Enhancement mechanism of corporate green investment efficiency from the perspective of group decision making

Abstract

Introduction:

Corporate green investment is a crucial initiative for promoting high-quality economic development. However, vague decision-making scenarios imply that large-scale green investment entails significant economic risks such as resource misallocation and homogenized development.

Methods:

This paper develops a theoretical framework grounded in collective intelligence and organizational information processing theories. Using a research sample comprising Chinese listed companies from 2010 to 2020, we conduct an empirical analysis employing fixed-effects models and system GMM estimators, and measure green investment efficiency using the Richardson model.

Results:

We find that the impact of peer effects on green investment efficiency is conditional. Under conditions of effective information-processing capabilities, peer effects significantly reduce the degree of green investment inefficiency. Peer effects may influence investment efficiency through two pathways: "blind imitation" and “rational information learning”. Only the latter effectively improves resource allocation, and this enhancing effect primarily depends on the support of enterprises' internal information transparency, information acquisition capability, and information analysis capability. Heterogeneity analysis further reveals that followers' imitation of leading enterprises does not necessarily enhance efficiency: simple replication of green-investment-scale leaders yields no significant benefits, whereas learning from corporate social responsibility (CSR) leaders significantly optimizes investment efficiency.

Discussion:

Moderation analysis indicates that enhancing information transparency, information acquisition capability, and information analysis capability significantly strengthens the positive impact of peer effects on green investment efficiency. From the dual perspectives of government regulation and corporate governance, we propose targeted policy recommendations for optimizing group decision-making regarding green investments and improving investment efficiency.

1 Introduction

According to the report of the 28th Conference of the Parties, global climate finance flows amount to one-third of what is required to achieve temperature control goals, constituting a “funding bottleneck.” Annual investments of $2.4 trillion are expected to be required by 2030. To implement its strategy for high-quality economic development, the Chinese government issued the Guiding Opinions on Accelerating the Establishment and Improvement of a Green, Low-Carbon, and Circular Economic System (State Council Document (2021) No. 4), which proposes the implementation of a series of green practices, including green investment, throughout the entire process. It emphasizes strengthening enterprises primary responsibility for transition and supporting green innovation and upgrades. However, corporate willingness and efficiency in green investment remain generally low (Costa et al., 2024). While such investments are intended to reduce resource consumption, restore the environment, and achieve ecological balance, they often generate limited short-term economic returns and entail considerable uncertainties. Under resource constraints, rational entrepreneurs who favor certain returns are reluctant to proactively allocate funds to green investments. The essence of this dilemma lies in the information asymmetry and return uncertainty that place corporate green investment decisions in a “dilemma”: each enterprise adopts a wait-and-see approach out of concern that it will bear the costs of the green transformation alone. Under the constraints of such a dilemma, rational enterprises find it difficult to muster sufficient incentives to invest, resulting in the socially optimal level of green investment failing to materialize spontaneously. It is against this backdrop that collective intelligence theory suggests that when most individuals within a group engage in coordinated action, they can generate substantial collective benefits. Conversely, the group may face issues, such as insufficient developmental momentum. Furthermore, the collective cognitive capacity of a group addresses the shortcoming in traditional research on the peer effect, which has paid little attention to internal information-processing processes within enterprises, thereby overcoming the information constraints and uncertainties faced by individual enterprises. Green investment by enterprises is not an isolated individual decision, but is embedded within a network of group decision-making comprising enterprises in the same industry, region or supply chain.

Moreover, the 14th Five-Year Plan for Digital Economy Development emphasizes that productivity transformations based on massive information learning are challenging traditional organizational design logic. In the digital economy era, business decision-making is shifting from reliance on individual experience and judgement to drawing on the aggregation and collaborative analysis of collective information, profoundly altering how corporate decision-makers operate and manage their businesses. Consequently, examining corporate green investment within the framework of group decision making not only represents a theoretical extension of the traditional assumption of independent decision-making, but also constitutes a targeted response to the practical challenges facing green investment. The information aggregation mechanism derived from the collective interactions among enterprises enables managers to estimate the expected returns and variance of green investments through rational information acquisition and analysis. Can leveraging the guiding and exemplary role of typical cases to establish a peer effect characterized by information learning among enterprises serve as an effective measure for addressing the core issue of low green investment willingness among businesses?

It is worth noting that enterprises face a dual pressure from central and local governments—economic growth targets and environmental governance requirements. This pressure creates a situation where competition and symbiosis coexist among enterprises, generating significant peer effects and spillover effects on green investment (Tallman et al., 2004; Marquis and Battilana, 2009). The ecological and environmental benefits generated by green investments exhibit strong externalities. Corporate managers rely on public and private information, that is, observable, tangible, or mandatorily disclosed green investment information within their peer groups, to make their own decisions (Mbanyele et al., 2022). In this context, complementary green investment information among enterprises compensates for the shortcomings of policy pressures and incentive effects, amplifies the variability of green investment, and influences corporate green investment efficiency (Grieser et al., 2022). Conversely, for newly emerging and promising green industries that have been verified, a group decision-making environment with excessive information noise makes it easy for corporate managers to form a false consensus and trigger the “bandwagon phenomenon.” This leads to the misallocation of resources, thereby creating a “crowding-out effect” that displaces genuinely efficient green investments; in particular, favorable economic and environmental policies have led enterprises to excessively align with policy orientations, preventing managers from objectively assessing industrial development prospects. Without a rational performance evaluation system, this may cause adverse economic consequences such as excessive green investments and overcapacity (Liu et al., 2022; Wang et al., 2023).

In summary, this study addressed the following questions: during the current accelerated phase of corporate green transformation, can decision-makers enhance their green investment efficiency by learning from or imitating the green investment information of peer enterprises? what heterogeneous characteristics do peer effects exhibit on corporate green investment from a group decision-making perspective? How can a rational peer effect, characterized by information learning, be fostered to optimize the group decision-making environment and improve green investment efficiency? This study theoretically analyzed the existence of peer effects in corporate green investment from a group decision-making perspective and their impact mechanisms on green investment efficiency and reveals the moderating mechanisms underlying the influence of these two factors from an information learning perspective. Based on this foundation, a scientific empirical analysis is conducted using sample data from listed Chinese enterprises. This study deviates from the traditional assumption of independent corporate decision-making. It embeds corporate green investment behavior within a group decision-making network characterized by peer relationships, revealing the impact of peer effects, marked by information learning, on corporate green investment efficiency. This study contributes theoretical insights into the factors influencing corporate green investment and provides empirical guidance for corporate green investment practices. Second, this study extends the existing research from equilibrium analysis to a deeper level of heterogeneous learning models. This approach provides a reference for formulating and refining green investment decisions among diverse corporate entities in the green investment decision-making community. Finally, complex information noise within the information learning environment leads to consensus bias in the peer effects of corporate green investment, making it difficult for unscientific guidance strategies to support the formation of collective intelligence. This study focuses on three dimensions, information disclosure, acquisition, and analysis, and proposes an efficiency enhancement mechanism for corporate green investment from a group decision-making perspective, thereby enriching management strategies to address group decision-making challenges.

2 Literature review and theoretical analysis

2.1 Literature review

Corporate green investment promotes the restoration of resources and environmental consumption and achieves an ecological balance De Oliveira et al., (2024). (Eyraud et al. 2013) define it as an investment that reduces or curbs the emissions of various pollutants, including greenhouse gases, through the research and development of green technologies without significantly reducing the production and consumption of non-energy products. Viewed from the socio-economic context in which corporate green investment is embedded, this is not an isolated action by individual enterprises, but rather a group decision-making phenomenon deeply rooted in regions, industries, and supply chains. Enterprises within the same administrative region share similar policy environments, resource endowments, and public scrutiny; geographical proximity facilitates mutual observation and behavioral benchmarking (Xu et al., 2024), giving rise to regional “clusters” of green investment. Within the same industry, enterprises face common technological trajectories, competitive landscapes, and compliance requirements; product homogeneity renders green investment a mutually interpretable competitive signal, and the green strategic transformation of industry leaders strongly influences the decisions of followers. Core enterprises and their upstream and downstream partners, meanwhile, form vertical transmission chains through procurement standards (Wei and Li, 2026), technical specifications and carbon neutrality commitments, propagating green investment requirements step by step along the value chain and thereby incorporating even previously unenthusiastic small and medium-sized enterprises into the green transition network. These three scenarios essentially correspond to three classic theoretical perspectives on the peer effect: institutional isomorphism (Tang et al., 2019) explains the convergence pressures faced by enterprises at the regional level due to a shared institutional environment; the information cascade (Cong and Xiao, 2024) explains the logic of following at the industry level, whereby enterprises update their beliefs by observing the behavior of pioneers; and social learning (Wang and Feng, 2023) explains the process of in-depth transfer of green knowledge achieved through long-term interaction at the supply chain level. However, while these three theoretical perspectives each emphasize different driving mechanisms of the peer effect, they fail to fully explain how peer information is processed within organizations to produce differentiated efficiency outcomes—a theoretical gap that the group decision-making perspective seeks to fill.

