Does Biodiversity Conservation Pay Off? An Empirical Analysis of Japanese Firms

Abstract

This study investigates the bidirectional relationship between biodiversity conservation, an increasingly important dimension of corporate social responsibility (CSR), and corporate financial performance (CFP). Specifically, it compares the manufacturing sector, which has substantial environmental impact and close ties to ecosystems, and the nonmanufacturing sector. The analysis draws on 1079 firm-year observations of Japanese companies from 2017 to 2022, employing the ratio of biodiversity-related expenditures to total environmental costs as the independent variable. CFP is measured by return on assets (ROA) and the price-to-book ratio (PBR). The results show that the effects on ROA significantly differ between manufacturing and nonmanufacturing sectors, with more positive impacts in manufacturing. In contrast, no clear sectoral differences are identified for the PBR. The reverse analysis suggests that, in the nonmanufacturing sector, firms with a higher PBR tend to allocate less to biodiversity conservation, whereas in manufacturing firms, both ROA and the PBR indicate positive effects, although statistical significance was not established. These findings indicate that biodiversity conservation in the manufacturing sector can be regarded as a strategic investment that contributes to profitability, and that its effects differ across industries. The study further suggests that investors and policymakers should consider industry-specific characteristics when evaluating corporate initiatives and designing institutional frameworks.

1. Introduction

In recent years, corporate social responsibility (CSR) has gained prominence as a framework for companies to address environmental and social challenges. Beyond mere legal compliance, CSR emphasizes the development of sustainable business models and contributions to society. CSR encompasses various dimensions, including environmental and social considerations, effective utilization of human resources, and corporate governance. The concept has its roots primarily in developments from the early 1950s to the present [1]. Bowen [2], often referred to as the “father of CSR,” emphasized in his book Social Responsibilities of the Businessman that corporations have an obligation to act in accordance with the goals and values of society, framing CSR as a form of social responsibility that goes beyond economic profit. In other words, CSR was originally regarded as the pursuit of social and environmental values based on a corporation’s ethical values. However, as CSR activities have evolved and expanded in response to changing societal contexts and emerging challenges—and as the resources allocated to these activities have increased—questions have arisen regarding whether such investments are justified by corresponding returns [3]. Aguinis and Glavas [4], in their comprehensive review of CSR research, argued that companies engage in CSR not merely out of moral obligation but primarily due to instrumental motives—such as the expectation of economic benefits, such as improved performance. In other words, CSR activities can be regarded as strategic investments undertaken by companies in anticipation of future benefits. Examining how companies allocate limited resources to CSR activities and how such allocations affect their corporate financial performance (CFP) is essential for evaluating the nature and effectiveness of CSR in today’s business environment.

On the other hand, it has also been pointed out that CFP may influence CSR engagement [5,6,7]. Firms with strong CFP may be more able to allocate slack resources to CSR, in which case the observed effect of CSR could be merely apparent. Examining whether CSR enhances CFP, whether CFP drives CSR, or whether both occur simultaneously is an important theoretical and practical task.

CSR initiatives encompass a wide range of areas, including environmental, social, and governance issues, but among them, environmental challenges have been attracting increasing attention in recent years. This study focuses on biodiversity conservation, one of the core issues in the environmental domain. The loss of biodiversity conservation is recognized as a major risk alongside climate change; however, it differs from other environmental issues in that it exerts long-term impacts on resource procurement and supply chains. Examining biodiversity conservation as an independent research subject and investigating its financial effects holds both academic and practical significance.

The term biodiversity is a coined word introduced in the late 1980s by Walter G. Rosen of the National Academy of Sciences [8], and global interest in the concept surged following the Earth Summit held in Rio de Janeiro in 1992 [9]. The Convention on Biological Diversity, adopted at the Earth Summit in 1992, defines biodiversity as “the variability among living organisms from all sources including, inter alia, terrestrial, marine and other aquatic ecosystems and the ecological complexes of which they are part; this includes diversity within species, between species and of ecosystems.” [10]. In other words, biodiversity refers to the diversity of genes, species, and ecosystems [11].

According to the World Economic Forum 2024 [12], biodiversity loss is ranked as the third most severe global risk over the next decade, following two climate-related risks. The reason biodiversity loss is considered a critical issue to be addressed is that we directly and indirectly benefit from the “services” provided by ecosystems, which are integral components of biodiversity. Costanza et al. [13] estimated the annual economic value of this services to be USD 33 trillion, which was 1.8 times the global gross national product (GNP) at that time. In 2021, commissioned by the UK government, The Economics of Biodiversity: The Dasgupta Review [14] was published by Professor Sir Partha S. Dasgupta, an economist at the University of Cambridge. The report criticizes conventional economics for underestimating the role of nature and proposes a new production function that incorporates nature as a factor in production. It also issues a strong warning against global biodiversity loss, noting, for instance, that sustaining our current standard of living would require the resources of 1.6 Earths. Against this backdrop, the impacts of biodiversity loss are expected to be particularly pronounced not only in industries that depend directly on natural resources, such as pharmaceuticals and agriculture, but also in manufacturing sectors that consume large amounts of raw materials and water.

Despite the growing recognition that biodiversity conservation is an issue that companies can no longer ignore, empirical research examining the relationship between such efforts and CFP remains limited. Many empirical studies rely primarily on corporate disclosure data, and few have evaluated the extent to which firms prioritize and strategically position biodiversity conservation based on actual expenditure data as reported in financial statements or integrated reports. Further investigation is needed into how the strategic prioritization of biodiversity conservation within a limited environmental budget affects CFP. Moreover, the possibility that such impacts may vary across industries has not been sufficiently examined.

Based on the above, this study examines whether corporate biodiversity conservation efforts enhance CFP and whether this relationship differs across industries. This study is grounded in the natural-resource-based view (NRBV), which provides a theoretical basis linking a firm’s dependence on natural resources and its environmental capabilities to competitive advantage and CFP. The NRBV assumes that the ability to adapt to environmental challenges leads to long-term competitive advantage [15]. Within this framework, the study empirically examines the impact of biodiversity conservation on CFP, focusing on Japanese firms for which detailed data on biodiversity-related expenditures are available. Specifically, it calculates the ratio of biodiversity-related expenditures and analyzes how these initiatives affect internal performance, measured by return on assets (ROA), and external performance, measured by the price-to-book ratio (PBR), defined as the ratio of stock price to book value per share. The study also considers the possibility that this impact varies across industries, comparing the manufacturing sector—closely linked to ecosystems through raw material procurement and production activities—with the nonmanufacturing sector. This comparison aims to show that the effects of biodiversity conservation are not uniform across industries and to provide both theoretical and practical implications. In addition, the study examines how CFP influences biodiversity conservation efforts, assessing whether a bidirectional relationship exists between the two. To address concerns regarding endogeneity and causality, this study employs time-lagged models and panel regression approaches.

The following section provides a brief overview of prior research on the impact of CSR and biodiversity conservation on CFP and presents the theoretical framework underpinning the hypotheses. Next, the research methodology and empirical results are outlined, followed by a discussion of the key findings, their theoretical and practical implications, and the limitations of the study.

2. Literature Review

In recent years, the concept of CSR has expanded over time. While it initially focused on addressing social issues such as working conditions, human rights, and local communities, it has come to include environmental protection as an important component [16]. The concept that focuses specifically on the environmental dimension is corporate environmental responsibility (CER). In MSCI’s ESG ratings, the environmental category is divided into four areas: climate change, natural capital, pollution and waste, and environmental opportunities. While all of these are components of CER, in recent years corporate efforts toward natural capital—and in particular, biodiversity conservation—have been attracting increasing attention.

Biodiversity conservation is conceptually related to CER but is not merely a subcategory of it. While many other environmental initiatives focus on reducing environmental impact in the short term, biodiversity conservation takes a long-term approach. It aims to secure the sustainability of ecosystem services and to create future opportunities for resource procurement and product differentiation. In industries that depend on natural resources, it has the characteristics of a strategic investment. At the same time, because it is difficult to measure and quantify its outcomes, it requires analytical considerations that differ from those applied to CER. Because of this difficulty in measurement, firms have a responsibility to clearly communicate the content and significance of their initiatives to external stakeholders. Hassan et al. [17] highlight the importance of corporate accountability for the impact that firms have on biodiversity. Many existing studies [17,18,19] focus on corporate reporting as a means of fulfilling this accountability and measure the extent of biodiversity conservation through such reports.

