Impact of the Top Management Teams’ Environmental Attention on Dual Green Innovation in Chinese Enterprises: The Context of Government Environmental Regulation and Absorptive Capacity

Abstract

Green innovation (GI) is a key measure for enterprises to realize green transformation and sustainable development. Top management teams’ environmental attention (TMTEA) plays a critical role in shaping organizational strategic direction, value orientation, management mode, and behavioral patterns, serving as a micro-foundation for GI. Based on exploring the relationship between TMTEA and GI, this study adopts the ambidexterity theory to categorize dual green innovation (Dual_GI) into breakthrough green innovation (BGI) and progressive green innovation (PGI), and examines the impact of TMTEA on Dual_GI from the perspectives of external government environmental regulation (GER) and internal absorptive capacity (AC). Drawing on the attention-based view (ABV), this study uses data samples of Chinese A-share listed companies from 2010 to 2022 and establishes a fixed-effect model to empirically test this relationship. The results show the following: (1) TMTEA has a positive impact on corporate Dual_GI, and the promotion effect on PGI is more significant. (2) Both GER and AC can positively moderate the impact of TMTEA on Dual_GI, and both have a stronger moderating effect on TMTEA on PGI. (3) Further analysis shows that this driving effect is more obvious in state-owned enterprises, non-heavy polluting enterprises and enterprise maturity, and TMTEA can also drive Dual_GI to improve sustainable development performance. This study deepens the research scope and boundary conditions of TMT’s micro-psychological cognition and GI. It provides new insights for managers in emerging economies to rebalance their companies’ economic benefits and environmental transformation.

1. Introduction

Over the past few decades, rapid development has led to severe environmental pollution, posing a threat to human survival and safety. Green development is now an imperative trend driving global sustainable economic growth [1]. China is the world’s largest carbon emitter and a rapidly growing economy. It has committed to the “dual-carbon” goals and actively promotes ecological civilization. China employs measures such as industrial restructuring, the deployment of clean technology, and stringent environmental policies, demonstrating its role as a responsible major player in global environmental governance [2]. Enterprises play a key role in the green transition. They now face growing scrutiny from governments, investors, and the public regarding their environmental strategies and practices [3]. Traditional production modes, characterized by high energy use and pollution, no longer meet the requirements of sustainable development. Firms must integrate ecological responsibility with economic performance and pursue transformation through green innovation [4].

Green innovation (GI) serves as a critical pathway for firms to achieve sustainable development [5]. It enhances efficiency and reduces emissions through the use of green technologies, thereby helping firms achieve both economic and environmental benefits [6]. Based on organizational ambidexterity theory, GI can be categorized into breakthrough green innovation (BGI) and progressive green innovation (PGI) [7]. BGI emphasizes disruptive technologies that break existing path dependencies and comprehensively transform a firm’s technologies, products, and services [8], whereas PGI focuses on improving and upgrading existing green technologies and products to facilitate the efficient translation of innovation outcomes into green productivity [9]. However, GI generally entails high Investments, long cycles, and significant uncertainties. It also exhibits positive externalities through knowledge spillovers and negative externalities from environmental governance, which collectively dampen firms’ motivation to innovate and exacerbate resource constraints [10]. Consequently, a single innovation mode can hardly strike a balance between short-term performance and long-term development. Firms thus need to pursue dual green innovation (Dual_GI) to orchestrate both BGI and PGI activities, thereby sustaining their competitive advantage in the market.

In this context, Dual_GI has become a critical strategy for firms to integrate internal and external resources and achieve green transformation. It refers to the synergistic advancement of both BGI and PGI under the guidance of sustainable development principles, aiming to optimize or reconfigure production modes and achieve comprehensive greening across technological systems, product portfolios, service frameworks, and management paradigms [11]. While existing studies have predominantly emphasized macro-level drivers such as environmental regulations [12], market competition [13], and stakeholder pressure [14], the pivotal role of managers as micro-level decision-makers within China’s institutional context has often been overlooked. As the core unit of strategic formulation and implementation, the top management team (TMT) exerts a profound influence on innovation decisions and type selection through its characteristics and cognitive structure [15]. Although prior research has examined the effects of TMT demographic attributes (e.g., age, gender, tenure) [16], personality traits (e.g., overconfidence, perfectionism) [17], or leadership styles (e.g., green transformational leadership) [18] on GI, studies exploring the relationship between executives’ psychocognitive factors and GI remain scarce. According to the Attention-Based View (ABV) [19,20], attention serves as a key reflection of cognition; when executives focus their attention on relevant issues, their cognitive processes deepen, thereby shaping subsequent decisions. TMT environmental attention (TMTEA), defined as the extent of TMTs’ focus on environmental issues and their solutions, including environmental concerns, impact assessment, and the formulation and implementation of environmental strategies [21], constitutes an essential cognitive mechanism that reveals how cognitively driven agency within TMTs coordinates GI practices through resource allocation [22,23]. Therefore, investigating the influence of TMTEA on Dual_GI in Chinese enterprises holds substantial theoretical and practical significance.

Existing research on TMTEA and GI has mainly focused on two aspects. First, the relationship between TMTEA and corporate green behavior and performance transformation, as exemplified by studies such as Zor (2023) [3], Wang and Liu (2024) [21], and Liu and Cao (2025) [24]. These studies, however, often conceptualize GI as a monolithic construct without disentangling its subtypes, thereby overlooking the differential impact of TMT characteristics on various categories of GI and lacking systematic comparative analysis. Different types of GI impose distinct resource and capability requirements, which may also lead to varied demands on the TMT. Second, another strand of literature examines TMTEA and the choice of corporate environmental strategy. Although these studies recognize differences in innovation patterns to some extent, most adopt a content-based classification, such as green product innovation and green process innovation [8]. Such categorization often results in dimensional overlap and impairs discriminative validity in measurement [8], thus failing to fully reveal how cognitive differences within TMTs translate into divergent investment strategies. In the context of China’s economic transition, TMTEA stems from executives’ ongoing perception of macro policies, industry competition, and changes in societal attitudes, combined with their internal green knowledge structure, thereby forming contextually embedded green cognition. Driven by different innovation motivations, cognitive differences within the TMT may emerge, leading to a comprehensive trade-off between opportunities and risks [25] and resulting in two distinct attention allocation strategies, namely “attention focus” or “attention distraction” [19,26], which ultimately shape differentiated investment decisions. Dual_GI imposes distinct demands on the TMTs’ ability to perceive and respond to environmental issues, making it particularly relevant to the institutional and resource constraints faced by firms in emerging economies, such as China. Nevertheless, the theoretical mechanisms underlying TMTs’ inclination toward promoting either BGI or PGI remain underexplored. Therefore, this study aims to systematically elucidate the influence mechanisms between TMTEA and Dual_GI to address this research gap.

Furthermore, the ABV emphasizes that executive attention allocation is influenced by situated attention [19]. Although this theory posits that the conversion of environmental attention into innovation behavior depends on the joint effects of internal and external conditions, existing studies have predominantly examined single perspectives, such as external regulation or internal resources. They fail to systematically reveal the differential impact of attention configuration on resource allocation and capability utilization under varying contextual conditions [27]. To address this gap, this study integrates external government environmental regulation (GER) and internal absorptive capacity (AC) into a research framework to investigate how they jointly moderate the effect of TMTEA on Dual_GI and to elucidate the underlying differentiated pathways. On the one hand, prior research indicates that GER influences corporate decision-making through dual mechanisms of pressure and incentive effects [28], prompting TMTs to weigh innovation risks and returns and thereby affecting their innovation mode choices. The question remains as to how GER guides TMT to adjust investment strategies under conditions of imbalanced pressure and incentives. On the other hand, AC relies on the organization’s existing knowledge base [29] and provides critical resource support for Dual_GI, helping firms overcome organizational inertia and mitigate innovation risks. However, when internal resources are scarce or capabilities are weak, whether AC can still effectively facilitate green innovation or even exert counterintuitive moderating effects remains underexplored. By integrating internal and external situational factors, this study aims to address empirical gaps in the mechanism that translates attention into innovation, particularly revealing how GER and AC collectively shape unique pathways for corporate differentiated green investment strategies within China’s institutional context.

In summary, although existing research has yielded substantial insights, three main limitations remain. First, studies exploring the relationship between executives’ psychocognitive factors and GI are still scarce. Second, there is a lack of research that reveals the influence mechanisms of TMTs’ cognitive differences on different categories of GI and their corresponding differentiated investment strategies. Third, the differential effects of attention allocation on resource distribution and capability utilization under varying contextual conditions have not been systematically identified. To address these gaps, this study utilized a sample of A-share listed companies in Shanghai and Shenzhen Stock Exchanges of China from 2010 to 2022. Grounded in the ABV and organizational ambidexterity theory, we employed multiple regression analysis to examine the impact mechanism of TMTEA on Dual_GI. Furthermore, GER and AC were introduced as moderators to clarify the boundary conditions under which TMTEA influences Dual_GI. This study addresses three key questions: (1) Can Dual_GI be effectively achieved when the TMT allocates attention to environmental issues? (2) Does TMTEA exert a differentiated impact on Dual_GI? (3) What roles do external GER and internal AC play in moderating the relationship between TMTEA and Dual_GI?

This study makes four main contributions. First, by examining the influence of TMTEA on Dual_GI, it reveals the micro-level cognitive mechanisms through which firms promote GI practices, thereby expanding the application scope of the ABV in environmental management research. Second, building on organizational ambidexterity theory, it distinguishes between BGI and PGI, elucidating the heterogeneous effects of TMTEA on these two categories and addressing a prior research gap regarding how TMTs’ cognitive differences shape GI types and investment strategies. Third, by introducing external GER and internal AC as moderating variables, it clarifies how attention allocation differentially influences resource distribution and innovation capability under varying contextual conditions, expanding the boundary conditions and contextual mechanisms of TMTEA-driven Dual_GI. Fourth, it systematically proposes concrete pathways for Chinese firms to implement differentiated environmental management and resource allocation strategies, identifies core risks during innovation implementation, and suggests corresponding mitigation mechanisms, thereby offering a theoretical and practical framework for the implementation of green innovation management in emerging economies.

The rest of this study is structured as follows: Section 2 elaborates the theoretical basis and research hypotheses, and constructs the theoretical model. Section 3 introduces the data and methods. Section 4 presents the empirical results. Section 5 elaborates on the heterogeneity analysis and economic consequence analysis. In Section 6, this study is summarized.

2. Theoretical Analysis and Research Hypotheses

2.1. Top Management Teams’ Environmental Attention and Dual Green Innovation

As the maker and executor of enterprise strategy, TMTs’ attention allocation determines the direction of enterprise’s future development [19,30]. According to the ABV, TMTEA refers to the degree of emphasis placed by the TMT on environmental issues and their associated solutions [31], which reflects the values, thinking patterns, and ideologies of TMT. The translation of TMTEA into Dual_GI is a process that encompasses perception, attention, interpretation, and action (as illustrated in Figure 1). In a complex and dynamic business environment, severe environmental challenges drive the TMT to continuously monitor changes in macro-policies, industry competition, and social perceptions [8,11]. Confronted with information complexity and attention scarcity, the TMT selectively screens and integrates information related to environmental innovation, allocating limited attention to the firm’s long-term sustainable development [32,33]. Through in-depth interpretation of environmental information, the TMT enhances its ability to identify opportunities and risks, which facilitates the systematic evaluation of the firm’s current and future benefits [34], thereby formulating appropriate green innovation strategies. Thus, as the translational outcome of TMTEA-driven decisions, the relationship between TMTEA and Dual_GI can be explained from at least two levels: internal and external to the organization.