Because the decision-making framework for corporate green investment remains immature and exhibits significant irrationality (Nagy et al., 2021), scholars have explored the factors that influence corporate green investment. Internally, corporate green investment is influenced by executive background (Jia et al., 2021), innovation capacity (Ning et al., 2025), managerial gender (Atif et al., 2020), risk preference (Peng and Kong, 2024), and information mining (Zhu et al., 2025). Externally, corporate green investments exhibit stronger characteristics of “government dependency,” whereas the gradual expansion of public scrutiny is a new trend driving corporate green development (Zhang et al., 2022b).

Most studies on the drivers of corporate green investment assume the independence of individual decision-makers and overlook the fact that, as members of a collective, enterprises are constrained by information asymmetry and profit uncertainty, leading to significant peer characteristics in their green investment behaviors; that is, they imitate or learn from the green investment behaviors of their peers. Generally, the peer effect refers to the tendency of individual decision-making behavior to align with peer behavior, as it is influenced by personal characteristics and the interconnected group context (De Oliveira et al., 2024). Regarding corporate investment behavior, decision-makers cannot exist in isolation from their social environments. They obtain additional information that is useful for investment decisions through non-market interactions, such as industry background characteristics, geographical proximity, and social network relationships, to make better forward-looking investment decisions. This causes convergence in collective corporate investment behavior (Hong et al., 2004; Peng et al., 2021). Scholars recognize the existence of peer behavior in various corporate decision-making processes. However, research on the peer effects on green investments remains limited. Although several studies have confirmed the presence of such effects, few have conducted detailed analyses on how rational information learning behaviors may amplify the variability of economic outcomes in green investment. A review of the international literature reveals that the aforementioned research gaps are equally present. International research on the peer effect has primarily focused on financial decisions such as capital structure and dividend policy (Fairhurst and Nam, 2020); discussions regarding green investment—a context characterized by limited short-term economic benefits and uncertainty—remain relatively limited. While the theory of institutional isomorphism has been widely applied to explain the adoption of corporate sustainability, it has paid little attention to how differences in corporate information capabilities lead to variations in the quality of learning; research on collective intelligence (Kudyba et al., 2020), has been largely confined to the performance of small groups in laboratory settings and has not yet been extended to the realm of inter-organizational strategic decision-making. In fact, this implies that examining corporate green investment within a group decision-making framework not only addresses a gap in the domestic literature but also represents a proactive contribution to the international theoretical dialogue—by introducing a conditional perspective on information-processing capabilities, it offers new theoretical possibilities for understanding the differentiated efficiency consequences of the peer effect.

2.2 . Theoretical analysis

Rational information learning behavior helps reduce decision-making costs and lowers the benefit uncertainty of green investment, thereby influencing its economic and environmental effects (Lieberman and Asaba, 2006). The theory of uncertainty suggests that business managers cannot accurately assess the benefits and risks of investments, leading them to favor imitation strategies (Bustamante and Frésard, 2021). As the level of competition in the green investment market intensifies, managers tend to mimic the behavior of peer enterprises to maintain a market advantage or alleviate regulatory pressure (Xu et al., 2025a). Corporate green investments exhibit peer behavior based on factors such as industry background, geographic location, and social relationships, which are characterized primarily by mutual imitation and learning (Costa et al., 2024).

From a cost-benefit perspective, rational entrepreneurs who favor certain returns are unlikely to proactively allocate funds to green projects under limited resource constraints (Tallman et al., 2004; Marquis and Battilana, 2009). Green investments fail to yield direct economic efficiency and the ecological and environmental benefits they generate exhibit strong externalities. Managers make decisions based on information obtained from reference points to reduce costs and uncertainty. Influenced by organizational inertia and institutional pressures, corporate managers typically rely on imperfect public and private information to optimize green investment decisions through investment information that is observable within the industry, whether tangible or subject to mandatory disclosure (Pan et al., 2025), thereby affecting the efficiency of corporate green investment. It is worth noting that information cascade theory (Tump et al., 2020; Cong and Xiao, 2024) tends to view enterprises as passive recipients of external signals. However, it falls short in explaining how internal information-processing structures selectively interpret and respond to peer signals. Social learning theory (Lieberman and Asaba, 2006) focuses on the outcomes of behavioral convergence, yet pays little attention to the process-level differences in how peer information is translated into organizational action; institutional isomorphism theory (Tang et al., 2019) emphasizes pressure-driven imitation, but struggles to distinguish between efficiency-enhancing learning and dysfunctional conformity, nor can it fully explain why peer effects vary in terms of efficiency. These limitations point to a key research gap: while existing theories can explain whether enterprises follow their peers, they fail to adequately explain how peer information is processed within organizations to produce differentiated efficiency outcomes. Specifically, existing theories face three difficult-to-answer questions: (1) Why do peer effects manifest differently across different information environments? (2) Why do followers benefit from learning from leaders in social responsibility, but not from learning from leaders with high investment levels? (3) How can one empirically distinguish between information learning and pure imitation?

We believe that the group decision-making perspective, based on collective intelligence and organizational information processing theory, provides a complementary framework capable of bridging this gap. Unlike the aforementioned theories, which treat enterprises as “black boxes,” the group decision-making perspective views corporate green investment as a collective information-processing event: the investment behavior of peer enterprises constitutes an external information pool, and whether, how, and to what extent an enterprise utilizes this pool depends on its internal group decision-making mechanisms—including the boards' network for information acquisition, the information-processing structure of senior management, and the transparency of information disclosure. The unique contribution of this perspective lies in the fact that it not only predicts the existence of the peer effect but also predicts the conditional nature of its effects—peer information can only be effectively translated into investment decisions when enterprises possess the corresponding information-processing capabilities. In fact, information complementarity in group decision-making helps to compensate for the shortcomings of policy pressures and incentives, amplify investment variability, and alter total factor productivity and output, thereby influencing the effectiveness of corporate green investment (Bustamante and Frésard, 2021). Furthermore, constrained by incomplete information and bounded rationality, moderate imitation and learning from peer enterprises' green investment decisions help to reduce decision-making costs and uncertainty, thereby enhancing the efficiency of green investment. Therefore, this paper proposes Hypothesis 1:

Hypothesis 1: Under conditions of effective information-processing capabilities, the peer effect in corporate green investment enhances the efficiency of green investment.

The asymmetry and complexity of information in the green investment market increase the cost of information seeking for corporate investment decision-makers (Xu et al., 2025a). Green investment behaviors within peer groups contain crucial information relevant to decision-making and serve as a source of information for target enterprises (Zhao and Wang, 2024). The disclosure of this green investment decision information to decision-makers is critical for the effectiveness of peer effects. Given the risks inherent in corporate green investment, when enterprises engage in uncertain practices—selectively disclosing positive information while concealing negative information, or even deliberately hiding high-risk investments—they undermine the positive impact of peer effects on green investment efficiency. From the perspective of analyst attention, enhancing corporate transparency and issuing high-quality forecasts can effectively curb inefficient investments (Chen et al., 2017; Elaoud and Jarboui, 2017). Analysts track and gather financial and non-financial information that affects corporate value (Chen et al., 2011). By leveraging professional knowledge and skills or conducting field research to interpret and extract insights from diverse information sources and disclose these insights to investors, thereby alleviating the imbalance between information supply and demand, while enhancing the efficiency of resource allocation. Hence, higher information transparency enables green investment decision-makers to obtain effective information from peer enterprises, thereby enhancing the efficiency of green investment. Therefore, Hypothesis 2 is proposed.

Hypothesis 2: Increased information transparency will amplify the positive effect of corporate green investment and the peer effect on the efficiency of green investment.

Corporate decision-makers make green investment decisions based on superior information acquisition capabilities, which help reduce corporate information search costs (He et al., 2025). The peer effect in corporate green investment primarily refers to the phenomenon in which different enterprises within the market facilitate social interaction, resource sharing, and demonstration effects among peer enterprises through certain social relationships. From the perspective of institutional arrangements for information disclosure by listed enterprises, the content of disclosures is influenced by regulatory requirements, management incentives, and the oversight role of the board of directors. The effectiveness of independent directors depends on their individual characteristics, such as educational and professional backgrounds and on their social networks (Khan et al., 2025). The opportunity to gain reputational capital, tacit knowledge, and strategic resources through an independent director's network depends on the director's position within the network and governance capabilities. Enterprises' investment efficiencies are determined by the centrality of their independent directors' networks, meaning that the higher the centrality, the higher the investment efficiency (Chen and Nie, 2023). Therefore, the information acquisition methods employed by independent directors reflect their information-gathering capabilities, enabling them to swiftly obtain valuable green investment information. Therefore, Hypothesis 3 is proposed.