However, these reports may be used as a tool for corporate impression management. Impression management refers to the act of shaping a company’s image to align with external expectations or evaluations or to morally justify inappropriate behavior [20]. In other words, biodiversity disclosure may become a formalistic response that is disconnected from actual practice, suggesting that analyses relying solely on disclosure scores carry the risk of overestimation or underestimation. While many prior studies focus on biodiversity-related reporting and stakeholder relationships, they do not explicitly examine the relationship between firms’ concrete actions or expenditures and CFP. In this context, behavior-based measures—indicators that capture actual corporate actions, have emerged as a valuable approach for assessing biodiversity conservation outcomes. In fact, Dutta and Dutta [21] examined the relationship between corporate environmental performance and biodiversity-related disclosure. They reported that firms with higher water consumption and waste emissions tend to disclose biodiversity-related information to maintain their legitimacy. This finding suggests that firms with low environmental performance may use disclosure as a self-justification tool to manage impressions, supporting the premise of this study that corporate environmental behavior cannot be fully evaluated based solely on disclosed information.

Amid this context, Elsayed [19] offers a study on biodiversity conservation. The study examines the relationship between biodiversity-related reporting and corporate financial performance, focusing on a sample of 100 companies listed in the Fortune Global 500. It uses multiple regression analysis with ROA and the PBR as dependent variables. Elsayed [19] found that the biodiversity-related disclosure score was positively and statistically significantly associated with ROA at the 5% level. In contrast, although a positive relationship was observed with the PBR, it was not statistically significant. This research makes a valuable contribution to the limited body of research on the relationship between biodiversity conservation efforts and CFP. However, the study focuses solely on information disclosed and does not examine whether the level of disclosure accurately reflects the extent of actual corporate efforts. Considering the potential gap between reporting and behavior—commonly referred to as greenwashing—analyses that rely solely on self-reported data have inherent limitations.

An empirical study by Hassan and Romilly [22] is an example that considers both the level of disclosure and behavior-based measures. Their study analyzed the relationships among greenhouse gas (GHG) emissions, environmental disclosure, and corporate economic performance. It finds a consistently negative causal relationship between GHG emissions and economic performance, while the relationship between environmental disclosure and economic performance varies by region. These findings suggest that actual environmental actions may have a clearer impact on CFP than disclosure alone, highlighting the need to place greater emphasis on behavior-based indicators in CER research. Moreover, the regional variation implies that findings based on one county may not directly generalize to others; since this study focuses on Japanese firms, this consideration is relevant.

Furthermore, Bach et al. [23] examine the relationship between biodiversity risk and financial performance. Using text mining of 10-K reports, they developed a biodiversity risk score and examined its association with financial performance. The results showed that biodiversity risk tends to have a negative effect on sales growth and ROA, with the effect being particularly pronounced in industries highly dependent on biodiversity. This finding supports the view that biodiversity loss can undermine corporate value and highlights the need for strategic actions to mitigate such risks.

Existing research on biodiversity conservation has mainly focused on disclosure, and few studies have directly measured the actual level of engagement. In Japan, manufacturing accounts for a large share of the economy, and much of its raw material procurement depends on natural resources. Despite the substantial risk that biodiversity loss could undermine supply chain resilience, empirical evidence in this area remains limited. This study aims to fill this gap by using actual expenditure data. In this study, “actual expenditure” refers to the monetary amounts reported by firms, as opposed to disclosure scores or qualitative reporting. To capture the actual level of corporate engagement, this study adopts the proportion of biodiversity conservation expenditures to total environmental protection expenditures. This is a relative indicator that reflects the strategic priority given to biodiversity conservation, differing from prior studies that have mainly relied on disclosure-based measures.

3. Theoretical Framework and Hypothesis Development

3.1. The Impact of Biodiversity Conservation Efforts on CFP

To understand the impact of corporate biodiversity conservation efforts on CFP, this study adopts the natural-resource-based view (NRBV) of a firm, proposed by Har [15], as its primary theoretical foundation. The NRBV is an extension of the resource-based view (RBV), originally proposed by Barney [24], in the context of the natural environment. While Hart [15] acknowledged that the RBV provides a more flexible and comprehensive framework than traditional theories of competitive advantage, he pointed out that it has not theoretically addressed the interactions between firms and the natural environment. This is because previously overlooked aspects of the natural environment have now come to be seen as potential constraints on firms’ ability to achieve a sustainable competitive advantage [25]. Based on this recognition, the NRBV asserts that “it is likely that strategy and competitive advantage in the coming years will be rooted in capabilities that facilitate environmentally sustainable economic activity” [15].

When engaging in environmental protection activities, companies face a wide range of resource allocation choices. With numerous environmental issues such as greenhouse gas emissions reduction, waste management, and water resource management, firms must make strategic decisions on how to allocate their limited resources. This study focuses on the extent to which companies allocate their environmental protection expenditures specifically to biodiversity conservation, in relative terms, as a means of capturing their strategic priorities. Compared to other environmental protection activities, such as compliance with environmental regulations, biodiversity conservation is inherently more complex and requires a longer-term perspective. In this sense, it can be seen as a more advanced form of environmental initiative. In other words, prioritizing the allocation of resources to biodiversity conservation suggests that such efforts are positioned as a strategic choice aimed at enhancing a company’s medium- to long-term sustainability and competitiveness. This aligns with the NRBV’s emphasis on developing environmental capabilities, which can lead to the attainment of competitive advantage. The NRBV categorizes firms’ environmental strategies into three types: (1) pollution prevention, (2) product stewardship, and (3) sustainable development.

From the perspective of pollution prevention, reducing hazardous substances in the production process and properly managing waste can mitigate negative impacts on surrounding ecosystems, thereby indirectly contributing to biodiversity conservation. When companies strategically prioritize biodiversity conservation within their environmental protection investments and engage in pollution prevention, they enhance their management capabilities to proactively prevent the emission of harmful substances and the overconsumption of natural resources. Such efforts can lead to reduced raw material and treatment costs through improved resource and energy efficiency, as well as lower future regulatory compliance costs. As a result, firms with a higher relative allocation of spending on biodiversity conservation are expected to achieve greater profitability (ROA) through enhanced operational efficiency.

Product stewardship is an environmental strategy aimed at minimizing negative environmental impacts throughout the entire product life cycle—from the design stage to manufacturing, use, and disposal. In the context of biodiversity conservation, this strategy includes the use of environmentally responsible raw materials—such as Forest Stewardship Council (FSC)-certified timber and Roundtable on Sustainable Palm Oil (RSPO)-certified palm oil—as well as the exclusion of suppliers involved in illegal logging or habitat destruction.

A key characteristic of this strategy is that it requires environmental consideration across the entire supply chain. In other words, efforts must be made at multiple stages—from upstream to downstream—which entails practical challenges such as addressing information asymmetry and ensuring traceability. In addition, procuring environmentally responsible materials can pose constraints in terms of cost and supply. Overcoming such challenges requires collaboration with a diverse range of external stakeholders, and the quality of this collaboration critically influences the effectiveness of the initiatives. Therefore, when companies place particular emphasis on biodiversity conservation and engage in product stewardship, it demonstrates their ability and willingness to implement environmental considerations across the entire supply chain from a long-term perspective. This serves as a clear signal to external stakeholders that the company is implementing a sustainable, biodiversity-conscious environmental strategy. Such environmentally conscious actions, undertaken in cooperation with stakeholders, are perceived as desirable, proper, or appropriate by society, thereby contributing to the attainment of legitimacy, as defined by Suchman [26]. Gaining environmental legitimacy helps reduce unsystematic risks [27] and tends to stabilize stock prices. Such stability enhances investor confidence and promotes long-term value creation, ultimately contributing to a higher PBR.

