False Stability? How Greenwashing Shapes Firm Risk in the Short and Long Run

Abstract

This study examines the relationship between greenwashing and firm risk among listed Australian firms from 2014 to 2023. We construct a firm-level greenwashing score as the residual based on regressions of composite ESG on Scope 1–2 CO₂ emissions; positive residuals indicate overstated sustainability relative to emissions. Using realized volatility as a measure of firm risk and applying the Generalized Method of Moments (GMM) regression framework, we uncover three key findings. First, contemporaneous greenwashing significantly lowers volatility, which is consistent with legitimacy and signalling theory, as overstated ESG credentials create a temporary perception of stability. Second, the risk-reducing effect is strongest with a one-period lag, likely reflecting delayed ESG and emissions reporting cycles and investor reaction times. Third, by the two-period lag, the effect reduces in magnitude, suggesting that markets eventually recognize the misalignment between ESG claims and environmental performance. Robustness checks with the E-pillar confirm these dynamics. Additional tests excluding the COVID-19 period (2020 and 2021) reveal that the risk-mitigating effects of greenwashing are even stronger during normal market conditions, implying that pandemic-related volatility may have muted the signalling power of ESG narratives. While firm fundamentals (e.g., book-to-market) explain part of risk variation, greenwashing-driven effects are economically meaningful yet short-lived. The findings underscore that greenwashing offers only temporary risk mitigation; as transparency improves and regulatory enforcement strengthens, firms relying on inflated ESG narratives face diminishing benefits and potential long-term risk penalties.

1. Introduction

Sustainability has become a strategic priority for corporations and investors, yet the risk of “greenwashing”, i.e., companies exaggerating or misrepresenting their environmental performance, is ever-present. A 2022 survey of executives found that while 80% rated their firms’ environmental initiatives as above average, 58% admitted to prevalent “green hypocrisy,” acknowledging that their companies overstated their sustainability efforts (Gregory, 2024). Greenwashing broadly refers to the dissemination of disinformation or misleading communications by an organization to present an environmentally responsible public image (Oppong-Tawiah & Webster, 2023). In other words, it is the discrepancy between a firm’s “green” talk and its actual “green” walk (He et al., 2024). This practice can take many forms, for example, selective disclosure (trumpeting positive environmental information while hiding negative aspects) or decoupling, wherein firms engage in symbolic environmental actions that do not reflect actual performance improvements (Gregory, 2024). The prevalence of greenwashing poses a serious challenge as stakeholders seek reliable indicators of corporate sustainability. It not only undermines trust but also may distort market mechanisms intended to reward genuine sustainability performance.

Greenwashing is not a victimless strategy; it has significant implications for investors, consumers, and regulators. By creating a veneer of sustainability, firms can hide underlying risks and vulnerabilities, leading investors to misprice risk and undercharge the cost of capital for greenwashing firms (G. Liu et al., 2024). Consumers, too, can be misled into favouring products or brands based on false eco-friendly claims, which, when revealed, erode brand credibility and customer loyalty (Paramitha et al., 2025; Chadichal et al., 2025). Indeed, prior studies in marketing find that perceived deception in environmental claims leads to lower corporate credibility, less favourable brand attitudes, and reduced purchase intentions (Nyilasy et al., 2014; Newell et al., 1998). Over time, repeated greenwashing can breed cynicism, making stakeholders more sceptical of all sustainability claims. Recognizing these dangers, regulators have begun to clamp down. In Australia, for instance, the securities regulator (ASIC) recently intensified enforcement against corporate greenwashing, issuing numerous corrective disclosures and infringement notices for misleading “sustainable” claims (Finpublica, 2014). This growing scrutiny reflects a broader trend: while greenwashing might offer short-term reputational or financial gains, it carries significant long-term risks as misrepresentations eventually come to light (KPMG, 2024).