Hypothesis 3: Increased information acquisition capabilities will amplify the positive effect of corporate green investment and the peer effect on green investment efficiency.

Due to opportunistic motives among the enterprise's management team, the enterprise may develop selective disclosure problems, such as sampling disclosures and concealed negative information, which could mislead the formation of a collective consensus. Corporate green investments require that executives interpret the information they obtain. Differences in environmental backgrounds lead to varying degrees of analysis among executives, which, in turn, affects the process of translating information into actionable guidance (Zhu et al., 2023). By comparison, executives with environmental backgrounds are highly attuned to environmental risks and possess specialized environmental knowledge, which influences corporate green investment strategies (Bansal and Roth, 2000). Executives with an environmental background can leverage their extensive environmental knowledge and experience to reduce the risks associated with corporate environmental investments, which will be beneficial for improving the efficiency of corporate green investments (Angriawan and Abebe, 2011). Therefore, higher information analysis capabilities enable decision-makers to accurately interpret the correct aspects of standardized disclosures and to uncover negative information that peer enterprises have attempted to conceal, thereby improving green investment efficiency. Therefore, Hypothesis 4 is proposed.

Hypothesis 4: Increased information analysis capabilities will amplify the positive effect of corporate green investment and the peer effect on green investment efficiency.

It should be noted that the effect of peer effects on efficiency is not unconditional. Theoretically, there exist two distinct pathways: first, imitative convergence driven by institutional pressure, whereby firms blindly follow their peers to gain legitimacy, which may lead to resource misallocation and the “bandwagon phenomenon”; second, a learning mechanism based on information processing, whereby firms reduce decision-making uncertainty and optimize investment allocation by collecting, interpreting, and utilizing peer information. This paper tends to argue that the effectiveness of the second pathway depends on firms' internal information-processing capabilities—when firms possess relatively high levels of information transparency, acquisition capability, and analytical capability, peer effects are more likely to translate into efficiency gains; otherwise, peer effects may remain at the level of mere behavioral convergence. Therefore, the analytical perspective this paper attempts to propose is: under what conditions do peer effects influence efficiency through the information-learning mechanism. The group decision-making perspective, grounded in collective intelligence and organizational information processing theory, may offer a complementary analytical approach to understanding this issue.

Hypotheses 1–4 above reveal the main effect of the peer effect on the efficiency of green investment and the moderating role of the information-learning mechanism. However, the efficiency consequences of the peer effect may vary depending on whom the followers are emulating. Leaders identified by the scale of green investment may be motivated by non-economic considerations, such as policy compliance and image-building. If followers simply copy this visible scale without discerning the underlying rationale, they are prone to blind imitation, which ultimately leads to resource misallocation. Conversely, leaders distinguished by corporate social responsibility (CSR) exhibit stronger social commitment, reflecting more robust long-term strategies and stakeholder relationships; followers who learn from them can acquire strategic information and optimize their investment allocation. Accordingly, the efficiency-enhancing effect of imitating green-investment-scale leaders is expected to be insignificant or even negative, whereas learning from CSR-oriented leaders is expected to exert a significant positive effect.

3 Research design

3.1 Data sources

This study selected 821 Chinese listed enterprises from 2010 to 2020, forming an unbalanced panel data comprising 4,065 observations. The research sample is processed as follows: (1) Listed enterprises in the financial sector were excluded; (2) listed enterprises marked as ST or ^*ST were excluded; (3) errors were minimized as much as possible; all missing data for relevant variables were excluded without any imputation; and (4) enterprises with only 1 year of data were excluded, as their data lacked continuity. Green investment data are sourced from the annual reports of listed enterprises, whereas all other data are sourced from the CSMAR and Wind databases. To mitigate the influence of extreme outliers, all continuous variables are winsorized at the 1st and 99th percentiles only in the CSR heterogeneity analysis (see Section 4.3 for details; see Appendix Table A1 for comparison). As reported in Table 2, the deviation variable exhibits severe right-skewness in the full sample, with a mean of 0.058, a standard deviation of 1.170, and a maximum value of 73.175, suggesting that OLS estimates may be driven by a few extreme observations. After winsorization, the variable distribution becomes substantially more reasonable. The regression results before and after winsorization are compared in Appendix Table A1, and the core conclusions remain unchanged.

This paper defines peer enterprises based on industry classification, primarily in light of the following considerations: Firstly, green investment exhibits significant industry homogeneity. Enterprises within the same industry face similar technological trajectories, environmental compliance standards, and competitive landscapes; green investment behavior constitutes mutually interpretable competitive signals and market standards, providing a natural foundation for information learning (Lieberman and Asaba, 2006). Secondly, in terms of information accessibility, industry classification is the most stable and observable dimension in the disclosure of information by listed companies; the disclosure standards for green investment data among enterprises within the same industry are relatively consistent in annual reports, facilitating horizontal comparison and learning by target enterprises. Thirdly, in the Chinese context, environmental regulations at the industry level exhibit a high degree of consistency, and the institutional pressures faced by enterprises within the same industry are converging, making the industry the most likely arena for the emergence of green investment peer effects.

It is worth noting that the definition of peer groups based on industry classification has certain limitations. Enterprises within the same industry may not necessarily regard one another as relevant benchmarks if they differ significantly in size, market position, or business model; conversely, enterprises in different industries may also exert peer influence through channels such as supply chains or geographical proximity. Due to data limitations, this paper is unable to fully capture these more nuanced peer channels—such as supply chain linkages or interlocking director networks—which may constitute closer information-learning pathways than industry classification alone.

3.2 Model Specification

This study employed a fixed-effects model to examine the impact of corporate green investment and peer effects on green investment efficiency, as shown in Equation 1.

where the dependent variable is GIE, which represents the green investment efficiency of listed enterprise i in year t. The primary explanatory variable is Deviation, which measures the absolute difference between a corporate green investment and the average green investment of its peer group. Controls refer to control variables; u_i represents the individual fixed effect, v_t is the time fixed effect; and ε_{i, t} is the random error term.

3.3 Variable definition

3.3.1 Corporate green investment efficiency

Following the methodology of Zhang et al. (2022a), we manually collected annual reports from listed enterprises and screened expenditures directly related to environmental protection in the notes to construction in progress, such as projects for wastewater treatment, exhaust gas control, dust removal, energy conservation, desulfurization, and denitrification. These expenditures were aggregated annually to calculate total green investment. Regarding the efficiency of corporate green investments, this study draws on Richardson model for estimation. Green investment remains, in essence, a capital allocation decision involving the deployment of financial resources into projects with expected future returns; it is influenced by financing constraints, growth opportunities, and adjustment costs, which is consistent with the mechanism for optimal investment decisions by enterprises as described in the Richardson framework. Although this model is originally developed for general capital investment, it has been extended to the green investment context by studies such as Liu et al. (2022). Corporate green investment is not entirely profit-driven; rather, it often serves to meet environmental regulations, respond to stakeholders, or maintain corporate reputation. Therefore, the “investment deviation” estimated here is more accurately described as the gap between actual investment and the economically optimal level conditional on enterprise fundamentals, rather than a value judgment on the investment itself. Green investment is characterized by long-term return cycles and significant externalities; short-term deviations from the optimal level are more likely to reflect genuine decision-making biases rather than normal strategic adjustments. Thus, measuring green investment inefficiency by the absolute value of the Richardson regression residual captures the extent to which the amount of corporate investment aligns with its fundamental characteristics, reflecting the degree of optimization in resource allocation, as shown in Equation 2.

Where i and t denote enterprise and year, respectively; GINV represents the amount of new green investment for that year; Growth represents the enterprise growth potential, measured by the revenue growth rate; Lev represents financial leverage, calculated as total liabilities divided by total assets; Cash represents the cash holding level, calculated by dividing the final cash and cash equivalent balance by total assets; Age represents the enterprise age since listing; Size represents the enterprise scale, measured by the logarithm of total assets; Return is the annual stock return rate for enterprises considering cash dividend reinvestment; ER represents the intensity of environmental regulation, measured by the ratio of the frequency of environmental protection-related terms to the overall word count in municipal government work reports; GINV_{i, t − 1} represents the new green investment amount for the period_t−1; ∑Year represents the fixed effect for year; and ∑Industry represents the industry fixed effect. The absolute value of the residuals from the ordinary least squares regression measures the degree of green investment inefficiency.

3.3.2 Corporate green investment and peer group indicators

According to the “2021 Third Quarter Industry Classification Results for Listed Enterprises” released by the China Securities Regulatory Commission, peer corporations are defined as those with the same industry category name, by adopting the first-level category industry standard. For manufacturing corporations, the scope of enterprises within the same cluster is determined using a secondary industry classification standard. The green investment peer group indicator for the target enterprise is the average green investment of the peer group. The green investment peer group difference is measured as the absolute value of the difference between the target corporate green investment and the average green investment of its peer group. As shown in Equations 3 and 4:

Where i and t denote enterprise and year, respectively; represents the sum of green investments made by all enterprises within the primary industry category to which the enterprise i belongs (using secondary industry standards for manufacturing), excluding enterprise i itself; n denotes the number of enterprises within the same group; Gis_{peer, i, t} captures the peer group's average green investment; and Gis_{j, t} represents green investment for enterprise i, and Deviation_{i, t} measures the absolute deviation of target enterprise i from the average green investment level of its peer group.