Sustainable development refers to efforts that contribute to the overall sustainability of society by balancing environmental protection with economic growth. It represents a broader and more long-term perspective than the aforementioned strategies of pollution prevention and product stewardship. In the context of biodiversity conservation, this strategy includes initiatives such as the establishment of nature reserves and development projects carried out in collaboration with local communities. Allocating a relatively large share of environmental protection expenditures to biodiversity conservation represents a concrete manifestation of a company’s adoption of a long-term sustainability-oriented strategy. It reflects a strategic decision that goes beyond short-term compliance obligations. Through improved resource procurement and favorable evaluations by investors, such efforts can influence a firm’s CFP. The sustainable procurement of resources contributes to a stable supply of raw materials, enabling more precise production planning, reduced inventory costs, and optimized supply chain management. These improvements, in turn, help stabilize and enhance ROA. Furthermore, with the recent expansion of ESG investing, investors increasingly regard corporate initiatives’ ESG areas as important indicators of long-term growth potential and risk management capability [28]. Proactive investment in sustainable development is seen as a demonstration of CSR, which can enhance investor evaluations and contribute to higher stock prices and an increase in the PBR.

Based on the above, proactive efforts toward biodiversity conservation are theoretically supported as having a positive impact on both the ROA and PBR. Based on the discussion thus far, the following general hypotheses are proposed.

Hypothesis 1 (H1). 

The proportion of biodiversity conservation project expenditures relative to total environmental expenditures has a positive relationship with ROA.

Hypothesis 2 (H2). 

The proportion of biodiversity conservation project expenditures relative to total environmental has a positive relationship with PBR.

3.2. Industry Differences: A Comparison Between Manufacturing and Nonmanufacturing Sectors

Financial effects may not manifest uniformly across all firms because societal evaluations of corporate environmental activities may not manifest uniformly across all firms. This is because societal evaluations of corporate environmental activities and stakeholder expectations vary significantly, depending on industry and business structure [29]. This context dependence can be explained through legitimacy theory [26]. Legitimacy theory posits that it is essential for organizations to gain and maintain legitimacy—that is, to be perceived as socially appropriate—in order to survive and thrive. From this perspective, firms operating in high-impact industries face greater pressure to secure social approval through environmental initiatives, making their disclosures and actions more likely to influence corporate evaluations and financial outcomes. The manufacturing sector is closely linked to ecosystems through its business activities—from raw material procurement and production processes to product disposal—and tends to generate significant environmental impacts. Therefore, biodiversity conservation efforts in manufacturing are more likely to be directly tied to corporate value, particularly in terms of environmental risk management, regulatory compliance, and the fulfillment of social responsibility, compared to those in the nonmanufacturing sector. For example, avoiding the use of illegally logged timber in the supply chain or preventing ecosystem pollution caused by factory wastewater can help manufacturing firms reduce risks that threaten business continuity. As mentioned earlier, prioritizing biodiversity conservation over other environmental protection activities represents a long-term strategy that goes beyond mere compliance with general environmental regulations. In the manufacturing sector, which tends to be subject to such regulations, this strategic emphasis is more likely to be positively evaluated by the market.

In contrast, the nonmanufacturing sector often has relatively lower direct environmental impacts compared to manufacturing. As a result, the effects of biodiversity conservation activities on profitability and market valuation may be less pronounced in these sectors. Indeed, both legitimacy theory and stakeholder theory emphasize the context-dependent nature of social approval and stakeholder expectations ([26,30]), suggesting that industry-specific differences are likely to influence the impact on CFP.

Based on the above, the conceptual framework of this study is grounded in legitimacy theory, which posits that firms must gain and maintain social approval to survive and thrive. Proactive biodiversity conservation is assumed to enhance environmental legitimacy, which in turn is positively associated with CFP. While stakeholder theory further highlights the context-dependent nature of legitimacy, this relationship is expected to be more pronounced in the manufacturing sector, given its stronger and more direct environmental impacts. Accordingly, we propose the following hypotheses:

Hypothesis 3 (H3). 

In the manufacturing sector, the proportion of biodiversity conservation project expenditures relative to total environmental expenditures has a stronger positive relationship with ROA than in the nonmanufacturing sector.

Hypothesis 4 (H4). 

In the manufacturing sector, the proportion of biodiversity conservation project expenditures relative to total environmental expenditures has a stronger positive relationship with PBR than in the nonmanufacturing sector.

3.3. The Impact of CFP on Biodiversity Conservation Efforts

While many previous studies [31,32] have focused on the impact of CSR on CFP, the reverse relationship—the influence of CFP on CSR activities—should not be overlooked. Analyzing this reverse direction allows for consideration of the possibility that CSR may be a consequence of corporate strategy, thereby providing a more comprehensive understanding of the CSR–CFP relationship beyond a one-way perspective.

As Waddock and Graves [7] demonstrated, strong financial performance may encourage corporate social performance. However, since CSR encompasses a wide range of practices, the extent to which CFP influences specific domains of CSR—particularly those that remain underexplored—has yet to be fully understood. This study seeks to contribute to the literature by extending this discussion to the relationship between biodiversity conservation activities and CFP. In examining the influence of CFP on corporate biodiversity conservation activities, this study integrates slack resources theory [7,33] and legitimacy theory [26]. It considers the logical basis for how financial success affects biodiversity investment from two complementary perspectives: the capacity to invest and the motivation to do so.

Slack resources theory [7,33] posits that strong financial performance allows a firm to secure surplus resources, or “slack,” which can then be allocated to various activities. Firms with robust financial results possess the capacity to direct capital toward socially valuable initiatives, such as environmental protection. Biodiversity conservation, in particular, demands significant complexity and a long-term perspective, making it feasible only when sufficient resources are available. Therefore, it is inferred that firms with higher profitability and market valuation have a greater capacity to allocate funds to such initiatives.

The motivation to translate this capacity into action is explained by legitimacy theory [26]. Firms that are financially successful face increased scrutiny and higher expectations from society and stakeholders, providing them with a strong incentive to maintain and enhance their legitimacy. Engaging in biodiversity conservation is an extremely effective means to achieve this, as it is a global social issue of growing importance. This demonstrates a firm’s strong commitment to environmental stewardship, thereby boosting its legitimacy.

This relationship may also vary across industries. As discussed earlier, manufacturing firms tend to exert greater environmental impact; thus, engagement in biodiversity conservation is more directly linked to business continuity and competitive advantage. Therefore, when financially capable, these firms are more likely to actively invest in biodiversity initiatives. In contrast, nonmanufacturing firms, even when possessing sufficient financial slack, may have fewer direct incentives to engage in such activities, resulting in a relatively weaker relationship.

As Wang et al. [34] point out, the relationship from CFP to CSR is inconclusive across previous studies. This divergence is likely due to the variety of CSR measurement indicators and research subjects. In this study, drawing on the classification by Endrikat et al. [5], we capture corporate biodiversity conservation efforts using a process-based indicator: the ratio of expenditures actually allocated to such initiatives. This approach allows for a more objective analysis in the field of biodiversity conservation, where the quantitative measurement of outcomes is often challenging.

Based on the above discussion, we argue that firms with high CFP gain the capacity to invest in biodiversity conservation through slack resources theory, while also having the motivation to meet social expectations as explained by legitimacy theory. When both the capacity and the motivation are present, firms are more likely to actively engage in biodiversity conservation. We further propose that this motivation is stronger in the manufacturing sector, leading to the following hypothesis:

Hypothesis 5 (H5). 

ROA influences the proportion of expenditure allocated to biodiversity conservation projects, and this effect is stronger in the manufacturing sector rather than nonmanufacturing sector.

Hypothesis 6 (H6). 

PBR influences the proportion of expenditure allocated to biodiversity conservation projects, and this effect is stronger in the manufacturing sector rather than nonmanufacturing sector.

This study explains the impact of biodiversity conservation activities on CFP through the NRBV, while theorizing the influence of CFP on biodiversity conservation activities by integrating slack resources theory and legitimacy theory. By combining these perspectives, the relationship between biodiversity conservation and CFP can be understood as a mutually reinforcing cyclical process, allowing for a deeper and more multifaceted understanding.