First, TMTEA enhances an organization’s adaptability to the external environment, enabling firms to seize opportunities, mitigate risks, and optimize resource allocation and performance output for Dual_GI. On one hand, TMTEA improves the TMTs’ ability to identify green opportunities amid policy, market, and technological changes [11]. By proactively investing in and advancing green innovation (e.g., developing new energy technologies or upgrading environmental processes), firms send external signals of their environmental commitments. Not only does it help build a responsible green image, gain consumer recognition, and secure support from environmentally oriented investors, but it also enables firms to acquire additional external resources and capital [35]. These resources provide a trial-and-error space for high-risk exploratory innovation, strengthen the iterative capacity of exploitative innovation, and ultimately transform environmental inputs into competitive advantages and economic returns, forming a virtuous cycle. On the other hand, amid increasingly stringent environmental regulations and social scrutiny, TMTEA enables the TMT to anticipate and address legitimacy pressures, thereby mitigating risks [23]. Through green innovation, firms proactively meet or even exceed environmental standards to obtain environmental legitimacy, preventing penalties, lawsuits, or reputational damage caused by non-compliance [36]. Furthermore, the early deployment of green technologies avoids costly emergency technological transformations or equipment upgrades when regulations tighten, resulting in long-term cost savings and providing stable support for continuous innovation. Thus, TMTEA can be regarded as an organization’s strategic response to the dynamism of the external environment.

Second, TMTEA facilitates the reshaping of an organization’s internal structure, providing support for Dual_GI through resource allocation and cultural development. On the one hand, TMTEA fosters a green culture and stimulates innovation momentum. TMT translates environmental attention into a clear green vision by demonstrating green transformational leadership and setting an example for others to follow. This action effectively enhances employees’ organizational identification, enabling them to internalize environmental goals and proactively contribute insights to innovation [37]. Meanwhile, TMTEA helps foster a green organizational culture that supports sustainability and tolerates trial-and-error [21], creating a favorable innovation atmosphere for Dual_GI. On the other hand, TMTEA guides resource allocation toward the green innovation system. TMT prioritizes the allocation of financial resources (e.g., R&D funds), human resources (e.g., researchers), and organizational capital (e.g., cross-departmental collaboration mechanisms) to green projects, solidifying the resource foundation required for innovation [24]. Additionally, by establishing a systematic green management system and internal control mechanisms, the consistency of innovation direction and the efficiency of resource utilization are ensured, supporting the coordinated development of Dual_GI.

Although TMTEA theoretically provides a cognitive premise for Dual_GI, firms tend to translate TMTEA into PGI rather than BGI under the practical constraints of China’s transitional economic context. As Likar et al. (2023) [38] noted, a “regional innovation paradox” exists in emerging economies, where higher R&D investment does not necessarily translate into greater returns in terms of scientific excellence and economic performance. This phenomenon arises because Chinese firms generally face the dual dilemma of weak internal innovation capabilities and external institutional pressures [39]. Driven by different innovation motivations, TMT develops cognitive differences that prompt it to thoroughly weigh opportunities and risks [25]. Consequently, TMT adopts two distinct attention allocation strategies, namely “attention focus” and “attention distraction” [19,26], which ultimately lead to variations in investment strategies.

First, driven by profit-seeking motivation, TMT pursues short-term economic performance and operational security, perceiving the green market as an opportunity rather than a threat [28]. Accordingly, TMT proactively adopts an “attention focus” strategy, directing its attention to knowledge domains that deliver rapid results and involve controllable risks. This leads to an obvious efficiency orientation in its investment strategy, whereby limited resources are prioritized for PGI, which improves existing processes and enhances resource efficiency. PGI is characterized by low risk and short cycles, making it easier to achieve sustainable development performance. In contrast, BGI features high technological barriers and substantial capital requirements, as it necessitates the integration of cross-domain technologies through distant knowledge search [40,41]. This easily causes the distraction of attention and resources, resulting in the marginalization of BGI in resource allocation.

Second, driven by responsibility motivation, TMT typically allocates attention to PGI projects that rapidly improve environmental performance (e.g., employee environmental training and partial optimization of production processes) in response to external ecological regulations and to shape a green image. Its investment strategy exhibits a distinct compliance orientation, focusing on marginal improvements to existing organizational technologies and routines while accumulating green achievements within established trajectories [42]. Admittedly, stable organizational culture and behavioral patterns facilitate the continuous implementation of innovation, helping firms quickly accumulate green reputation and social recognition [43]. However, these factors also create organizational inertia toward disruptive changes during resource allocation, making it difficult for high-risk experiments required for BGI to secure sufficient resource support.

Therefore, we propose the following:

H1. 

TMTEA can exert a significantly positive impact on corporate Dual_GI, and its promotional effect on PGI is significantly stronger than that on BGI.

2.2. The Moderating Effect of Government Environmental Regulation

The situational attention principle indicates the external environment influences TMTs’ attention preference. As highlighted earlier, against the backdrop of insufficient innovation motivation and scarce resources among Chinese firms, government support and guidance for enterprises become particularly critical. As a core source of formal environmental legitimacy, government environmental regulation (GER) is recognized as a pivotal external factor driving corporate ecological strategies. How, then, does GER guide TMTs to promote Dual_GI?

This study argues that GER facilitates the relationship between TMTEA and Dual_GI. Firstly, as an external pressure, GER can promote corporate environmental responsibility through regulation [44]. This pressure mitigates TMTs’ inertial constraints while enforcing sustained ecological vigilance. Given internal resource limitations, TMT initiates cross-boundary collaboration to access heterogeneous external resources. Integrating these with internal assets enables strategic resource allocation toward Dual_GI, thereby enhancing environmental legitimacy. Secondly, as an incentive measure, GER can promote TMT to increase environmental attention through subsidies [45]. Therefore, to obtain the support of the government and stakeholders, TMT must shift its resource allocation to Dual_GI. Thirdly, GER can promote public green participation and stimulate their green demand [28]. TMT must prioritize dual innovation capabilities to boost green output, thereby fulfilling the public’s demand for green products and maintaining its green reputation.

However, environmental policies are inherently characterized by significant uncertainty, and the difficulty in effectively balancing pressures with incentives may lead enterprises to adopt a conservative stance in green innovation decision-making [21]. On one hand, in recent years, the Chinese government has progressively intensified GER through a series of policy measures, including central environmental inspections and mandatory environmental information disclosure, thereby significantly raising the requirements for corporate environmental behavior. Ecological protection has become a crucial determinant of operational viability. In the face of mounting institutional compliance costs, TMTs that prioritize ecological issues are more likely to perceive such pressures and experience cognitive anxiety, which narrows their strategic focus toward addressing external environmental legitimacy challenges. Consequently, TMTs are compelled to allocate limited resources to PGI, which can more rapidly enhance both operational profits and environmental performance, thereby avoiding operational risks and compliance sanctions [42,46]. This behavioral mechanism aligns with the core tenets of threat rigidity theory [47].

On the other hand, the strengthening of GER means the expansion of incentive support policies [48]. TMT leverages such opportunities to enhance PGI with limited resources, rapidly establishing compliant green management systems to secure substantial subsidies. In contrast, BGI is characterized by long cycles and high difficulty [49,50]. Even if TMT intends to promote it, insufficient innovation capacity often makes it challenging to achieve short-term returns or alleviate legitimacy pressure. Therefore, facing resource constraints and profit considerations, TMT prioritizes PGI implementation due to loss aversion, thereby establishing stronger resource foundations for future BGI development [25].

Therefore, we propose the following:

H2. 

GER positively moderates the relationship between TMTEA and Dual_GI, with a stronger moderating effect on the relationship between TMTEA and PGI.

2.3. The Moderating Effect of Absorptive Capacity

Growing external pressures are compelling firms to increasingly prioritize the social and environmental impacts of their operations [51]. Addressing ecological challenges often demands extensive knowledge and significant internal operational changes [52], making Dual_GI contingent on the integration of external knowledge and internal capabilities. Absorptive capacity (AC), serving as a critical bridge between internal and external knowledge, refers to a firm’s ability to identify, assimilate, transform, and exploit knowledge. It comprises two dimensions: potential AC (identification and assimilation) and realized AC (transformation and exploitation) [53,54]. AC not only facilitates knowledge exploration and exploitation, but also acts as a key driver of Dual_GI [55]. By leveraging AC, firms establish a knowledge base that can be applied to green technology, product, and process innovations, thereby enhancing green performance [56]. Furthermore, AC functions as an organizational learning mechanism that strengthens the integration of heterogeneous knowledge, both internal and external, thereby promoting the achievement of Dual_GI [57,58].

This study argues that AC enhances the positive effect of TMTEA on Dual_GI. When TMTs direct attention to environmental issues, strong AC enables firms to conduct forward-looking external knowledge searches, cross organizational boundaries to acquire heterogeneous green resources, and reconstruct resource bases through internalization and absorption [59]. Meanwhile, higher AC strengthens TMTs’ effectiveness in integrating internal and external green knowledge, facilitating knowledge transfer and circulation within organizations, and driving improvements and innovations in green technologies [60]. For instance, firms can leverage AC to capture market knowledge from customers, suppliers, and competitors [54], or expand innovation pathways through industry-university-research collaborations [60]. Ultimately, such external knowledge is transformed into intellectual capital embedded in organizational structures [55]. This process contributes to the development of learning-oriented organizations and innovation cultures, enhances employees’ willingness to engage in green initiatives, and promotes the integration and utilization of knowledge [61].

From the perspective of the differentiated effect of TMTEA on Dual_GI, this raises the question of whether the moderating role of AC also exhibits heterogeneity. Theoretically, AC inherently emphasizes the acquisition, transformation, and exploitation of external knowledge, which aligns closely with the leapfrog searches and disruptive advancements required for BGI [41]. Thus, AC should be more conducive to promoting BGI. Yet, in the Chinese context, AC may instead strengthen the relationship between TMTEA and PGI.

To elaborate, potential AC (PAC) pertains to external knowledge exploration and technological disruption, while realized AC (RAC) focuses on the exploitation and utilization of existing knowledge [62]. PAC serves as the foundation for RAC, yet RAC often emerges as the practical key for firms to achieve innovation. As Crescenzi and Gagliardi (2018) [63] noted, even when firms have sufficient resources to establish internal knowledge bases, their PAC still differs substantially. Specifically, some firms demonstrate excellence in knowledge transformation and utilization but only moderate performance in knowledge acquisition and assimilation [64].

Internally, the structural shortage of human capital in Chinese firms (e.g., R&D talent) prevents them from fully absorbing emerging technologies during industry-university-research collaborations. Additionally, the risk of knowledge leakage exacerbates the loss of core critical technologies when communicating with external partners, making it difficult for firms to leverage PAC to acquire and transform more breakthrough green knowledge, primarily that dependent on distal searches and tacit knowledge. Externally, while AC can enhance organizational adaptability to the environment [54], intense institutional pressures increase operational costs, prompting firms to prioritize the application of RAC. Firms then rely on explicit knowledge bases and proximal search strategies to advance PGI in a pressure-response mode, thereby achieving short-term adaptation and survival goals.

Therefore, we propose the following:

H3. 

AC positively moderates the relationship between TMTEA and Dual_GI, with a stronger moderating effect on the relationship between TMTEA and PGI.

Thus, we construct a hypothesized theoretical model, as shown in Figure 2, which serves as a reference for the subsequent theoretical analysis and hypothesis testing.