A substantial body of research in the literature has established a link between ESG performance and firm risk. Numerous studies demonstrate that superior ESG performance can mitigate firm risk, leading to lower stock return volatility and a reduced likelihood of stock price crashes (Engelhardt et al., 2021; D. Liu et al., 2023; M. Liu et al., 2024). However, the evidence is not universally consistent. The impact of ESG on risk varies across contexts, and some critics contend that any apparent reduction in risk may simply reflect other firm characteristics that are correlated with ESG rather than ESG itself. In the Australian context, early evidence broadly echoes international results. The S&P/ASX 200 ESG Index has shown marginally lower volatility than the broader market over the past decade (Ung & Abburu, 2021), suggesting that ESG-focused Australian firms may be exposed to lower risk. Similarly, ESG disclosure has been found to reduce firm-specific (idiosyncratic) risk in Australia (Gholami et al., 2023). Overall, existing studies imply that stronger ESG performance should be associated with lower firm-level volatility. The prevailing explanation is that high ESG firms are perceived as more transparent, better managed, and less prone to negative stakeholder shocks or reputational crises. However, this stream of research largely treats a high ESG score as a proxy for genuine sustainability performance. It does not sufficiently account for the phenomenon of greenwashing, where a high ESG score may reflect savvy disclosure and public relations rather than substantive environmental improvement. This conflation is a critical limitation. If high ESG scores can be achieved through symbolic means, then the established negative relationship between ESG and risk may be misleading, masking a latent vulnerability for firms that are overstating their credentials. Our study addresses this conceptual and empirical gap by shifting the focus from ESG performance level to ESG misrepresentation. By linking this greenwashing component rather than raw ESG scores to realized volatility, we separate the stabilizing effects of true ESG investment from the short-lived reputation effects of ESG inflation. This distinction provides an identification advantage over existing ESG–risk studies and helps explain mixed empirical findings in the literature, particularly why ESG sometimes reduces risk, sometimes increases it, and sometimes is insignificant.

Despite mounting concerns, the relationship between greenwashing and firm risk remains underexplored in the academic literature. Much of the existing research on corporate greenwashing has focused on qualitative aspects or market valuation effects, for example, how greenwashing influences stock prices or brand value (Walker & Wan, 2012; Testa et al., 2018; Ghitti et al., 2023), firm size (Marquis & Toffel, 2012), financial flexibility (D. Zhang, 2022), financial profits (E.-H. Kim & Lyon, 2015), rather than direct risk measures (Gregory, 2024). Studies have shown that revelations of greenwashing tend to erode shareholder value, with firms accused of greenwashing experiencing declines in both stock price and reputation (Gregory, 2024). For example, event studies show that firms caught making misleading environmental claims suffer negative abnormal returns (Jones & Rubin, 2001; Lundgren & Olsson, 2009). Researchers have also begun to identify firm characteristics associated with greenwashing behaviour. Greenwashing has been observed more frequently in larger firms and those facing greater pressure or constraints. Companies under such circumstances may have more resources to devote to image management or a greater incentive to “hide” poor environmental performance (Gregory, 2024). There is evidence that some firms with high ESG ratings do not actually have lower carbon emissions, suggesting that strong ESG scores can sometimes be achieved through disclosure and PR rather than real carbon reductions (M. Zhang et al., 2025; Treepongkaruna et al., 2024). Such findings support the notion of greenwashing as “cheap talk,” where firms enjoy a positive reputation for sustainability without substantively curbing their environmental impact. Overall, the literature raises a critical question: if a firm succeeds in portraying itself as environmentally responsible without corresponding performance, what are the implications for its risk profile? This question is especially salient in the Australian market, where investor appetite for ESG investments is growing, and where improved transparency (e.g., mandated climate-risk disclosures) is making it increasingly difficult for firms to maintain a false green image over time.

While a substantial literature has examined the impact of ESG performance on firm value and risk, relatively few studies have directly examined how greenwashing (i.e., overstated ESG performance) affects firm risk. In this paper, we address this gap by examining the impact of greenwashing on firm risk, using a novel measurement approach and focusing on Australian companies from 2014 to 2023. We construct a firm-level “greenwashing score” that captures the misalignment between a company’s ESG score and its actual environmental performance. In practical terms, this score is derived by regressing firms’ ESG scores against their carbon CO₂ emissions and taking the residual, effectively isolating the portion of a firm’s ESG performance that is unexplained by its objective environmental impact (i.e., a high ESG rating that is not justified by low emissions). This residual-based metric reflects the degree to which a firm’s ESG standing may be inflated or disconnected from its real carbon footprint—a quantitative proxy for greenwashing behaviour. Using this measure on a comprehensive sample of Australian listed firms, we then analyze the relationship between greenwashing and firm risk, measured by realized stock return volatility (RV).