4. Methodology

4.1. Data and Sample

In this study, panel data on Japanese companies were utilized to empirically examine the relationship between corporate biodiversity conservation efforts and CFP. Data on biodiversity conservation activities were obtained from CSR Company Data: Environment Edition [35]. Specifically, information was collected from items on environmental conservation expenditures and expenditures on biodiversity conservation projects. The number of firms for which both environmental conservation and biodiversity project expenditure data were available amounted to 178 firms in the 2017 edition, 183 in 2018, 179 in 2019, 189 in 2020, 180 in 2021, and 170 in 2022, resulting in a total of 1079 firm-year observations. In addition, the financial performance indicators—ROA and PBR—were obtained from the Nikkei NEEDS-FinancialQUEST database.

4.2. Definition of Variables

4.2.1. Independent Variables

Building on the discussion in Section 3, this study adopts the ratio of biodiversity conservation expenditures to total environmental conservation expenditures (BioDiv Ratio) as the independent variable. This ratio reflects relative degree to which a firm prioritizes biodiversity conservation within its overall environmental initiatives. In other words, it captures the strategic positioning of biodiversity conservation within a company’s broader environmental strategy. Although the BioDiv Ratio does not fully capture a firm’s commitment to biodiversity conservation, a high relative ratio—even with a small absolute expenditure—may serve as a signal to external stakeholders that the firm is strategically emphasizing biodiversity conservation.

4.2.2. Dependent Variables

ROA is an indicator calculated by dividing net income by total assets, and it reflects a firm’s profitability. As an accounting-based measure of internal efficiency [36], ROA is suitable for capturing the internal impact of CSR activities on firm operations.

PBR is calculated by dividing a company’s stock price by its book value per share. It serves as a benchmark for evaluating whether a stock is over- or undervalued. As a market-based indicator of corporate value [37], it reflects investors’ assessment of a firm’s future profitability and intangible assets. Therefore, it is useful for examining how CSR initiatives influence market perceptions.

This study employs both ROA and PBR to examine the effects of CSR activities on a firm’s internal efficiency and market valuation. In addition, this study also analyzes the reverse direction, namely the influence of CFP on biodiversity conservation efforts. In this case, ROA and PBR are treated as independent variables, while the BioDiv Ratio serves as the dependent variable.

4.3. Analytical Method

This study employs a combination of multiple regression analysis methods to examine the impact of biodiversity conservation efforts on CFP from multiple perspectives. All analyses were conducted using the statistical software R version 4.5.0.

First, to test H1 and H2, ordinary least squares (OLS) were conducted using the full sample of firms, with BioDiv Ratio as the independent variable and ROA and PBR as the dependent variables. To control for the effects of industry and year, industry dummy variables and year dummy variables were included in the model.

Next, to test H3 and H4, we estimated a regression model that included an interaction term between the BioDiv Ratio and a manufacturing dummy. This analysis examines whether the impact of the BioDiv Ratio on ROA and PBR differs between manufacturing and nonmanufacturing firms.

Furthermore, considering that the effects of expenditures on biodiversity conservation may materialize over time, we conducted a regression analysis using a one-year lagged variable of the BioDiv Ratio. While this analysis does not directly demonstrate causality, it focuses on the temporal sequence between variables to examine the direction of influence and assess the robustness of the results. To ensure the reliability of the estimation results, we conducted multicollinearity diagnostics using the variance inflation factor (VIF) in each regression model and confirmed that multicollinearity was not a concern. Additionally, sensitivity analyses were performed to examine the possibility that unobserved factors may create spurious correlations.

In the reverse-direction analysis, we conducted regression analyses to examine how ROA and PBR affect the following year’s BioDiv Ratio. Specifically, the previous year’s ROA and PBR were used as independent variables, and the subsequent year’s BioDiv Ratio was set as the dependent variable. To examine whether the relationship differs between manufacturing and nonmanufacturing sectors, we conducted a regression analysis that includes interaction terms between industry type and financial indicators.

To improve the accuracy of causal inference and eliminate bias from unobserved firm-specific factors, Section 5.2, Section 5.3 and Section 5.4 additionally describe panel data analyses that accounted for both firm and year fixed effects. Because the fixed effects (FE) model removes persistent differences across firms, it allowed us to estimate effects based on within-firm annual variation, in contrast to OLS, which reflects long-term structural differences across industries.

The Shapiro–Wilk test indicated non-normality of the residuals. However, given the large sample size of 1079, the central limit theorem suggests that the distribution of the estimated coefficients can be approximated by a normal distribution. Furthermore, considering the potential presence of heteroskedasticity in the residuals, we used robust standard errors (HC1) to ensure the reliability of the estimates.

5. Results

5.1. Correlation Analysis and Results of the Basic Model

To understand the relationship between the BioDiv Ratio and financial indicators, correlation coefficients among the main variables were calculated (

Table 1. Correlation coefficients between BioDiv Ratio and ROA/PBR. Source: authors’ calculations using R.

	ROA		PBR	BioDiv Ratio
ROA	1
PBR	0.264	***	1
BioDiv Ratio	−0.013		0.047	1

*** p < 0.001, n = 1079.

). The results showed a significant positive correlation between PBR and ROA (r = 0.264, p < 0.001). It suggests that firms with higher market valuation also tend to exhibit higher operational efficiency. A slight positive correlation was observed between PBR and the BioDiv Ratio (r = 0.047), while a slight negative correlation was found between ROA and the BioDiv Ratio (r = −0.013); however, neither of these was statistically significant. These results suggest that the BioDiv Ratio is unlikely to exert a direct impact on firms’ PBR and ROA.

Next, a multivariate regression analysis was conducted by introducing industry and year dummy variables to control for industry and temporal effects (

Table 2. Regression results: impact of BioDiv Ratio on ROA. Source: authors’ calculations using R.

Term	Estimate	Std. Error	t. Value	p. Value	Signif.
(Intercept)	5.257	1.600	3.286	0.001	**
BioDiv Ratio	−0.190	0.335	−0.567	0.571
Electricity and Gas	−2.979	1.460	−2.040	0.042	*
Commerce	−1.526	1.430	−1.067	0.286
Construction	1.913	1.384	1.382	0.167
Manufacturing	1.255	1.338	0.938	0.348
Mining	6.265	2.078	3.015	0.003	**
Real Estate	−2.586	1.589	−1.627	0.104
Transportation and Information and Communications	−0.489	1.403	−0.348	0.728
Year 2018	0.696	0.415	1.676	0.094
Year 2019	0.410	0.418	0.982	0.326
Year 2020	−0.295	0.412	−0.716	0.474
Year 2021	−0.857	0.417	−2.053	0.040	*
Year 2022	0.112	0.423	0.264	0.792
Firm Size	0.002	0.009	0.225	0.822
Adjusted R-squared:	0.094

** p < 0.01, * p < 0.05, n = 1079.

and

Table 3. Regression results: impact of BioDiv Ratio on PBR. Source: authors’ calculations using R.

Term	Estimate	Std. Error	t. Value	p. Value	Signif.
(Intercept)	2.223	0.660	3.368	0.001	***
BioDiv Ratio	0.257	0.138	1.857	0.064
Electricity and Gas	−0.749	0.602	−1.243	0.214
Commerce	−0.373	0.590	−0.632	0.528
Construction	−0.571	0.571	−1.001	0.317
Manufacturing	−0.127	0.552	−0.231	0.818
Mining	−1.248	0.857	−1.456	0.146
Transportation and Information and Communications	−0.220	0.579	−0.381	0.704
Real Estate	1.508	0.656	2.300	0.022	*
Year 2018	0.032	0.171	0.185	0.853
Year 2019	−0.054	0.172	−0.313	0.754
Year 2020	0.048	0.170	0.285	0.775
Year 2021	0.124	0.172	0.719	0.472
Year 2022	−0.097	0.175	−0.553	0.580
Firm Size	−0.006	0.004	−1.789	0.074
Adjusted R-squared:	0.027

*** p < 0.001, * p < 0.05, n = 1079.

). The results showed that the BioDiv Ratio had a negative coefficient with ROA (β = −0.190) and a positive coefficient with PBR (β = 0.257); however, neither was statistically significant. These findings suggest that, when analyzing the entire sample, the BioDiv Ratio does not have a clear impact on CFP. Therefore, Hypothesis 1 and Hypothesis 2—“The higher the proportion of biodiversity conservation project expenditures to total environmental expenditures, the higher the ROA and PBR”—were not supported by this model. However, considering that the adjusted R-squared values were only 0.094 for the ROA model and 0.027 for the PBR model, it is possible that the current model does not fully capture the effects of biodiversity-related expenditures. Hence, there remains a possibility that the hypotheses could be supported under alternative model specifications, such as nonlinear models, or by taking industry-specific characteristics into account.

