Environmental Project Financing and Insolvency Risk: Moderating Effects of Public Relations Function and Climate Policy Uncertainty

Abstract

Despite growing emphasis on financial firms’ engagement in environmental project financing (EPF), its impact on insolvency risk remains underexplored. Drawing on signaling theory, this study investigates how EPF influences financial institutions’ insolvency risk and examines the moderating roles of the existence of a public relations function and global climate policy uncertainty. By analyzing data of 291 financial firms from major developed economies over 2014–2023, this study finds that environmental project financing reduces insolvency risk. Furthermore, the presence of a public relations function strengthens this effect, while the rising global climate policy uncertainty weakens it. This study contributes to existing literature while offering both strategic insights and practical implications.

1. Introduction

Climate shocks are increasingly becoming recurring realities, compelling financial markets to reconfigure their assumptions about risk, value, and resilience. Consequently, markets are placing a premium on green financial stewardship more than ever, not only to improve firm legitimacy but also to enhance resilience and long-term viability [1,2]. For financial institutions (banks and insurance firms), green finance lending is more than ethical posturing; it provides powerful signals to external markets regarding their adaptability, credibility, and foresight [3,4]. These signals, when properly decoded by investors and markets, can shield institutions against unexpected financial decline and default [5,6]. Therefore, understanding how green finance lending affects insolvency risk is fundamental to the stability of the global financial ecosystem.

A burgeoning literature examining this relationship has emerged. In China, ref. [7] shows that green finance notably improves firm performance among heavily polluting firms, primarily through enhanced bank monitoring of accounting quality and capital structure. A recent study by [8] covering bank-year observations across BRIC countries finds that greater environmental engagement by SME banks significantly improves their performance—an outcome that aligns with signaling theory, where credibly green behavior reflects ethical leadership and transparency. Ref. [9] reveals that green finance improves banks’ environmental performance. Similarly, a positive effect of green lending in lowering crash risk is observed by [10].

Despite the valuable contributions of these studies, the literature remains largely silent on the nuanced effects of green financing on the insolvency risk of financial institutions themselves. In particular, there is limited clarity on how green finance lending within financial institutions shapes investor perceptions and influences institutional insolvency risk. This paper aims to address this research gap. We shift the focus to financial institutions, systematically examining how green finance decisions operate as strategic signals and affect institutional insolvency risk. By integrating signaling theory with firm-level financial data, we aim to clarify how and why green lending matters for financial stability, thereby advancing both theory and practice in green finance.

While recent studies such as [11] on green loans and firm performance, [12] on green loans and bank risk, and [13] on credible green commitment and bank financing have advanced our understanding of green financing, our study is theoretically distinct along several important dimensions. First, prior work predominantly treats green lending either as an outcome (e.g., access to financing, firm performance) or examines its implications in non-financial firms, whereas we conceptualize environmental project financing by financial institutions themselves as a forward-looking signal of underlying asset quality and risk discipline, directly linking it to insolvency risk—a core stability metric largely overlooked in this literature. Second, and most importantly, our study introduces dual boundary conditions, a firm-level communication mechanism (public relations function) and a macro-level interpretive environment (climate policy uncertainty)—thereby capturing both signal transmission and signal reception contingencies, which remain largely absent in prior research.

This makes two important contributions to the existing literature. Our first contribution lies in extending the conversation on institutional insolvency risk into the domain of financial institutions themselves. Although earlier work has explored how environmental practices affect non-financial firms’ vulnerability to crash risk or poor performance, financial institutions (banks and insurance companies) have been largely overlooked, despite their central role in allocating green capital. By focusing directly on these institutions, this research brings them from the periphery to the center of inquiry. The focus on financial institutions is very important because the dynamics of risk transmission in financial institutions are different from those in manufacturing firms: the investment risk of financial institutions can ripple across entire markets. In positioning green finance lending as both a financial decision and a strategic signal, we enrich the literature on crash risk by offering insights that are tailored to the distinctive mechanisms of financial institutions. Secondly, we identify the boundary conditions under which the impact of green finance lending can raise or decrease the insolvency risk. We show that signals do not operate in a vacuum; their reception depends on both firm-level capacities and country-level environments. We show that green lending initiatives can mitigate insolvency risk when financial institutions have public relations functions. However, when they operate in an environment that has a higher global climate policy uncertainty index, the positive effect of EPF on insolvency risk is attenuated. This nuanced perspective advances the literature by clarifying the contingencies that condition the impact of adopting green initiatives.

2. Theoretical Framework

Signaling theory [14,15] has been widely used to explain organizational communication with external audiences under conditions of information asymmetry. Signaling theory posits that internal actors possessing private information—such as firms, managers, or employees—send observable signals to those on the outside who lack access to this knowledge. In the financial literature, it has been used to show how firms reduce uncertainty and convey otherwise unobservable qualities—whether it is managerial competence, firm performance potential, or commitment to social and environmental responsibility [11,16]. These signals enable receivers such as investors, regulators, and other stakeholders to make a better judgment about the quality of the firm.

The main crux of signaling theory is about how and why signals are transmitted. The basic assumption underneath the transmission of signals is the presence of information asymmetry [17]. In capital markets, firms and banks generally know more about the quality of their projects, risk management practices, and long-term strategic orientations than investors or regulators [18]. This imbalance creates a need for signaling. A second assumption is that signalers are motivated by self-interest; they transmit signals to gain access to resources, gain favorable evaluations, or reduce monitoring costs. A third assumption is that signals must carry some cost. The last important assumption is that receivers interpret signals based on earlier beliefs and observed outcomes, indicating that only those signals consistently backed by real quality will remain effective [19].

This theory also suggests that not every action qualifies as a meaningful signal. For a signal to function properly, it must satisfy several important characteristics. First, it must be observable: stakeholders should be able to clearly identify and evaluate the signal [20]. In the case of banks, environmental project financing must be reported in ways that are transparent and verifiable, rather than hidden in vague disclosures. Second, the signal must correlate with the unobservable quality it claims to represent [21]. If green financing is truly tied to effective risk management and sustainability outcomes, then investors can legitimately interpret it as a credible sign of long-term resilience [22]. Third, the signal should be sufficiently costly to prevent opportunistic imitation [23]. Undertaking large-scale environmental financing projects, for example, requires not only financial commitment but also institutional expertise and long-term dedication—barriers that low-quality or insincere banks are less likely to bear. We acknowledge that the existence of a separating equilibrium where the costs of engaging in environmental project financing are sufficiently higher for lower-quality firms to deter imitation, which cannot be directly observed in our empirical setting and is therefore treated as a maintained theoretical assumption underpinning our analysis. Fourth, effective signals must be reliable over time and resistant to manipulation. If stakeholders detect that signals are superficial or opportunistic, their effectiveness diminishes quickly. Finally, a signal must be clear, as unclear signals can be misinterpreted [11,17,18].

