A Systems Perspective on Corporate Social Responsibility Decoupling and Investment Efficiency: Evidence from Chinese Listed Firms

Abstract

This study examines the impact of corporate social responsibility (CSR) decoupling on investment efficiency through the lens of systems thinking, using 34,143 firm-year observations from Chinese listed firms between 2009 and 2022. CSR decoupling is conceptualized as a systemic misalignment between two interrelated governance subsystems: the externally facing legitimacy subsystem and the internally embedded strategic action subsystem. Drawing on legitimacy theory and systems thinking, we find that CSR decoupling significantly reduces investment efficiency, primarily through overinvestment, with no consistent evidence of underinvestment. Furthermore, this effect is amplified in tightly coupled supply chain systems and is especially pronounced in foreign-owned firms. The findings contribute to the integration of systems thinking into CSR and corporate governance research, emphasizing the role of structural coupling strength in shaping the consequences of symbolic–substantive misalignment. The study also offers managerial and policy implications for improving the alignment between external CSR communication and internal strategic execution to enhance investment discipline and long-term value creation.

1. Introduction

In recent years, the intersection of corporate governance, environmental sustainability, and corporate social responsibility (CSR) has received growing attention. In China, recent policy initiatives have accelerated the adoption of mandatory sustainability reporting, positioning CSR disclosure as a critical signal of corporate legitimacy and ethical commitment [1]. While such disclosures are intended to enhance transparency and reduce information asymmetry [2], concerns have emerged regarding the authenticity of these reports. Specifically, stakeholders question whether reported CSR efforts genuinely reflect substantive practices, giving rise to the phenomenon of CSR decoupling—a misalignment between symbolic CSR communication and substantive CSR performance [3].

CSR decoupling occurs when firms engage in CSR symbolically to satisfy external legitimacy pressures or regulatory requirements, without embedding these practices into their strategic and operational core [4]. While such symbolic actions may offer short-term legitimacy benefits, they risk damaging long-term stakeholder trust and undermining firm value if not backed by real change [5].

This study conceptualizes CSR decoupling as a form of systemic misalignment between two interrelated governance subsystems: the externally facing legitimacy subsystem and the internally embedded strategic action subsystem. Firms operate within multi-layered systems—comprising institutional pressures, supply chain networks, as well as internal strategies and operations—and CSR decoupling reflects a failure in coordinating signals and actions across these layers. This conceptualization strengthens the logical link between legitimacy pressures and governance failures in capital allocation. This misalignment can distort stakeholder perceptions, disrupt capital allocation, and erode long-term firm value.

We adopt systems thinking to examine how the misalignment between external and internal CSR actions—measured as CSR decoupling—affects corporate investment efficiency. From a governance perspective, CSR decoupling may impair firms’ ability to allocate capital effectively, by weakening stakeholder trust and increasing informational noise. Prior research has shown that CSR can influence firms’ capital allocation. For example, Bhandari and Javakhadze [6] find that CSR engagement weakens the sensitivity of corporate investment to growth opportunities, suggesting potential distortions in investment efficiency. However, limited studies have specifically examined how the misalignment between external CSR signaling and internal CSR implementation influences capital allocation efficiency, particularly from a systems perspective [1,7]. Unlike prior studies that focus on aggregated CSR scores, this paper explicitly measures the misalignment between symbolic signaling and substantive implementation. Understanding this misalignment is particularly important in emerging markets like China, where regulatory shifts and stakeholder expectations are rapidly evolving [8,9]. The broader impact of CSR on long-term corporate strategy and investment decisions—often reflected in actual CSR actions—requires further investigation within China’s regulatory environment [10]. Understanding the inconsistencies between external CSR disclosures and internal CSR performance in this systems context is crucial for assessing its influence on corporate strategy and evaluating its potential to foster sustainable value creation.

The systems lens emphasizes not only internal–external alignment, but also structural coupling within inter-firm networks. Supply chain concentration thus provides a natural context to examine how system-level interdependencies magnify the effects of CSR decoupling. Therefore, we explore how supply chain concentration—the extent to which a firm is dependent on a limited number of dominant customers or suppliers—moderates the effect of CSR decoupling. From a systems perspective, we interpret supply chain concentration as a proxy for system coupling strength: higher concentration implies tighter interdependence and more transparent information flows within the supply chain network. Such structural coupling may either help align CSR across firms or magnify the visibility and consequences of CSR inconsistencies [11,12]. Accordingly, we test whether stronger coupling amplifies the performance effects of CSR decoupling on investment efficiency.

To empirically evaluate these relationships, we construct a novel measure of CSR decoupling using third-party CSR evaluation data, capturing the absolute gap between external CSR disclosures and internal CSR practices. Based on a large sample of 34,143 firm-year observations from Chinese listed firms over the period 2009–2022, we document three key findings. First, CSR decoupling significantly reduces investment efficiency, primarily through overinvestment. Second, this negative effect is more pronounced in firms with high supply chain concentration, reflecting the role of system coupling strength in amplifying misalignment. Third, heterogeneity analysis reveals that CSR decoupling negatively affects investment efficiency across different ownership structures, with particularly strong effects in foreign-owned firms, possibly due to dual legitimacy expectations and weaker integration with domestic regulatory systems. These findings are robust to firm- and region-level controls, as well as industry and year fixed effects, and hold across multiple robustness checks.

This study contributes to the literature in several ways. Theoretically, we frame CSR decoupling as a form of subsystem misalignment within a firm’s broader governance system, offering a systems-based interpretation of CSR inconsistencies. Empirically, we provide robust evidence that CSR decoupling undermines investment efficiency, particularly under conditions of strong system coupling strength. Practically, our findings emphasize the need for firms to strengthen coordination between CSR disclosure and implementation and for regulators to ensure alignment between symbolic commitments and substantive actions. Together, these insights contribute to a more holistic understanding of CSR not just as a reporting obligation, but as a multi-layered governance system shaped by both internal and external dynamics.