5.2. Interaction Model: Manufacturing vs. Nonmanufacturing Sector

Next, to examine the possibility that the effect of the BioDiv Ratio on CFP varies by industry, a regression model including a manufacturing dummy variable (manufacturing) and its interaction term (BioDiv Ratio × manufacturing) was estimated (

Table 4. Interaction model results: manufacturing vs. nonmanufacturing sector for ROA. Source: authors’ calculations using R.

Term	Estimate	Std. Error	t. Value	p. Value	Signif.
(Intercept)	6.307	0.833	7.573	0.000	***
BioDiv Ratio	−0.339	0.308	−1.101	0.271
Manufacturing	1.290	0.284	4.536	0.000	***
Size	−0.012	0.009	−1.345	0.179
Year 2018	0.638	0.363	1.756	0.079
Year 2019	0.327	0.391	0.836	0.403
Year 2020	−0.324	0.407	−0.796	0.426
Year 2021	−0.893	0.424	−2.105	0.036	*
Year 2022	0.039	0.457	0.086	0.932
BioDiv Ratio: manufacturing	3.289	1.094	3.008	0.003	**
Adjusted R-squared	0.043

*** p < 0.001, ** p < 0.01, * p < 0.05, n = 1079.

and

Table 5. Interaction model results: manufacturing vs. nonmanufacturing sector for PBR. Source: authors’ calculations using R.

Term	Estimate	Std. Error	t. Value	p. Value	Signif.
(Intercept)	1.874	0.466	4.022	0.000	***
BioDiv Ratio	0.163	0.334	0.487	0.627
Manufacturing	0.175	0.126	1.389	0.165
Size	−0.006	0.005	−1.306	0.192
Year 2018	0.035	0.104	0.337	0.736
Year 2019	−0.049	0.108	−0.455	0.649
Year 2020	0.049	0.247	0.198	0.843
Year 2021	0.118	0.123	0.956	0.339
Year 2022	−0.111	0.116	−0.957	0.339
BioDiv Ratio: manufacturing	0.620	0.377	1.646	0.100
Adjusted R-squared	0.004

*** p < 0.001, n = 1079.

). In the PBR model, the coefficient of the BioDiv Ratio for the nonmanufacturing sector was 0.163 (p = 0.627), indicating no statistical significance. However, the interaction term with the manufacturing dummy was 0.620 (p = 0.100), which, although not statistically significant at the conventional levels, approached the 10% threshold. This suggests that the effect of the BioDiv Ratio may be relatively greater in the manufacturing sector than in the nonmanufacturing sector. This effect was not observed in the nonmanufacturing sector, prompting a reconsideration of sector-specific factors, such as the relatively higher environmental burden and stakeholder pressure typically associated with manufacturing.

Furthermore, this trend was more pronounced in the ROA model. The coefficient of the interaction term was 3.289 (p = 0.003), indicating a statistically significant positive effect. This result demonstrates the effect of the BioDiv Ratio was significantly stronger in the manufacturing sector compared to the nonmanufacturing sector. As shown in Figure 1, the slope of the BioDiv Ratio is clearly positive for manufacturing firms, whereas it is nearly flat for nonmanufacturing firms, making the difference in effects apparent. This provides important evidence that environmentally conscious management not only improves corporate reputation but also translates into tangible financial returns.

These findings indicate that the economic significance of biodiversity conservation activities varies across industries. In particular, such efforts may carry strategic value for firms in the manufacturing sector. This suggests that the effects of CSR should not be viewed uniformly across all firms but rather considered in light of industry-specific contexts and conditions.

However, the panel data analysis that accounted for firm and year fixed effects showed trends different from those in the main analysis (

Table 6. FE model results: manufacturing sector. Source: authors’ calculations using R.

Dependent Variable	Term	Estimate	Std. Error	t. Value	p. Value	Signif.
ROA	BioDiv Ratio	−0.248	0.646	−0.383	0.702
	Size	0.084	0.062	1.368	0.172
	Adjusted R-squared	−0.308
PBR	BioDiv Ratio	−0.474	0.229	−2.064	0.039	*
	Size	0.010	0.019	0.511	0.610
	Adjusted R-squared	−0.309

* p < 0.05, n = 777.

and

Table 7. FE model results: nonmanufacturing sector. Source: authors’ calculations using R.

Dependent Variable	Term	Estimate	Std. Error	t. Value	p. Value	Signif.
ROA	BioDiv Ratio	−0.380	0.162	−2.345	0.020	*
	Size	0.123	0.073	1.683	0.094
	Adjusted R-squared	−0.296
PBR	BioDiv Ratio	−0.070	0.043	−1.623	0.106
	Size	−0.045	0.056	−0.799	0.425
	Adjusted R-squared	−0.331

* p < 0.05, n = 302.

). In the ROA model, the coefficient of BioDiv Ratio was negative for both manufacturing and nonmanufacturing firms, with a significant negative relationship observed in the nonmanufacturing sector (p = 0.020). In the PBR model, a significant negative coefficient (β = –0.380, p = 0.039) was found for manufacturing firms, whereas no significant effect was observed for nonmanufacturing firms. These results indicate that when firm-specific factors and common year shocks were strictly removed, the positive relationship between biodiversity conservation and financial performance observed in the main analysis was no longer confirmed. In particular, the characteristics inherent to the manufacturing sector were eliminated under the firm fixed effects, likely causing this discrepancy. The positive relationship for manufacturing observed in the main analysis is therefore likely attributable to environmental factors common to the manufacturing sector as a whole.

5.3. Robustness Check Using Lagged Models

Given the possibility that the effects of biodiversity conservation may materialize over time, we employed a one-year lagged variable (BioDiv Ratio_Lag1) and performed a regression analysis incorporating a manufacturing dummy and its interaction term (

Table 8. Lagged analysis results: impact of BioDiv Ratio on ROA. Source: authors’ calculations using R.

Term	Estimate	Std. Error	t. Value	p. Value	Signif.
(Intercept)	6.847	1.005	6.812	0.000	***
Biodiv_Ratio_Lag1	−0.109	0.295	−0.369	0.712
Manufacturing	1.372	0.337	4.069	0.000	***
Size	−0.012	0.011	−1.094	0.274
Year 2019	−0.274	0.397	−0.690	0.490
Year 2020	−1.051	0.428	−2.455	0.014	*
Year 2021	−1.490	0.439	−3.394	0.001	***
Year 2022	−0.470	0.470	−1.001	0.317
Biodiv_Ratio_Lag1: manufacturing	3.668	1.258	2.916	0.004	**
Adjusted R-squared	0.046

*** p < 0.001, ** p < 0.01, * p < 0.05, n = 826.

and

Table 9. Lagged analysis results: impact of BioDiv Ratio on PBR. Source: authors’ calculations using R.

Term	Estimate	Std. Error	t. Value	p. Value	Signif.
(Intercept)	2.156	0.655	3.291	0.001	**
Biodiv_Ratio_Lag1	0.203	0.355	0.571	0.568
Manufacturing	0.168	0.167	1.003	0.316
Size	−0.009	0.007	−1.450	0.148
Year 2019	−0.003	0.118	−0.023	0.981
Year 2020	0.090	0.281	0.319	0.750
Year 2021	0.156	0.133	1.172	0.241
Year 2022	−0.070	0.124	−0.561	0.575
Biodiv_Ratio_Lag1: manufacturing	0.545	0.415	1.313	0.189
Adjusted R-squared	0.004

** p < 0.01, n = 826.

). This analysis aims to verify the robustness of the results obtained in the previous section and to examine the potential impact of biodiversity conservation expenditures on CFP in the following fiscal year.