Our findings reveal a counterintuitive yet insightful pattern that challenges the simplistic view of a uniformly negative ESG–risk relationship: greenwashing appears to be negatively associated with firm risk in the short term. In other words, firms with higher greenwashing scores, those that look sustainably virtuous on paper despite poor environmental performance, tend to exhibit lower market risk (e.g., lower stock return volatility and/or risk premiums) in the contemporaneous period. This suggests that, at least initially, corporate greenwashing can create a false signal of stability or lower risk, presumably because investors and stakeholders respond to the firm’s strong ESG credentials and perceive the company as safer or more responsible. This result aligns with the idea that a burnished sustainability image can confer reputational benefits and investor confidence, even if unwarranted. However, when we probe the dynamics over time, the effect tells a more cautionary tale. The results from lagged models indicate that the risk-reducing benefits of greenwashing are most substantial with a one-period lag, likely reflecting delayed ESG and emissions reporting cycles, as well as investor reaction times. By the two-period lag, however, the effect decreases (becomes less negative), indicating that market participants eventually recognize discrepancies between ESG claims and actual environmental conditions. In the long run, firms that engage in greenwashing do not enjoy sustained risk reduction; in fact, they may ultimately face corrective impacts or higher risk once the exaggeration is exposed, consistent with market discipline and regulatory scrutiny catching up. These findings, drawn from the Australian context, underscore that greenwashing can be a double-edged sword: it offers short-lived risk mitigation or lower perceived risk, but this strategic benefit is transient and likely to reverse as transparency improves.

To reinforce our findings, we conducted a robustness check by constructing the greenwashing score using only the Environmental (E) pillar score, a narrower measure of firm environmental credentials, and again found similar results. This consistency strengthens the validity of our greenwashing proxy. It underscores that the risk implications are not a function of composite ESG branding alone, but specifically of the disconnect between stated environmental performance and actual CO₂ outcomes. We also re-estimate the models excluding the COVID-19 period (2020–2021) and find that the greenwashing–risk relationship remains intact, with even stronger short-term effects, further supporting the robustness of our results under varying market conditions.

In summary, our study makes a significant contribution to the growing ESG literature in two key ways. First, we develop a robust quantitative measure of greenwashing at the firm level, addressing previous challenges in capturing this elusive construct. Second, we provide empirical evidence on how greenwashing behaviour affects firm risk, an aspect neglected mainly in prior research, revealing that any risk reduction achieved through greenwashing is temporary. The results carry important implications for managers, investors, and policymakers. They suggest that while deceptive ESG practices might “buy” a company a short-term perception of safety, such practices are not a viable long-term risk management strategy. As stakeholders become more adept at detecting superficial ESG claims and as regulatory regimes penalize misrepresentation, the market is likely to correct overpriced or “un-risked” firms that have relied on greenwashing. The introduction of stricter disclosure standards and anti-greenwashing enforcement (e.g., in Australia’s recent policies (Finpublica, 2014) will further limit the ability of firms to sustain a misaligned ESG narrative. Our research thus reinforces the broader notion that authentic sustainability performance, not just high ESG scores, is what ultimately counts for lowering firm risk in a durable way.

2. Literature Review

2.1. Greenwashing in Corporate Practice

Greenwashing, broadly defined as the discrepancy between a firm’s environmental claims and its actual environmental performance, has emerged as a critical issue in corporate sustainability discourse (Oppong-Tawiah & Webster, 2023; He et al., 2024). Companies often employ symbolic actions, selective disclosure, or promotional campaigns to highlight environmental achievements while concealing negative practices (Gregory, 2024). Prior studies have documented that such behaviour allows firms to reap reputational benefits without implementing substantive environmental changes (Testa et al., 2018; Ghitti et al., 2023). Larger firms appear more prone to greenwashing due to resource availability and heightened stakeholder pressure to demonstrate sustainability (Marquis & Toffel, 2012). Despite its prevalence, detecting greenwashing remains challenging, as ESG scores may not always accurately reflect true environmental performance, and firms with strong ESG ratings can still exhibit high emissions (M. Zhang et al., 2025; Treepongkaruna et al., 2024).

2.2. Financial and Market Implications of Greenwashing

Most empirical research on greenwashing has examined its effects on market value, brand perception, and investor behaviour. Event studies consistently show that revelations of greenwashing lead to declines in stock price, reputational damage, and erosion of investor trust (Jones & Rubin, 2001; Lundgren & Olsson, 2009). Marketing literature similarly finds that misleading environmental claims reduce consumer confidence, purchase intentions, and long-term brand credibility (Nyilasy et al., 2014; Newell et al., 1998). At the firm level, greenwashing has been linked to financial outcomes, including profitability (E.-H. Kim & Lyon, 2015), financial flexibility (D. Zhang, 2022), and market reputation (Walker & Wan, 2012). However, the literature is less developed regarding the implications of greenwashing for firm risk. While ESG performance has been widely studied in relation to risk, few studies directly examine whether overstated or misrepresented ESG credentials distort risk pricing or create hidden vulnerabilities (Gregory, 2024).