• (1) Control variables. the main variables include enterprise size (Size), debt-to-asset ratio (Lev), return on assets (ROA), cash flow (Flow), shareholding concentration (Shareholding), and board independence (D_size). Specific variable definitions and descriptive statistics are presented in Tables 1 and 2, respectively.

This paper selects information transparency, information acquisition capacity and information analysis capability as proxy variables for the information learning mechanism. The underlying logic is that information learning is not an instantaneous process, but rather unfolds sequentially: enterprises first observe peer behavior, then acquire and interpret relevant information, and finally adjust their decisions accordingly. These three variables correspond to key links in this chain: information transparency determines whether enterprises can observe their peers' green investment behavior; information acquisition capacity influences whether enterprises can effectively collect these signals; and information analysis capability determines whether enterprises can transform external information into a basis for internal decision-making.

Specifically, firstly, the KV index measures the quality of corporate disclosure, which determines whether green investment information from enterprises within the same group can be effectively disseminated to the market and observed by other enterprises. Within the information learning framework, it is only when information is fully disclosed that a target enterprise can extract effective signals from the investment behavior of its peers, thereby reducing the uncertainty in its own decision-making; if information disclosure is inadequate, the peer effect is more likely to stem from blind imitation driven by institutional pressure, rather than learning behavior based on information.

Secondly, the network centrality of independent directors measures an enterprise's ability to obtain external heterogeneous information. Independent directors are embedded in broader social networks through their concurrent roles; those occupying structural holes are able to obtain and transmit green investment information from other enterprises within the same industry. Within the information learning framework, a higher intermediary centrality implies that an enterprise is able to obtain green investment developments from peer enterprises more rapidly and comprehensively, thereby effectively incorporating peer information into its own decision-making reference set.

Third, the number of executives with an environmental background measures an enterprise's ability to analyze green investment information. Executives with an environmental background possess specialized environmental knowledge and risk identification capabilities, enabling them to discern the content of genuine strategic value amidst the complex and diverse investment information from peer enterprises (Angriawan and Abebe, 2011). Within the framework of information learning, this analytical capability determines whether an enterprise can transform the peer information it has acquired into effective investment decisions, rather than simply replicating the investment scale of its peers.

It should be noted that, while the aforementioned proxy variables cannot directly observe the micro-level process of “what Enterprise A has learnt from Enterprise B,” they respectively capture the preconditions, channel conditions and transformation conditions necessary for information learning to occur. At the empirical level, if these variables do indeed reinforce the positive impact of the peer effect on the efficiency of green investment, this indirectly supports the theoretical proposition that “the peer effect enhances efficiency through the information-learning mechanism”—because if the peer effect were solely attributable to institutional pressure or blind imitation, enterprises' information capabilities should not significantly moderate this relationship.

4 Analysis of Results

4.1 Benchmark regression and asymmetric analysis

Table 3 reports the benchmark regression results and the asymmetric test. Column (1) presents the benchmark result without control variables. Column (2) adds enterprise-level control variables, and the coefficient of Peer Green Investment Deviation remains significantly positive at the 1% level (0.0033), providing supporting evidence for Hypothesis 1 that the peer effect enhances corporate green investment efficiency. Columns (3) and (4) further examine the asymmetric nature of this effect. The results in Column (3) show that the coefficient on Negative Peer Deviation is significantly positive (0.0034, p < 0.01), whereas the coefficient on Positive Peer Deviation is statistically insignificant (-0.0007). Column (4) confirms this asymmetry using Signed Deviation (−0.0033, p < 0.01). In terms of economic significance, the coefficient of Deviation is 0.0033. Combined with its standard deviation of 1.17, a one-standard-deviation increase in deviation leads to approximately a 0.0039 change in the green investment inefficiency index. This value corresponds to approximately 26% of the mean green investment inefficiency (0.015) during the sample period, indicating that peer convergence behavior can alleviate approximately one-quarter of the efficiency deviation. Considering that green investment decisions are simultaneously influenced by multiple factors, including financing constraints, institutional pressure, and enterprise characteristics, this marginal effect carries practical economic significance.

These results indicate that negative peer deviation significantly reduces corporate green investment efficiency; that is, the lower the green investment level of peer enterprises, the lower the investment efficiency of the focal enterprise. Positive peer deviation, however, does not produce a significant effect. This suggests that the adverse effect of the peer effect stems primarily from the low green investment levels of peer enterprises, rather than from overinvestment by the focal enterprise. When peer enterprises' green investment is insufficient, the focal enterprise struggles to optimize its investment decisions through intra-industry information learning and resource coordination, leading to a decline in investment efficiency. In other words, the peer effect operates through a directional information-learning mechanism: efficiency losses from deviation occur only when peer enterprises' green investment falls below that of the focal enterprise.

4.2 Robustness tests and endogeneity treatment

We conducted robustness tests to verify the robustness of the above conclusions. First, we replaced the indicators of green investment disparities among enterprises. For all enterprises, we adopted the primary industry classification standard to determine the scope of the peer groups. The green investment peer group difference indicators for enterprises were recalculated using the methods specified in Equations 3 and 4, and the regression results are shown in Table 4. Second, we re-estimated corporate green investment efficiency using the Generalized Method of Moments (GMM model), and used the results as a substitute indicator for GIE. The regression results are shown in Table 4. When we replaced the enterprise-level green investment peer group differences and GIE indicators, column (3) showed the regression results. These results collectively confirm the robustness of the regression findings. To control the impact of extreme values on empirical results, all variables are winsorized at the 1% and 99% percentiles. Column (4) shows the results, which are consistent with those of the benchmark effect analysis. Additionally, although fixed-effects models can mitigate endogeneity issues caused by unobservable factors that do not change over time, they struggle to address the endogeneity that stems from unobservable variables that evolve over time. Considering that corporate green investment efficiency exhibits pronounced dynamic effects, this study constructed a dynamic panel model using the GMM approach, in which the one-period lagged dependent variable GIE_{i, t − 1} is incorporated to verify the impact of green investment and the peer effect on corporate GIE. The regression results in Column (5) of Table 4 further confirm the robustness of the benchmark effect analysis findings.

In peer-effect research, the behavioral interdependence inherent in peer effects makes it inherently difficult to identify a fully exogenous and strongly correlated instrumental variable (Ross and Shi, 2022; Caeyers and Fafchamps, 2024). This paper attempted to use the one-period lagged average green investment of peer enterprises as an instrumental variable. However, the first-stage F-statistic failed to pass the weak instrumental variable test (Stock-Yogo critical value), indicating that this IV does not satisfy the relevance condition. Under weak instruments, 2SLS estimates may produce severe bias; therefore, this paper does not adopt these estimation results. Instead, this paper employs a multi-dimensional identification strategy: (1) a two-way fixed-effects model to absorb unobserved heterogeneity that does not vary over time; (2) a system GMM dynamic panel estimator to address dynamic endogeneity; and (3) cross-validation through robustness strategies including alternative peer-group definitions, lagged moderating variables, and exclusion restriction tests. Although the aforementioned methods cannot fully eliminate omitted variable bias, the consistent pattern of results supports the core findings of this paper.

Furthermore, to distinguish between the information-learning mechanism and the regulatory imitation mechanism, this paper further examines the moderating role of the intensity of environmental regulation. As shown in Table 5, column (2), the coefficient of the interaction term between industry deviation and the intensity of environmental regulation is 4.739, which is significantly positive at the 1 percent level, indicating that the higher the intensity of environmental regulation, the stronger the positive impact of the peer effect on the efficiency of green investment. This result provides supporting evidence for the information-learning mechanism—if the peer effect were solely attributable to passive imitation by enterprises seeking to comply with regulatory requirements, environmental regulation should not reinforce the efficiency-enhancing effect of the peer effect. Specifically, the higher the regulatory intensity, the greater the need for enterprises to identify truly effective green investment strategies through information learning, thereby enhancing investment efficiency. To address potential reverse causality issues with the moderating variable, this paper treats the intensity of environmental regulation as lagged by one period. As shown in Table 5, column (3), the coefficient of the interaction term between lagged environmental regulation and sectoral deviation is 2.989, which remains significantly positive at the 1% level. This indicates that the estimated moderating effect in the current period is not driven by reverse causality, further enhancing the credibility of the identified mechanism. To distinguish between the relative importance of industry-level and geographical-level peer effects, this paper incorporates both into the model simultaneously. As shown in Table 5, column (4), after controlling for the geographical-level peer effect, the coefficient for industry deviation remains 0.00229, which is significantly positive at the 5% level, while the coefficient for geographical deviation is 0.00128 and fails to pass the significance test. These results suggest that, after controlling for industry factors, the independent effect of geographical proximity weakens. In other words, technological trajectories and institutional pressures among enterprises in the same industry are the main channels through which peer effects operate; geographical distance may also play a role, but it largely overlaps with industry clustering and is difficult to isolate.