The analysis revealed that the effect of BioDiv Ratio_Lag1 in the nonmanufacturing sector was not statistically significant for either ROA or PBR (ROA: β = −0.109, p = 0.712; PBR: β = 0.203, p = 0.568). In the ROA model, the interaction term was positive and statistically significant (β = 3.668, p = 0.004), indicating that the BioDiv Ratio had a significantly greater effect in the manufacturing sector than in the nonmanufacturing sector. In the PBR model, the interaction term was also positive (β = 0.545), suggesting a greater effect in the manufacturing sector than in the nonmanufacturing sector; however, this difference was not statistically significant (p = 0.189).

These results confirm the trend observed in Section 5.2. Specifically, in the ROA model, the interaction term was statistically significant and positive, indicating that the effect of biodiversity conservation expenditures was significantly greater in the manufacturing sector than in the nonmanufacturing sector, even in the lagged model. In the PBR model, the interaction term was also positive, suggesting a larger effect in the manufacturing sector; however, this difference was not statistically significant. In the nonmanufacturing sector, no significant effects were observed in either model, supporting the robustness of the industry differences identified in the previous section. A multicollinearity check using VIF in the regression models presented in Section 5.2 and Section 5.3 confirmed that all explanatory variables had VIF values below 2, indicating no concern regarding multicollinearity.

Furthermore, to examine the possibility that the observed relationship might be explained by unmeasured factors, we conducted a sensitivity analysis, drawing on the idea of Rosenbaum bounds. The analysis indicated that, in the ROA model, an unmeasured factor would need to explain more than 10.23% of the residual variance to nullify the estimated effect, and approximately 3.85% to render it statistically insignificant. It implies that the results are reasonably robust, as statistical significance was maintained even when assuming confounding of a magnitude comparable to that of firm size.

On the other hand, in the PBR model, the confounding required to nullify the estimated effect was as small as 3.65%, and statistical significance was not observed under the current conditions. Based on the above, the effect of BioDiv Ratio_Lag1 on ROA is estimated to be relatively robust, whereas its effect on PBR remains only suggestive and should be interpreted with caution.

In this section as well, we conducted tests using a lagged model with firm and year fixed effects (

Table 10. Lagged analysis results (FE model): manufacturing sector. Source: authors’ calculations using R.

Dependent Variable	Term	Estimate	Std. Error	t. Value	p. Value	Signif.
ROA	BioDiv Ratio	−1.570	0.845	−1.859	0.064
	Size	0.048	0.073	0.652	0.515
	Adjusted R-squared	−0.399
PBR	BioDiv Ratio	−0.284	0.358	−0.792	0.429
	Size	−0.024	0.020	−1.203	0.230
	Adjusted R-squared	−0.387

n = 594.

and

Table 11. Lagged analysis results (FE model): nonmanufacturing sector. Source: authors’ calculations using R.

Dependent Variable	Term	Estimate	Std. Error	t. Value	p. Value	Signif.
ROA	BioDiv Ratio	−0.058	0.138	−0.419	0.676
	Size	0.176	0.076	2.329	0.021	*
	Adjusted R-squared	−0.404
PBR	BioDiv Ratio	0.000	0.011	−0.035	0.972
	Size	0.001	0.010	0.114	0.909
	Adjusted R-squared	−0.453

* p < 0.05, n = 232.

). As a result, the positive and significant effects observed in the OLS models disappeared entirely, consistently showing negative impacts. This finding suggests that the effects observed in the main analysis were likely capturing structural differences between firms with inherently high levels of biodiversity conservation efforts and those with lower levels.

In light of the results presented in Section 5.2 and Section 5.3, this study further examined differences across subindustries within the manufacturing sector, focusing on the impact of the BioDiv Ratio on both PBR and ROA. The results indicated that in the pharmaceutical and precision instruments industries, the PBR model showed statistically significant positive coefficients (p < 0.001). Furthermore, in the pharmaceutical industry, a significant positive effect was also observed in the ROA model. These findings may not be unrelated to the high level of environmental measures required in these industries. It should be noted, however, that this analysis is presented as supplementary due to limitations in the sample size for each subindustry. The detailed regression results are provided in Appendix A

Table A1. Regression results within the manufacturing sector for ROA and PBR. Source: authors’ calculations using R.

Dependent Variable	Term	Estimate	Std. Error	t. Value	p. Value	Signif.
ROA	(Intercept)	7.363	1.260	5.846	0.000	***
	BioDiv Ratio	2.281	0.856	2.666	0.008	**
	Chemicals	1.124	0.957	1.174	0.241
	Electrical Machinery	0.275	0.949	0.290	0.772
	Fabricated Metal Products	−2.086	1.242	−1.680	0.093
	Food Products	−0.099	1.011	−0.098	0.922
	Iron and Steel	−3.719	1.201	−3.098	0.002	**
	Machinery	−0.356	1.001	−0.355	0.723
	Nonferrous Metals	−1.516	1.212	−1.250	0.212
	Pharmaceuticals	3.652	1.052	3.473	0.001	***
	Precision Instruments	0.732	1.227	0.597	0.551
	Pulp and Paper	−2.921	1.313	−2.224	0.026	*
	Rubber Products	−0.647	1.161	−0.557	0.577
	Textile Products	−1.873	1.624	−1.154	0.249
	Transportation Equipment	−0.567	0.980	−0.579	0.563
	Miscellaneous Products	−1.334	1.156	−1.154	0.249
	Size	−0.009	0.010	−0.919	0.358
Adjusted R-squared	0.118
PBR	(Intercept)	1.136	0.375	3.029	0.003	**
	BioDiv Ratio	0.613	0.255	2.408	0.016	*
	Chemicals	0.437	0.285	1.534	0.125
	Electrical Machinery	0.312	0.283	1.104	0.270
	Fabricated Metal Products	−0.474	0.370	−1.283	0.200
	Food Products	0.535	0.301	1.777	0.076
	Iron and Steel	−0.895	0.357	−2.502	0.013	*
	Machinery	0.076	0.298	0.256	0.798
	Nonferrous Metals	−0.644	0.361	−1.786	0.075
	Pharmaceuticals	1.194	0.313	3.812	0.000	***
	Precision Instruments	1.285	0.365	3.520	0.000	***
	Pulp and Paper	−0.669	0.391	−1.711	0.087
	Rubber Products	−0.548	0.346	−1.585	0.113
	Textile Products	−0.491	0.483	−1.016	0.310
	Transportation Equipment	−0.494	0.292	−1.692	0.091
	Miscellaneous Products	−0.279	0.344	−0.811	0.418
	Size	0.003	0.003	0.994	0.321
Adjusted R-squared	0.177

*** p < 0.001, ** p < 0.01, * p < 0.05, n = 777.

.

5.4. Reverse Direction Analysis

This section examines the reverse direction of the relationship analyzed in Section 5.3, namely the effect of CFP on corporate biodiversity conservation expenditures. Specifically, we conducted a regression analysis that included a manufacturing dummy variable and its interaction terms (ROA_Lag1 × Manufacturing, PBR_Lag1 × Manufacturing) (

Table 12. Interaction model results: impact of prior-year ROA on BioDiv Ratio. Source: authors’ calculations using R.

Term	Estimate	Std. Error	t. Value	p. Value	Signif.
(Intercept)	0.005	0.054	0.097	0.923
ROA_Lag1	−0.005	0.005	−1.001	0.317
Manufacturing	−0.137	0.033	−4.127	0.000	***
Size	0.002	0.000	3.282	0.001	**
Year 2019	0.002	0.020	0.118	0.906
Year 2020	0.014	0.023	0.620	0.536
Year 2021	0.018	0.023	0.757	0.449
Year 2022	0.013	0.023	0.589	0.556
ROA_Lag1:manufacturing	0.010	0.005	1.774	0.077
Adjusted R-squared	0.038

*** p < 0.001, ** p < 0.01, n = 826.

and

Table 13. Interaction model results: impact of prior-year PBR on BioDiv Ratio. Source: authors’ calculations using R.