3. Materials and Methods

3.1. Sample

A-share listed companies on the Shanghai and Shenzhen Stock Exchanges in China from 2010 to 2022 are selected as the research subjects. The samples were screens according to the following principles: first, due to the unique financial structure of the economic, insurance, and market services industry, and financial services industry, such samples were removed; second, abnormal enterprise samples such as ST and *ST were eliminated; third, any samples with missing variables were removed; and fourth, Winsorization was applied to all continuous variables. After data filtering, the final sample consists of 4094 companies and 31,357 practical observations. The unbalanced panel data were obtained after the data were processed.

The data sources are as follows: (1) TMTs’ environmental attention data were sourced from the Management Discussion and Analysis (MD&A) section of the listed company’s annual reports, extracted and analyzed using Python-based (version 3.11.1) keyword text analysis. (2) Data about dual green innovation, specifically green patents, were obtained from the China Research Data Service (CNRDS) database. Patent classification codes for inventions and utility models of listed companies were first acquired from CNRDS. Subsequently, these data were matched against the 2010 “International Patent Classification Green Inventory” published by the World Intellectual Property Organization (WIPO), a United Nations agency, to identify and categorize the patents into green invention patents and green utility model patents. (3) Government environmental regulation data were measured using information from the China Statistical Yearbook and other official statistical yearbooks. The level of GER was assessed for 30 provinces in China, excluding Tibet, Taiwan, Hong Kong, and Macau. (4) Data on R&D personnel, used to measure absorptive capacity, were sourced from the China Stock Market and Accounting Research (CSMAR) database. (5) Control variables, including financial indicators and corporate governance measures, were collected from both the CSMAR and Wind databases.

3.2. Measurement of Variables

3.2.1. Dependent Variable

Our dependent variable is Dual_GI, including BGI and PGI. Previous studies on Dual_GI measurement mainly used the following three methods: (1) questionnaire surveys, criticized for subjectivity in scale design and sample selection [8]; (2) expensed and capitalized R&D expenditures, which are susceptible to distortions from outsourcing, financial misreporting, and issues of data inaccuracy, incompleteness, and missing values; (3) green patents, which integrate environmental benefits with technological advancement. Moreover, patent counts offer a relatively objective measure less influenced by managerial discretion, thereby effectively capturing firms’ innovation capabilities and strategic decisions [65]. Compared to utility-type patents, invention-type patents are characterized by their focus on novel product and technology development, featuring breakthrough advancements that exhibit higher technological sophistication and innovation value [66]. Consequently, green patents are widely recognized as a valid and appropriate indicator for measuring Dual_GI. Therefore, we utilized green patents to measure Dual_GI, as described in Liao et al. (2020) [14]. Green invention-type and green utility-type patents were used to measure BGI and PGI, respectively. In the empirical study, we used the natural logarithm of the number of green patent applications plus 1.

3.2.2. Independent Variable

Our independent variable is TMTEA. Previous studies primarily employ questionnaire surveys and textual analysis to measure TMTEA. Given the limitations of questionnaire surveys, we opt for textual analysis to measure TMTEA, as recommended by Wang and Liu (2024) [21]. The MD&A sections of corporate annual reports serve as our primary textual source, with environmental keywords quantified by frequency. According to the ABV, lexical frequency effectively captures cognitive focus [67]. Empirical consensus confirms that annual reports serve as reliable proxies for managerial cognition, including attentional focus [68]. Liao et al. (2022) [20] establish that publicly available annual reports accurately map corporate environmental engagement and attention allocation. MD&A text constitutes a high-validity source for executive cognition, directly capturing TMTs’ strategic deliberations on organizational development and innovation trajectories [69].

The steps of text analysis are as follows: (1) Using python crawled the annual reports of Shanghai and Shenzhen A-share listed companies from 2010 to 2022, and further screened out the content of the MD&A section. (2) All tables within the MD&A sections were eliminated, as they predominantly contain numerical values with limited analytical utility for textual measurement. (3) Referring to Wang and Liu (2024) [21], we constructed and expanded the environmental attention word frequency table, as shown in

Table 1. TMTEA keywords.

Environmental Keywords

Safe Production, Protection, Exceedance, Ozone Layer, Dusting, Atmosphere, Low Carbon, Carbon Dioxide, Prevention and Control, Abandonment, Exhaust, Wastewater, Waste, Waste Residue, Dust, Wind Energy, Boiler, Boilers, Eco-Friendly, Environmental Protection, Recycling, Methane, Emission Reduction, Reduced Consumption, Reduced Noise, Energy Efficiency, Conservation, Purification, Sustainability, Renewable, Air, Waste, Wastage, Process Re-engineering, Green, Green, Energy Consumption, Energy, Emissions, Ventilation emissions, destruction, habitat, clean, fuel, waste, ecology, biomass, water treatment, acidic, solar, natural gas, soil, desulphurisation, denitrification, tailpipe, greenhouse gases, pollution, effluent, harmless, paperless, species, depletion, recycling, soot, smoke, flue gas, liquefied petroleum gas, poisonous, organics, waste heat, reuse, noise, heavy metals, natural resources

. (4) The “jieba” library of Python was used to count the frequency of environmental keywords in the MD&A. Furthermore, stopwords were purged to mitigate measurement interference from extraneous lexical content. Then, we used the natural logarithm of the sum of the frequency of environmental keywords plus 1 to measure TMTEA.

3.2.3. Moderating Variable

Our moderating variables are government environmental regulation (GER) and absorptive capacity (AC) selected from outside and inside the enterprise, respectively.

First, GER signifies a governmental dedication to pollution control, representing external forces that shape corporate environmental behavior. Scholarly consensus on measuring GER remains elusive, with standard metrics including regulatory outcomes (e.g., pollutant emissions) and regulatory intensity (e.g., corporate environmental investment). However, these approaches often overlook internal firm factors, such as firm size and managerial environmental awareness, which compromise the objective assessment of governmental regulatory impact. Addressing these limitations, this study employs provincial pollutant discharge fee intensity, measured by the ratio of total collected pollutant discharge fees to gross industrial output value, as a proxy for GER, following Shi et al. (2025) [70] and Wang et al. (2022) [71]. As a well-established flexible instrument in China’s environmental governance framework, pollutant discharge fees serve as a primary regulatory mechanism [71]. This approach internalizes corporate environmental costs into production expenses to address externalities, thereby stimulating technological innovation and improvement [70]. Its capacity to provide an objective measure independent of firm-specific characteristics constitutes a critical methodological advantage. Additionally, to mitigate inflated estimation coefficients, the GER intensity values are scaled by a factor of 100.

Second, as discussed, AC determines the effectiveness of converting external resources into internal propulsion. Previous studies have predominantly employed R&D intensity as a proxy variable for AC [72]. However, this metric primarily reflects an organization’s commitment level to R&D activities and innovation, rather than providing an objective representation of its internal capabilities. We contend that AC depends on the critical tacit resource of innovative human capital reserve [57]. This is because R&D personnel possess enhanced capabilities to identify, utilize, and assimilate external knowledge, aligning directly with the inherent nature of AC. Consequently, innovative human capital constitutes a more suitable indicator for reflecting objective AC. Therefore, this study used the proportion of R&D employees to total employees as a measure of AC.

3.2.4. Control Variable

Considering that Dual_GI may be affected by the basic characteristics of enterprises, corporate governance, and management characteristics, we selected the following control variables with reference to previous research: firm age, firm size, financial leverage, proportion of female executives, board size, proportion of independent board, and duality of chief executive officer and chairman [73,74]. In addition, due to the high-dimensional heterogeneity of property rights, pollution, region, individual, time, and industry, we also set dummy variables for these variables. The variable names and descriptions are shown in

Table 2. Variable names and descriptions.

Variable Name		Symbol	Variable Description	Data Sources
Breakthrough green innovation		BGI	The natural logarithm of the number of green invention-type patent applications plus 1	CNRDS
Progressive green innovation		PGI	The natural logarithm of the number of green utility-type patent applications plus 1
Dual green innovation		Dual_GI	The natural logarithm of the number of green patent applications plus 1
Top management team environmental attention		TMTEA	The natural logarithm of the sum of the frequency of environmental keywords plus 1	The annual report of listed companies
Government environmental regulation		GER	The proportion of sewage charges to the total industrial output value	CNRDS
Absorptive capacity		AC	The proportion of R&D employees in total	CSMAR
Controlled variable	Firm size	Size	Total assets of the business plus the natural log of 1
	Firm age	FirmAge	The natural logarithm of the difference between the current year and the listed year of the enterprise	CSMAR and Wind database
	Financial leverage	Lev	Total liabilities
	Female executives	Female	Total number of female executives / total number of top management team
	Board size	BOA	Total number of board members
	Independent board	DIR	Total number of independent directors/board of directors
	Duality of chief executive officer and chairman	Dual	Duality of the general manager and the chairman is assigned to 1, otherwise 0.
	Property right	SOE	State-owned enterprises are assigned to 1, otherwise 0.
	Pollution	Pollute	The heavily polluting enterprises supervised by the government are assigned to 1, otherwise 0.
	District	District	The eastern district is assigned to 1, otherwise 0.
	Individual	Firm	Firm fixed effect
	Time	Year	Year fixed effect
	Industry	Industry	Industry fixed effect

.

3.3. Model Design

3.3.1. Baseline Model

To test our research hypotheses concerning the influence of TMTEA on Dual_GI and the moderating effects of GER and AC, this study employs a panel data fixed effects model. This model selection is justified for the following reasons: (1) Panel data models provide a robust framework for causal inference by incorporating both cross-sectional and time-series dimensions. Controlling for fixed effects enhances the ability to discern the statistical relationships between independent and dependent variables. (2) Our dataset consists of panel data from listed companies, exhibiting substantial heterogeneity across firms, years, and industries. The fixed effects model effectively controls for these unobserved time-invariant factors, thereby improving estimation accuracy [3]. (3) It allows for the implementation of cluster-robust standard errors to address potential heteroskedasticity and autocorrelation. Consequently, to explicitly account for potential heteroskedasticity and within-firm autocorrelation, we estimate all specifications using firm-level cluster-robust standard errors within a panel model that includes three-way fixed effects for firm, year, and industry. To verify the hypothesis of this study, that is, the influence of TMTEA on Dual_GI, we constructed a multiple linear regression model as shown in Equations (1)–(3):

$${DualGI}_{it}={\alpha }_0+{\alpha }_1{TMTEA}_{it}+{\alpha }_K\sum {Controls}_{it}+\sum Firm+\sum Year+\sum Industry+{\epsilon }_{it}$$

(1)

$${BGI}_{it}={\alpha }_0+{\alpha }_1{TMTEA}_{it}+{\alpha }_K\sum {Controls}_{it}+\sum Firm+\sum Year+\sum Industry+{\epsilon }_{it}$$

(2)

$${PGI}_{it}={\alpha }_0+{\alpha }_1{TMTEA}_{it}+{\alpha }_K\sum {Controls}_{it}+\sum Firm+\sum Year+\sum Industry+{\epsilon }_{it}$$

(3)

where ${\mathrm{D}\mathrm{u}\mathrm{a}\mathrm{l}\_\mathrm{G}\mathrm{I}}_{\mathrm{i}\mathrm{t}}$ represents the number of green patent applications of company i in year $\mathrm{t}$, BGI_it represents the number of green invention-type patent applications of company $\mathrm{i}$ in year $\mathrm{t}$, PGI_it represents the number of green utility-type patent applications of company $\mathrm{i}$ in year $\mathrm{t}$, TMTEA_it represents the TMT environmental attention of company $\mathrm{i}$ in year $\mathrm{t}$, $\sum {\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}\mathrm{s}}_{\mathrm{i}\mathrm{t}}$ is the selected control variable, Firm, Year and Industry represent the fixed effect of the firm, year and industry, and ε_it represents the disturbance term.