The robustness tests described above were primarily based on the Richardson residual model; however, as this model captures only the degree of deviation from optimal investment, this paper further employs Data Envelopment Analysis (DEA) for cross-validation, a method that explicitly incorporates both environmental and economic performance dimensions. This paper utilizes the VRS-DEA output-oriented model to calculate efficiency scores by industry and year, and uses 1 minus the efficiency score as an indicator of green investment inefficiency. Table 6 reports the results. In the extended DEA model (Column 2)—which incorporates both environmental and economic input-output dimensions—the deviation coefficient is 0.0161 (p < 0.05), providing supporting evidence that the peer effect significantly enhances enterprises' green investment efficiency, consistent with the benchmark findings. By contrast, in the pure green DEA model (Column 3), which relies solely on a single environmental output dimension, the coefficient is not significant (0.0250, p > 0.1); this suggests that a single environmental indicator is insufficient to comprehensively capture the economic-environmental trade-offs inherent in green investment efficiency, thereby confirming the necessity of the extended model.

Table 7 further examines moderating effects within the DEA framework. The interaction term between deviation and environmental executives remains significantly positive (0.0958, p < 0.01), consistent in direction with the results of the Richardson model, thereby supporting the information-learning mechanism. The interaction term between deviation and directors' network centrality is not significant in the DEA model, which may be attributed to sample loss (the sample size decreased from 4,065 to 3,515) and reduced estimation efficiency resulting from DEA grouping calculations; nevertheless, the significance of the interaction term between deviation and environmental executives indicates that the information analysis channel remains robust within the DEA framework, and the overall pattern continues to support the core conclusions of the information-learning mechanism. Taking these results together, the core conclusion of this paper—that the peer effect of corporate green investment has a positive impact on green investment efficiency—is supported. The estimation results of the extended DEA model are consistent in direction with the benchmark Richardson model (the deviation coefficients are both positive), but its explanatory power (R² = 0.042) is lower than that of the benchmark model, and the DEA model relying solely on the environmental dimension fails to achieve statistical significance. On the one hand, this may stem from the sensitivity of the DEA model to input-output indicator selection; on the other hand, it also indicates that multidimensional measurement of green investment efficiency still faces certain methodological limitations. Overall, the DEA analysis provides auxiliary directional support for the benchmark findings.

4.3 Heterogeneity analysis

The theoretical analysis presented earlier indicates that the efficiency implications of the peer effect depend on whether followers engage in rational information learning or blind imitation. This paper tests this distinction by classifying enterprises into leaders (the top third) and followers (the bottom two-thirds) along two different dimensions: the scale of green investment and corporate social responsibility (CSR). If followers merely replicate the visible scale of investment by leaders without discerning the underlying efficiency drivers, this is likely to lead to distortions in resource allocation; By contrast, learning from CSR-oriented leaders—who typically demonstrate stronger long-term strategies and stakeholder relationships—may optimize investment allocation.

Table 8 presents the subsample regression results. Columns (1) and (3) show that, regardless of whether leaders are identified by green investment scale or CSR scores, the peer effect within the leader group is significantly positive, indicating that effective information exchange mechanisms have formed among leading firms. However, when shifting the perspective to followers' imitation of leaders, the efficiency consequences diverge: Column (2) shows that the coefficient for followers imitating green investment scale leaders is insignificant (−0.0003), suggesting that such simple replication does not improve efficiency; Column (4) shows that following CSR leaders generates a significantly positive effect (0.5660, p < 0.01), and this effect exceeds the peer effect within the CSR leader group itself (0.4520, p < 0.01). The interaction term test further indicates that the follower group exhibits significantly higher sensitivity to peer deviation than the leader group (difference=-0.042, p < 0.05). A possible explanation for this result is as follows. If followers simply imitate green investment scale leaders, they tend to focus only on the “visibility” of investment scale while neglecting the efficiency-driven factors behind it, which easily triggers a bandwagon effect and crowds out truly efficient green investment. In contrast, in the CSR dimension, followers—due to information disadvantages and insufficient green investment experience—are more sensitive to peer deviations. They tend to interpret peer firms' green investment deviations as signals of industry trend changes, thereby exhibiting stronger imitation learning motivations. CSR leaders, however, possess more sufficient information channels and independent decision-making frameworks, relying relatively less on peer deviations. Therefore, simple replication of green investment scale leaders constitutes blind imitation that fails to improve resource allocation efficiency, whereas learning from CSR leaders represents rational information learning in which followers exhibit higher sensitivity and can effectively optimize investment allocation.

It should be noted that, as some observations were excluded due to missing values during the sub-sample regression analysis, the combined size of the sub-samples (3,778) is slightly smaller than that of the full sample (3,794); however, this does not affect the reliability of the conclusions.

5 Further analysis

The empirical analysis results in the preceding section confirm the positive peer effect on corporate GIE. Notably, in a group decision-making environment aggregating multi-source green investment information, the information learning process of corporate green investment reflects how managers adjust their green investment decisions by acquiring green investment information from peer enterprises and analyzing its content. Therefore, rather than directly observing information-learning behavior among enterprises, this section examines the conditions required for enterprises to engage in collective information learning through the three stages of information release, information acquisition and information analysis. Based on the research hypotheses put forward earlier, it explores the moderating role of information learning in the relationship between the green investment peer effect and GIE.

5.1 Model settings

Corporate decision-makers engage in the rational learning of green investment information and form peer effects. This study focused on the causal role of information learning in peer effects by employing a moderation model to examine the moderating effects of information transparency, information acquisition capability, and information analysis capability on the relationship between corporate green investment peer effects and efficiency. The specific model is shown in Equation 5.

Where M_{i, t} represents the moderating variable (i.e., KV Index, Median_betcentrl, or EP_execu) denoting information transparency, acquisition, and analysis capability, respectively. Deviation × M_{i, t}represents the interaction term between the within-group variance indicator and the moderator variable, with the other variables specified in Equation 1.

5.2 Variable definition

• (1) Information transparency. The coefficient measuring the impact of stock trading volume on stock returns is termed the KV Index, and reflects the market's dependence on volume information and the level of corporate disclosure. In the investment market, the KV Index reflects investors' objective assessment of the degree of market information asymmetry, that is, the effectiveness of corporate disclosure by listed enterprises. This encompasses mandatory and voluntary disclosures (Hu et al., 2024). This study selects the KV Index as a proxy variable for information transparency; the KV Index reflects the market's comprehensive assessment of the quality of a company's information disclosure; in essence, it is a comprehensive indicator of a company's information processing capabilities—in companies operating in a high-quality information environment, managers demonstrate greater sensitivity to and discernment regarding external investment signals (including green investment signals released by peer companies), and posit that higher information disclosure quality is correlated with greater information transparency. According to Kim and Verrecchia (2001), if a listed enterprise maintains high information disclosure quality, investors' stock returns will exhibit weaker dependence on its trading volume. Consequently, the KV Index relates not only to a company's own disclosure of information, but also reflects the overall information environment in which the company utilizes external information for decision-making. Additionally, the KV Index avoids the issues associated with accounting variables, such as discretionary accruals of profits and earnings management; therefore, this study adopts the approach of Kyle (1985) and (Chordia et al., 2001) using Equation 7 to calculate the KV Index.

Where P_t is the closing price of the stock on day t and Volt represents the trading volume (number of shares) for the day. Vol₀ represents the average trading volume for all the trading days during the study period. The KV Index is constructed using the λ values obtained from regression analysis of each listed enterprise via ordinary least squares (excluding cases where λ is negative). The smaller the value of λ, the more comprehensive the disclosure of information.

• (2) Information acquisition capability. To protect their personal reputations, independent directors are motivated to process and synthesize external heterogeneous information before conveying it to the enterprise. An independent director network is a social network formed by independent directors serving concurrent roles. These concurrently serving independent directors occupy structural hole positions within the network, possessing informational and control advantages that enable them to access and control new heterogeneous information and knowledge resources (Zhang et al., 2024; Xu et al., 2025b). Independent directors and enterprise management form a social network that facilitates frequent social interactions and greater information sharing. In particular, independent directors possess information sources unavailable to other directors; maintain extensive social networks; and bring greater expertise, experience, and skills to the table, thereby demonstrating heightened sensitivity to policy developments. Simultaneously, owing to their unique position within the enterprise, independent directors provide an external perspective on business decision-making, which profoundly influences the formulation of the enterprise green development strategy. Accordingly, this study employs the independent director betweenness centrality (Betcentrl) as a proxy variable for corporate information acquisition capability, as shown in Equation (8).