Term	Estimate	Std. Error	t. Value	p. Value	Signif.
(Intercept)	0.050	0.050	0.997	0.319
PBR_Lag1	−0.059	0.020	−2.897	0.004	**
Manufacturing	−0.177	0.036	−4.984	0.000	***
Size	0.002	0.000	3.359	0.001	***
Year 2019	0.004	0.020	0.218	0.827
Year 2020	0.014	0.024	0.609	0.543
Year 2021	0.013	0.023	0.574	0.566
Year 2022	0.010	0.022	0.453	0.650
PBR_Lag1:manufacturing	0.075	0.022	3.364	0.001	***
Adjusted R-squared	0.047

*** p < 0.001, ** p < 0.01, n = 826.

). As a result, the interaction term between ROA_Lag1 and manufacturing was not statistically significant (β = 0.010, p = 0.077), whereas the interaction term between PBR_Lag1 and manufacturing showed a positive relationship that was statistically significant at the 1% level (β = 0.075, p = 0.001). These results suggest that, in the manufacturing sector, firms with higher market valuations tended to increase their biodiversity conservation expenditure ratios in the following year. In contrast, for nonmanufacturing firms, the effect of lagged ROA was not statistically significant, whereas lagged PBR showed a significant negative effect (β = −0.059, p = 0.004). In other words, firms with higher market. Value tended to exhibit a lower BioDiv Ratio in the following year. In the nonmanufacturing sector, production activities are less closely tied to natural resources than in the manufacturing sector. As a result, firms with higher market valuations may have lower incentives to increase biodiversity-related spending. This may be because companies with strong market positions prioritize other strategic investments or feel less need to further enhance their environmental image. These results support Hypotheses 6, but not 5.

VIF values for all explanatory variables in each regression model were below 4, indicating no concern regarding multicollinearity. Furthermore, sensitivity analyses were conducted to evaluate the robustness of these interaction terms. In the ROA model, it was found that an unobserved confounder would need to explain more than 7.1% of the residual variance in order to reduce the estimated effect of the interaction term (ROA_Lag1 × Manufacturing) to zero. This indicates that the effect itself is relatively robust. However, only 0.5% confounding would be sufficient to eliminate its statistical significance at the 5% level, suggesting some fragility in terms of maintaining significance and requiring cautious interpretation. In the PBR model, eliminating the estimated effect of the interaction term (PBR_Lag1 × Manufacturing) would require 11.2% confounding, while eliminating statistical significance would require 4.9% confounding. These thresholds are higher than those in the ROA model, indicating a greater degree of robustness in the estimation. While this analysis does not provide direct evidence of causality, it offers important insight into the potential bidirectionality of the relationship between CSR and CFP. Further research, including qualitative investigations, is needed to explore the underlying reasons for the differences across industries.

In the supplementary analysis using the FE model (

Table 14. FE model results: manufacturing sector, impact of prior-year CFP on BioDiv Ratio. Source: authors’ calculations using R.

Dependent Variable	Term	Estimate	Std. Error	t. Value	p. Value	Signif.
ROA	BioDiv Ratio	−0.004	0.006	−0.656	0.512
	Size	0.005	0.006	0.909	0.364
	Adjusted R-squared	−0.387
PBR	BioDiv Ratio	−0.018	0.014	−1.317	0.188
	Size	0.005	0.005	0.997	0.320
	Adjusted R-squared	−0.382

n = 594.

and

Table 15. FE model results: nonmanufacturing sector, impact of prior-year CFP on BioDiv Ratio. Source: authors’ calculations using R.

Dependent Variable	Term	Estimate	Std. Error	t. Value	p. Value	Signif.
ROA	BioDiv Ratio	−0.004	0.005	−0.793	0.429
	Size	0.002	0.004	0.503	0.616
	Adjusted R-squared	−0.449
PBR	BioDiv Ratio	−0.053	0.038	−1.399	0.164
	Size	0.002	0.004	0.690	0.491
	Adjusted R-squared	−0.437

n = 232.

), the positive and significant effect for manufacturing observed in the main analysis was not confirmed. Specifically, in both the ROA and PBR models, the coefficients of ROA_Lag1 and PBR_Lag1 were negative for both manufacturing and nonmanufacturing firms, but none were statistically significant. Although the main claim of this study—that the relationship between CSR and financial performance is bidirectional and varies by industry—was not supported by the more rigorous causal examination based on within-firm variation in the FE model, it remains an important issue for understanding macro-level characteristics across firms.

6. Discussion and Conclusions

6.1. Key Findings

The primary objective of this study was to empirically examine how corporate efforts toward biodiversity conservation influence CFP in the context of Japanese firms. While a substantial body of prior CSR research [22,38,39] has investigated the relationship between general environmental practices and corporate value, empirical studies focusing specifically on biodiversity remain scarce. By addressing this gap, the present study analyzes how biodiversity conservation affects financial outcomes, thereby providing a more nuanced understanding of its effectiveness. Furthermore, this study contributes by examining the reverse direction of influence—that is, how CFP affects the proportion of expenditures allocated to biodiversity conservation—thereby offering insights into the potential directionality of the relationship.

The analysis utilized data from the CSR Company Data: Environment Edition [35]. The independent variable was the BioDiv Ratio, defined as the proportion of expenditures on biodiversity conservation to total environmental conservation expenditures. This measure was used to quantitatively capture the strategic positioning of biodiversity conservation within a company’s overall environmental conservation efforts. The dataset is particularly robust because it is based on actual monetary expenditures. In contrast, self-reported CSR ratings or disclosure indices may be subject to impression management or selective reporting.

The dependent variables were the PBR, representing market valuation, and ROA, indicating profitability. Multiple regression analyses were conducted while controlling for year, firm size, and industry. In addition, drawing on theoretical suggestions that the effectiveness of CSR initiatives may vary by industry characteristics, as well as practical considerations, such as the relatively high environmental burden of the manufacturing sector, this study analyzes the differences between manufacturing and nonmanufacturing firms. Furthermore, this study contributes by examining the reverse direction of influence—that is, how CFP affects the BioDiv Ratio—thereby offering insights into the potential directionality of the relationship.

A series of regression analyses was conducted stepwise. The first model, based on the entire sample, revealed no statistically significant relationship between the BioDiv Ratio and the financial performance indicators ROA and PBR.

To examine potential industry-specific differences, a model incorporating an interaction term between the BioDiv Ratio and a manufacturing dummy was employed.

With respect to ROA, the results for the nonmanufacturing sector were consistently negative and statistically insignificant across all models, whereas the interaction term was positive and statistically significant, clearly indicating a difference between manufacturing and nonmanufacturing sectors. This finding suggests the potential presence of a positive effect in the manufacturing sector. In contrast, for the PBR, the results for the nonmanufacturing sector were positive but statistically insignificant, and the interaction term was also not significant. Accordingly, no significant difference between the two sectors was observed, and the possibility of a positive effect in the manufacturing sector remains only tentative.

In the analysis focusing on the reverse direction, the impact of the previous year’s ROA and PBR on the subsequent year’s BioDiv Ratio was examined. In the nonmanufacturing sector, the ROA model revealed a negative but statistically insignificant effect, while the PBR model showed a negative and statistically significant effect. This suggests firms with higher market valuations tend to allocate a smaller proportion of their environmental expenditures to biodiversity conservation. In contrast, in the manufacturing sector, the interaction term was positive in both models. The ROA model indicated a favorable effect, and the PBR model demonstrated a positive and statistically significant effect. This is consistent with slack resources theory, which suggests that firms with greater financial capacity are more likely to engage in biodiversity initiatives.

To rigorously examine causality, panel data analyses incorporating firm and year fixed effects were conducted for all analyses. As a result, clear and consistent differences were observed between the OLS estimates and FE estimates. These results were attributable to differences in the nature of the variation captured by the two methods. The OLS analysis reflects both persistent differences across firms and year-to-year variation within firms, whereas the FE model removes firm-specific fixed factors and estimates effects based solely on within-firm annual variation. Therefore, the positive effects observed in the OLS analysis were likely reflecting structural differences between groups of firms with persistently high levels of both the BioDiv Ratio and CFP and those without such characteristics. In the FE analysis, even when the same firm increased its BioDiv Ratio in a given year, no tendency for CFP to improve in the following year was observed. This suggests that the effects of environmental investments may not materialize in the short term, or that the associated costs of such investments may temporarily depress financial indicators. These findings indicate that the relationship between biodiversity conservation efforts and CFP should not be understood as a simple, one-way causal link. Rather, it needs to be captured within a multi-layered model that takes into account firm characteristics, industry structure, and the time lag in investment effects.