2. Literature Background and Hypothesis

2.1. CSR Decoupling and Its Theoretical Foundations

Prior studies have consistently shown that non-financial information plays a crucial role in reducing information asymmetry and offering a more comprehensive view of a firm’s value and performance [13]. CSR disclosure, as a key form of non-financial information, serves as an important signal to stakeholders about a firm’s commitment to sustainability and ethical practices [14]. While the early literature emphasized CSR’s legitimacy-enhancing function, debates persist over its tangible contribution to financial outcomes and firm value [15].

In early 2024, China’s corporate governance landscape was reshaped by the introduction of mandatory sustainability disclosure rules issued by the China Securities Regulatory Commission and major stock exchanges. These guidelines mandate listed companies—especially those included in key indices or with dual listings—to publicly disclose sustainability-related data. This regulatory shift reflects growing institutional pressure for transparency and accountability, embedding CSR reporting more deeply into firms’ governance structures while simultaneously raising concerns about its authenticity.

CSR practices are grounded in legitimacy theory, which posits that a firm’s long-term survival depends on its alignment with societal norms, values, and expectations [16]. To meet institutional pressures and gain legitimacy, firms take actions to align with sustainability standards, such as implementing CSR initiatives and adhering to environmental, social, and governance (ESG) criteria [17]. In doing so, companies communicate their compliance with societal expectations, reinforcing their “license to operate”. Firms employ CSR activities as a strategy to mitigate risks that could jeopardize their legitimacy and reduce information asymmetry with stakeholders, particularly within the context of sustainable development [18,19]. Research has shown that CSR disclosures foster trust among stakeholders, which can lead to reduced operational costs, increased efficiency, and enhanced reputational capital [20,21].

While CSR offers legitimacy benefits, an ongoing question remains: Do CSR activities genuinely promote sustainability, or do they obscure profitability and complicate investor decision-making [22]? Some studies, such as Chen et al. [23], have found that CSR spending may reduce profitability in the short term while improving environmental outcomes. Most empirical evidence support a positive relationship between CSR and financial performance. For instance, Al Amosh et al. [24] demonstrated that improved CSR reporting enhances financial performance in Levant capital markets, and Kalia and Aggarwal [25] found similar effects in healthcare firms in developed economies. Additionally, firms with strong ESG performance are less prone to stock price crashes [26], and tend to outperform peers in both market and accounting metrics over the long term [27].

However, emerging evidence indicates that CSR disclosure may not uniformly enhance firm value. Zamir et al. [28] provide evidence that CSR disclosures in emerging markets can reduce informational gaps and strengthen stakeholders’ perception of transparency without necessarily improving intrinsic resource allocation efficiency. Additional evidence from emerging markets also highlights the dual nature of CSR reporting: while accounting-based measures may show improvements, market-based indicators can deteriorate, suggesting that CSR disclosures often serve symbolic purposes without enhancing intrinsic firm value [29]. Such findings reveal the dual role of CSR disclosure: while it can reduce information gaps and support financing, it may also remain symbolic if not backed by substantive practices, reinforcing the risk of CSR decoupling.

CSR decoupling, therefore, has become a central concern. It is defined as the misalignment between external CSR communication and internal CSR implementation [30,31]. From a legitimacy perspective, external CSR activities emphasize policy compliance and public image management [32], while from a strategic perspective, substantive CSR actions reflect genuine efforts to embed sustainability into core operations [13]. Together, these two orientations form interdependent subsystems within the broader governance framework: the external legitimacy subsystem, which comprises policy compliance, symbolic reporting, and media visibility, and the internal capability subsystem, which reflects the actual integration of sustainability practices into business strategies and operations.

Coupling in systems theory refers to the dynamic interdependence between subsystems [33], maintained through feedback loops that circulate information and resources. When applied to CSR, the alignment of the external legitimacy and internal capability subsystems is essential for system coherence. CSR decoupling emerges when these feedback mechanisms weaken or collapse, leading to persistent misalignment between symbolic communication and substantive implementation. Figure 1 illustrates this conceptual systems model of CSR decoupling.

Prior research shows that such decoupling frequently manifests in selective CSR disclosure, reflecting institutional pressures that prioritize legitimacy over operational integration [19,32]. Firms may emphasize symbolic CSR to meet external expectations while failing to translate those commitments into substantive action [30,34]. Governance mechanisms such as board-level CSR committees are critical in realigning subsystems, while network-level structures such as supply chain concentration can amplify the visibility and consequences of misalignments. When stakeholders detect inconsistencies, the informational noise undermines investor confidence, restricts access to capital markets, and impairs firms’ ability to allocate resources efficiently [35,36].

Recent evidence further highlights the importance of distinguishing between symbolic and substantive CSR actions in shaping investor perceptions. Amel-Zadeh and Serafeim [37] highlight that ESG disclosures are most valuable when they reflect substantive integration into corporate strategy and operations. Symbolic or unstandardized disclosures, by contrast, are perceived as less reliable and less decision-useful, thereby weakening their capacity to guide capital allocation. This evidence reinforces our systems perspective: when the external legitimacy subsystem (symbolic CSR communication) is decoupled from the internal strategic subsystem (substantive CSR implementation), the informational feedback loop breaks down, leading investors to misinterpret a firm’s sustainability trajectory. Consequently, CSR decoupling not only distorts stakeholder evaluations but also undermines capital allocation efficiency, particularly in market environments where CSR signals are increasingly used to benchmark long-term corporate value.

In this study, CSR decoupling is operationalized as the absolute difference between external CSR actions and internal CSR actions. Although this measure cannot distinguish between over-reporting and under-reporting, it captures the degree of systemic misalignment that lies at the heart of our research focus. From a systems perspective, the severity of misalignment—rather than its direction—is what determines whether coordination between subsystems breaks down. By emphasizing the magnitude, this measure reflects the extent to which CSR disclosures diverge from substantive practices, thereby aligning with our systems-based conceptualization of governance failure. Importantly, by capturing the degree of subsystem misalignment, this approach links directly to our central outcome of interest—investment efficiency—since higher misalignment increases informational noise and weakens both market perceptions and internal resource allocation. This enables a system-level analysis of how governance failures disrupt capital allocation and affect long-term corporate performance. In essence, such subsystem misalignments represent a breakdown of system coupling, which directly obstructs efficient resource allocation and ultimately undermines investment efficiency.