2.3. ESG, Carbon Emissions, and Greenwashing

Greenwashing occurs when firms project an image of strong ESG performance while their underlying environmental practices, such as carbon emissions, contradict these claims. Recent studies (e.g., Treepongkaruna et al., 2024) highlight that discrepancies between reported ESG scores and actual emissions performance provide a valid and quantifiable basis to identify potential greenwashing. This aligns with legitimacy theory (Suchman, 1995), which posits that firms strategically disclose sustainability credentials to maintain social acceptance while potentially concealing unsustainable practices.

2.4. Firm Risk Measurement and Its Relevance for Greenwashing

Recent work shows that firm-level total risk is multidimensional, encompassing systematic volatility, idiosyncratic volatility, tail risk, and price crash risk (Stoja et al., 2024). A significant portion of this total volatility, particularly for individual stocks, is idiosyncratic risk, risk that is unique to the firm and uncorrelated with the broader market (Ang et al., 2006). Theories suggest that in markets where information asymmetry is high, firm-specific news and events are a primary driver of this idiosyncratic volatility (IV) (Engle et al., 2020). The connection to greenwashing is direct: if greenwashing represents a form of information asymmetry or a latent negative signal, its revelation would be a quintessential firm-specific shock, likely increasing IV and, by extension, total RV.

Beyond continuous volatility, a critical strand of research focuses on extreme negative outcomes, namely stock price crash risk. Crash risk occurs when the accumulation of bad news, previously hidden by managerial obfuscation (e.g., through complex or misleading reporting), reaches a tipping point and is released to the market all at once, leading to a severe, sudden price drop (Habib et al., 2017). This is often measured using the negative skewness of return distributions or the frequency of extreme negative returns (Chen et al., 2001; J.-B. Kim & Zhang, 2014). Tail risk, a related concept, specifically quantifies the risk of very large losses, focusing on the extreme left end (“tail”) of the return distribution (Stoja et al., 2024). Greenwashing naturally fits the conditions that give rise to both crash risk and tail risk. By overstating ESG performance, firms effectively conceal their true environmental exposures. When the misrepresentation is eventually uncovered through regulatory scrutiny, investigative reporting, or an adverse environmental event, it can produce a substantial negative information shock, potentially leading to a price crash. Prior evidence shows that poor disclosure quality, earnings manipulation, and opaque reporting increase crash risk (Hutton et al., 2009). Greenwashing, as a form of information distortion, is therefore likely to heighten a firm’s exposure to extreme downside risk.

While idiosyncratic volatility, crash risk, and tail risk are crucial and theoretically linked to corporate misconduct, we select RV as our primary risk measure for several compelling reasons. First, the RV, which aggregates high-frequency squared returns (Alfeus et al., 2024), has been shown to be strongly associated with both idiosyncratic volatility and tail-risk dynamics (Fan et al., 2016) because it incorporates intra-period price jumps and discontinuities. Thus, RV captures a broader spectrum of firm-level risk than annualized return volatility or GARCH-based conditional variance, which may miss intraday information. Second, an increase in the latent bad news associated with greenwashing should, in theory, manifest as increased general volatility before it culminates in a specific crash event. RV is sensitive enough to capture this building turbulence. Third, from a methodological standpoint, RV can be cleanly and directly calculated from high-frequency return data, ensuring reliability and comparability across our sample.

2.5. Greenwashing and Firm Risk in the Australian Context

The Australian market provides a unique setting for studying the risk implications of greenwashing. On one hand, investor demand for ESG-aligned portfolios has grown rapidly, incentivising firms to emphasize sustainability credentials. On the other hand, regulators such as the Australian Securities and Investments Commission (ASIC) have intensified their scrutiny of misleading ESG claims, issuing corrective disclosures and infringement notices (Finpublica, 2014). This dual context, with rising demand and stricter oversight, increases both the temptation to engage in greenwashing and the likelihood of detection. Recent evidence suggests that while firms may temporarily benefit from portraying themselves as environmentally responsible, market corrections and regulatory actions eventually expose misaligned ESG narratives (KPMG, 2024). This evolving institutional environment makes Australia an ideal empirical setting to investigate whether greenwashing provides only short-term risk mitigation before eventual correction.