6.2. Discussion

The finding that the BioDiv Ratio has a statistically significant positive effect on ROA in the manufacturing sector is likely related to the fact that this sector is subject to greater societal scrutiny and regulatory pressure regarding environmental impacts compared to other sectors. However, in the FE model based on within-firm variation, this positive effect was not statistically significant, suggesting that short-term increases in biodiversity conservation expenditures do not necessarily lead to improved profitability in the following year. On the other hand, although no statistically significant effect was observed for the PBR, this result may reflect the inherent characteristics of the PBR, which are based on market expectations and external evaluations. Given its forward-looking nature, the PBR is more susceptible to various exogenous factors, potentially diluting the observable impact of biodiversity conservation efforts. For instance, factors beyond the core implementation of biodiversity conservation—such as the timing and content of related disclosures, investor attention to themes unrelated to biodiversity, stock price volatility, or media coverage—may have a substantial influence on market valuation. The study did not control for these factors, as detailed data on them were unavailable. The result may also reflect how firms communicate their biodiversity initiatives to external stakeholders. Market valuation could be influenced not only by financial resources devoted to biodiversity initiatives but also by how these initiatives are framed in corporate reports, which can shape investor perceptions.

The finding that biodiversity conservation efforts positively affect CFP in the manufacturing sector was more pronounced in ROA than in the PBR. This result is consistent with prior research. For instance, Elsayed (2023) [19] reported that biodiversity disclosure is positively associated with ROA at a significance level below 5%, whereas its association with the PBR, although positive, is not statistically significant. This study reinforces the theoretical trends identified in previous research on the effects of biodiversity conservation on CFP by providing support based on expenditure-oriented data. Furthermore, disaggregating the analysis into manufacturing and nonmanufacturing sectors contributes to refining the understanding of the effectiveness of biodiversity conservation initiatives through a contextualized examination grounded in industry-specific characteristics.

What distinguishes biodiversity conservation from other environmental preservation activities lies in its long-term approach to the foundation of a firm’s business operations. While many other environmental initiatives—such as CO₂ reduction—primarily aim to mitigate immediate environmental burdens, biodiversity conservation focuses on securing the sustainability of ecosystem services. This approach aligns with risk management strategies, as the loss of biodiversity can directly threaten supply chain stability and expose firms to regulatory or reputational risk. At the same time, it offers long-term benefits, such as opportunities for product differentiation, development of eco-friendly products, or creation of innovative business models. Particularly in the manufacturing sector, which relies heavily on natural resources, these combined risk mitigation and benefits makes conservation efforts more likely to be viewed as strategic investments.

The observed tendency for biodiversity conservation efforts to exert a stronger positive effect on ROA than on the PBR suggests that such initiatives may first be reflected in internal operational efficiency and profitability, while their influence on external market evaluations may be more indirect and time-lagged. This staged manifestation of effects offers a potential pathway for explaining the causal link between biodiversity conservation and CFP, a relationship that is not directly established in this study.

From a practical perspective, this finding may inform managerial decision-making by indicating that, even if CSR initiatives appear to be mere costs in the short term, they may lead to financial benefits over the long run. For initiatives such as biodiversity conservation, where the effects take time to materialize, it is advisable to place less emphasis on short-term stock price movements or market valuations. Instead, improvements in internal efficiency, such as ROA, can serve as leading performance indicators. For investors, such internal improvements can signal a firm’s long-term growth potential and thus warrant attention beyond immediate market valuations.

The finding that biodiversity conservation efforts exert a positive effect on both profitability and market valuation in the manufacturing sector—particularly in industries such as pharmaceuticals and precision instruments—may reflect industry-specific business environments and characteristics. Both the pharmaceutical and precision instrument industries share a strong connection with the natural environment. For instance, in the pharmaceutical sector, compounds derived from plants, animals, and microorganisms are frequently used as raw materials for drug development. Thus, biodiversity richness serves as a critical source for future research and development, as well as for product differentiation. In the precision instrument sector, firms are expected not only to minimize environmental burdens in sourcing materials and manufacturing processes but also to address stringent requirements for product recycling and post-use management. In this respect, environmental considerations are required throughout the entire product life cycle—an approach that aligns with the concept of product stewardship, one of the key environmental strategies under the NRBV. When corporate efforts toward biodiversity conservation are closely linked to the structural characteristics of a firm’s industry, such efforts may constitute not merely acts of social contribution but strategic investments that enhance competitive advantage, as posited by the NRBV. This study extends the theoretical scope of the NRBV by conceptualizing biodiversity conservation as a strategic driver of competitiveness grounded in industry-specific resource dependencies.

6.3. Limitations

This study has several limitations. First, as this research is based on archival data and employs an empirical design, none of the analyses can clearly establish causality. This limitation arises from the inability to isolate exogenous variation in biodiversity conservation efforts within the available dataset.

Second, the scope of this study is limited by both the number of firms and the period covered, constraining the ability to examine medium- to long-term effects. Because of the reduced sample size, only a one-year lag could be applied, which may not fully capture the longer-term market responses implied by this potential pathway. Moreover, as the analysis is based solely on Japanese firms, the generalizability of the findings to other contexts remains limited.

Third, firms engage in various CSR activities beyond biodiversity conservation, such as climate initiatives or community investment, which may be correlated with biodiversity initiatives. This overlap makes it challenging to isolate the specific effort of biodiversity conservation on CFP. In addition, the adjusted R² values from the regression analyses in this study were generally low, ranging from approximately –0.02 to 0.05, indicating that the models explained only a limited proportion of the variation in ROA and the PBR. This reflects the fact that many factors influence CFP and suggests that the models in this study capture only part of that variation.

Fourth, the indicator BioDiv Ratio may be affected by fundamental differences in environmental compliance costs among industries. For example, industries facing heavy burdens from environmental regulation compliance tend to incur higher overall environmental costs. Consequently, even if biodiversity conservation expenditures remain constant, the BioDiv Ratio may appear relatively small in such cases. This indicator is meaningful in that it signifies the relative priority accorded to biodiversity conservation within a company’s environmental strategy; however, it is difficult to completely eliminate the influence of industry-specific structural factors.

Finally, the use of the FE model has the advantage of allowing causal inference by removing persistent firm-specific characteristics. However, it also has the limitation of excluding from the analysis long-term and persistent effects based on structural differences across firms. As a result, the FE model focuses specifically on the impact of short-term fluctuations and is not suitable for capturing long-term patterns or strategic characteristics observed in cross-firm comparisons. This characteristic is an important consideration when interpreting the differences between the OLS and FE results.

6.4. Future Research Directions

First, this study did not conduct a rigorous identification of causality. Future research should adopt more robust identification strategies, such as using instrumental variable methods or exploiting exogenous shocks like regulatory changes related to biodiversity conservation.

Second, expanding the sample size and extending the observation period would make it possible to examine medium- to long-term effects, particularly those with time lags in market-based indicators. Since long-term investments such as biodiversity conservation may require several years for their effects to materialize, the use of long-term time-series data would be especially valuable. In addition, future research could incorporate cross-country data or diverse institutional contexts to assess whether the patterns observed in this study are consistent across different environments.

Third, by appropriately controlling for other CSR activities mentioned in the study’s limitations, future research could better isolate the effect of biodiversity conservation efforts. An approach that clarifies the unique impact of biodiversity conservation— as a specific domain of CSR—on CFP would be desirable.

Fourth, this study used OLS as the main analysis and employed the FE model as a supplementary approach. Because the FE model removes persistent firm-specific characteristics, it cannot capture structural effects. Future research would benefit from employing methods that can integratively analyze both cross-firm differences and within-firm variation.

Finally, compared with other environmental issues, biodiversity conservation still lacks a widely adopted standardized disclosure framework, and quantitative indicators remain underdeveloped. As a result, corporate narratives in this area tend to reflect a company’s individuality and position to a greater extent. By combining narrative analysis and examining how biodiversity conservation activities are positioned and communicated in integrated reports and sustainability reports, future research could gain a deeper understanding of how such efforts are conveyed to external stakeholders. This approach could also provide richer insights into how these communications shape market response mechanisms.