Based on signaling theory, environmental project financing serves as a particularly appropriate and powerful signal because it combines visibility, credibility, and cost. By channeling capital toward socially responsible projects with tangible environmental benefits, banks communicate to investors, regulators, and the wider public that they possess both the foresight and capacity to manage long-term risks. The strong commitment to sustainability ensures that only banks with genuine capabilities and a strong orientation toward sustainability are willing to send such a signal [24]. At the same time, the outcomes of these projects are observable—whether in the form of renewable energy capacity, reduced emissions, or improved environmental resilience—which allows stakeholders to verify the authenticity of the signal over time. By using EPF as a signal, it not only reduces information asymmetry by showing otherwise hidden characteristics such as managerial competence and risk management quality but also enhances investor confidence by demonstrating alignment with broader social and policies goals.

2.1. Environmental Project Financing and Financial Institutions’ Insolvency Risk

Firm reputation and market confidence provide one of the most compelling mechanisms through which EPF can shape a bank’s insolvency risk. From a signaling perspective, when a bank commits substantial resources to green lending, it conveys to the market that it is forward-looking, responsible, and willing to bear costs that only genuinely resilient institutions can sustain [11,16,25,26]. This visible commitment builds reputational capital, which functions as a protective buffer in times of uncertainty. Investors, regulators, and even the general public interpret such financing as a credible signal of stability and managerial prudence, which reduces the likelihood of sudden withdrawals of trust that often precede default [27,28]. A strong reputation, once established, also nurtures long-term relationships with institutional investors, who are less prone to short-term speculation and herd behavior, thereby reducing the bank’s insolvency risk. Banks that consistently demonstrate credibility through verifiable environmental financing are more likely to retain market confidence during stressful periods, reducing the chances that negative shocks spiral into insolvency risk. Importantly, reputational capital does not emerge in isolation; it is co-produced with enhanced transparency and credible disclosures, which jointly reinforce stakeholders’ confidence in the bank’s underlying quality.

Risk management capabilities are another mechanism through which EPF can influence a bank’s insolvency risk. Engaging in environmental lending is not a routine financial decision; it demands a sophisticated framework for evaluating long-term risks, including regulatory shifts, climate exposures, and industry transitions [29]. By undertaking such financing, banks send a powerful signal that they possess advanced internal systems and expertise for identifying and mitigating complex, often hidden risks. This matters greatly to investors because poorly managed risks are one of the primary triggers of insolvency risk [12]. When banks demonstrate through costly and observable commitments that they can effectively screen projects, monitor borrowers, and anticipate environmental contingencies, investors interpret this as evidence of strong managerial foresight and resilience. The credibility of the signal lies in it being difficult to imitate; only banks with genuine capabilities can afford to embed environmental considerations into lending decisions without undermining profitability. Thus, EPF becomes more than a lending strategy, it evolves into a visible and credible marker of institutional discipline in risk management, one that enhances market confidence and dampens the bank’s insolvency risk. These capabilities also enable more accurate and credible disclosures, thereby reducing information asymmetry and further strengthening the overall signaling effect.

Information asymmetry is considered one of the main challenges in the financial market, as investors often lack insight into the true quality of a bank’s lending portfolio and, hence, its underlying risk profile [17]. Considering that EPF is highly scrutinized by regulators, rating agencies, and stakeholders, banks are compelled to disclose additional information about their monitoring processes and impact assessments [29]. These disclosures serve as observable signals that reduce opacity, reassuring investors that the bank is not hiding adverse information. Moreover, transparency here is not costless; developing credible reporting systems, auditing environmental outcomes, and maintaining consistent disclosures requires significant effort and resources [16], which ensures that only banks with genuine commitments and capabilities can sustain such practices. Consequently, investors interpret these disclosures as credible signals of managerial integrity. Importantly, EPF serves as a platform for transparency through which banks not only differentiate themselves from opportunistic peers but also build the legitimacy of transparency, which helps them to maintain the trust that often precedes financial default. This reduction in information asymmetry simultaneously reinforces reputational capital and constrains managerial discretion, illustrating how transparency, reputation, and governance effects are jointly produced within the signaling process.

Defaults are often linked to managerial opportunism, particularly when top management conceal negative information for extended periods and then release it all at once and trigger sharp downward corrections [17]. EPF serves as a costly and credible signal that managers are oriented toward long-term responsibility rather than short-term opportunism [13]. Unlike symbolic signals or superficial ESG statements, committing substantial funds to environmental lending ties the bank’s reputation and financial resources to projects that unfold over years, often under the watchful eye of regulators, auditors, and civil society [30]. This visibility constrains managers from engaging in opportunistic behaviors such as withholding bad news or overstating performance, since inconsistencies between their signals and project outcomes would quickly erode credibility [18]. Investors, recognizing the difficulty of faking such costly commitments, interpret environmental lending as a sign that management is less likely to engage in opportunism, thereby reducing the perceived risk of hidden information surfacing suddenly [16,31]. Over time, the discipline imposed by costly signaling helps to stabilize investor expectations, lower suspicion of managerial misconduct, and reduce the probability of financial collapse. By doing so, EPF operates as a governance mechanism that protects banks from managerial withholding of important financial information, which could otherwise erode investor confidence and lead to financial default when revealed abruptly. Notably, reduced managerial opportunism further enhances disclosure credibility and strengthens reputational capital, completing a reinforcing cycle among governance discipline, transparency, and stakeholder trust.

Taken together, these mechanisms—reputational capital, risk management capabilities, information asymmetry reduction, and managerial opportunism deterrence—should not be viewed as independent channels but as mutually reinforcing components of a unified signaling process. EPF, as a costly and observable commitment, simultaneously enhances internal capabilities and external perceptions, with each mechanism amplifying the others to produce a stronger and more credible signal of bank stability. Based on the above discussion, we hypothesized the following.

H1. 

Environmental project financing reduces financial institutions’ insolvency risk.

Building on the above relationship, signaling theory suggests that the effectiveness of a signal is inherently contingent on both the firm’s ability to communicate the signal and the external environment in which the signal is interpreted. Accordingly, we incorporate two theoretically grounded moderators that capture these distinct yet complementary dimensions. First, the existence of a public relations function reflects the firm’s signal transmission capability, enhancing the visibility, clarity, and credibility of environmental project financing (EPF) to external stakeholders. Second, global climate policy uncertainty represents the signal interpretation context, shaping how investors and creditors decode and evaluate the informational value of EPF under varying levels of environmental policy ambiguity.

By jointly considering these moderators, we move beyond a direct-effect framework and provide a more nuanced explanation of when and why EPF serves as an effective signal in reducing default risk—capturing both firm-level communication mechanisms and macro-level interpretive conditions that are central to signaling processes. Importantly, we deliberately do not incorporate conventional board-level characteristics (e.g., independence, size, or diversity) because while they capture internal monitoring and governance quality, they do not directly shape the observability, clarity, or external credibility of signals; in contrast, the public relations function uniquely represents a firm’s dedicated signaling infrastructure, making it theoretically consistent with signaling theory as a moderator that directly conditions how effectively environmental project financing is communicated and interpreted by external stakeholders.