3.3.2. Moderating Model

In order to verify the moderating effect of government environmental regulation and absorptive capacity, we also constructed a multiple linear regression model as shown in Equations (4) and (5):

$$Y_{it}={\beta }_0+{\beta }_1{TMTEA}_{it}+{\beta }_2{ER}_{it}+{\beta }_3{{TMTEA}_{it}\times GER}_{it}+{\beta }_K\sum {Controls}_{it}+\sum Firm+\sum Year+\sum Industry+{\epsilon }_{it}$$

(4)

$$Y_{it}={\beta }_0+{\beta }_1{TMTEA}_{it}+{\beta }_2{AC}_{it}+{\beta }_3{{TMTEA}_{it}\times AC}_{it}+{\beta }_K\sum {Controls}_{it}+\sum Firm+\sum Year+\sum Industry+{\epsilon }_{it}$$

(5)

where ${\mathrm{G}\mathrm{E}\mathrm{R}}_{\mathrm{i}\mathrm{t}}$ represents the proxy variable of government environmental regulation, ${\mathrm{A}\mathrm{C}}_{\mathrm{i}\mathrm{t}}$ represents the proxy variable of absorptive capacity.

3.4. Statistical Methods

The statistical methods employed in this study are as follows: (1) Descriptive statistics were calculated to assess sample distribution characteristics and evaluate compliance with model assumptions. (2) Variance Inflation Factors (VIF) were computed to detect potential multicollinearity issues. (3) Beyond estimating the baseline model using the fixed effects specification, we conducted four distinct robustness checks for the main effect: subsample regression, alternative regression model specifications, substitution of the dependent variable, and substitution of the core independent variable. Furthermore, to address endogeneity concerns, we implemented an instrumental variables (IV) approach to mitigate reverse causality and employed propensity score matching (PSM) to correct for potential sample selection bias. (4) To mitigate multicollinearity between interaction terms and their constituent variables, independent and moderating variables were mean-centered before constructing interaction terms for testing the moderation effects. (5) Heterogeneity analyses were performed across subgroups defined by property rights, pollution intensity, and firm lifecycle. The reliability of these subgroup comparisons was confirmed using the Chow test. Additionally, we further investigated the economic consequences arising from the influence of TMTEA on Dual_GI.

4. Results

4.1. Descriptive Statistics and Correlation Analysis

4.1.1. Descriptive Statistics

Table 3. Descriptive statistics of the main variables.

Variables	Obs	Mean	Std.dev.	Min	Median	Max	VIF
Dual_GI	31,357	0.806	1.129	0.00	0.00	4.98
BGI	31,357	0.526	0.914	0.00	0.00	4.49
PGI	31,357	0.543	0.888	0.00	0.00	4.04
TMTEA	31,357	1.975	0.656	0.00	1.95	3.37	1.25
AC	31,357	13.441	12.801	0.00	13.00	74.73	1.56
GER	31,357	0.058	0.050	0.00	0.05	0.27	1.03
FirmAge	31,357	2.923	0.324	1.10	2.94	3.64	1.08
Size	31,357	22.211	1.276	19.91	22.02	26.16	2.15
Lev	31,357	0.418	0.205	0.05	0.41	0.88	1.43
Female	31,357	19.532	11.320	0.00	18.18	54.55	1.12
BOA	31,357	2.119	0.197	1.61	2.20	2.71	1.70
DIR	31,357	37.638	5.352	25.00	36.36	60.00	1.51
Dual	31,357	0.290	0.454	0.00	0.00	1.00	1.14
SOE	31,357	0.348	0.476	0.00	0.00	1.00	1.39
District	31,357	0.721	0.449	0.00	1.00	1.00	1.09
Pollute	31,357	0.321	0.467	0.00	0.00	1.00	1.22

presents descriptive statistics for key variables. All variance inflation factors (VIF) are below 5, indicating no severe multicollinearity. Dual_GI exhibits a mean of 0.806 (max = 4.98; SD = 1.129), revealing substantial inter-firm variation and predominantly low levels of green innovation. BGI and PGI show means of 0.526 (SD = 0.914) and 0.543 (SD = 0.888), respectively, suggesting a significant divergence between innovation modes. TMTEA demonstrates a mean of 1.975 (max = 3.37; SD = 0.666), reflecting pronounced cross-firm heterogeneity in TMTEA, albeit predominantly at moderate-to-high levels.

The divergent characteristics of TMTEA and Dual_GI metrics suggest potential complexity in their relationship. Chinese firms currently encounter two primary internal constraints in GI: First, insufficient innovation motivation persists. Despite increasing environmental pressures, firms widely perceive environmental investments as competing with productive investments. At the same time, executives lack the strategic mindset to integrate sustainability, creating challenges in reconciling ecological and economic development. Second, inadequate innovation capacity exists. Scarcity of innovation resources and R&D capabilities, combined with high costs and technological barriers in Dual_GI (particularly BGI), leads to conservative approaches toward GI. Figure 3 demonstrates that although institutional improvements have driven upward trends in TMTEA and Dual_GI, Dual_GI levels remain low with substantial practical constraints. Significant variations in AC and GER across samples further support this observation. Other variables exhibit patterns consistent with established literature.

4.1.2. Correlation Analysis

Table 4. Correlation coefficients of the main variables.

Variables	Dual_GI	BGI	PGI	TMTEA	AC	GER	FirmAge	Size
Dual_GI	1.000
BGI	0.918 ***	1.000
PGI	0.900 ***	0.710 ***	1.000
TMTEA	0.220 ***	0.156 ***	0.250 ***	1.000
AC	0.132 ***	0.159 ***	0.056 ***	−0.125 ***	1.000
GER	0.033 ***	0.037 ***	0.031 ***	0.052 ***	0.037 ***	1.000
FirmAge	0.062 ***	0.055 ***	0.052 ***	0.071 ***	0.043 ***	−0.021 ***	1.000
Size	0.409 ***	0.390 ***	0.381 ***	0.127 ***	−0.122 ***	0.096 ***	0.183 ***	1.000
Lev	0.202 ***	0.177 ***	0.216 ***	0.104 ***	−0.170 ***	0.034 ***	0.159 ***	0.506 ***
Female	−0.115 ***	−0.102 ***	−0.114 ***	−0.138 ***	0.102 ***	−0.016 ***	0.055 ***	−0.187 ***
BOA	0.065 ***	0.062 ***	0.061 ***	0.073 ***	−0.126 ***	0.042 ***	0.039 ***	0.262 ***
DIR	0.014 ***	0.022 ***	0.010 ***	−0.045 ***	0.051 ***	−0.006 ***	−0.003 ***	−0.009 ***
Dual	−0.038 ***	−0.027 ***	−0.046 ***	−0.077 ***	0.113 ***	−0.062 ***	−0.100 ***	−0.189 ***
SOE	0.091 ***	0.098 ***	0.083 ***	0.109 ***	−0.152 ***	0.107 ***	0.167 ***	0.358 ***
District	0.032 ***	0.039 ***	0.021 ***	−0.082 ***	0.103 ***	−0.083 ***	−0.033 ***	−0.038 ***
Pollute	−0.101 ***	−0.117 ***	−0.077 ***	0.325 ***	−0.202 ***	0.073 ***	0.026 ***	0.051 ***
Variables	Lev	Female	BOA	DIR	Dual	SOE	District	Pollute
Lev	1.000
Female	−0.129 ***	1.000
BOA	0.155 ***	−0.174 ***	1.000
DIR	−0.018 ***	0.075 ***	−0.558 ***	1.000
Dual	−0.147 ***	0.134 ***	−0.184 ***	0.113 ***	1.000
SOE	0.295 ***	−0.233 ***	0.277 ***	−0.071 ***	−0.311 ***	1.000
District	−0.082 ***	0.108 ***	−0.099 ***	0.031 ***	0.101 ***	−0.181 ***	1.000
Pollute	−0.039 ***	−0.087 ***	0.095 ***	−0.050 ***	−0.043 ***	0.061 ***	−0.176 ***	1.000

Note: *** indicates statistical significance at the 1% level.

presents the Pearson correlation coefficients among the key variables. As shown, TMTEA exhibits positive correlations with Dual_GI, BGI, and PGI. Both GER and AC are significantly positively correlated with the three independent variables. The control variables also show significant effects to varying degrees. Overall, the correlation coefficients among the variables, excluding the three dependent variables (Dual_GI, BGI, and PGI), are relatively low, suggesting that multicollinearity is not a significant concern. Since multiple linear regression analyses are conducted for each dependent variable separately, meaning that the three are never simultaneously included in a single model, the validity of the regression results remains intact. Nevertheless, the relationships among these variables warrant further validation through multivariate regression techniques.

4.2. Results of Baseline Regression

Table 5. Results of baseline regression.

Variables	(1)	(2)	(3)
	Dual_GI	BGI	PGI
TMTEA	0.052 ***	0.029 ***	0.046 ***
	(0.012)	(0.010)	(0.010)
FirmAge	0.136	0.082	0.161
	(0.124)	(0.112)	(0.102)
Size	0.329 ***	0.243 ***	0.230 ***
	(0.019)	(0.017)	(0.016)
Lev	−0.090	−0.101 **	−0.048
	(0.058)	(0.047)	(0.050)
Female	−0.003 ***	−0.002 **	−0.003 ***
	(0.001)	(0.001)	(0.001)
BOA	−0.013	0.013	−0.041
	(0.066)	(0.055)	(0.057)
DIR	0.002	0.001	0.002
	(0.002)	(0.002)	(0.002)
Dual	0.004	0.029 **	−0.027 *
	(0.017)	(0.014)	(0.014)
SOE	0.057	0.052	0.022
	(0.038)	(0.034)	(0.031)
District	0.257	0.307 **	0.093
	(0.170)	(0.140)	(0.144)
Pollute	0.129	0.111	0.247
	(0.485)	(0.278)	(0.544)
Constant	−7.207 ***	−5.464 ***	−5.177 ***
	(0.588)	(0.510)	(0.491)
Firm	Yes	Yes	Yes
Year	Yes	Yes	Yes
Industry	Yes	Yes	Yes
N	31,357	31,357	31,357
adj. R2	0.705	0.677	0.658

Note: Clustering standard errors at company level in parentheses. *, **, *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

shows the Results of Baseline Regression. Columns (1) to (3) report the impact of TMTEA on Dual_GI, BGI, and PGI. It can be seen that the regression coefficients for columns (1) to (3) are significant at the 1% level, with values of 0.2514, 0.1479, and 0.2183, respectively. TMTEA increases by one standard deviation, and Dual_GI, BGI, and PGI increased by 20.9%, 18.4% and 26.4%, respectively, indicating that TMTEA can significantly promote Dual_GI, and the promotion effect of TMTEA on PGI is considerably more substantial than that of BGI. Therefore, H1 are verified.

Research substantiates that the influence on organizational behavior hinges on the TMT’s attention to the external environment, rather than on changes themselves [19]. Environmentally oriented TMTs exhibit higher sensitivity to the external environment, assisting firms in seizing opportunities, mitigating risks, and optimizing resource allocation and performance output for Dual_GI. Meanwhile, they reshape the internal organizational structure and provide support for Dual_GI through resource allocation and cultural development. Furthermore, the impact of TMTEA on Dual_GI presents a differentiated effect. Owing to the dual dilemma of weak internal innovation capabilities and external institutional pressures commonly faced by Chinese firms, TMTs develop cognitive differences driven by distinct innovation motivations, which in turn lead to preferences for different innovation models. The profit-seeking motivation drives TMTs to prioritize short-term interests, while the responsibility motivation compels them to enhance environmental legitimacy. Compared with PGI, BGI emphasizes disruptive technological innovation, featuring high investment, high risk, and long cycles. Due to organizational inertia, TMTs tend to focus their attention on specific resources and conduct marginal improvements to existing organizational technologies and routines to enhance PGI.