Betweenness centrality measures the extent to which a director in a director network controls the paths to other directors. Specifically, let g_ab denote the minimum number of edges required to connect directors a and b; let denote the number of directors i that lie on the shortest path between directors a and b; and let n denote the total number of directors in the listed enterprise director network for the current year. To eliminate differences in network scale across years, this study uses (n−1) (n−2) as a scaling factor. Based on the annual calculation of each director's betweenness centrality within the director network, the median betweenness centrality (Median_betcentrl) of all independent directors on an enterprise's board is selected as the enterprise-level centrality metric.

• (3) Information analysis capability the soundness of corporate green investment decisions depends on whether executives can interpret the information they receive thoroughly and accurately. Executives with environmental backgrounds can leverage their unique experiences to mitigate the risks associated with corporate green investments, thereby driving increased spending on green initiatives (Chen et al., 2024). According to corporate internal control theory, to mitigate green investment risks, corporate executive teams will become more cautious under peer influence, leading to equally pronounced disagreements among executives. If executives with environmental expertise hold senior positions within a team and wield significant influence, they will possess a greater capacity to sway other executives or even lead them to comply with their directives, thereby increasing the likelihood of the enterprise pursuing green investment. Therefore, an environmental background determines whether executives can fully interpret peer enterprise information, thereby affecting the GIE. This study measures corporate information interpretation capabilities based on the number of executives with environmental backgrounds in listed enterprises. The key to determining an executive's environmental background lies in examining whether their resume includes keywords such as ecology, environment, low carbon, sustainability, environmental protection, energy conservation, green initiatives, new energy, and clean energy. If present, the executive is deemed to have an environmental background. The number of executives with environmental expertise in each enterprise is tallied annually. An enterprise's information analysis capability is measured by the number of executives with environmental expertise (EP_execu).

5.3 Analysis of results

Due to varying degrees of missing values across the moderator variables, the effective sample sizes for the three moderation effect models differed slightly (see Table 9); however, the coefficients and significance levels of the core explanatory variables remained robust. Table 9 shows that the Deviation × KV coefficient is significantly negative. As the KV index is a contrarian indicator (a lower value indicating more comprehensive disclosure), the economic implication of this result is that the higher the level of information transparency, the stronger the positive impact of the peer effect on an enterprise's green investment efficiency; in other words, information transparency positively moderates the relationship between the peer effect and green investment efficiency. Hypothesis 2 is thus supported. The main effect of information transparency (KV Index) is negative but not significant; the main effect of information access (Median_betcentrl) is negative but not significant; and the main effect of information analysis (EP_execu) is positive and significant. This pattern of results supports the following interpretation: information transparency and information acquisition capacity do not directly enhance GIE in their own right, but they exert an indirect influence by reinforcing the information-learning mechanism of the peer effect; information analysis capability, meanwhile, exerts both direct and indirect effects. This is consistent with the theoretical expectation that executives in an environmental context can directly optimize green investment decisions, while also supporting the logic that they enhance the efficiency of the peer effect by strengthening their peers' ability to interpret information. Higher information transparency enables corporate green investment decision-makers to obtain effective information from peer enterprises, thereby reducing investment risks arising from uncertainty and enhancing the efficiency of green investments. This aligns with the findings of Elaoud and Jarboui (2017), who discovered that analysts enhance corporate transparency, issue high-quality forecasts, and curb inefficient investments.

Table 9 shows that the coefficient for Deviation × Median_betcentrl is significantly positive, indicating that increased information acquisition capacity reinforces the positive impact of green investment and peer effects on corporate GIE, validating Hypothesis 3. The information acquisition methods reliant on independent director networks demonstrate strong information-gathering capabilities among investment decision-makers. This facilitates the swift acquisition of valuable information from peer enterprises, reduces information acquisition costs, and enhances corporate GIE. This aligns with the findings of Chen and Nie (2023), namely that the higher the network centrality of independent directors, the more effective their governance role becomes, as evidenced by higher investment efficiency in their respective enterprises.

Table 9 shows that the coefficient for Deviation × EP_execu is significantly positive, indicating that enhanced information analysis capabilities amplify the positive impact of green investment and peer effects on corporate GIE, confirming Hypothesis 4. Enhanced information analysis capabilities enable corporate green investment decision-makers to accurately interpret the correct aspects of standardized disclosures and to identify negative information that peer enterprises have concealed. This facilitates informed decision-making and improves the GIE. This conclusion aligns with the findings of Angriawan and Abebe (2011), indicating that executives with environmental backgrounds can mitigate the risks associated with corporate environmental investments by leveraging their unique environmental knowledge and experience.

To further mitigate potential reverse causality issues associated with the moderating variables, this paper treats all moderating variables as one period lagged in the robustness tests (see Table 10), The results show that the interaction term between the centrality of the director network and the environmental background of senior executives remains significantly positive, indicating that the moderating effects of these two information-learning channels are not driven by reverse causality; whereas the interaction term for the KV index is no longer significant after applying a one-period lag, suggesting that the moderating effect of information transparency may be more sensitive to the information environment in the same period, and its intertemporal stability is weaker than that of the former two. Furthermore, this paper attempts to introduce competitive pressure (the industry Herfindahl index) and regulatory imitation pressure (the intensity of regional environmental regulations) as alternative moderating variables, and finds that their interaction terms are all insignificant; this to some extent weakens the competitive explanation that “the peer effect enhances efficiency through competitive pressure or institutional imitation rather than through information learning.”

5.4 Robustness tests

To further mitigate potential issues of reverse causality in the moderating variables, this paper conducts robustness tests using one-period lagged moderating variables. Theoretically, one-period lagged director network centrality, environmental executive background, and disclosure of key audit matters satisfy the exogeneity condition (i.e. current-period investment efficiency cannot exert a reverse influence on the network structure or executive composition in the previous period), while still retaining their moderating effects on the relationship between current-period deviation and investment efficiency.

The moderating effect of one-period lagged director network centrality remains significantly positive (coefficient 9.7086, p < 0.01), while the moderating effect of one-period lagged environmental executive background also remains significant (coefficient 0.0033, p < 0.05), indicating that these results are robust across periods. However, the moderating effect of the one-period lagged KV index is no longer significant (coefficient−0.0100, p > 0.10), suggesting that the originally significant results for the KV index may have been influenced by reverse causality or omitted variables within the same period, and that its moderating effect does not hold once time-series endogeneity is controlled for. This limitation is relatively common in information disclosure research—as regulatory disclosures and corporate behavior are often adjusted simultaneously, identifying a fully exogenous and intertemporally stable moderating variable is inherently challenging.

The main text of this paper uses contemporaneous moderators as the benchmark, and supplements them with tests using one-period lagged moderators to alleviate concerns regarding reverse causality. In addition, the credibility of causal inference is enhanced from various perspectives through a range of robustness strategies, including alternative co-group definitions (geographical proximity), dual peer effects, and dynamic panel GMM estimates. The above results indicate that the positive effect of information analysis capability (background as an executive with environmental backgrounds) on enhancing the efficiency of the peer effect is highly robust; the moderating effects of information acquisition capability (centrality of the directors' network) and information transparency (KV index) exhibit a certain degree of sensitivity under different testing conditions, which may be related to the measurement methods of the proxy variables, sample characteristics and the time dimension, but do not affect the validity of the information learning mechanism as an overall explanatory framework.

6 Conclusions and implications

This study theoretically analyzes the mechanism through which corporate green investment peer effects influence GIE. Based on the sample of listed enterprises, an empirical analysis model is constructed to examine the impact of corporate green investment peer effects on GIE. Furthermore, the heterogeneous effects of the peer effect on corporate GIE were analyzed across different learning approaches. Finally, this study revealed the moderating roles of information transparency, access capability, and analysis capability in the relationship between corporate green investment peer effects and their impact on GIE. It clarifies the specific pathways for enhancing corporate green investment peer effects from a group decision-making perspective. The theoretical analysis and empirical results indicate that the effect of peer effects in corporate green investment on investment efficiency is conditional. Its positive enhancing effect depends on a “rational information learning” mechanism supported by enterprises' internal information-processing capabilities, rather than on simple investment convergence. Using the scale of corporate green investment or social responsibility scores as the criterion for distinguishing information learning methods, the peer effect among leading enterprises enhanced corporate GIE. However, for followers, when enterprises imitated the top one-third of the enterprises in green industry investments, the peer effect did not significantly improve GIE. When enterprises follow the top one-third of industry peers in terms of corporate social responsibility (CSR), the peer effect also significantly improves green investment efficiency, and this promoting effect (coefficient = 0.5660) even exceeds the peer effect within the CSR leader group itself (coefficient = 0.4520).