2.2. CSR Decoupling and Investment Efficiency

Building on this operationalization, the degree of CSR decoupling serves not only as an indicator of governance misalignment but also as a direct mechanism through which capital allocation can be distorted. A growing body of literature suggests that well-executed CSR practices contribute to better capital allocation and enhanced investment efficiency by mitigating information asymmetry and fostering trust among stakeholders [38,39,40]. Firms with strong CSR performance and transparent reporting attract capital more easily by presenting a positive corporate image and demonstrating sustainable long-term competitiveness, thereby reducing the risks of adverse selection in capital markets [41].

However, CSR decoupling fundamentally undermines this mechanism by creating informational distortions that mislead investors and decision-makers. As Amel-Zadeh and Serafeim [37] show, investors increasingly rely on ESG disclosures to inform capital allocation, but these assessments are only reliable when disclosures reflect substantive integration into strategy and operations. When disclosures are merely symbolic, investors struggle to differentiate genuine sustainability commitments from impression management, which weakens their ability to correctly price firm value and allocate resources.

This distortion intensifies information asymmetry and amplifies agency problems such as managerial opportunism and moral hazard. In the short term, symbolic CSR may still provide firms with temporary legitimacy benefits—such as alleviating stakeholder pressure, improving public image, and easing financing constraints [34]. For instance, greenwashing practices allow firms to secure favorable lending terms by leveraging ESG labels and selective media framing [42]. Yet, over time, the misalignment between symbolic and substantive CSR actions disrupts the accuracy of external market signals and the efficiency of internal capital allocation, leading to persistent inefficiencies [43].

From a systems thinking perspective, CSR decoupling disrupts the internal coordination and external feedback loops that support effective investment decision-making [31]. Symbolic actions misrepresent operational reality, causing investors and analysts to overestimate a firm’s sustainability trajectory. This results in a systematic bias toward overinvestment in unproductive projects, while at the same time weakening internal discipline to allocate resources to genuinely productive opportunities [44]. Consequently, CSR decoupling impairs overall capital allocation efficiency, with particularly pronounced effects on overinvestment rather than underinvestment.

While prior studies have documented the financial risks of CSR decoupling, they have largely relied on backward-looking, accounting-based indicators such as return on assets (ROA) or return on equity [36]. For instance, Walker and Wan [45] found that symbolic actions in corporate environmental responsibility are negatively associated with financial performance. Liu et al. [46] highlighted that firms catering to distracted investors tend to prioritize external ESG activities over internal ones, resulting in high ESG decoupling and reduced valuations. Although useful for measuring operational outcomes, these metrics fail to fully capture how inconsistencies in CSR affect external perceptions, market value, or strategic foresight over the long-term [47]. This limits their applicability in assessing CSR’s long-term economic implications. Furthermore, CSR decoupling is more than a financial distortion—it is a governance misalignment between external legitimacy claims and internal resource deployment.

By contrast, investment efficiency serves as a forward-looking, market-based indicator that reflects how effectively firms allocate capital to projects that generate sustainable returns [48,49]. Unlike static profitability ratios, investment efficiency incorporates both market perceptions and firm-level strategic discipline, making it a more appropriate metric for understanding how CSR decoupling disrupts the coordination between internal governance and external stakeholder expectations [43].

Given this theoretical and empirical context, we conceptualize CSR decoupling as a structural fault in the firm’s governance system, a systemic breakdown in governance feedback loops, which directly translates into inefficient investment outcomes. CSR decoupling can exacerbate agency problems by creating distorted signals that management may strategically exploit. When symbolic CSR disclosures succeed in attracting legitimacy, firms often gain easier access to external financing. However, this financing may be directed toward low-return or opportunistic projects, increasing the risk of overinvestment. Thus, CSR decoupling not only reflects symbolic reporting but also actively contributes to excessive capital allocation through inflated legitimacy and reduced investor scrutiny.

At the same time, CSR decoupling can also heighten informational uncertainty, leading to constrained financing and underinvestment. When stakeholders perceive inconsistencies between symbolic disclosures and substantive practices, trust erodes, financing costs rise, and firms may lose access to valuable external resources. Internally, misalignment between governance subsystems may also prompt managers to adopt overly conservative investment strategies to avoid external criticism. Consequently, CSR decoupling may discourage firms from pursuing value-enhancing projects, resulting in underinvestment.

Accordingly, we propose the following hypotheses:

H1: 

CSR decoupling is negatively associated with investment efficiency.

H1a: 

CSR decoupling increases over-investment.

H1b: 

CSR decoupling increases under-investment.

3. Data and Descriptive Statistics

3.1. Sample and Data Sources

This study constructs a panel of Chinese A-share listed firms from 2009 to 2022, yielding 34,143 firm-year observations. The sample selection follows standard practices in the CSR and investment efficiency literature and applies the following exclusion criteria: (1) firms in the financial, educational, and real estate sectors. These industries are characterized by highly distinctive regulatory regimes, governance models, and CSR disclosure practices. For example, financial firms operate under stringent capital adequacy and risk management rules, education companies are heavily policy-driven with strong public service attributes, and real estate firms are subject to cyclical regulatory interventions and unique disclosure norms. Including such firms would introduce structural biases and reduce the comparability of results; therefore, their exclusion enhances the robustness and validity of our analysis; (2) firms subject to special treatment (ST) or particular treatment (PT) status, as these are typically associated with financial distress or operational irregularities; (3) firms listed for fewer than two years, to avoid immature governance bias; and (4) observations with missing values for key financial or CSR data.

CSR data are obtained from the Sino-Securities Index (SNSI) database, which offers comprehensive ESG ratings and CSR evaluation indicators for all A-share firms [50]. The SNSI database is widely recognized for its longitudinal coverage and methodological consistency, making it particularly suitable for long-term CSR decoupling analysis. Firm-level financial and governance variables are sourced from the China Stock Market and Accounting Research (CSMAR) database, a leading platform for academic research in corporate finance and accounting in China [51].