4.3. Robustness Test

4.3.1. Subsample Regression

In 2015, China promulgated a series of environmental protection policies, such as the “New Environmental Protection Law”, which led to increased environmental pressure on enterprises. In view of this, we selected the sub-samples from 2015 to 2022 for regression analysis, and the regression results are presented in

Table 6. Results of subsample regression and Tobit Regression.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Dual_GI	BGI	PGI	Dual_GI	BGI	PGI
Model	Subsample Regression			Tobit Regression
TMTEA	0.045 ***	0.022 *	0.044 ***	0.2241 ***	0.2067 ***	0.2593 ***
	(0.014)	(0.012)	(0.011)	(0.0197)	(0.0217)	(0.0203)
Constant	−7.554 ***	−5.545 ***	−5.909 ***	−16.6858 ***	−16.8602 ***	−14.5799 ***
	(0.803)	(0.716)	(0.683)	(0.4393)	(0.4715)	(0.4376)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Firm	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	24,446	24,446	24,446	31,357	31,357	31,357
adj. R2	0.739	0.713	0.700	-	-	-

Note: Clustering standard errors at company level in parentheses. *, *** indicate statistical significance at the 10% and 1% levels, respectively. The Tobit model uses the maximum likelihood method for parameter estimation, so there is no R².

. Column (1) to column (3) show that TMTEA is thus positively correlated with Dual_GI, BGI, and PGI (p < 0.01), consistent with the above result.

4.3.2. Replacing the Regression Model

The number of green patent applications has more zero values, and the left-truncated feature is evident. Therefore, we used the Tobit model to replace the multiple linear regression model. As shown in Table 6, from column (4) to (6) of Table 6, the impact of TMTEA is positive (p < 0.01), which indicates that TMTEA can still promote Dual_GI, BGI, and PGI of enterprises by Tobit regression.

4.3.3. Replacing the Dependent Variable and Core Independent Variable

To ensure robustness, we first replaced the dependent variable. We used green patent grants to replace green patent applications for measurement. The number of green utility-type patent grants was used as a proxy variable for PGI, and the number of green invention-type patent grants was used as a proxy variable for BGI. As shown in

Table 7. Results of replacing the dependent variable and core independent variable.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Dual_GI2	BGI2	PGI2	Dual_GI	BGI	PGI
TMTEA	0.057 ***	0.022 ***	0.050 ***
	(0.010)	(0.007)	(0.010)
GreenTMT				0.313 ***	0.202 ***	0.294 ***
				(0.085)	(0.067)	(0.082)
Constant	−5.931 ***	−2.678 ***	−5.337 ***	−7.187 ***	−5.457 ***	−5.162 ***
	(0.525)	(0.359)	(0.512)	(0.592)	(0.512)	(0.495)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Firm	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes

Note: Clustering standard errors at company level in parentheses. *** indicates statistical significance at the 1% level.

, columns (1) to (3) show that the regression results have not changed.

Second, we also replace the core independent variables. TMT with a green background has received green education and functional experience, which will affect their environmental awareness and cause them to pay more attention to environmental issues and sustainable development [75]. Therefore, we selected the proportion of executives with a green background to the total number of executives as the proxy variable for TMTEA. Columns (4) to (6) in Table 6 report the robustness test results that replace the core independent variable, wherein GreenTMT still positively affects GI, BGI, and PGI (p < 0.01).

4.3.4. Endogeneity Test

First, there may be a reverse causal relationship between TMTEA and Dual_GI. This study employed the mean value of TMTEA for other listed companies in the same industry in each year as the instrumental variable for TMTEA. It used 2SLS to address the endogenous problem. The mean value of TMTEA of other listed companies in the same industry in each year does not influence Dual_GI of the enterprise. It exhibits strong exogeneity and is closely related to the TMTEA of the enterprise, thus meeting the requirement of an instrumental variable. As shown in

Table 8. Results of instrumental variable method.

Variables	(1)	(2)	(3)	(4)
	First Stage	Second Stage
	TMTEA	Dual_GI	BGI	PGI
TMTEA_Mean	0.266 ***
	(0.037)
TMTEA		0.579 ***	0.297 *	0.534 ***
		(0.219)	(0.178)	(0.182)
Controls	Yes	Yes	Yes	Yes
Firm	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes
Kleibergen-Paap rk LM	51.854
Cragg-Donald Wald F	100.675
Kleibergen-Paap Wald F	53.096
10% maximal IV size	16.38
N	31,357	31,357	31,357	31,357
adj. R2	0.003	−0.062	−0.008	−0.095

Note: Clustering standard errors at the company level in parentheses. *, *** indicate statistical significance at the 10% and 1% levels, respectively.

, column (1) indicates that TMTEA_Mean is significantly correlated with TMTEA at the 1% level in the first stage. Columns (2) to (4) show that TMTEA in the second stage positively affects Dual_GI, BGI, and PGI at the 1% level, and the F value is greater than 16.38, which satisfies the weak instrumental variable. The regression results are still consistent.

Second, there may be sample self-selection bias between TMTEA and Dual_GI. This study used PSM for the test. Prior studies often categorized samples based on median or mean values, an approach that is straightforward but overly mechanistic and lacking solid theoretical justification. Converting continuous variables into binary categories in this manner results in a significant loss of variation information and fails to account for unobserved factors. Following the method of Ma et al. (2025) [76], we directly employ a dichotomous variable for group classification. Firms that have obtained ISO 14001 certification are designated as the treatment group the (ISO = 1), while those without certification form the control group. We include all control variables as covariates, estimate propensity scores using a logit model, and match each treated firm with the most comparable control firm. The adoption of ISO certification is motivated by its close association with environmental attributes and green innovation, as well as its relatively exogenous nature, which more closely approximates a randomized experimental setting. As shown in

Table 9. Results of the balance test.

Variables	Unmatched	Mean		Bias (%)	t-Test
	Matched	Treated	Control
FirmAge	U	2.942	2.917	7.7	5.57 ***
	M	2.942	2.940	0.5	0.29
Size	U	22.227	22.206	1.6	1.18
	M	22.227	22.235	−0.7	−0.39
Lev	U	0.397	0.424	−13.7	−9.78 ***
	M	0.397	0.399	−0.9	−0.55
Female	U	19.347	19.584	−2.1	−1.54
	M	19.347	19.266	0.7	0.43
BOA	U	2.114	2.120	−2.9	−2.08 **
	M	2.114	2.113	0.9	0.53
DIR	U	37.653	37.634	0.4	0.26
	M	37.653	37.746	−1.7	−1.01
Dual	U	0.319	0.28144	8.3	6.14 ***
	M	0.319	0.328	−1.8	−1.06
SOE	U	0.292	0.363	−15.2	−11.02 ***
	M	0.292	0.288	0.7	0.43
District	U	0.765	0.709	12.7	9.17 ***
	M	0.765	0.757	1.7	1.02
Pollute	U	0.359	0.311	10.2	7.56 ***
	M	0.359	0.362	−0.6	−0.37

Note: **, *** indicate statistical significance at the 5% and 1% levels, respectively.

, after matching, all covariates exhibit standardized bias below 5%, and t-tests indicate no statistically significant differences between the treatment and control groups, confirming that balance was achieved across all variables. Figure 4 demonstrates sufficient overlap in the propensity score distributions, with all treated observations lying within the common support. The average treatment effect on the treated (ATT) is positive and statistically significant at the 1% level. As shown in

Table 10. Results of PSM.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Dual_GI	BGI	PGI	Dual_GI	BGI	PGI	Dual_GI	BGI	PGI
Model	Nearest Neighbor Matching			Kernel Matching			Radius Matching
TMTEA	0.054 ***	0.033 *	0.046 **	0.052 ***	0.029 ***	0.046 ***	0.060 ***	0.034 ***	0.053 ***
	(0.020)	(0.017)	(0.018)	(0.012)	(0.010)	(0.010)	(0.012)	(0.010)	(0.010)
Constant	−8.217 ***	−5.987 ***	−6.490 ***	−7.230 ***	−5.482 ***	−5.194 ***	−7.122 ***	−5.408 ***	−5.057 ***
	(0.975)	(0.839)	(0.841)	(0.589)	(0.511)	(0.491)	(0.602)	(0.533)	(0.478)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Firm	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	10,384	10,384	10,384	31,332	31,332	31,332	31,332	31,332	31,332
adj. R2	0.726	0.703	0.672	0.705	0.677	0.658	0.704	0.676	0.657
ATT	0.022 ***	0.018 ***	0.017 ***	0.016 ***	0.014 ***	0.013 ***	0.016 ***	0.014 ***	0.013 ***

Note: Clustering standard errors at the company level in parentheses. *, **, *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

, columns (1) to (3) report the regression results of nearest neighbor matching. It can be seen that high-level TMTEA has a positive effect on Dual_GI, BGI, and PGI, which demonstrates the robustness of the results. In addition, similar results are obtained by using kernel matching and radius matching, which further support H1.

4.4. Results of Moderating Effect

4.4.1. Moderating Effect of Government Environmental Regulation

To avoid the collinearity problem, we centralize both the independent variable and the moderating variable to construct the interaction term.

Table 11. Results of the moderating effect.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Dual_GI	BGI	PGI	Dual_GI	BGI	PGI
TMTEA	0.052 ***	0.028 ***	0.045 ***	0.054 ***	0.030 ***	0.049 ***
	(0.012)	(0.010)	(0.010)	(0.012)	(0.010)	(0.010)
GER	0.321	0.405 **	0.326 *
	(0.220)	(0.201)	(0.192)
TMTEA_c*GER_c	0.607 ***	0.517 ***	0.600 ***
	(0.222)	(0.192)	(0.194)
AC				0.004 ***	0.004 ***	0.001 **
				(0.001)	(0.001)	(0.001)
TMTEA_c*AC_c				0.003 ***	0.002 **	0.004 ***
				(0.001)	(0.001)	(0.001)
Constant	−7.199 ***	−5.453 ***	−5.168 ***	−7.082 ***	−5.292 ***	−5.163 ***
	(0.588)	(0.509)	(0.490)	(0.594)	(0.515)	(0.493)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Firm	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	31,357	31,357	31,357	31,357	31,357	31,357

Note: Clustering standard errors at the company level in parentheses. *, **, *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

reports the results of the regulating effect test. Columns (1) to (3) report the moderating effect of GEA. The regression coefficient of TMTEA_c * GER_c in column (1) is significantly positive (β= 0.607, p < 0.01), indicating that GEA positively moderates the relationship between TMTEA and Dual_GI. The interaction term coefficients in column (2) and column (3) are 0.517 and 0.600 (p < 0.01), respectively. It can be seen that compared with BGI, GER has a more significant moderating effect on the relationship between TMTEA and PGI. H2 is verified. Figure 5 presents the moderating impact of GER. Under high GER conditions, the positive relationships between TMTEA and all three green innovation types (Dual_GI, BGI, PGI) demonstrate steeper slopes. Critically, GER exerts a more substantial moderating influence on the TMTEA to PGI relationship (K3 = 0.171) than on TMTEA to BGI (K2 = 0.137), providing empirical validation for hypothesis H2.