This suggests that followers' asymmetric imitation of different leader types differentially affects corporate GIE. In particular, enterprises pursue higher green investment levels to narrow the gap, leading to blind green investment and a “bandwagon phenomenon.” This triggers resource misallocation and creates a “crowding-out effect” on efficient green investments. By contrast, when following CSR leaders, followers—due to information disadvantages and insufficient experience in green investment—are more sensitive to peer deviations and exhibit stronger information cascades and learning motivations. CSR leaders, however, possess more sufficient information channels and independent decision-making frameworks, relying relatively less on peer deviations. Therefore, learning from CSR leaders constitutes rational information learning, through which followers can effectively optimize investment allocation. Finally, information transparency exerts a positive moderating effect on the peer effect, as improved transparency provides a higher-quality external information environment for information learning. However, this effect exhibits weak intertemporal robustness after controlling for reverse causality, suggesting that the role of information transparency may be influenced by endogeneity within the same period; greater caution is therefore required in the causal identification of its moderating effect. Enhanced information acquisition capabilities enable enterprises to swiftly obtain valuable information from their peer groups, thereby reducing information acquisition costs. This strengthens the positive effect of the green investment peer effect on corporate GIE. Information analysis capabilities enable decision-makers to accurately interpret standardized disclosures and to detect concealed negative information from peer enterprises, thereby facilitating better decision-making and improving GIE. The policy implications of the findings are as follows:

First, corporate green investment is shaped not solely by external regulations or firm-level attributes, but also by the actions of peer enterprises, reflecting a departure from the conventional view of isolated decision-making. The green investment activities of peer enterprises significantly influence increases in corporate green investment. Therefore, given the current inadequacy of green investment decision-making systems, effectively leveraging the guiding role of leading enterprises in green investment and promoting rational learning behaviors among enterprises are crucial for harnessing the impact of green investment peer effects on investment efficiency.

Second, complementarity in investment decisions compensates for the shortcomings of policy pressure and incentives, amplifying investment variability. Unlike the coercive intervention of policy systems, the peer effect serves as an implicit mechanism that promotes precautionary green investments by enterprises, thereby reducing resistance to policy implementation. When formulating policies related to green development and environmental subsidies, the state should consider the spillover effects of corporate green investments, utilizing them as supplementary measures within the policy framework to enhance the precision and effectiveness of such policies. For example, by establishing model and pilot enterprises, policies can be formulated to benefit enterprises in advantageous positions, thereby stimulating other enterprises in the same industry to actively participate in green investment activities. This leverages the implicit influence of peer effects to enhance the efficiency of corporate green investment. Policymakers should strengthen oversight mechanisms involving the government and the public to prevent the compliant “bandwagon phenomenon” in green investments.

Finally, corporate managers must maintain sensitivity to green investment clusters, establish scientific green investment decision-making systems, and objectively analyze the returns and losses associated with green investment clusters to mitigate decision-making risks in this area. Enterprises should enhance their capabilities of resource allocation, value creation, and multi-dimensional investment performance evaluation to mitigate the indiscriminate and excessive investment of green investments. Simultaneously, enterprises should enhance their information coordination capabilities through information disclosure, acquisition, and analysis. This fosters rational information learning behaviors, preventing enterprises from blindly following trends or imitating others. This enables them to make green investment decisions that better align with their actual conditions, thereby improving their GIE.

This study focuses on the role of information learning in the peer effect by examining the moderating effects of information disclosure, acquisition, and analysis on the relationship between peer behavior and corporate GIE. However, several limitations remain and warrant further refinement in subsequent research:

Firstly, it should be noted that the Deviation variable only captures the degree of convergence between a corporate green investment and its peer benchmark, and cannot fully distinguish whether this convergence stems from active information learning or passive institutional imitation. Although the moderating effects of information transparency, information acquisition capability, and information analysis capability provide indirect support for the information-learning mechanism, we cannot rule out the possibility that enterprises may also align their investment behavior with peers due to policy compliance pressures or competitive anxiety. Moreover, the reflection problem and common-shock problem inherent in peer-effect research are difficult to fully resolve with existing methods. Future research that leverages exogenous policy shocks—such as environmental regulations or green finance pilot programs rolled out in batches—would help disentangle these two mechanisms.

Second, limitations regarding the measurement of peer relationships. This paper defines peer enterprises based on industry classification, failing to fully capture more direct channels of information transmission such as supply chain linkages, interlocking director networks and geographical proximity. These network relationships may constitute closer channels for information learning than industry classification alone. Future research could incorporate social network analysis methods to construct multidimensional inter-enterprise networks, thereby providing a more detailed portrayal of the pathways through which green investment information spreads within peer groups.

Third, regarding the measurement limitations of the micro-level information learning process. The moderating variables in this paper reflect information capabilities at the enterprise level, rather than the micro-level learning process of “what specifically Enterprise A has learnt from Enterprise B.” Future research could utilize text analysis methods—such as semantic similarity analysis of statements relating to green investment in corporate annual reports and corporate social responsibility reports, or micro-level survey data—to observe or verify information learning behavior among enterprises more directly, thereby identifying the pathways through which information learning mechanisms operate with greater precision.

References

• 1

  AngriawanA.AbebeM. (2011). Chief executive background characteristics and environmental scanning emphasis: an empirical investigation. J. Bus. Strateg.28, 1–21. doi: 10.54155/jbs.28.1.1-21

  • CrossRef
  • Google Scholar

• 2

  AtifM.AlamM. S.HossainM. (2020). Firm sustainable investment: are female directors greener?Bus. Strategy Environ.29, 3449–3469. doi: 10.1002/bse.2588

  • CrossRef
  • Google Scholar

• 3

  BansalP.RothK. (2000). Why companies go green: a model of ecological responsiveness. Acad. Manag. J.43, 717–736. doi: 10.2307/1556363

  • CrossRef
  • Google Scholar

• 4

  BustamanteM. C.FrésardL. (2021). Does firm investment respond to peers' investment?Manag. Sci.67, 4703–4724. doi: 10.1287/mnsc.2020.3695

  • CrossRef
  • Google Scholar

• 5

  CaeyersB.FafchampsM. (2024). Exclusion bias and the estimation of peer effects. J. Hum. Res.61, 1120–11337R2. doi: 10.3368/jhr.1120-11337R2

  • CrossRef
  • Google Scholar

• 6

  ChenC.YanY.JiaX.WangT.ChaiM. (2024). The impact of executives' green experience on environmental, social, and governance (ESG) performance: evidence from China. J. Environ. Manag.366:121819. doi: 10.1016/j.jenvman.2024.121819

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 7

  ChenF.HopeO.-K.LiQ.WangX. (2011). Financial reporting quality and investment efficiency of private firms in emerging markets. Account. Rev.86, 1255–1288. doi: 10.2308/accr-10040

  • CrossRef
  • Google Scholar

• 8

  ChenS.NieY. (2023). Director network, governance of independent directors and efficiency of enterprise capital allocation. Entrepreneursh. Res. J.13, 163–184. doi: 10.1515/erj-2020-0201

  • CrossRef
  • Google Scholar

• 9

  ChenT.XieL.ZhangY. (2017). How does analysts' forecast quality relate to corporate investment efficiency?J. Corp. Financ.43, 217–240. doi: 10.1016/j.jcorpfin.2016.12.010

  • CrossRef
  • Google Scholar

• 10

  ChordiaT.RollR.SubrahmanyamA. (2001). Market liquidity and trading activity. J. Financ.56, 501–530. doi: 10.1111/0022-1082.00335

  • CrossRef
  • Google Scholar

• 11

  CongL. W.XiaoY. (2024). Information cascades and threshold implementation: Theory and an application to crowdfunding. J. Financ.79, 579–629. doi: 10.1111/jofi.13294

  • CrossRef
  • Google Scholar

• 12

  CostaH.DemmouL.FrancoG.LampS. (2024). The role of financing constraints and environmental policy on green investment. Econ. Lett.239:111741. doi: 10.1016/j.econlet.2024.111741

  • CrossRef
  • Google Scholar

• 13

  De OliveiraU. R.MenezesR. P.FernandesV. A. (2024). A systematic literature review on corporate sustainability: contributions, barriers, innovations and future possibilities. Environ. Dev. Sustain.26, 3045–3079. doi: 10.1007/s10668-023-02933-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 14

  ElaoudA.JarbouiA. (2017). Auditor specialization, accounting information quality and investment efficiency. Res. Int. Bus. Financ.42, 616–629. doi: 10.1016/j.ribaf.2017.07.006

  • CrossRef
  • Google Scholar

• 15

  EyraudL.ClementsB.WaneA. (2013). Green investment: trends and determinants. Energy Policy60, 852–865. doi: 10.1016/j.enpol.2013.04.039

  • CrossRef
  • Google Scholar

• 16

  FairhurstD.NamY. (2020). Corporate governance and financial peer effects. FinancManag.49, 235–263. doi: 10.1111/fima.12240