3.2. Measurement of CSR Decoupling

The core explanatory variable, CSR decoupling (CSR_D), captures the misalignment between external CSR communications and internal CSR practices, operationalized as the absolute difference between standardized scores of external and internal CSR actions, following Hawn and Ioannou [52].

External CSR actions are symbolic, aiming to satisfy regulatory expectations and shape public perception, often without inducing substantive organizational change. These actions are designed for external stakeholders such as regulators, customers, and the general public. In contrast, internal CSR actions are substantive in nature, embedded in corporate strategy, operations, and governance processes. They engage internal stakeholders such as employees and reflect the firm’s actual commitment to sustainability.

Unlike previous approaches combining multiple data sources—for example, using Bloomberg scores for CSR reporting and Refinitiv for CSR performance [35,53], which may introduce measurement bias [54], we rely solely on SNSI data. Using the internal–external classification scheme from Hawn and Ioannou [52], indicators are grouped into two subsystems: the external CSR actions and the internal CSR actions.

3.2.1. External CSR Actions

The external CSR component includes seven indicators aimed at signaling responsibility to external stakeholders: (1) EnvReduce index measures whether a firm discloses pollution reduction and governance statements for six pollutants (e.g., gas emissions, water, dust). A score of 0 indicates no disclosure, 1 indicates qualitative disclosure, and 2 indicates quantitative disclosure. The total score ranges from 0 to 12, representing the sum of disclosures for each pollutant. (2) GRI index is a binary variable (0 or 1) that indicates whether a firm follows GRI (Global Reporting Initiative) for its CSR disclosure. (3) ShareholdersProtection, CreditorProtection, SupplierProtection, CustomerProtection, PublicRelations are five binary variables (0 or 1) that indicate whether the firm addresses the rights of shareholders, creditors, suppliers, customers, and public relations, respectively. The composite score for external CSR actions is computed as the sum of the seven indicators above.

3.2.2. Internal CSR Actions

The internal CSR component consists of ten indicators reflecting actual sustainability integration into internal governance and operations: (1) Env_Emission index represents whether a firm discloses emissions information for six pollutants. Scores are assigned as 0 for no disclosure, 1 for qualitative disclosure, and 2 for quantitative disclosure, with a total possible score ranging from 0 to 12. (2) WorkSafety and StaffProtection, and SystemConstruction indexes are three binary variables (0 or 1) that indicate whether the firm addresses workplace safety, employee rights protection, and the development of CSR systems. (3) ISO 140012 and ISO 900012 are two binary variables (0 or 1) that indicate whether the firm holds these environmental and quality management certifications. (4) There are four indexes related to the composition of the firm’s board structure: the proportion of female executives, the proportion of female board members, the proportion of female management, and the proportion of independent directors. The composite score for internal CSR actions is the sum of all ten indicators, with proportion-based measures standardized prior to aggregation.

Both external and internal CSR scores are first standardized (z-scores) to ensure comparability in scale. Following prior CSR literature [52], we adopt an equal-weight aggregation of indicators when constructing composite scores. This approach has the advantages of transparency, replicability, and consistency with earlier studies, while also avoiding the introduction of arbitrary weights in the absence of a universally accepted standard. The CSR decoupling index (CSR_D) is calculated as the absolute difference between the standardized external and internal CSR action scores. This absolute difference captures the degree of misalignment between a firm’s external symbolic CSR communication and its internal substantive CSR practices. A larger CSR_D value indicates a greater degree of decoupling, implying weaker integration between CSR disclosure and operational commitment.

3.3. Investment Efficiency Measures

There is no universally accepted metric for measuring investment efficiency. However, a widely adopted approach in the literature is to assess efficiency by evaluating the deviation between a firm’s actual investment and the level predicted by its growth opportunities. This deviation is interpreted as a proxy for inefficiency, with larger residuals indicating greater divergence from optimal investment behavior [43,55,56].

Consistent with this stream of research, we employ two established models to construct our dependent variables: (1) Inv_E: Based on the framework proposed by Biddle, Hilary and Verdi [56], this measure captures investment inefficiency as the absolute value of the residual term from a firm-level investment prediction model. The model regresses actual investment on growth opportunity proxies (e.g., lagged sales growth, return on assets), and the residual represents unexplained deviation—interpreted as inefficiency. (2) Inv_EC: As a robustness check, we adopt an alternative specification developed by Chen, Hope, Li and Wang [43], which includes additional firm characteristics to better capture expected investment behavior. The methodology similarly defines investment inefficiency as the residual from an expected investment model.

Both proxies quantify how far a firm’s investment deviates from its predicted optimal level. In this study, higher values of Inv_E or Inv_EC indicate greater investment inefficiency, reflecting either overinvestment or underinvestment. Conversely, smaller values imply more efficient capital allocation. For consistency with prior literature, we refer to these measures as “investment efficiency” throughout.

3.4. Control Variables

To account for firm-level heterogeneity and mitigate potential omitted variable bias, we include a set of control variables that have been widely used. These variables are defined as follows:

(1) Firm age (Age), refers to the number of years since a company was established; (2) Firm Size (Size), measured by the firm’s total assets; (3) Firm growth (Growth), measured by the growth of operating revenue; (4) Leverage (Lev), measured by a firm’s debt deflated by total assets, used to assess the financial risk of a company; (5) Board size (B_Size), defined as the total number of members on a company’s board of directors. (6) Regulation intensity of governance (R_Env), measured by the proportion of investment in waste gas and wastewater pollution control in the firm’s location relative to the region’s industrial output value for the given year.

To mitigate the influence of outliers, all continuous variables are winsorized at the 1st and 99th percentiles. Additionally, variables that exhibit skewed distributions are log-transformed to improve normality and enhance the robustness of regression estimates.

3.5. Descriptive Statistics

Table 1. Summary statistics.