2.2. Moderating Role of the Public Relations Function

Thus far, we argue that banks and insurance firms rely on credible signals to reduce the information asymmetry between themselves and external stakeholders. Against this backdrop, EPF is one such signal indicating that the institution is not only attuned to emerging environmental risks but is also proactively allocating resources toward sustainable projects. Yet, the strength and clarity of this signal are contingent on how effectively it is communicated and perceived by the market. This is where public relations functions step in, serving as the channel that amplifies, clarifies, and legitimizes the signal. Hence, we contend that public relations (PR) functions play a moderating role in the relationship between EPF and insolvency risk in financial institutions. Prior research in corporate communication and reputation management emphasizes that firms do not merely “send signals,” but actively construct them through mediated communication systems that influence stakeholder interpretation [32,33].

We argue that environmental project financing decisions may remain muted if not actively communicated to investors, regulators, and clients. PR activities ensure that these efforts are not buried in annual reports or technical disclosures but are presented as deliberate strategic choices [20]. By packaging the narrative around environmental financing in ways that resonate with stakeholders’ concerns about climate risk, regulatory trends, and long-term sustainability, PR departments enhance the credibility and visibility of the bank’s signal. The strategic communication literature suggests that visibility is a precondition for reputational impact, as uncommunicated actions generate weak or noisy signals in capital markets [34,35]. A stronger, amplified signal reduces information asymmetry, thereby reducing the uncertainty that often leads to insolvency risk.

The public relations department plays a crucial role in authenticating the bank’s green signals by backing them with transparent metrics, third-party validations, and consistent messaging across platforms [36]. By embedding these assurances into press releases, media engagements, and stakeholder dialogues, PR functions reduce doubts about opportunism and improve market trust [37]. Research in corporate legitimacy and impression management shows that credibility is reinforced when firms align voluntary disclosures with external verification mechanisms such as audits, certifications, or ESG ratings [38,39]. Greater authenticity makes the signal more costly to mimic, aligning well with the fundamental assumptions of signaling theory and thus dampening the probability of insolvency risk initiated by investor reaction or reputational shocks [17]. In this way, PR functions amplify the deterring effect of EPF on insolvency risk by authenticating the signal to the market.

Lastly, signaling is not a one-off event; its effectiveness depends on consistency over time [19]. PR activities sustain and reinforce the signal of environmental lending by ensuring that it is not a sporadic announcement but part of an ongoing dialogue with stakeholders [40]. Regular media updates, sustainability reports, community engagements, and digital storytelling continuously remind markets of the firm’s green orientation. This continuous signal enforcement helps to prevent sudden erosion of trust or surprise revelations, both of which can drive the financial downturn of the financial institutions. This leads to the following hypothesis.

H2. 

The impact of environmental project financing on insolvency risk is stronger when a financial institution has a public relations function.

2.3. Moderating Role of Climate Policy Uncertainty

The effectiveness of any signal depends not only on the action itself but also on the interpretive context in which it is received [21]. In this context, banks’ and insurers’ environmental project financing decisions are not evaluated in isolation; they are filtered through broader perceptions of the environmental performance and environmental policies of the country in which these institutions operate [29]. This contextual lens influences the credibility, strength, and reception of the signal, thereby altering the extent to which environmental project financing reduces insolvency risk.

This study argues that in countries with less uncertain climate policies, stakeholders interpret EPF as an authentic, costly, and consistent signal of long-term commitment to sustainability. National-level consistent policies related to emissions control, renewable energy adoption, and environmental governance lend credibility to firms’ environment-related initiatives by creating a supportive institutional and regulatory environment [41]. Investors are more likely to believe that such lending is genuine and aligned with the broader national agenda rather than symbolic or opportunistic [42]. Strong national environmental policies are often accompanied by stable and forward-looking policies, which reassure investors that EPF is not a passing fad but a strategically viable choice [43]. The predictability of future regulatory support reinforces the durability of the green signal. Investors interpret such lending as not only a current commitment but also a reliable pathway toward future returns, lowering the probability of stock price crashes driven by uncertainty [44]. Enhanced credibility makes the green lending signal more costly to mimic, which reduces information asymmetry and strengthens its default mitigating effect.

However, major developed countries such as the G7 increasingly face rising global climate policy uncertainty despite their strong institutional frameworks and significant global carbon footprint. This uncertainty stems from their security-driven energy transitions, inconsistent long-term regulatory commitments, and disruptive global events such as the Russia–Ukraine war and COVID-19 [45,46]. Climate policy uncertainty persists when regulatory frameworks lack predictability, consistency, and transparency, particularly in policy alignment aimed at transitioning towards a low-carbon global economy [47]. In such contexts, stakeholders face a fundamental tension between two opposing effects of climate policy environments on EPF signaling effectiveness.

On the one hand, stable and coherent climate policy environments enhance the informational value of EPF by improving signal credibility and interpretability. In contexts where climate policy certainty is high, when a bank or insurer’s green lending aligns with the strong environmental track record of its home country, the signal gains added resonance. Conversely, in environments with weak climate policy certainty, investors perceive a misalignment between firm-level actions and national climate priorities, diminishing the authenticity of corporate efforts [43]. As a result, when climate policies are clear and stable, this amplified signaling effect reassures investors that firms’ EPF is part of a structural shift, thereby lowering uncertainty and reducing insolvency risk.

On the other hand, we acknowledge that highly ambitious or rapidly evolving climate policy environments may simultaneously generate transition risk uncertainty, particularly for financial institutions engaged in long-term green lending. Specifically, stringent or frequently changing regulatory commitments may increase compliance costs, reversal risk, and the credit risk of financed green projects, thereby introducing new sources of financial fragility [30,48]. This suggests a countervailing mechanism in which climate policy uncertainty may weaken the financial stability benefits of EPF by increasing perceived regulatory and transition risk exposure.

We reconcile these two competing mechanisms by arguing that, in the context of climate policy uncertainty, the dominant effect operates through the deterioration of signal interpretability rather than through transition risk exposure. When policy environments are unstable, investors are less able to distinguish between genuine long-term commitments and short-term strategic positioning, thereby weakening the signaling value of EPF more than it affects its operational cost structure. At the same time, a country with consistent climate policies creates reputational spillovers that elevate the status of all domestic institutions operating within it [49]. For banks and insurers, this means that their EPF activities are perceived as a positive backdrop that amplifies reputational benefits. Such reputational capital functions as insurance in times of financial stress; even when shocks occur, markets are less likely to penalize institutions severely when they are associated with a country that has demonstrated credible environmental stewardship. Conversely, when firms operate in countries with high policy volatility, any perceived inconsistency in EPF is more likely to be penalized due to weaker institutional anchoring and higher interpretive ambiguity. Hence, based on this discussion, we formulate the following hypothesis.

H3. 

The impact of environmental project financing on insolvency risk is weaker in environments with greater global climate policy uncertainty.