Research further indicates that TMTs develop adaptive capabilities in response to external contingencies, particularly policy shifts [77]. This adaptability enables GER to recalibrate TMTs’ environmental cognition through amplified salience of ecological imperatives. Environmentally oriented TMTs proactively restructure cognitive schemas by assimilating external information, strategically align operations with policy mandates, and allocate resources toward Dual_GI. GER simultaneously generates an opportunity-threat duality; regulatory stringency elevates perceived pressure through the internalization of non-compliance costs, whereas resource subsidies foster innovation willingness through the prioritization of environmental benefits. Compared to BGI, PGI exhibits reduced risk exposure and compressed return cycles. Consequently, under resource constraints and return imperatives, TMTs prioritize PGI investments to mitigate legitimacy threats and enhance environmental performance. Such strategic focus, however, may inadvertently divert resources from BGI initiatives, potentially limiting radical innovation pathways.

4.4.2. Moderating Effect of Absorptive Capacity

Columns (4) to (6) in Table 11 report the regulating effect of AC. The regression coefficient of TMTEA_c * AC_c in column (4) is significantly positive (β = 0.003, p < 0.01), indicating that AC positively moderates the relationship between TMTEA and Dual_GI, and H3a is verified. The coefficients of TMTEA_c * AC_c in column (2) and column (3) are 0.002 (p < 0.05) and 0.004 (p < 0.01), respectively. Compared with BGI, AC has a more significant moderating effect on the relationship between TMTEA and PGI; thus, H3b is not confirmed. Figure 6 presents the moderating effects of AC. Under high AC conditions, the positive relationships between TMTEA and all three green innovation types (Dual_GI, BGI, PGI) demonstrate steeper slopes. However, AC exerts a more substantial moderating influence on the TMTEA to PGI relationship (K6 = 0.286) than on TMTEA to BGI (K5 = 0.125); only H3a has been confirmed.

The reason is that the regulating effect of AC on TMTEA and Dual_GI is rooted in the synergistic adaptation of the knowledge base and search strategy. When the AC is stronger, TMT will continue to pay attention to environmental problems and more systematically extract and digest green knowledge, thus promoting Dual_GI. However, Chinese firms confront distinctive constraints: structural human capital shortages and knowledge leakage risks jointly impair PAC efficacy in acquiring and transforming radical green knowledge, especially via distal search and tacit knowledge integration; concurrently, stringent institutional pressures (e.g., compliance mandates) prioritize RAC, leveraging explicit knowledge reservoirs through proximal search for compliance-driven incremental adaptation. Consequently, AC augmentation predominantly intensifies RAC-driven exploitation efficiency along entrenched knowledge paths, elevating PGI. While AC holistically facilitates Dual_GI, PAC remains restricted by these barriers and institutionally driven short-termism, thus failing to enable the disruptive exploration essential for BGI sufficiently.

5. Further Analysis

5.1. Heterogeneity Analysis

5.1.1. Property Right Heterogeneity Analysis

The heterogeneity of enterprise property rights leads to a difference between business objectives and resource capture capabilities, thereby affecting the promotional effect of TMTEA on Dual_GI. Therefore, we divided the sample data into groups according to state-owned and non-state-owned enterprises (SOEs and NSOEs), and conducted a Chow test to assess the significance of coefficient differences between the two groups. The results were shown in

Table 12. Results of property right heterogeneity analysis.

Variables	State-Owned			Non-State-Owned
	(1)	(2)	(3)	(4)	(5)	(6)
	Dual_GI	BGI	PGI	Dual_GI	BGI	PGI
TMTEA	0.059 ***	0.042 ***	0.057 ***	0.050 ***	0.033 ***	0.038 ***
	(0.020)	(0.016)	(0.016)	(0.014)	(0.012)	(0.012)
Constant	−7.974 ***	−5.987 ***	−6.025 ***	−7.731 ***	−5.886 ***	−5.525 ***
	(1.122)	(1.020)	(0.916)	(0.661)	(0.559)	(0.548)
Firm	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	10,991	10,991	10,991	20,366	20,366	20,366
adj. R2	0.752	0.731	0.697	0.678	0.642	0.638
Chow test	13.37 ***	14.02 ***	11.33 ***	13.37 ***	14.02 ***	11.33 ***

Note: Clustering standard errors at the company level in parentheses. *** indicates statistical significance at the 1% level.

. The differences in coefficients between SOEs and NSOEs regarding the effects of TMTEA on Dual_GI, BGI, and PGI are all statistically significant at the 1% level, with p values of 13.37, 14.02, and 11.33, respectively. The results indicate that TMTEA has a significantly positive influence on both groups of firms, with a more pronounced effect on PGI. Notably, the regression coefficients for SOEs are significantly larger than those for NSOEs, implying that TMTEA in SOE exerts a stronger promoting effect on Dual_GI, resulting in higher levels of both PGI and BGI.

The reason is that SOEs possess a dual nature encompassing both social and market imperatives. Although their relatively rigid organizational structures may reinforce the tendency of TMTs to adopt a “security-oriented mindset when prosperous” by prioritizing PGI, SOEs are driven not only by economic objectives but also by the mandate to deliver social services. Owing to their stronger political connections, SOEs inherently benefit from superior resource endowments and fiscal subsidies. This inherent advantage positions them to achieve breakthroughs in green technology and maintain a high level of BGI. [78]. In addition, they exhibit heightened resilience against regulatory pressures and market volatility while demonstrating lower sensitivity to performance feedback. Consequently, SOEs leverage their inherent advantages to sustain TMT’s environmental commitment, utilize abundant capital for green project development, and implement eco-oriented decisions flexibly. These attributes facilitate their superior attainment of dual balance. In contrast, the heightened marketization compels NSOEs to prioritize profit maximization, rendering them more responsive to performance feedback. Their weaker political ties, resource constraints, and inferior risk resilience prompt a strategic emphasis on progressive innovation. This approach allows NSOEs to navigate institutional pressures and competitive markets at lower cost while securing broader stakeholder support.

5.1.2. Pollution Heterogeneity Analysis

The heterogeneity of enterprise pollution leads to different challenges in the green transformation of enterprises, thereby affecting the promotional effect of TMTEA on Dual_GI. Therefore, we categorize samples into heavy pollution and non-heavy pollution enterprises (HPEs and NHPEs) based on the classification standard defined by China’s Ministry of Environmental Protection in 2008, and the results are as shown in

Table 13. Results of pollution heterogeneity analysis.

Variables	Heavy Pollution			Non-Heavy Pollution
	(1)	(2)	(3)	(4)	(5)	(6)
	Dual_GI	BGI	PGI	Dual_GI	BGI	PGI
TMTEA	0.044 **	0.019	0.036 **	0.052 ***	0.037 ***	0.044 ***
	(0.019)	(0.015)	(0.017)	(0.014)	(0.012)	(0.012)
Constant	−6.072 ***	−4.401 ***	−4.591 ***	−7.461 ***	−5.604 ***	−5.280 ***
	(0.992)	(0.843)	(0.844)	(0.708)	(0.624)	(0.573)
Firm	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	10,125	10,125	10,125	21,232	21,232	21,232
adj. R2	0.623	0.593	0.562	0.731	0.642	0.638
Chow test	9.97 ***	16.84 ***	10.10***	9.97 ***	16.84 ***	10.10 ***

Note: Clustering standard errors at the company level in parentheses. **, *** indicate statistical significance at the 5% and 1% levels, respectively.

. The differences in coefficients between HPEs and NHPEs regarding the effects of TMTEA on Dual_GI, BGI, and PGI are all statistically significant at the 1% level, with p values of 9.97, 16.84, and 10.10, respectively. Overall, the coefficient of TMTEA is significantly larger in NHPEs than in HPEs, indicating a more pronounced promoting effect of TMTEA on Dual_GI, BGI, and PGI among NHPEs. In terms of differential effects, although TMTEA in both groups exerts a stronger influence on PGI, it exhibits a significant positive correlation with BGI at the 1% level in NHPEs, whereas its impact on BGI is statistically insignificant in HPEs.

The reason is that HPEs face stringent regulatory oversight and acute environmental legitimacy pressures, necessitating environmental remediation to address historical liabilities while balancing economic growth. However, prohibitive transition costs and path dependency often lead to symbolic innovations, such as compliance cost reduction or greenwashing. Conversely, NHPEs experience lower external pressure and environmental sensitivity, granting TMT greater latitude in environmental responses. These firms prioritize brand reputation and market competitiveness, enabling sustainability-focused TMT to secure green investments while addressing stakeholder demands. This reduces innovation costs and risks, facilitating breakthrough innovations and green competitive advantage. Thus, while TMTEA positively affects Dual_GI in both groups, its promoting effect is significantly stronger in NHPEs.

5.1.3. Enterprise Lifecycle Heterogeneity Analysis

Enterprises exhibit distinct resource bases and innovation capabilities across life cycles, shaping TMT risk preferences, innovation motivation, and strategic objectives. This heterogeneity ultimately generates strategic dynamism in Dual_GI. Therefore, referring to Yang and Deng (2023) [79], we used the cash flow method to divide the life cycle. The results are shown in

Table 14. Results of enterprise lifecycle heterogeneity analysis.

Variables	Growth Period			Mature Period			Recession Period
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	Dual_GI	BGI	PGI	Dual_GI	BGI	PGI	Dual_GI	BGI	PGI
TMTEA	0.036 **	0.012	0.036 ***	0.123 ***	0.119 ***	0.067 **	0.014	0.008	0.034
	(0.015)	(0.012)	(0.012)	(0.037)	(0.030)	(0.031)	(0.027)	(0.022)	(0.021)
Constant	−6.917 ***	−5.484 ***	−4.926 ***	−7.686 ***	−5.286 ***	−5.768 ***	−6.805 ***	−4.888 ***	−4.706 ***
	(0.715)	(0.618)	(0.616)	(1.365)	(1.190)	(1.184)	(1.280)	(1.069)	(1.035)
Firm	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	21,387	21,387	21,387	3807	3807	3807	6032	6032	6032
adj. R2	0.718	0.692	0.665	0.690	0.652	0.659	0.683	0.652	0.657
Chow test	2.79 ***	2.66 ***	2.80 ***	2.79 ***	2.66 ***	2.80 ***	2.79 ***	2.66 ***	2.80 ***

Note: Clustering standard errors at the company level in parentheses. **, *** indicate statistical significance at the 5% and 1% levels, respectively.

. The coefficient differences in the effects of TMTEA on Dual_GI, BGI, and PGI across the growth, maturity, and decline periods are all statistically significant at the 1% level, with test statistics of 2.79, 2.66, and 2.80, respectively. Overall, TMTEA during the growth and maturity period is significantly positively correlated with Dual_GI at the 5% and 1% levels, respectively. In terms of differential effects, TMTEA in growth-phase enterprises shows a significant influence only on PGI, whereas in mature-phase enterprises, it significantly affects both BGI and PGI at the 1% and 5% levels, respectively. This indicates that growth-stage firms exhibit a stronger inclination toward PGI, whereas mature-stage firms display a more balanced green dual capability, with a pronounced focus on BGI. Additionally, the TMTEA coefficient during the recession period is positive but not statistically significant.

The reason is that growth-period enterprises prioritize market expansion and survival. Heightened sensitivity to external regulations drives TMT to pursue GI for environmental legitimacy and market breakthroughs. However, severe resource constraints and limited innovation capacity suppress their engagement in large-scale, high-quality BGI. Inadequate resource foundations and risk management capabilities prevent them from undertaking BGI-associated risks. Despite recognizing the imperatives of green development, survival pressures compel TMT to allocate resources toward niche-oriented PGI for performance feedback.