  • CrossRef
  • Google Scholar

• 17

  GrieserW.LeSageJ.ZekhniniM. (2022). Industry networks and the geography of firm behavior. Manag. Sci.68, 6163–6183. doi: 10.1287/mnsc.2021.4133

  • CrossRef
  • Google Scholar

• 18

  HeY.ZhangY.ZhangX. (2025). How does voluntary environmental information disclosure affect corporate green financing capability? evidence from Chinese heavy polluting industries. J. Environ. Manag.373:123495. doi: 10.1016/j.jenvman.2024.123495

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 19

  HongH.KubikJ. D.SteinJ. C. (2004). Social interaction and stock-market participation. J. Financ.59, 137–163. doi: 10.1111/j.1540-6261.2004.00629.x

  • CrossRef
  • Google Scholar

• 20

  HuY.YeY.YuX.PiaoX.HuangL.LiB. (2024). Managerial overconfidence and corporate information disclosure. Borsa Istanbul Rev.24, 263–279. doi: 10.1016/j.bir.2023.12.011

  • CrossRef
  • Google Scholar

• 21

  JiaF.LiG.LuX.XieS. (2021). CEO given names and corporate green investment. Emerg. Markets Rev.48:100808. doi: 10.1016/j.ememar.2021.100808

  • CrossRef
  • Google Scholar

• 22

  KhanJ.FengyunW.FawadA. (2025). Corporate environmentalism and value creation: investigating the role of shared independent directors in green technology adoption and financial performance. Corp. Soc. Respons. Environ. Manag.32, 118–137. doi: 10.1002/csr.2935

  • CrossRef
  • Google Scholar

• 23

  KimO.VerrecchiaR. E. (2001). The relation among disclosure, returns, and trading volume information. Account. Rev.76, 633–654. doi: 10.2308/accr.2001.76.4.633

  • CrossRef
  • Google Scholar

• 24

  KudybaS.FjermestadJ.DavenportT. (2020). A research model for identifying factors that drive effective decision-making and the future of work. J. Intellect. Capital21, 835–851. doi: 10.1108/JIC-05-2019-0130

  • CrossRef
  • Google Scholar

• 25

  KyleA. S. (1985). Continuous auctions and insider trading. Econometr. J. Econ. Soc.53, 1315–1335. doi: 10.2307/1913210

  • CrossRef
  • Google Scholar

• 26

  LiebermanM. B.AsabaS. (2006). Why do firms imitate each other?Acad. Manag. Rev.31, 366–385. doi: 10.5465/amr.2006.20208686

  • CrossRef
  • Google Scholar

• 27

  LiuL.ZhaoZ.ZhangM.ZhouD. (2022). Green investment efficiency in the Chinese energy sector: overinvestment or underinvestment?Energy Policy160:112694. doi: 10.1016/j.enpol.2021.112694

  • CrossRef
  • Google Scholar

• 28

  MarquisC.BattilanaJ. (2009). Acting globally but thinking locally? the enduring influence of local communities on organizations. Res. Org. Behav.29, 283–302. doi: 10.1016/j.riob.2009.06.001

  • CrossRef
  • Google Scholar

• 29

  MbanyeleW.HuangH.LiY.MuchenjeL. T.WangF. (2022). Corporate social responsibility and green innovation: Evidence from mandatory CSR disclosure laws. Econ. Lett.212:110322. doi: 10.1016/j.econlet.2022.110322

  • CrossRef
  • Google Scholar

• 30

  NagyR. L.HagspielV.KortP. M. (2021). Green capacity investment under subsidy withdrawal risk. Energy Econ.98:105259. doi: 10.1016/j.eneco.2021.105259

  • CrossRef
  • Google Scholar

• 31

  NingW.SaeedU. F.KongkuahM. (2025). Saving the environment in emerging markets: the synergistic roles of corporate ownership structure, financing strategy, and innovation capacity. Bus. Strat. Environ.34, 5114–5138. doi: 10.1002/bse.4245

  • CrossRef
  • Google Scholar

• 32

  PanX.WeiY.ManglaS. K.SongM. (2025). How do ESG practices promote collaboration on green investment for supply chain enterprises in China?Transp. Res. Part E Logist. Transp. Rev.198:104118. doi: 10.1016/j.tre.2025.104118

  • CrossRef
  • Google Scholar

• 33

  PengD.KongQ. (2024). Corporate green innovation under environmental regulation: The role of ESG ratings and greenwashing. Energy Econ.140:107971. doi: 10.1016/j.eneco.2024.107971

  • CrossRef
  • Google Scholar

• 34

  PengZ.LianY.ForsonJ. A. (2021). Peer effects in RandD investment policy: Evidence from China. Int. J. Financ. Econ.26, 4516–4533. doi: 10.1002/ijfe.2028

  • CrossRef
  • Google Scholar

• 35

  RossS. L.ShiZ. (2022). Measuring social interaction effects when instruments are weak. J. Bus. Econ. Stat.40, 995–1006. doi: 10.1080/07350015.2021.1895811

  • CrossRef
  • Google Scholar

• 36

  TallmanS.JenkinsM.HenryN.PinchS. (2004). Knowledge, clusters, and competitive advantage. Acad. Manag. Rev.29, 258–271. doi: 10.5465/amr.2004.12736089

  • CrossRef
  • Google Scholar

• 37

  TangP.FuS.YangS. (2019). Do peer firms affect corporate social performance?J. Clean. Product.239:118080. doi: 10.1016/j.jclepro.2019.118080

  • CrossRef
  • Google Scholar

• 38

  TumpA. N.PleskacT. J.KurversR. H. (2020). Wise or mad crowds? the cognitive mechanisms underlying information cascades. Sci. Adv.6:eabb0266. doi: 10.1126/sciadv.abb0266

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 39

  WangH.YeS.ChenH.YinJ. (2023). The impact of carbon emission trading policy on overcapacity of companies: evidence from China. Energy Econ.126:106929. doi: 10.1016/j.eneco.2023.106929

  • CrossRef
  • Google Scholar

• 40

  WangJ.FengT. (2023). Supply chain ethical leadership and green supply chain integration: a moderated mediation analysis. Int. J. Logist. Res. Appl.26, 1145–1171. doi: 10.1080/13675567.2021.2022640

  • CrossRef
  • Google Scholar

• 41

  WeiH.LiK. (2026). Green transformation along the supply chain: spillover effects from backward vertical partnerships. China J. Account. Res.19:100479. doi: 10.1016/j.cjar.2026.100479

  • CrossRef
  • Google Scholar

• 42

  XuJ.ZhongL.BiR.WangC. (2025a). Board networks and firms' technological innovation output: the moderating roles of shareholder networks and CEO networks. Syst.13:414. doi: 10.3390/systems13060414

  • CrossRef
  • Google Scholar

• 43

  XuR.HeC.LiY.KongL. (2024). Peer effects of corporate green innovation: evidence from China's listed firms. Financ. Res. Lett.61:105025. doi: 10.1016/j.frl.2024.105025

  • CrossRef
  • Google Scholar

• 44

  XuY.JiangL.WanY. (2025b). Jack of all trades or master of one? analyst specialization and corporate investment efficiency. Financ. Res. Lett.85:108327. doi: 10.1016/j.frl.2025.108327

  • CrossRef
  • Google Scholar

• 45

  ZhangH.LiuZ.ZhongW. (2024). Network centrality and dissent voting by directors on boards. Manag. Org. Rev.20, 60–84. doi: 10.1017/mor.2023.29

  • CrossRef
  • Google Scholar

• 46

  ZhangH.ShaoY.HanX.ChangH.-L. (2022a). A road towards ecological development in China: the nexus between green investment, natural resources, green technology innovation, and economic growth. Res. Policy77:102746. doi: 10.1016/j.resourpol.2022.102746

  • CrossRef
  • Google Scholar

• 47

  ZhangH.XuT.FengC. (2022b). Does public participation promote environmental efficiency? evidence from a quasi-natural experiment of environmental information disclosure in China. Energy Econ.108:105871. doi: 10.1016/j.eneco.2022.105871

  • CrossRef
  • Google Scholar

• 48

  ZhaoT.WangH. (2024). The industry peer effect of enterprise ESG performance: the moderating effect of customer concentration. International review of economics and finance92, 1499–1525. doi: 10.1016/j.iref.2024.03.018

  • CrossRef
  • Google Scholar

• 49

  ZhuC.LiN.MaJ. (2023). Environmental backgrounds of CEOs and corporate environmental management information disclosure: the mediating effects of financing constraints and media attention. Corp. Soc. Res. Environ. Manag.30, 2885–2905. doi: 10.1002/csr.2522

  • CrossRef
  • Google Scholar

• 50

  ZhuJ.ZhiW.FangY. (2025). Ethical leadership, organizational learning, and corporate ESG performance: a moderated mediation model. Int. Rev. Econ. Financ.98:103966. doi: 10.1016/j.iref.2025.103966

  • CrossRef
  • Google Scholar