3. Data and Methodology

This study focuses on 291 financial firms from six developed economies, including Canada, Germany, France, Japan, the United Kingdom, and the United States, over the period 2014 to 2023. Italy is excluded from the G7 sample due to the unavailability of country-level GCPU data for the full study period. These countries are selected due to their substantial global economic importance, collectively accounting for approximately 45% of global GDP, and for possessing highly developed and sophisticated financial markets compared to the rest of the world [50,51]. Following the 2008 global financial crisis, these economies implemented significant financial and institutional reforms aimed at strengthening their financial systems, including the banking, insurance, mortgage sectors, etc. [52]. Moreover, their industrial and economic advancement has contributed to a notable rise in carbon emissions, which collectively represented about 33% of total global emissions in 2019 [53]. Given this evolving landscape, being the political and most dominant economic leaders in the world and being responsible for a major share of global pollution, these major economies have increasingly prioritized climate action initiatives and leveraged their financial sectors as key instruments for achieving sustainability goals [54]. Therefore, this context provides a compelling and relevant setting for conducting this research.

To construct the study sample, all publicly listed financial firms from the selected economies that engaged in environmental project financing at least once during the sample period were first identified. From this initial pool, firms with at least five consecutive years of available data, including complete information for all key variables, were selected. The sample excludes firms that never undertake environmental project financing and focuses on firms with variation in EPF engagement over time, allowing the key independent variable to switch between 0 and 1 within firms. To mitigate the influence of outliers and ensure data consistency, all financial variables were winsorized at the 1% and 99% levels. This selection process yielded a final balanced sample of 291 financial firms. Though this sampling approach is standard in panel data studies as it enhances comparability and consistency across observations and ensures data reliability, it may exclude firms with incomplete reporting histories or those that entered or exited the market during the sample period. As such, the results should be interpreted with this consideration in mind. To extract firm- and country-level information, this study utilizes data from Refinitiv, firms’ annual reports, official websites, Figshare, the World Bank, and the World Governance Indicators databases.

Table 1. Sample Distribution.

Countries	Number of Firms
Canada	42
Germany	21
France	15
Japan	56
United Kingdom	71
United States	86
Total	291
Sectors	Number of Firms
Insurance	53
Capital Markets	77
Banks	99
Consumer Finance	22
Financial Services	40
Total	291

provides the sample distribution across countries and industries.

Measures and Empirical Strategy

The dependent variable of this study is insolvency risk (Insol_Risk). This study follows [55,56,57] and employs the Z-score as a measure of financial firms’ financial stability. While some studies [55] refer to this measure as default risk, the Z-score is more precisely interpreted as a measure of distance to insolvency. Specifically, the Z-score is calculated as (ROA + (Total Equity/Total Assets)) divided by the standard deviation of ROA. The Z-score is inversely related to bankruptcy, as it measures a firm’s distance to default. Therefore, a higher Z-score indicates a lower probability of default, while a lower Z-score suggests a higher likelihood of default [55].

The key independent variable in this study is environmental project financing (Env_Proj). To measure this variable, the study uses data from Refinitiv, which provides a dummy indicator taking the value of 1 if a firm evaluates a project based on biodiversity and environmental risks prior to offering funding, and 0 otherwise [13]. Furthermore, this study examines the moderating effect of two variables. First, the study incorporates a firm-level moderating variable, namely the existence of a public relations function (Pub_Rel). To measure the Pub_Rel function (for further detail, please see [20]), in line with established archival content analysis approaches, the study follows [20,58] and constructs a dummy variable that takes the value of 1 if a firm reports or maintains a formal public relations structure—such as a department, unit, team, or dedicated communication office—and 0 otherwise. This identification is based on a systematic review of annual reports, in which we search for references to communication-related organizational functions, including public relations, regulatory affairs, government relations, and external affairs. This procedure allows us to capture the presence of an institutionalized communication capacity within firms. The presence of a public relations function enables firms to effectively bridge the gap between key stakeholders and the organization [20].

Second, the country-level moderating variable is the Global Climate Policy Uncertainty (GCPU) Index (Please see [44] for further detail on the construction of this index). Following [44,59], this study obtains the GCPU index from Figshare. The index is constructed using the methodology of [60], based on climate policy-related news coverage from leading newspapers across twelve major economies, including Canada, Japan, Germany, France, the United States, and the United Kingdom. The dataset provides country-specific GCPU series derived from this multi-country news corpus. In this study, we use the country-level GCPU series corresponding to the six developed economies in our sample. A higher GCPU value reflects a greater degree of ambiguity and unpredictability concerning future climate-related regulations, policies, and governmental actions. This measure effectively captures the macro-level uncertainty that may influence firms’ strategic environmental financing decisions. Furthermore, to enhance the credibility and robustness of the empirical model, the study includes several control variables in line with prior research [56,57,61]. The firm-level controls comprise financial leverage (Lever) measured as the debt-to-equity ratio, firm size (Size) measured as the logarithm of total assets, profitability (Prof) measured as the return on assets, board size (Bor_Size) as the total number of board members, and board gender diversity (Bor_Gend) measured as the ratio of female to total board members [62,63,64,65,66]. Additionally, this study incorporates country-level control variables, including the logarithm of GDP per capita (GDP_Cap) and institutional regulatory quality (Regu_Qua), measured using the World Governance Indicators (WGI) regulatory quality index [67,68].

To empirically examine the impact of environmental project financing on insolvency risk, as well as the moderating effects of the existence of public relations functions and the Global Climate Policy Uncertainty Index, this study employs the two-step system Generalized Method of Moments (GMM) estimator. This panel estimation technique is particularly suitable for addressing potential econometric challenges such as heteroskedasticity arising from the heterogeneous nature of the sample, serial correlation, and endogeneity [69,70,71]. The two-step system GMM utilizes internal instruments, such as lagged levels and differences in endogenous variables, to mitigate biases associated with simultaneity and omitted variable concerns, thereby producing consistent and asymptotically efficient estimates [18,62,72]. To further validate the robustness of the results, the study reports the Arellano–Bond AR(2) and Hansen tests, where statistically insignificant p-values indicate the absence of serial correlation and overidentification, confirming the overall reliability of the estimated model.

4. Empirical Results

4.1. Descriptive and Correlation

Table 2. Descriptives.

Variable	Acronyms	Obs	Mean	Std. Dev.	Min	Max
Insolvency risk	Insolv_Risk	2910	12.700	6.454	3.607	30.896
Environmental Project Financing	Env_Proj	2910	0.346	0.476	0.000	1.000
Public Relations Function	Pub_Rel	2910	0.340	0.474	0.000	1.000
Global Policy Uncertainty Index	GCPU	2910	2.657	1.124	0.000	4.824
Leverage	Lever	2910	1.229	0.847	0.300	4.982
Firm Size	Size	2910	20.954	2.182	11.962	26.609
Profitability	Prof	2910	0.379	0.839	−0.141	3.100
Board Size	Bor_Size	2910	11.210	3.538	3.000	28.000
Board Gender Diversity	Bor_Gend	2910	24.392	14.437	0.000	100.000
GDP Per Capita	GDP_Cap	2910	10.792	0.224	10.427	11.324
Regulatory Quality	Regu_Qua	2910	1.504	0.202	1.059	1.880

presents the descriptive statistics of all included variables. The mean value of the z-score (Insolv_Risk) is 12.7, suggesting a moderate-to-low insolvency risk for the sample firms. The average value of environmental project financing (Env_Proj) is 0.346. For the public relations function (Pub_Rel), the mean value is 0.34, indicating that approximately 99 firms in the sample have a public relations function. The mean value of GCPU is 2.65, indicating a moderate-to-low global climate policy certainty prevailing in this region across the sample period. On average, the sample firms have a leverage ratio (Lever) of 1.2, a size of 20.9, a profitability (Prof) of 0.37, a board size (Bor_Size) of 11 members, and a board gender diversity ratio (Bor_Gend) of 24.4. The means of the financial variables are close to their respective standard deviations, suggesting their standardization.