Variables	N	Mean	Std. Dev.	Min	Max
Inv_E	34,143	0.044	0.045	0.000	0.445
CSR_D	34,143	−0.007	0.610	−2.607	3.236
Age	34,143	2.894	0.350	0.693	4.174
Size	34,143	22.196	1.265	19.531	26.386
Growth	34,143	0.171	0.474	−6.629	9.137
Lev	34,143	0.424	0.204	0.007	1.957
B_Size	34,143	2.241	0.179	0.000	2.944
R_Env	34,143	0.002	0.002	0.000	0.028

presents descriptive statistics. The mean investment inefficiency (Inv_E) is 0.044, suggesting that, on average, 4.4% of total assets are invested inefficiently. The mean CSR_D is −0.007, indicating a slight tendency for overstated external CSR relative to internal implementation. Substantial variation in CSR_D (−2.607 to 3.236) supports its suitability for testing heterogeneity. Variance Inflation Factor (VIF) tests were also conducted to assess multicollinearity (test results are shown in Appendix A). The results show that all VIF values are close to 1, with a mean VIF of 1.12, well below the conventional threshold of 10. This indicates that multicollinearity is not a concern in our models.

4. Empirical Model and Results

4.1. Research Design

We estimate panel regressions to test whether higher CSR decoupling is associated with reduced investment efficiency. All models control for industry and year fixed effects to address sector-specific heterogeneity and macroeconomic fluctuations. The baseline model is specified as follows:

$$\begin{array}{c}{Inv\_E}_{it}={\alpha }_0+{\beta }_1{CSR\_D}_{it}+{\beta }_2{Age}_{it}+{\beta }_3{Growth}_{it}+{\beta }_4{Size}_{it}+{\beta }_5{Lev}_{it}+{\beta }_6{B\_Size}_{it}\\ +{\beta }_7{R\_Env}_{it}+\sum Industry+\sum Year+{\epsilon }_{it}\end{array}$$

(1)

where ${Inv\_E}_{it}$ denotes the investment efficiency of firm $i$ in year $t$, measured following Biddle, Hilary and Verdi [56]. The key explanatory variable, CSR decoupling (CSR_D), is defined in Section 3.2. Control variables (Section 3.4) include firm age, growth, size, leverage, board size, and regional environmental regulation intensity. We hypothesize a positive association between CSR_D and investment inefficiency (Inv_E), indicating that symbolic CSR actions decoupled from substantive implementation distort capital allocation.

Table 2. Matrix of correlation.

Variables	(1)	(2)	(3)	(4)	(5)	(7)	(8)	(9)
(1) Inv_E	1.000
(2) CSR_D	0.034	1.000
(3) Age	−0.093	0.003	1.000
(4) Growth	0.019	0.002	−0.021	1.000
(5) Size	−0.041	−0.095	0.163	−0.056	1.000
(6) Lev	0.025	−0.034	0.126	0.000	0.450	1.000
(7) B_Size	−0.009	−0.025	−0.001	−0.019	0.260	0.146	1.000
(8) R_Env	0.024	−0.014	−0.131	0.008	−0.003	0.064	0.105	1.000

reports the Pearson correlation coefficients among the main variables. As expected, CSR_D exhibits a positive correlation with Inv_E, consistent with our hypothesis. Additionally, no severe multicollinearity is detected.

4.2. Main Regression Results

Table 3. Main regression results.

	(1)	(2)	(3)	(4)	(5)
Variables	Inv_E	Inv_E	Inv_E	Inv_Over	Inv_Under
CSR_D	0.0014 ***			0.0029 **	0.0004
	(3.6282)			(3.0213)	(1.7880)
CSR_E		−0.0013 ***
		(−3.8877)
CSR_I			−0.0007
			(−1.3958)
Controls	yes	yes	yes	yes	yes
Industry FE	yes	yes	yes	yes	yes
Year FE	yes	yes	yes	yes	yes
Adj. R2	0.0473	0.0474	0.0470	0.0410	0.1531
N	34,143	34,143	34,143	12,034	22,109

All models control for firm age (Age), firm growth (Growth), firm size (Size), leverage (Lev), board size (B_Size), and regional environmental regulation intensity (R_Env), and Controls have the same definitions across all following tables. The full results of main regression are displayed in Appendix A. t-statistics in parentheses; ***, ** denote 1% and 5% significance levels.

reports the regression results of the relation between CSR_D and Inv_E. Model (1) reports the baseline regression. The coefficient of CSR_D is positive and statistically significant at the 1% level, supporting Hypothesis 1 and indicating that greater CSR decoupling increases investment inefficiency.

Model (2) investigates the influence of external CSR actions (CSR_E) on investment efficiency. The coefficient is negative and significant at the 1% level, indicating that stronger external CSR efforts are associated with lower investment inefficiency. This result aligns with prior research, e.g., Benlemlih and Bitar [41], which finds that transparent and credible CSR disclosures enhance investor trust and reduce adverse selection, thereby improving capital allocation efficiency.

Model (3) examines the effect of internal CSR actions (CSR_I). Although the coefficient is negative, it is not statistically significant, implying no robust direct effect of internal CSR actions on investment efficiency. While this may suggest a potential efficiency benefit from substantive CSR practices, the evidence remains inconclusive. Several factors may explain this null result. First, as internal CSR is proxied by heterogeneous indicators (e.g., employee welfare, internal governance, and environmental practices), which may not fully capture the depth or quality of substantive CSR engagement. Second, the benefits of internal CSR are often long-term in nature: initiatives such as green technology investment typically require years to translate into measurable financial outcomes, which reduces short-term statistical significance. Third, internal CSR practices are less visible to outside investors compared to external disclosures, making their value harder to incorporate into capital allocation decisions. Finally, in emerging markets like China, institutional mechanisms for monitoring and rewarding substantive CSR remain underdeveloped, further constraining the ability of internal CSR to directly improve investment efficiency. This is consistent with Walker and Wan [45], who argue that substantive environmental initiatives may not yield immediate financial benefits, and their effect may be contingent on broader contextual factors.