Mature-period enterprises focus on sustaining competitive advantages. Abundant resources, financing capacity, and stable profitability empower them to pursue high-risk BGI. Established reputations reduce external financing costs, thereby diminishing incentives for PGI that targets subsidies. Accumulated R&D experience lowers innovation failure risks while enhancing TMTEA, thereby redirecting resources toward BGI. Consequently, TMTEA significantly strengthens Dual_GI sustainability in mature-period enterprises.

Recession-period enterprises seek new growth avenues to regain viability. They typically encounter operational difficulties, managerial rigidity, and a lack of innovation awareness. Shrinking market share intensifies cash flow pressures, depletes capital reserves, diminishes profitability, and increases financing barriers. Although symbolic green innovation may signal positive prospects to stakeholders, a deficient innovation capability impedes the translation of TMTEA into actionable strategies. TMT adopts conservative stances, prioritizing routine operations over green innovation for survival. Thus, TMTEA exerts no significant effect on Dual_GI in the recession period.

5.2. Economic Consequence Analysis

Sustainable development performance (SDP) is a comprehensive measure that assesses the economic, environmental, and social sustainability of an enterprise’s development goals, ensuring its sustainable competitive advantage. Previous research on TMTEA and Dual_GI on SDP has been mostly limited to direct effects or single dimensions, such as environmental and economic performance. The mechanism of action between the three needs requires further study. Therefore, we further analyze whether TMTEA can promote SDP through Dual_GI.

Elkinton (1998) [80] proposed the “Triple Bottom Line” theory of sustainable development, which asserts that corporate growth must integrate the three pillars of economic, environmental, and social performance. Building on this framework, we deconstruct SDP into environmental, economic, and social responsibility dimensions. Some scholars have measured SDP using the respective dimensions of ESG (Environmental, Social, and Governance) ratings [81]. Following this approach, environmental performance (EP) is primarily measured by the E-score from third-party ESG ratings [82], while social responsibility performance (SRP) is assessed using the S-score derived from ESG evaluations [83]. The ESG rating for this article is sourced from the China Securities Database. The return on assets (ROA) is used to measure economic performance [24]. SDP is calculated from EP, ROA, and SRP using the entropy weight method.

The results are shown in

Table 15. Results of economic consequence analysis.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	SDP	Dual_GI	SDP	BGI	SDP	PGI	SDP
TMTEA	0.0038 ***	0.0523 ***	0.0037 ***	0.0288 ***	0.0038 ***	0.0457 ***	0.0037 ***
	(0.001)	(0.012)	(0.001)	(0.010)	(0.001)	(0.010)	(0.001)
Dual_GI			0.0015 ***
			(0.000)
BGI					0.0027 ***
					(0.001)
PGI							0.0032 ***
							(0.001)
Constant	0.2228 ***	−7.2070 ***	0.2419 ***	−5.4642 ***	0.2245 ***	−5.1766 ***	0.2263 ***
	(0.037)	(0.588)	(0.037)	(0.510)	(0.038)	(0.491)	(0.038)
Firm	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	31,357	31,357	31,357	31,357	31,357	31,357	31,357
adj. R2	0.537	0.710	0.537	0.677	0.579	0.658	0.579

Note: Clustering standard errors at the company level in parentheses. *** indicate statistical significance at the 1% level, respectively.

. Columns (1) and (2) show that TMTEA positively affects Dual_GI and SDP. Column (3) shows that the positive effect of TMTEA on SDP is reduced under the control of Dual_GI. The Dual_GI plays a mediating role, which confirms the conclusion that TMTEA can promote the SDP through Dual_GI. Results from columns (4) to (7) further indicate that both BGI and PGI exhibit a significant mediating effect between TMTEA and SDP. Furthermore, the aforementioned mediating effects were examined using a non-percentile bootstrap method with 1000 resamples. As summarized in

Table 16. Mediation test results of Bootstrap.

Transmission Channel	Effect Size		Standard Error	Boot 95% CI	Relative Effect Size
TMTEA → Dual_GI → SDP	Total Effect	0.0218	0.00062	[0.0206,0.0230]
	Direct Effect	0.0186	0.00065	[0.0173,0.0199]
	Indirect Effect	0.0031	0.00017	[0.0028,0.0035]	14.60%
TMTEA → BGI → SDP	Total Effect	0.0218	0.00061	[0.0206,0.0230]
	Direct Effect	0.0197	0.00062	[0.0185,0.0209]
	Indirect Effect	0.0021	0.00012	[0.0018,0.0024]	9.61%
TMTEA → PGI → SDP	Total Effect	0.0218	0.00062	[0.0206,0.0230]
	Direct Effect	0.0190	0.00066	[0.0177,0.0202]
	Indirect Effect	0.0029	0.00019	[0.0025,0.0032]	13.06%

, all bootstrapped 95% confidence intervals excluded zero, and the relative effect sizes were 14.60%, 9.61%, and 13.06%, respectively. Overall, these results confirm the significant mediating roles of Dual_GI, BGI, and PGI. Notably, TMTEA appears to enhance SDP more substantially through PGI than through other pathways.

The results demonstrate that, amid increasingly severe environmental issues, TMT continuously perceives macro policies, industrial competition, and shifts in social perception, integrating these with its own green knowledge structure to form a distinctive green cognition. This cognitive process compels the TMT to focus its limited attention on enhancing the organization’s green ambidextrous capability to gain a competitive advantage, thereby successively improving SDP. Notably, the fact that TMTEA further promotes SDP through PGI not only corroborates the role of PGI in enabling organizations to swiftly adapt to environmental demands and achieve competitive advantage and value creation through marginal technological and managerial improvements but also reflects a pragmatic choice under the current constraints of insufficient innovation capability and motivation among Chinese enterprises. On the one hand, uncertainty in environmental policies—particularly the imbalance between punitive and incentive measures—has dampened corporate enthusiasm for innovation. On the other hand, internal innovation capability gaps and a shortage of innovation capital have heightened concerns around core knowledge leakage. Consequently, the TMTs’ contextually shaped green cognition further reinforces a stable innovation culture and organizational rigidity, leading most Chinese firms to rely on limited greening practices to sustain operations. Although PGI facilitates short-term value creation, only BGI can achieve disruptive transformation and sustain first-mover advantages in the long run. Thus, governmental macro-regulation is essential to maintain environmental stability, while enterprises should not only leverage PGI to establish a resource base but also consciously pursue BGI for technological breakthroughs. The dialectical synergy between BGI and PGI is crucial for achieving corporate sustainable development.

6. Conclusions and Implications

6.1. Results Discussion

Based on the ABV, this study uses the data of 1722 listed companies in China from 2010 to 2022 to not only test the relationship between TMTEA and Dual_GI but also reveal the moderating role of government environmental regulation and absorptive capacity. We obtain the following core conclusions:

(1)

TMTEA directly impacts Dual_GI. Research confirms that this impact does not stem from environmental changes themselves but rather depends on TMTs’ attention to the external environment [19], underscoring that environmental attention serves as a key driver of green innovation [3]. When TMT consistently focuses on ecological issues, it enables strategic responses to external changes, facilitates timely adjustments in strategic direction to seize opportunities and mitigate risks, and supports the restructuring of internal processes through resource allocation and cultural shaping, thereby providing necessary support for Dual_GI.

Notably, this facilitating effect is amplified for PGI. Although direct empirical evidence remains limited, existing studies suggest such a tendency [38]. Grounded in the context of an emerging economy such as China, this study highlights the dual challenges faced by Chinese firms: insufficient internal innovation capacity and external institutional pressures [39]. These practical constraints trigger innovation motivations, leading to cognitive divergence within the TMT. This leads to an attention allocation strategy characterized by resource optimization and risk control, which consequently favors investments in more secure and efficient PGI. This is because cognitive divergence is directly associated with preferences for innovation modes [25]. When processing information, TMT must strategically balance opportunities and threats to ensure effective decision-making. Different innovation motivations lead to distinct strategic trade-offs: profit-seeking motives drive TMT to prioritize short-term benefits, while responsibility motives compel improvements in environmental legitimacy [28]. As BGI emphasizes disruptive technological change with longer cycles and higher risks, TMT tends to focus more attention on PGI, which involves marginal improvements to existing technologies and organizational routines that can more promptly enhance firm performance.

(2)

GER strengthens the positive relationship between TMTEA and Dual_GI, and has a stronger moderating role on PGI. Extensive research suggests that effective environmental regulation fosters innovation through both coercive pressures and incentives, aligning with the Porter Hypothesis [84]. When policy institutions change, TMTs adapt promptly and make responsive decisions [77]. GER heightens the salience of environmental issues in TMT cognition. Consequently, TMTs proactively align with policies, redirecting firm resources towards Dual_GI.

However, uncertainty in environmental policies makes it challenging to balance pressures and incentives, leading TMTs of Chinese firms to remain relatively conservative in green innovation decisions [21]. In recent years, the continuous rollout of mandatory policies has significantly increased firms’ institutional costs, confining TMTs’ attention to addressing legitimacy threats over the long term. Concurrently, to capture opportunities embedded in policies, TMTs strive to quickly establish compliant green management systems, thereby securing more green subsidies. Under resource constraints and benefit trade-offs, TMTs, driven by loss aversion, have to allocate limited resources to PGI to achieve short-term benefits or mitigate legitimacy pressures. This practice helps firms consolidate their resource bases, creating conditions for the implementation of long-cycle BGI in the future.

(3)

AC positively moderates the relationship between TMTEA and Dual_GI, consistent with Pacheco et al. (2018) [85], indicating AC can positively moderate the link between organizational internal factors and Dual_GI. AC enables executives to better identify green resources during the knowledge reconnaissance and cognition phase, and facilitates TMTs in assimilating green knowledge during the internalization phase, ultimately converting it into outputs [57].

Notably, this study finds that AC can better moderate the relationship between TMTEA and PGI, which contrasts with previous studies. AC, serving as a critical bridge that integrates internal and external knowledge, has often been considered more conducive to BGI, which demands substantial heterogeneous resources [41]. In fact, AC consists of two dimensions: potential AC and realized AC, each focusing on the exploration and exploitation of external knowledge, respectively [62]. From this perspective, China’s unique resource and institutional context helps elucidate the observed discrepancy. On the one hand, facing structural deficiencies in specialized human capital and potential knowledge leakage risks, firms struggle to fully leverage their potential AC to acquire and transform green knowledge critical for BGI. On the other hand, under intense institutional environmental pressures, firms prioritize utilizing their realized AC, integrating existing explicit knowledge bases with local search strategies to rapidly meet short-term demands through a “pressure-response” mode of PGI. Thus, the role of AC is contingent upon the synergistic alignment between knowledge bases and search strategies within specific contexts.

(4)

Heterogeneity analysis reveals that property rights, pollution intensity, and firm lifecycle moderate the influence of TMTEA on Dual_GI. This positive effect is significantly more substantial in state-owned (SOEs), heavily polluting (HPEs), and mature-period enterprises. Economic consequences analysis further demonstrates that TMTEA enhances sustainable development performance (SDP) through Dual_GI. We propose four mechanisms. First, the prioritization of social services by SOEs, coupled with robust political ties, substantial resource endowments, and heightened risk resilience [78], empowers their TMTs to implement environmental decisions flexibly. Second, NHPEs experience attenuated external pressures and environmental scrutiny, affording TMTs greater strategic discretion in environmental responses. At the same time, their market-driven focus on reputation and competitiveness enables them to address environmental demands and secure green financing. Third, mature-period enterprises leverage abundant resources, superior access to funding, stable profitability, and risk-mitigating R&D experience [21], collectively bolstering the TMTs’ propensity for high-risk, disruptive innovation. Fourth, SDP emerges as an outcome of TMTEA and Dual_GI, wherein environmental awareness enables TMTs to assess externalities accurately, formulate context-appropriate environmental strategies, and mobilize resources for green ambidexterity, ultimately capturing market opportunities and securing enduring competitive advantages [11].