Table 3. Correlation.

No.	Variable	1	2	3	4	5	6	7	8	9	10	11	12	13
1	Insolv_Risk	1
2	Env_Proj	−0.00	1
3	Pub_Rel	−0.01	−0.03	1
4	GCPU	0.02	−0.12	0.01	1
5	Pub_Rel × Env_Proj	−0.02	0.32	0.21	−0.05	1
6	GCPU × Env_Proj	0.00	0.21	−0.01	0.21	0.28	1
7	Lever	−0.02	0.09	−0.00	−0.16	0.09	0.02	1
8	Size	0.01	0.30	−0.01	−0.12	0.14	0.26	0.16	1
9	Prof	0.02	0.01	0.06	−0.14	0.06	−0.02	−0.06	0.06	1
10	Bor_Size	−0.02	0.20	0.01	−0.18	0.14	0.15	0.12	0.26	0.11	1
11	Bor_Gend	−0.02	0.04	−0.05	0.28	0.08	0.12	−0.07	−0.02	0.02	0.05	1
12	GDP_Cap	0.02	−0.09	0.08	0.25	−0.05	0.11	−0.23	−0.04	−0.04	0.00	0.06	1
13	Regu_Qua	0.01	−0.07	−0.01	0.09	0.01	−0.02	−0.02	−0.17	−0.08	−0.15	0.19	−0.05	1

presents the correlation matrix for the key variables of interest. The correlation between insolvency risk (Insolv_Risk) and environmental project financing (Env_Proj) is negative, indicating that firms engaging more extensively in environmental project financing tend to exhibit lower insolvency risk. Likewise, the interaction term between the public relations function and environmental project financing (Pub_Rel × Env_Proj) is negatively correlated with Insolv_Risk, suggesting that the presence of a public relations function may further enhance this risk-reducing effect. In contrast, the interaction between the Global Climate Policy Uncertainty Index and environmental project financing (GCPU × Env_Proj) shows a positive correlation with Insolv_Risk, implying that higher climate policy uncertainty may weaken the beneficial impact of environmental project financing on firm insolvency risk. Nonetheless, correlations only indicate associations, not causality; therefore, this study employs regression analysis to rigorously assess these relationships. Overall, the correlation values are relatively small, indicating no evidence of multicollinearity within the empirical model.

4.2. Regression Analysis

Table 4. Impact of environmental project financing on insolvency risk.

Variables	1	2	3	4	5
Env_Proj	−2.901 **	−1.685 **	−2.246 **	−2.260 ***	−3.157 ***
	(1.306)	(0.847)	(1.027)	(0.449)	(1.084)
Pub_Rel		−0.867 *	−2.110 *		−0.497 *
		(0.412)	(1.272)		(0.271)
GCPU		0.366 **		0.402 ***	0.281 *
		(0.175)		(0.0496)	(0.167)
Pub_Rel × Env_Proj			−3.993 **		−3.436 ***
			(1.548)		(0.796)
GCPU × Env_Proj				0.694 ***	0.724 *
				(0.137)	(0.375)
Lever	0.0536	0.382 **	0.846	0.281 **	0.467 *
	(0.491)	(0.147)	(0.666)	(0.139)	(0.252)
Size	0.745 *	0.561 ***	0.649 **	0.0055	0.0949
	(0.414)	(0.200)	(0.263)	(0.0582)	(0.115)
Prof	−1.366 *	−0.759	−0.399	−0.0608 *	−0.765 ***
	(0.797)	(0.521)	(0.512)	(0.0307)	(0.139)
Bor_Size	−0.323 *	−0.438 ***	−0.253 *	0.0532	−0.140 *
	(0.172)	(0.125)	(0.138)	(0.0354)	(0.0812)
Bor_Gend	−0.0724	−0.0058	−0.0036	−0.0110 *	−0.0026 *
	(0.0498)	(0.0216)	(0.0239)	(0.0058)	(0.0013)
GDP_Cap	4.451 **	2.675 **	0.398	0.202	−0.704
	(2.076)	(1.107)	(1.143)	(0.351)	(0.619)
Regu_Qua	−0.629	0.375	0.343	−0.567 *	0.804
	(1.660)	(1.321)	(1.417)	(0.312)	(0.853)
Constant	−4.334 *	−2.260 *	−0.347	9.703 **	1.969 ***
	(2.260)	(1.293)	(1.379)	(3.969)	(0.7101)
Observations	2910	2910	2910	2910	2910
No. of instruments	61	47	73	56	78
AR(2)	0.560	0.510	0.825	0.547	0.689
Hansen	0.321	0.309	0.377	0.590	0.747

Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

reports the regression estimates obtained using the two-step system GMM, following a hierarchical pattern in which the independent variable, individual moderating variables, and their interaction terms are introduced in sequence. The results of the AR(2) and Hansen tests show insignificant p-values, indicating the absence of second-order serial correlation and over-identification issues. These outcomes confirm that the model is appropriately specified and that the instruments employed are valid [70].

We employ a hierarchical modeling approach to systematically assess the direct and moderating effects. Column 1 of Table 4 reports the coefficients of the key independent and control variables. The coefficient of environmental project financing (Env_Proj) is negative and statistically significant at the 1% level. This result provides empirical support for Hypothesis 1, indicating that environmental project financing contributes to a significant reduction in financial institutions’ insolvency risk. The coefficient of –2.901 indicates that engaging in environmental project financing is associated with a reduction of approximately 2.9 points in insolvency risk. Given its sample mean of 12.7, this represents roughly a 23% change and suggests that the effect is not only statistically significant but also economically meaningful in improving financial stability and reducing insolvency risk. This finding aligns with the theoretical arguments of [16,31]. With respect to moderating effects, Column 2 presents the results for the individual moderating variables, public relations functions (Pub_Rel) and global climate policy uncertainty (GCPU). Column 2 introduces the moderators (PRF and climate policy uncertainty) without interaction terms to examine their direct effects on insolvency risk. Notably, PRF exhibits a positive and significant coefficient, indicating that firms with a public relations function are associated with higher Z-scores (i.e., lower insolvency risk). Columns 3 and 4 subsequently report the results for the interaction terms, Pub_Rel × Env_Proj and GCPU × Env_Proj, respectively, alongside the independent and control variables.