Models (4) and (5) explore the channels of investment inefficiency by distinguishing between overinvestment and underinvestment. Model (4) shows that CSR decoupling significantly increases overinvestment inefficiencies, with a positive coefficient significant at the 5% level. This suggests that firms exhibiting higher levels of CSR decoupling are more prone to allocate capital beyond optimal levels, possibly due to increased agency problems and distorted stakeholder perceptions. In contrast, Model (5) indicates no significant effect of CSR decoupling on underinvestment, offering no support for Hypothesis 1b (H1b). This asymmetry can be explained through an agency theory lens. When managers face legitimacy pressures but lack substantive CSR engagement, they may use symbolic disclosures to attract external financing and ease stakeholder scrutiny. With enhanced access to capital and weaker monitoring, managers are incentivized to channel funds into projects that maximize private benefits (e.g., political connections, or reputation management), even when such projects yield limited returns. This opportunistic behavior naturally leads to excessive rather than insufficient investment. By contrast, underinvestment is less likely to result directly from CSR decoupling, because symbolic disclosures rarely constrain financing capacity—instead, they inflate it by projecting a favorable but misleading image to investors. These findings indicate that CSR decoupling primarily contributes to excessive rather than insufficient capital allocation.

These results reinforce the importance of alignment between symbolic and substantive CSR actions. While both internal and external CSR initiatives have the potential to enhance firm legitimacy and financial outcomes, a disconnect between them can erode stakeholder trust, heighten information asymmetry, and diminish capital allocation efficiency [57]. In this regard, reliable CSR disclosures and the use of verifiable metrics are essential to reduce opacity and strengthen stakeholder confidence [54]. Moreover, lower levels of CSR decoupling may improve firms’ access to external financing, reduce financing costs, and ultimately enhance investment efficiency [58,59].

In summary, consistent and credible CSR engagement is a valuable tool for improving investment efficiency, particularly by reducing overinvestment. However, such engagement appears to have limited impact on alleviating underinvestment, underscoring the asymmetric nature of CSR decoupling’s effects on firm behavior.

4.3. Robustness Tests

To ensure the robustness of our baseline findings, we conduct several supplementary analyses, including instrumental variable estimation, alternative variable specifications, sample restrictions, and introducing lagged variables.

4.3.1. Instrumental Variable Estimation

To further address potential endogeneity concerns, we adopt an instrumental variable (IV) approach using the lagged industry–province–year average CSR decoupling of peer firms (excluding the focal firm) as an instrument for CSR decoupling. This approach is consistent with prior studies that employ peer-based or regional average measures as instruments for firm-level governance constructs [60]. This peer-based instrument captures exogenous variation in CSR decoupling stemming from local institutional and industry practices, while being plausibly exogenous to the focal firm’s investment efficiency. We implement the IV estimation with firm fixed effects and year fixed effects, clustering standard errors at the firm level.

The IV results are shown in

Table 4. Instrumental variable regression results.

	(1)	(2)	(3)
Variables	Inv_E	Inv_Over	Inv_Under
CSR_D	0.0160 ***	0.0349 **	0.0116 ***
	(3.8281)	(3.2452)	(4.0901)
Controls	yes	yes	yes
Industry FE	yes	yes	yes
Year FE	yes	yes	yes
N	32,815	11,575	21,240

t-statistics in parentheses; ***, ** denote 1% and 5% significance levels.

, which confirm the robustness of our baseline findings. CSR decoupling is positively and significantly associated with investment inefficiency at the 1% level, with a stronger effect on overinvestment, while the effect on underinvestment also becomes significant. These findings indicate that CSR decoupling not only reflects symbolic reporting but also actively distorts capital allocation, primarily by inflating legitimacy and easing access to external financing, which encourages overinvestment. At the same time, misalignment erodes stakeholder trust and may constrain financing channels, thereby contributing to underinvestment.

4.3.2. Alternative Measures

We replace the baseline variables with alternative measures of investment efficiency and CSR decoupling.

Table 5. Robustness test by using alternative variables.

	(1)	(2)	(3)	(4)
Variables	Inv_E	Inv_EC	Inv_E	Inv_EC
CSR_D	0.0014 ***	0.0013 ***
	(3.6282)	(3.3727)
CSR_DA			0.0007 **	0.0008 *
			(3.0791)	(2.7495)
Controls	yes	yes	yes	yes
Industry FE	yes	yes	yes	yes
Year FE	yes	yes	yes	yes
Adj. R2	0.0473	0.0454	0.0477	0.0463
N	34,143	34,143	34,125	34,125

t-statistics in parentheses; ***, **, * denote 1%, 5%, and 10% significance levels.

reports the results of using alternative measures. In Model (1), we replicate the main regression for reference. In Model (2), we substitute the investment efficiency measure (Inv_E) with an alternative metric (Inv_EC) developed by Chen, Hope, Li and Wang [43]. Consistent with expectations, the coefficient of CSR decoupling (CSR_D) remains positive and statistically significant, confirming the robustness of the association.

In Models (3) and (4), we use an alternative CSR decoupling measure (CSR_DA), constructed following Talpur, Nadeem and Roberts [31]. Specifically, we use the institutional CSR performance score from the SNSI database as a proxy for internal CSR, and the external CSR disclosure score as a proxy for symbolic CSR. The difference between these two indicators forms CSR_DA. Model (3) uses Inv_E, and Model (4) uses Inv_EC as the dependent variable. In both models, the results remain positive and significant, reinforcing the validity of our main conclusions under alternative CSR decoupling constructions.

4.3.3. Excluding Firms Without Environmental Data

To address potential bias from firms lacking environmental disclosure, we estimate the models using only firms reporting environmental performance.

Table 6. Robustness test by excluding samples.

	(1)	(2)	(3)
Variables	Inv_E	Inv_E	Inv_EC
CSR_D	0.0014 ***	0.0012 **	0.0011 **
	(3.6282)	(3.1645)	(2.9783)
Controls	yes	yes	yes
Industry FE	yes	yes	yes
Year FE	yes	yes	yes
Adj. R2	0.0473	0.0441	0.0429
N	34,143	27,140	27,140

t-statistics in parentheses; ***, ** denote 1% and 5% significance levels.

presents the findings. Model (1) reproduces the baseline results. Models (2) and (3) present regression results using the restricted sample, applying Inv_E and Inv_EC, respectively.