6.2. Implications

6.2.1. Theoretical Implications

The theoretical implications of this study are manifested in several aspects:

First, this study advances micro-level cognitive mechanisms in green innovation by identifying how TMTEA translates into Dual_GI through resource allocation. While existing research has predominantly focused on macro-level drivers such as environmental regulations and market competition, or examined the relationship between static TMT characteristics (e.g., demographics, personality traits) and GI, it has inadequately explored the dynamic role of agency-driven TMT cognitive differences in shaping GI practices. Grounded in the Attention-Based View, this research systematically elucidates the mechanism through which TMTEA affects Dual_GI from the dual perspectives of external environmental adaptation and internal organizational restructuring, uncovering the micro-process by which cognitive focus is transformed into strategic green outcomes. Thereby, it directly responds to and empirically validates the call by Wu et al. (2024) [10], Wang and Liu (2024) [21], and Liu and Cao (2025) [24] for further investigation into the micro-psychological cognitive mechanisms driving GI, providing an important theoretical foundation at the micro level.

Second, this study fills the research gap regarding how TMT cognitive differences shape distinct GI categories and their associated investment strategies. Existing research has primarily treated GI as a unitary construct, emphasizing TMTEA’s macro-level effects on corporate green behavior and performance while overlooking its multidimensional heterogeneity. Another strand of research, which examines the relationship between TMTEA and environmental strategy selection, relies on content-based taxonomies (e.g., green product vs. process innovation). However, these taxonomies suffer from dimensional overlap and an insufficient mechanistic explanation. To address these limitations and contextualize China’s institutional and resource constraints, this study applies ambidexterity theory to decompose Dual_GI into BGI and PGI, clarifying their distinct theoretical mechanisms and resource demands. It further reveals how TMT cognitive divergence under these conditions drives the development of differentiated investment strategies. Specifically, TMTs develop motivation-driven cognitive differences through environmental sensing, leading to strategic trade-offs: profit-seeking motives prioritize short-term gains, while responsibility motives focus on mitigating legitimacy threats. Given BGI’s high investment and risk profile, TMTs adopt an “attention focus” strategy, allocating scarce cognitive resources to PGI, leveraging specialized knowledge to enhance innovation efficiency. Thus, clarifying Dual_GI drivers and interactions enables firms to strategically implement environmental strategies tailored to contextual demands, fostering sustainable development in China.

Third, this study extends understanding of boundary conditions for TMT micro-cognition driving Dual_GI, empirically supporting the situational attention principle. While this principle posits that translating TMTEA into Dual_GI requires internal and external catalysts, research on differential resource allocation and capability effects during cross-context attentional shifts remains limited [27]. Integrating GER and AC, we demonstrate their joint amplification of TMTEA’s influence on Dual_GI. Crucially, GER steers utility-maximizing firms towards PGI through deterrence and subsidies, while firms constrained by China-specific institutional barriers (e.g., deficient human capital, knowledge leakage risks) exhibit a preference for PGI when utilizing AC. Our results bridge an empirical gap regarding the differential moderating roles of contextual factors in the attention-innovation nexus and delineate the mechanism whereby GER pressure and internal AC collectively foster a corporate PGI bias over BGI in China. These findings deepen our understanding of policy efficacy in guiding corporate environmental strategy and innovation selection, providing critical managerial implications for optimizing green innovation under prevailing GER and AC constraints.

Fourth, offer actionable guidance for innovation management practices. It delineates how firms in China can implement effective environmental management and adopt differentiated resource allocation strategies. The research clearly identifies the various innovation risks enterprises encounter and proposes concrete solutions to mitigate them. These insights directly respond to the scholarly calls by Javed et al. (2025) [55], Pacheco et al. (2018) [85], and Ritter-Hayashi et al. (2021) [86] to integrate emerging economies into mainstream theoretical frameworks. Ultimately, this study offers a valuable framework for firms in emerging economies to better adapt to dynamic institutional pressures and successfully implement green innovation practices.

6.2.2. Managerial Implications

Managerial implications for enterprises: First, enhance TMTEA. Given the scarcity of TMTEA, TMT must transcend rhetorical commitments to integrate proactive environmental concerns into strategic actions. Concrete measures include: (a) delivering customized training workshops during senior executive development to enhance green cognition; (b) regularly convening industry experts for cutting-edge practice seminars to strengthen environmental decision-making capabilities; (c) embedding environmental objectives into long-term strategies via quantifiable time-bound green innovation plans to secure resource allocation and implementation.

Second, focus on Dual_GI construction. Effective Dual_GI necessitates sustained strategic attention from TMT, adapting dynamically based on phased outcomes and environmental contingencies. Specifically, (a) TMTs should establish cross-departmental environmental information-sharing mechanisms to enhance Dual_GI cognition and mitigate cognitive inertia in decision-making systematically. (b) TMTs must precisely evaluate their operational status and resource endowments to allocate innovation investments scientifically; this involves dedicated funding, balanced resource allocation between BGI and PGI to avoid excessive pursuit of difficulty or quantity, and distinct strategies tailored to each innovation mode to enhance green ambidextrous capability through phased adaptation. (c) Crucially, under resource constraints, strategically leveraging the competitive value of corporate social responsibility strengthens stakeholder synergy to secure innovation resources.

Third, proactively formulating environmental strategies that respond dynamically while operating beyond compliance with government policies. GER provides support for green innovation while simultaneously imposing significant legitimacy pressure and public scrutiny. Specifically, (a) TMT must internalize national green directives, enhance environmental literacy, and formulate Dual_GI strategies aligned with resource endowment. (b) Given policy uncertainty, firms must establish policy monitoring and rapid response mechanisms to maintain strategic flexibility against potential shocks. (c) Tailored environmental disclosure systems should address industry-institutional-firm contexts. Superior environmental performance yields legitimacy relief, reputational benefits, and green investment opportunities. Therefore, firms should transcend short-term PGI through long-term operational transformations that concurrently address institutional pressures. This requires proactive TMT implementing systematic BGI for sustainable advantage.

Fourth, improve AC. AC is a key variable affecting Dual_GI, enabling firms to integrate knowledge. However, Chinese firms that enhance AC under resource constraints and environmental pressures often encounter structural bottlenecks, fears of knowledge leakage, and stress, leading to a bias towards PGI. Overcoming this requires building structured, risk-managed AC: (a) Strengthen internal knowledge management (e.g., green repositories, cross-departmental communication, training, environmental scanning teams) with phased small-scale piloting to reduce risk; (b) Expand institutionally guaranteed collaborative networks (e.g., government-endorsed industry-university-research partnerships) with dedicated IP protection and risk controls; (c) Implement differentiated resource allocation and incentives. TMTs should guide AC to optimize PGI while allocating resources to agile units that explore the recombination of assimilated and peripheral knowledge in controlled settings to foster BGI potential, enabling synergistic co-evolution of Dual_GI.

Fifth, heterogeneity-driven proactive adaptation and active restructuring dynamically configure eco-innovative transformations. The impact of TMTEA on Dual_GI varies across enterprise types. (a) From ownership, SOEs embed eco-imperatives strategically via institutional advantages, launch TMT-led breakthrough mechanisms, while supporting PGI systematically; NSOEs link eco-actions to profits via incentives, prioritize cost-effective PGI, and bridge resource gaps externally. (b) From pollution intensity, HPEs pursue substantive transformation through transparent disclosure and green tech initiatives, converting regulations into systemic innovation sans greenwashing; NHPEs leverage TMTEA for sustainable value chains, integrating green investments with BGI. (c) From the lifecycle, growth firms drive PGI via legitimacy pressures; mature firms convert slack resources into BGI engines; recession firms activate TMTEA only when aligned with strategic restructuring.

Managerial implications for policymakers: The government should pay closer attention to environmental issues and actively guide enterprises to undertake environmental innovation. First, implement precisely targeted policy incentives. Precision-targeted policy incentives must optimize green subsidies, credit, and tax instruments to alleviate specific resource constraints and talent shortages, thereby enhancing corporate green capabilities and core talent cultivation. Concurrently, governments must establish rapid mechanisms for protecting intellectual property related to green innovation to mitigate the risks of knowledge leakage.

Second, adopt differentiated environmental supervision and performance evaluation frameworks. Differentiated supervision requires routine regulatory oversight and integration of substantive green innovation outputs into evaluation frameworks. Authorities must enforce stringent penalties for fraudulent environmental claims, accompanied by robust traceability mechanisms. Environmental regulatory standards should be tiered by enterprise proper rights, pollution intensity, and lifecycle, augmented with tailored compliance guidance to improve policy alignment.

Third, advance public–private collaboration and cultivate collective efficacy. Collaborative capacity building necessitates leveraging industry-academic platforms and sectoral innovation alliances for knowledge sharing. Governments must tailor environmental sustainability awareness campaigns and systematic capacity-building initiatives to firm heterogeneity, such as providing startups with market access guidance while focusing on deep emission reduction synergies for mature firms. This fosters environmental awareness, enhancing the efficacy of governance.

6.3. Limitations and Future Research Directions

We consider the study’s limitations and the direction of future research from several aspects.

First, this study relies exclusively on quantitative data from Chinese listed companies. On the one hand, it focuses solely on the aggregate characteristics of the sample firms, potentially overlooking more nuanced firm-level influences and contextual factors that shape innovation decisions, which may undermine the persuasiveness of the conclusions. Future research should adopt qualitative approaches (e.g., longitudinal case studies, executive interviews) to investigate the micro-level mechanisms of Dual_GI longitudinally, particularly TMT cognitive processes and decision-making logics, thereby constructing more robust causal pathways. On the other hand, the exclusive focus on Chinese listed firms may limit the generalizability of the findings, as regional, institutional, and industry-specific cultural factors could influence the outcomes. Future studies should expand the sample coverage to include countries or industries not examined in this research, thereby testing whether the conclusions hold across different cultural and contextual settings.

Second, although TMTEA is gauged through corporate annual reports consistent with extant literature, this approach risks social desirability bias and affective distortions. It fails to distinguish between substantive disclosures and symbolic greenwashing. Refined measurements, including field experiments, are necessary to dissect TMTEA dimensions such as attention depth and breadth, thereby elucidating how heterogeneous environmental cognitions distinctly shape Dual_GI.

Third, this paper discusses the direct effect of TMTEA on Dual_GI, without examining the mechanism of action between the two. Although we theoretically elaborate how TMTEA may affect resource allocation at the organizational level through certain factors, no specific mediating variables were proposed or hypothesized. Therefore, future research could further investigate how TMTEA facilitates Dual_GI by shaping internal governance mechanisms such as resource allocation and employee behavior, and empirically test the mediating effects of these variables through hypothesis testing.

Fourth, our contingency analysis of GER and AC overlooks their internal heterogeneity. While it offers a general overview of GER’s moderating effect and briefly discusses the mechanistic roles of AC dimensions from a theoretical perspective, future research should rigorously dissect how specific types of GER (e.g., command-based, market-based) and forms of AC (e.g., potential, realized) moderate the relationship between TMTEA and Dual_GI. This examination should be supported by empirical modeling to validate the underlying theoretical mechanisms. Furthermore, in the digital transformation era, exploring how big data capabilities reconfigure the effects of TMTEA allocation on Dual_GI represents a critical frontier.