The coefficient of the moderating effect Pub_Rel × Env_Proj, presented in Column 3, is negative and statistically significant at the 5% level. This finding supports the second hypothesis and highlights that the impact of environmental project financing on insolvency risk becomes stronger when a financial institution has an established public relations function. This result aligns with [37,40] and reinforces the notion that the presence of PR functions strengthens the negative effect of EPF on insolvency risk by authenticating the signal to the market. Regarding the moderating effect of global climate policy uncertainty, the coefficient of GCPU × Env_Proj is statistically significant at the 1% level and positive. This finding supports the third hypothesis and indicates that greater global climate policy uncertainty weakens the negative effect of environmental project financing on financial firms’ insolvency risk. This result is consistent with [30,45]. Our findings are more consistent with a signal interpretation mechanism, whereby higher climate policy uncertainty weakens the informational value and credibility of environmental project financing as perceived by stakeholders, rather than with a transition risk channel operating through the underlying riskiness of green investments.

Regarding the control variables, we interpret the results in line with the definition of the dependent variable, where higher Z-scores indicate lower insolvency risk. Leverage (Lever) and firm size (Size) exhibit positive and significant coefficients, suggesting that they are associated with higher Z-scores and thus lower insolvency risk. This may reflect that more leveraged and larger financial institutions in our sample operate under strict regulatory oversight and possess greater risk absorption capacity. In contrast, profitability (Prof), board size (Bor_Size), and board gender diversity (Bor_Gend) show negative and significant coefficients, indicating an association with lower Z-scores and thus higher insolvency risk. While these findings may appear counterintuitive, they should be interpreted with caution, as these control variables are included primarily to account for firm heterogeneity, and their economic interpretation is not the central focus of this study. For GDP per capita and regulatory quality, the results remain inconclusive.

To ensure the robustness of our findings, we employ an alternative measure of insolvency risk, namely the solvency ratio, to reassess the impact of environmental project financing. Following [56], we calculate the solvency ratio as the ratio of total capital to total assets and re-estimate the empirical model by using a two-step system GMM estimation technique. The choice of system GMM is guided by standard specification considerations, including the presence of persistence in the dependent variable and potential simultaneity bias. We present the Arellano–Bond test for serial correlation AR(2) and the Hansen test of overidentifying restrictions to assess instrument validity. The results confirm the absence of second-order serial correlation and support the validity of the instruments. In addition, we account for heteroskedasticity by using robust (Windmeijer-corrected) standard errors.

Table 5. Robustness test—alternative measure of insolvency risk (solvency ratio).

Variables	1	2	3	4	5
Env_Proj	−0.0965 ***	−0.0566 **	−0.0526 **	−0.0345 **	−0.0454 *
	(0.0354)	(0.0224)	(0.0266)	(0.0134)	(0.0256)
Pub_Rel		−0.0445 **	−0.0630 ***		−0.0467 ***
		(0.0193)	(0.0238)		(0.0106)
GCPU		0.0059 *		0.0082 ***	0.0052 *
		(0.0031)		(0.0021)	(0.0026)
Pub_Rel × Env_Proj			−0.0793 **		−0.0426 **
			(0.0370)		(0.0169)
GCPU × Env_Proj				0.0276 ***	0.0129 *
				(0.0044)	(0.0076)
Lever	−0.00509	0.0223 *	0.0287 **	0.0176 ***	0.0194 ***
	(0.0107)	(0.0114)	(0.0116)	(0.00375)	(0.0063)
Size	0.0200 *	0.0111 **	0.0155 ***	0.0020	0.0004
	(0.0103)	(0.0047)	(0.0049)	(0.0013)	(0.0019)
Prof	−0.0583 ***	0.0029	−0.0213 **	−0.0141 ***	−0.0233 ***
	(0.0206)	(0.0072)	(0.0090)	(0.0016)	(0.0028)
Bor_Size	−0.0080 ***	−0.0054 **	−0.0058 ***	−0.0094 ***	−0.0096 ***
	(0.0027)	(0.0022)	(0.0022)	(0.0008)	(0.0011)
Bor_Gend	−0.0001	0.0002	−0.0003	−0.0003 **	0.0002 *
	(0.0011)	(0.0004)	(0.0004)	(0.0001)	(0.0001)
GDP_Cap	0.0531	0.0238	0.0916 ***	0.106 ***	0.0538 ***
	(0.0410)	(0.0226)	(0.0211)	(0.00974)	(0.0133)
Regu_Qua	0.0480	0.0314	−0.0741 ***	−0.0655 ***	−0.0304 **
	(0.0329)	(0.0216)	(0.0229)	(0.00641)	(0.0150)
Constant	−0.781 *	0.159	−0.529 **	−0.986 ***	−0.366 **
	(0.417)	(0.250)	(0.225)	(0.107)	(0.147)
Observations	2910	2910	2910	2910	2910
No. of instruments	57	73	69	83	48
AR(2)	0.371	0.458	0.362	0.215	0.263
Hansen	0.270	0.369	0.348	0.316	0.344

Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

presents this robustness result, which remains consistent with the findings reported in Table 4. This consistency indicates that our model is not affected by measurement bias.

Furthermore, we employ the fixed-effects (FE) method as an alternative estimation technique to assess the robustness of our findings. The empirical model is re-estimated using FE, and the results are reported in

Table 6. Robustness test—regression results of fixed-effect method.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5
Env_Proj	−0.199 *	−0.219 **	−0.367 ***	−0.296 ***	−0.471 *
	(0.083)	(0.088)	(0.105)	(0.099)	(0.214)
Pub_Rel		−0.437 *	−0.595 *		−0.586 *
		(0.173)	(0.308)		(0.301)
GCPU		0.121 *		0.304 *	0.201 *
		(0.072)		(0.151)	(0.091)
Pub_Rel × Env_Proj			−0.589 *		−0.578 **
			(0.324)		(0.224)
GCPU × Env_Proj				0.032 ***	0.036 *
				(0.0105)	(0.0205)
Lever	0.103	0.124 ***	0.117 ***	0.409 *	0.323 *
	(0.231)	(0.013)	(0.039)	(0.211)	(0.141)
Size	0.162 *	0.171 ***	0.146 *	0.181	0.165
	(0.080)	(0.012)	(0.070)	(0.212)	(0.212)
Prof	−0.295 *	−0.285 *	−0.301 **	−0.282	−0.388 ***
	(0.170)	(0.150)	(0.110)	(0.170)	(0.120)
Bor_Size	−0.017 ***	−0.016 ***	−0.0204 *	−0.0153 *	−0.0188 *
	(0.0012)	(0.0011)	(0.0117)	(0.0069)	(0.0108)
Bor_Gend	−0.0160	−0.0136	−0.0140	−0.0163 *	−0.0173 *
	(0.0143)	(0.0102)	(0.0112)	(0.0073)	(0.0109)
GDP_Cap	−0.605	1.207	0.713 *	1.103 *	1.205
	(1.229)	(1.426)	(0.330)	(0.526)	(1.027)
Regu_Qua	0.226	−0.418	−0.250	−0.413 *	−0.536 *
	(0.311)	(0.698)	(0.283)	(0.201)	(0.301)
Constant	2.269 *	2.891 *	2.371 *	2.790 *	2.886 *
	(1.332)	(1.531)	(1.333)	(1.531)	(1.532)
Observations	2910	2910	2910	2910	2910
R-squared	0.281	0.263	0.294	0.302	0.379
F-Statistics	2.48 ***	2.69 ***	2.77 ***	3.43 ***	4.68 ***
Industry FE	Yes	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes	Yes

Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

. In this estimation, we include both country and industry fixed effects to account for unobserved heterogeneity across institutional and sectoral contexts. Importantly, the key explanatory variable, GCPU, varies over time within countries, ensuring that its effect is identified from within-country temporal variation. Coefficient and significance levels remain broadly consistent with the main results presented earlier in Table 4, reinforcing the reliability of our findings. Column 1 of Table 6 reports the coefficients of the key independent and control variables. The coefficient of environmental project financing is negative and statistically significant. Column 2 introduces the moderators (PRF and climate policy uncertainty) without interaction terms to examine their direct effects on insolvency risk. Columns 3 and 4 subsequently report the results for the interaction terms (moderating effects), respectively, alongside the independent and control variables. The results provide support to our hypothesized relationships. Column 5 shows the complete model results, which are consistent with earlier findings. This consistency suggests that the original model is not affected by estimation bias, and the observed relationships between environmental project financing, moderating variables, and insolvency risk are robust across different panel estimation approaches.

Further, given that the sample period (2014–2023) spans major global disruptions, including the COVID-19 pandemic and the Russia–Ukraine conflict, we conduct additional robustness checks to ensure that our results are not driven by crisis-period dynamics. Specifically, we introduce two dummy variables: a COVID-19 dummy that takes the value of 1 from 2019 onwards and 0 otherwise, and a Russia–Ukraine war dummy that takes the value of 1 from 2022 onwards and 0 otherwise. These variables capture the potential impact of major global shocks on financial institutions’ behavior and climate policy uncertainty. We re-estimated the empirical model including these crisis indicators.

Table 7. Robustness test—regression results with controlling crisis effects.

Variables	1	2	3	4	5
Env_Proj	−2.406 *	−1.625 *	−2.478 **	−3.093 ***	−4.286 ***
	(1.316)	(0.880)	(1.061)	(0.590)	(1.204)
Pub_Rel		−0.756 *	−1.961 *		−0.501 *
		(0.316)	(0.886)		(0.275)
GCPU		0.363 *		0.520 ***	0.513 **
		(0.188)		(0.109)	(0.228)
Pub_Rel × Env_Proj			−2.961 *		−2.309 ***
			(1.758)		(0.806)
GCPU × Env_Proj				0.836 ***	1.031 **
				(0.190)	(0.413)
Lever	0.018	0.355	0.839	0.249 **	0.394 *
	(0.491)	(0.550)	(0.667)	(0.124)	(0.221)
Size	0.910 **	0.561 ***	0.605 **	0.056	0.138
	(0.431)	(0.200)	(0.270)	(0.065)	(0.124)
Prof	−1.276 *	−0.776	−0.421	−0.150 *	−0.904 ***
	(0.691)	(0.531)	(0.513)	(0.076)	(0.148)
Bor_Size	−0.366 **	−0.425 ***	−0.207	−0.092 **	−0.182 **
	(0.186)	(0.135)	(0.148)	(0.038)	(0.087)
Bor_Gend	−0.062	−0.006	−0.012	−0.009 *	−0.029 *
	(0.055)	(0.023)	(0.025)	(0.004)	(0.016)
GDP_Cap	4.752 **	2.666 **	0.262	−0.0201	−0.894
	(2.188)	(1.107)	(1.153)	(0.373)	(0.628)
Regu_Qua	−1.158	0.348	0.923	−0.806 **	−1.896 *
	(1.895)	(1.396)	(1.577)	(0.324)	(0.993)
COVID_Dum	0.298	0.031	0.289	0.285	0.510
	(0.427)	(0.344)	(0.332)	(0.231)	(0.476)
Russ_Dum	0.870	1.033	0.931	1.338	0.844
	(0.592)	(1.001)	(0.733)	(1.120)	(0.690)
Constant	−4.901 **	−2.259 *	−0.226	1.280 ***	2.143 ***
	(2.400)	(1.293)	(1.388)	(0.425)	(0.720)
Observations	2910	2910	2910	2910	2910
No. of instruments	60	87	79	55	68
AR(2)	0.254	0.517	0.844	0.575	0.700
Hansen	0.479	0.305	0.472	0.563	0.344

Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

presents these findings, and the results remain qualitatively unchanged and consistent with the main findings. This suggests that our findings are not driven by crisis-specific effects.

Last, to address potential multicollinearity arising from interaction terms, we construct mean-centered versions of the constituent variables prior to forming the interaction terms. This transformation reduces non-essential multicollinearity between the interaction term and its underlying variables without affecting the interpretation of the coefficients. We re-estimated the empirical model, and

Table 8. Robustness test—results with mean-centered interactive variables.

Variables	1	2	3	4	5
Env_Proj	−3.429 **	−2.374 *	−3.414 *	−1.920 **	−2.364 **
	(1.632)	(1.029)	(1.550)	(0.770)	(1.205)
Pub_Rel		−2.013 *	−2.620 *		−0.674 *
		(1.016)	(1.470)		(0.334)
GCPU		0.334 *		0.295 ***	0.331 *
		(0.153)		(0.107)	(0.175)
Pub_Rel × Env_Proj			−1.687 *		−1.546 **
			(0.968)		(0.525)
GCPU × Env_Proj				0.679 ***	1.152 ***
				(0.214)	(0.350)
Lever	1.113	0.166 *	0.667	0.269 *	0.870 ***
	(0.771)	(0.032)	(0.676)	(0.156)	(0.275)
Size	0.852 *	0.411 *	0.386 *	0.001	0.018
	(0.437)	(0.222)	(0.218)	(0.072)	(0.146)
Prof	−1.479 **	−0.529	−0.202	−0.271 *	−1.438 ***
	(0.600)	(0.469)	(0.641)	(0.115)	(0.216)
Bor_Size	−0.216 *	−0.252 *	−0.270 *	−0.127	−0.188 ***
	(0.100)	(0.140)	(0.119)	(0.040)	(0.0609)
Bor_Gend	−0.081	−0.021	−0.016	−0.005 **	−0.014 *
	(0.052)	(0.021)	(0.018)	(0.002)	(0.007)
GDP_Cap	5.355 **	2.102 *	0.931	0.189	−0.434
	(2.234)	(1.028)	(1.071)	(0.436)	(0.666)
Regu_Qua	−0.117	1.338	1.696	−0.998 **	0.501
	(1.795)	(1.469)	(1.454)	(0.478)	(0.788)
Constant	−4.179 *	−1.714	−0.362	1.417 ***	1.470 *
	(2.286)	(1.567)	(1.361)	(0.487)	(0.796)
Observations	2910	2910	2910	2910	2910
No. of instruments	55	81	59	85	56
AR(2)	0.495	0.531	0.522	0.771	0.561
Hansen	0.372	0.376	0.384	0.837	0.580

Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

reports the result of this robustness test. The results remain qualitatively unchanged after mean-centering, confirming the robustness of our findings.