Across all models, the coefficient of CSR_D remains positive and statistically significant, indicating that our findings are not driven by low-transparency firms. Although this restriction reduces the sample size, the results remain statistically robust, suggesting that the reduced sample does not materially undermine the power of our tests. Instead, it reinforces the conclusion that CSR decoupling has consistent effects on investment efficiency even among firms with higher levels of environmental transparency.

4.3.4. Introducing Lagged Variables

To address potential reverse causality—where firms with higher investment inefficiency might be more likely to engage in CSR decoupling—we introduce a lagged dependent variable. In this test, we use a one-period lagged investment efficiency variable (Inv_EL) to reduce endogeneity concerns.

Table 7. Robustness test by introducing lagged variables.

	(1)	(2)	(3)
Variables	Inv_E	Inv_E	Inv_EC
CSR_D	0.0014 ***	0.0013 **	0.0013 **
	(3.6282)	(3.2171)	(3.1193)
Controls	yes	yes	yes
Industry FE	yes	yes	yes
Year FE	yes	yes	yes
Adj. R2	0.0473	0.0471	0.0452
N	34,142	30,034	30,034

t-statistics in parentheses; ***, ** denote 1%, and 5% significance levels.

presents the results. Model (1) provides the benchmark. Models (2) and (3) incorporate a lagged investment efficiency variable (Inv_EL), using Inv_E and Inv_EC as dependent variables, respectively. The coefficient for CSR decoupling remains positive and statistically significant at the 5% level in both specifications, suggesting that the directionality of the effect is not driven by endogeneity or reverse causality.

In summary, all robustness tests consistently confirm the positive relationship between CSR decoupling and investment inefficiency. These results remain stable across: different measures of investment efficiency and CSR decoupling, subsamples excluding firms with poor environmental disclosure, and endogeneity checks using lagged dependent variables. This strengthens our conclusion that CSR decoupling poses a significant risk to firms’ capital allocation efficiency.

5. Additional Analysis

This section further examines the impact of CSR decoupling by first testing the moderating role of system coupling strength and then analyzing ownership-based heterogeneity among state-owned, foreign, and private firms.

5.1. The Moderating Effect of System Coupling Strength

In complex organizational systems, firms do not operate in isolation but are embedded within broader ecosystems—particularly in supply chain networks—where inter-firm dependencies and resource linkages critically shape strategic behavior and performance outcomes. In this study, we conceptualize supply chain concentration—the extent to which a firm relies on a few dominant customers or suppliers—as a proxy for system coupling strength (SCS). According to systems theory, coupling strength reflects the intensity of interdependence and the transparency of information flows between subsystems [61]. Firms with highly concentrated supply chains face tighter monitoring and stronger reciprocal influence from key partners, thereby operating in a more tightly coupled network structure.

From a systems perspective, SCS functions as a structural coupling mechanism that governs the degree of interdependence and information exchange between the focal firm and its external stakeholders. In such networks, the actions of individual firms are more visible and more easily scrutinized, meaning that any misalignment between internal CSR practices and external CSR disclosures is more likely to be detected and to propagate adverse signals across the system. This amplification effect heightens the risks of resource misallocation, stakeholder distrust, and reputational damage.

Empirical research supports this mechanism. Prior studies indicate that supply chain concentration plays a critical role in shaping CSR engagement and sustainability performance [62,63]. For example, Liu et al. [64] suggest that firms with close ties to key partners tend to align CSR priorities through shared codes, collaborative initiatives, and mutual monitoring mechanisms. Li, Li and Xue [63] show that the protective effect of ESG during crises depends critically on a firm’s supply chain concentration, this finding highlights supply chain concentration as a structural moderator that conditions how external sustainability signals are interpreted, thereby empirically supporting its use as a proxy for system coupling strength (SCS).

To empirically test this moderating mechanism, system coupling strength (SCS) is proxied by supply chain concentration, which reflects the extent to which a firm depends on a limited number of dominant customers or suppliers. This coupling strength influences how internal governance signals and external CSR representations interact, thus affecting the degree to which CSR decoupling translates into observable investment inefficiencies. We hypothesize that higher SCS intensifies the negative effect of CSR decoupling on investment efficiency, consistent with the behavior of tightly coupled systems in complex adaptive networks, where misaligned inputs can cascade into systemic inefficiencies.

The results of this moderating effect test are in

Table 8. Moderating effect of system coupling strength.

	(1)	(2)	(3)
Variables	Inv_E	Inv_E	Inv_E
CSR_D	0.0014 ***	0.0012 **	0.0013 **
	(3.6282)	(3.1218)	(3.1825)
SCS		0.0002 ***	0.0002 ***
		(12.9053)	(12.9000)
CSR_D×SCS			0.0001 *
			(2.3749)
Controls	yes	yes	yes
Industry FE	yes	yes	yes
Year FE	yes	yes	yes
Adj. R2	0.0473	0.0497	0.0498
N	34,143	32,477	32,477

t-statistics in parentheses; ***, **, * denote 1%, 5%, and 10% significance levels.

. Model (3) estimating the interaction between SCS and CSR decoupling (CSR_D). The positive and significant coefficient on the interaction term indicates that higher system coupling strength strengthens the negative effect of CSR decoupling on investment efficiency. Consistent with systems theory, where stronger coupling heightens the diffusion and consequences of local misalignments across interconnected subsystems. In practical terms, greater SCS increases the visibility and traceability of CSR misalignment within the supply chain system, thereby reducing flexibility and amplifying reputational and financial risks.

These findings extend the literature on CSR decoupling by introducing a system-theoretic perspective, emphasizing how network-level structural characteristics moderate firm-level outcomes. In tightly coupled systems, internal-external misalignments in CSR behavior are more likely to be penalized by key stakeholders, making alignment between symbolic and substantive CSR not only desirable but necessary for maintaining investment discipline and long-term strategic coherence.

