rof/rfw018]

96. Miller, G.S. The Press as a Watchdog for Accounting Fraud. J. Account. Res.; 2006; 44, pp. 1001-1033. [DOI: https://dx.doi.org/10.1111/j.1475-679X.2006.00224.x]

97. Dyck, A.; Volchkova, N.; Zingales, L. The Corporate Governance Role of the Media: Evidence from Russia. J. Financ.; 2008; 63, pp. 1093-1135. [DOI: https://dx.doi.org/10.1111/j.1540-6261.2008.01353.x]

98. Bednar, M.K. Watchdog or lapdog? A behavioral view of the media as a corporate governance mechanism. Acad. Manag. J.; 2012; 55, pp. 131-150. [DOI: https://dx.doi.org/10.5465/amj.2009.0862]

99. Cui, W.; Chen, X.; Xia, W.; Hu, Y. Influence of Media Attention on the Quality of Environmental, Social, and Governance Information Disclosure in Enterprises: An Adjustment Effect Based on the Shareholder Relationship Network. Sustainability; 2023; 15, 13919. [DOI: https://dx.doi.org/10.3390/su151813919]

100. He, F.; Guo, X.; Yue, P. Media coverage and corporate ESG performance: Evidence from China. Int. Rev. Financ. Anal.; 2024; 91, 103003. [DOI: https://dx.doi.org/10.1016/j.irfa.2023.103003]

101. Ayres, I.; Braithwaite, J. Responsive Regulation: Transcending the Deregulation Debate; Oxford University Press: New York, NY, USA, 1992.

102. Holmström, B. Managerial incentive problems: A dynamic perspective. Rev. Econ. Stud.; 1999; 66, pp. 169-182. [DOI: https://dx.doi.org/10.1111/1467-937X.00083]

103. Masulis, R.W.; Reza, S.W. Agency Problems of Corporate Philanthropy. Rev. Financ. Stud.; 2015; 28, pp. 592-636. [DOI: https://dx.doi.org/10.1093/rfs/hhu082]

104. Hu, Y.; Chen, L. A review of internet-based communication research in China. Online Media Glob. Commun.; 2022; 1, pp. 124-163. [DOI: https://dx.doi.org/10.1515/omgc-2022-0009]

105. Hope, O.-K.; Li, Y.; Liu, Q.; Wu, H. Newspaper Censorship in China: Evidence from Tunneling Scandals. Manag. Sci.; 2021; 67, pp. 7142-7166. [DOI: https://dx.doi.org/10.1287/mnsc.2020.3804]

106. Delmas, M.A.; Burbano, V.C. The Drivers of Greenwashing. Calif. Manag. Rev.; 2011; 54, pp. 64-87. [DOI: https://dx.doi.org/10.1525/cmr.2011.54.1.64]

107. Liu, Y.; Li, W.; Wang, L.; Meng, Q. Why greenwashing occurs and what happens afterwards? A systematic literature review and future research agenda. Environ. Sci. Pollut. Res.; 2023; 30, pp. 118102-118116. [DOI: https://dx.doi.org/10.1007/s11356-023-30571-z]

108. Huang, Y.; Francoeur, C.; Brammer, S. What drives and curbs brownwashing?. Bus. Strateg. Environ.; 2022; 31, pp. 2518-2532. [DOI: https://dx.doi.org/10.1002/bse.3041]

109. Diwan, H.; Amarayil Sreeraman, B. From financial reporting to ESG reporting: A bibliometric analysis of the evolution in corporate sustainability disclosures. Environ. Dev. Sustain.; 2023; 26, pp. 13769-13805. [DOI: https://dx.doi.org/10.1007/s10668-023-03249-2]

110. Sun, Y.; Xu, C.; Li, H.; Cao, Y. What drives the innovation in corporate social responsibility (CSR) disclosures? An integrated reporting perspective from China. J. Innov. Knowl.; 2022; 7, 100267. [DOI: https://dx.doi.org/10.1016/j.jik.2022.100267]

111. Helfaya, A.; Morris, R.; Aboud, A. Investigating the Factors That Determine the ESG Disclosure Practices in Europe. Sustainability; 2023; 15, 5508. [DOI: https://dx.doi.org/10.3390/su15065508]

112. Luo, L.; Tang, Q. The real effects of ESG reporting and GRI standards on carbon mitigation: International evidence. Bus. Strateg. Environ.; 2023; 32, pp. 2985-3000. [DOI: https://dx.doi.org/10.1002/bse.3281]