5. Discussion and Conclusions

The results we found show that EPF significantly reduces the insolvency risk of financial institutions in major developed countries. Signaling theory points out the transmission of costly, credible signals that reduce information asymmetry between firms and their stakeholders. Our findings suggest that EPF by banks and insurance firms operates precisely in this way—it conveys commitment, foresight, and transparency to the market, thereby lowering the probability of sudden negative price corrections. Moreover, the positive role of EPF in reducing insolvency risk is strengthened when banks and insurance firms have PR functions and weakened when operating in environments of higher uncertainty related to global climate policies.

Our study makes important contributions to the related literature. First, it extends signaling theory into the novel context of financial institutions’ lending behavior. While there is abundant literature that has used signaling arguments to corporate disclosure, dividends, or CSR activities, our study shows that lending portfolios themselves can serve as strategic signals (e.g., [7,8,10,12]). Put alternatively, the allocation of capital toward environmentally friendly projects is not only a financial act but also a communicative one, with interpretive weight in the eyes of investors. This insight broadens the scope of signaling theory by suggesting that signals need not be limited to conventional disclosures but may also arise from resource allocation decisions that are visible and costly to reverse.

The second contribution of this study lies in identifying the contingency factors that influence the impact of EPF on the insolvency risk of financial institutions. Specifically, we identify public relations functions and institutional global climate policy uncertainty as boundary conditions of this relationship. Our analysis demonstrates that both firm-level capacities (such as public relations functions) and national-level factors (such as a country’s climate policy uncertainty) influence the effectiveness of green lending as a signal. This finding nuances the classic assumptions of signaling theory, which often treats signals as universally interpretable once their costliness is established (e.g., [17,19,29]). Our results suggest a more dynamic view and propose that signals must not only be costly but also intelligible and reinforced by their institutional environment. This theoretical enhancement brings fresh depth to signaling theory and offers a richer understanding of how markets interpret socially responsible initiatives.

Lastly, the context of developed countries (G7 countries excluding Italy) enables us to highlight that socially responsible signals are particularly potent in settings with mature financial markets, active investors, and strong regulatory oversight. The context of major developed economies provides a fertile ground for signals to be both noticed and scrutinized. The fact that our results hold consistently across these countries reinforces the generalizability of signaling theory to high-capacity financial systems. At the same time, these results encourage future researchers to examine whether these dynamics hold in less mature environments where institutional frameworks are less well developed. Our findings also open new avenues for future research to study how non-traditional signals, such as capital allocation decisions, shape investor perceptions across diverse institutional contexts.

5.1. Practical Implications

Our findings also hold practical implications. Importantly, to managers of banks and insurance firms, it suggests that EPF is not only about regulatory compliance or reputational gains—it is a risk management tool. The results emphasize the strategic importance of embedding green lending into both investment portfolios and communication strategies. By allocating capital to sustainable projects, managers send credible signals that reduce uncertainty among investors and mitigate the likelihood of financial downturn. This insight should encourage managers to view EPF as a strategic investment in financial stability. For risk officers and portfolio managers, incorporating sustainability considerations into lending policies is therefore not just ethically desirable but also financially prudent. Another implication is the centrality of communication. The results show that the effectiveness of environmental projects’ lending as a signal depends heavily on how it is conveyed to external stakeholders. Managers must ensure that these initiatives are not buried in technical disclosures but are actively and transparently communicated through credible reporting, investor briefings, and engagement with stakeholders. When effectively communicated, EPF builds reputational capital that cushions the institution in times of market turbulence. This requires closer coordination between sustainability teams, risk management, and public relations departments. The PR department may perform and monitor these required tasks.

For policymakers, the results highlight the significance of consistency in national climate policies in amplifying the effectiveness of firm-level EPF. Strong consistent environmental policies, reliable regulatory frameworks, and visible progress at the national level provide the fertile ground in which firm-level signals take root. Policymakers in developed markets should therefore recognize that by strengthening their own environmental policies, through emission targets, renewable energy incentives, and transparent reporting, they indirectly augment the stability of their financial sectors. Lastly, our work implies the significance of alignment between firm-level initiatives and national strategies. Policymakers and regulators can play a catalytic role by setting clear standards for what constitutes credible EPF and by discouraging symbolic actions that dilute market trust. This collaboration could help to create a cycle where sustainable finance builds market trust and environmental progress.

5.2. Limitations and Future Research Directions

This study also has some limitations. First, our analysis considered the context of major developed countries. Although this selection helps us to show how EPF operates in strong regulatory environments, it inevitably limits the generalizability of our findings to other settings. Emerging and developing economies, where institutional frameworks are weaker and investor protections are poor, may interpret EPF signals differently. Future research may consider including less developed contexts and compare whether the same mechanisms hold or whether unique dynamics emerge in less mature financial systems. Second, although our study provides strong evidence that EPF reduces insolvency risk, it does not fully disentangle the underlying mechanisms beyond the broad signaling perspective. For example, whether investors react more strongly to transparency in disclosures, reputational gains, or perceived regulatory alignment remains an open question. Future research could employ survey-based data using interviews with fund managers to identify the exact mechanisms through which EPF influences insolvency risk.

Moreover, in this study, a public relations function is operationalized as a binary indicator capturing only the existence of a PR unit. This measure does not fully reflect heterogeneity in PR capacity, such as differences in resources, sophistication, or effectiveness across firms. Future research could employ more granular proxies—such as PR budgets, investor relations team size, media coverage intensity and tone, or communication quality indices—to better capture the depth and strategic role of corporate communication functions. Additionally, one limitation of this study relates to the construction of the Z-score, which incorporates ROA in both its numerator and its volatility component. Although widely used in the literature, this may introduce a potential mechanical overlap when ROA is also included as an explanatory variable. Future research could address this concern by constructing alternative Z-score measures based on lagged ROA or excluding contemporaneous ROA from the rolling standard deviation to further mitigate potential mechanical correlations.

It is important to mention that our moderator, a public relations function, may be subject to endogeneity, as firms with stronger financial conditions or lower insolvency risk are more likely to establish dedicated communication functions. In our empirical strategy, we treat PRF as a predetermined firm characteristic rather than strictly exogenous. The use of the system GMM estimator helps to mitigate potential endogeneity concerns by controlling unobserved heterogeneity, dynamic effects, and reverse causality through internal instrumentation. Moreover, as our primary interest lies in the moderating role of PRF, concerns of direct reverse causality are comparatively less pronounced. Nevertheless, we acknowledge that residual endogeneity may persist, and future research could employ more exogenous or granular measures of communication capacity. Finally, although we used both firm-level and country-level moderators, our scope inevitably leaves out other potentially important moderating factors. Variables like the role of institutional investors, activist pressures in the country, and societal green orientation could further explain how markets interpret environmentally friendly signals. Future studies could investigate the moderating role of these factors to enrich our understanding of when EPF is most effective.