5.2. Heterogeneity Analysis by Ownership Structure

This section investigates how ownership structure moderates the relationship between CSR decoupling and investment efficiency, drawing on the system coupling perspective. Ownership structure is not only a firm-level characteristic but also an institutional arrangement that conditions how external pressures and internal governance interact [65]. Firms with different ownership types—state-owned enterprises (SOEs), private firms, and foreign firms—are embedded in distinct institutional environments, which shape their internal governance mechanisms, external stakeholder expectations, and the coupling strength between internal and external CSR subsystems [66]. In this way, ownership structure provides a crucial boundary condition for understanding the systemic consequences of CSR misalignment.

In the Chinese market, SOEs are subject to strong political oversight and are expected to meet policy-driven legitimacy goals. This institutional environment can reduce tolerance for symbolic CSR, but it may also create incentives to emphasize compliance-oriented disclosure rather than substantive integration. By contrast, private firms operate in a market-driven environment with weaker regulatory constraints. Their CSR choices are often more flexible and strategic, but this same discretion increases the risk that symbolic disclosures diverge from actual practices. Foreign firms face the most complex pressures, as they are simultaneously embedded in domestic regulatory systems and international stakeholder networks. This dual embeddedness often fragments CSR governance, increasing the likelihood of decoupling and amplifying its adverse effects on investment efficiency.

Foreign firms face unique pressures as they are embedded in both domestic and international institutional systems. This dual embeddedness creates a structurally fragmented CSR system, where internal practices must conform to global corporate standards, while external disclosures must meet local stakeholder expectations. Such cross-system misalignment increases the likelihood of CSR decoupling and amplifies the negative consequences on investment efficiency due to inconsistent stakeholder expectations and conflicting institutional logics.

The results in

Table 9. Results of heterogeneity analysis.

	(1)	(2)	(3)
Variables	State-Owned	Private	Foreign
CSR_D	0.0013 *	0.0013 *	0.0062 *
	(2.3510)	(2.2076)	(2.4785)
Controls	yes	yes	yes
Industry FE	yes	yes	yes
Year FE	yes	yes	yes
Adj. R2	0.0662	0.0466	0.0733
N	12,959	18,888	1061

t-statistics in parentheses; * denotes 10% significance levels.

show that CSR decoupling is positively associated with Inv_E across all ownership types, with statistical significance at the 10% level. However, there are notable differences in the magnitude of the effect. The coefficients for SOEs and private firms are relatively similar, suggesting that despite greater regulatory oversight, SOEs are not significantly more aligned in their CSR practices than private firms. This may imply that formal regulatory mechanisms alone are insufficient to prevent symbolic CSR behavior if internal incentive structures and monitoring remain weak.

For foreign firms, the coefficient is larger (0.0062), indicating a stronger negative impact of CSR decoupling on investment efficiency. This finding supports the notion that foreign firms are particularly vulnerable to CSR misalignment due to their exposure to fragmented institutional environments and divergent stakeholder expectations [67]. The decoupling they experience across jurisdictions may heighten reputational risks, reduce stakeholder trust, and impair their ability to make efficient capital allocation decisions.

6. Implications and Conclusions

6.1. Theoretical Implications

This study contributes to the literature by conceptualizing CSR decoupling as a systemic coordination failure between two interrelated subsystems: external legitimacy seeking and internal capability building. By integrating systems thinking with legitimacy theory, we move beyond traditional firm-level analyses and emphasize the dynamic interdependence between symbolic and substantive CSR. This framing highlights how decoupling generates informational distortions, disrupts feedback loops, and ultimately weakens investment efficiency. In doing so, the study extends prior research that has mainly focused on CSR’s direct link with firm performance, offering instead a systems-based interpretation of how governance misalignment undermines capital allocation efficiency. Furthermore, by examining the moderating role of supply chain concentration as a form of system coupling strength, we enrich the dialogue on network-level contingencies in shaping CSR outcomes, thereby advancing systems theory applications in corporate governance.

6.2. Practical Implications

From a managerial perspective, the findings underscore the need for firms to enhance coordination between external communication and internal CSR practices. Symbolic CSR may provide temporary legitimacy benefits, but its long-term risks outweigh such gains when not matched with substantive implementation. Embedding CSR metrics into strategic decision-making, strengthening internal feedback loops, and promoting CSR alignment throughout supply chain partnerships can reduce system-level inconsistencies. For policymakers, the results underscore the limitations of disclosure-based regulation alone. Effective CSR governance requires mechanisms for verifying the substantive alignment between reported commitments and actual practices to foster transparency, accountability, and capital efficiency.

6.3. Limitations and Future Research

Several limitations should be acknowledged. First, the analysis focuses on Chinese listed firms, which may limit the generalizability of the findings to other institutional environments. Future research could extend this framework to cross-country comparative settings, particularly contrasting state-led economies and market-led economies, to examine whether differences in regulatory enforcement and state involvement condition the effects of CSR decoupling. Second, the equal-weighted approach to CSR indicators may overlook differential importance across dimensions; future research could explore alternative weighting strategies, such as factor analysis or expert-based weighting, to refine the construction of CSR decoupling measures. Finally, researchers may examine multi-channel CSR signaling, including discrepancies between annual reports, ESG ratings, and social media communications, to better capture the dynamic interaction between symbolic representation and actual behavior across corporate systems. Such research would deepen our understanding of CSR as a system of institutional signaling, strategic execution, and stakeholder interpretation.

6.4. Conclusions

Overall, this study provides robust evidence that CSR decoupling significantly impairs investment efficiency, primarily through overinvestment. The findings demonstrate that the adverse effects of CSR decoupling are more severe in tightly coupled systems, such as those characterized by high supply chain concentration, and are particularly pronounced in foreign firms facing dual legitimacy pressures. By integrating systems theory into the study of CSR decoupling, this paper not only deepens theoretical understanding but also offers practical insights into the governance mechanisms required to ensure alignment between symbolic CSR communication and substantive action.