3. Theoretical Framework and Hypothesis

The interplay between greenwashing and firm risk can be elucidated through an integration of classic organization theories and modern ESG-centric perspectives. Legitimacy theory posits that firms may engage in greenwashing to conform to societal expectations and maintain legitimacy, thereby temporarily mitigating perceived risk (Suchman, 1995). In parallel, signalling theory suggests that inflated ESG disclosures serve as signals of responsibility and resilience, which may lower perceived risk among investors, albeit briefly, as these signals erode in credibility over time (Spence, 1973).

While these theories provide an initial understanding, institutional theory offers deeper insight. This means organizations operate within broader fields and often adopt sustainability practices symbolically, coerced by regulations, mimicked from peers, or normatively aligned with professional expectations to gain legitimacy (DiMaggio & Powell, 1983; Scott, 2008; He et al., 2024). This behaviour mirrors greenwashing, especially when compliance becomes symbolic rather than substantive.

Impression management further clarifies this behaviour by framing greenwashing as a strategic communication tactic that employs selective disclosure or rhetorical framing to craft perceptions of environmental responsibility, with short-term reputational and risk-reducing benefits, although increasingly fragile (Lyon & Montgomery, 2015; Cho et al., 2015).

Additionally, stakeholder theory emphasizes that while greenwashing may initially placate investors, regulators, and consumers, the resulting erosion of trust and heightened scrutiny over time invariably increases firm risk (Freeman & McVea, 2001). Agency theory complements this narrative by highlighting how managers may pursue greenwashing to gain short-term advantages, such as reduced financing costs or reputational gain, even at the expense of long-run shareholder value due to information asymmetry (Jensen & Meckling, 1976).

Together, these perspectives suggest that greenwashing can serve as an effective short-term risk mitigation tactic through legitimacy enhancement, signalling, approval from institutional frameworks, and improved impressions, but such benefits are inherently transient. As stakeholders become more discerning and transparency improves, the misalignment between ESG appearances and environmental reality becomes increasingly evident, eroding trust and increasing firm risk.

This theoretical framework underpins our hypotheses:

Hypothesis 1 (H1).

Greenwashing reduces perceived firm risk in the short term by enhancing legitimacy and conveying false-positive signals of stability.

Hypothesis 2 (H2).

The negative impact of greenwashing on firm risk fades over time, as credibility diminishes under increased scrutiny and exposure.

4. Research Design

4.1. Data and Sample

This study uses panel data for Australian-listed firms from 2014 to 2023, obtained from the London Stock Exchange Group (LSEG, formerly Refinitiv) ESG database.

• ESG Scores: LSEG provides standardized ESG scores ranging from 0 to 100, aggregated across environmental, social, and governance pillars.
• Carbon Emissions: LSEG reports firms’ Scope 1 and Scope 2 CO₂ emissions (metric tons), which represent operational and energy-related emissions.
• Firm-Level Risk: Realized stock return volatility (RV) is calculated from daily price data sourced from LSEG Eikon.
• Controls: Key firm-level controls include size, profitability (ROA), leverage, and book-to-market ratio.

We begin the sample in 2014, the year ESG reporting became more widely adopted in Australia following the ASX Corporate Governance Council’s 3rd Edition guidelines requiring disclosure of ESG-related risks (Corporate governance principles and recommendations, 2014). The sample ends in 2023 because many firms had not yet released their 2024 ESG reports. We begin with all firm-year observations for non-financial ASX-listed firms during this period. Financial firms are excluded because their capital structure, reporting conventions, and regulatory requirements differ fundamentally from those of non-financial companies, which would otherwise bias both emissions-based greenwashing measurement and volatility estimation. After excluding financial firms, we first removed firm years having no emission data and no ESG scores. Finally, we checked for the inclusion criteria to retain only firms with at least three consecutive years of non-missing ESG, emissions, and return data. This requirement is necessary for constructing lagged variables (GW_i,t−1, GW_i,t−2) and for system-GMM estimation, which relies on internal instruments derived from lagged levels and differences. However, we did not have to exclude any firm at this stage since all firms were found to have more than three years of consecutive data.

The final panel is unbalanced, reflecting the heterogeneity in listing duration, ESG disclosure initiation dates, and emissions-reporting consistency. An unbalanced structure is appropriate because GMM does not require balanced data and because forced balancing would disproportionately exclude younger, smaller, and newly reporting firms, thus worsening representativeness. A CONSORT-style flowchart summarizing the sample construction process is presented in Figure 1.

Figure 2 below illustrates the sector-wise distribution of our sample. The Materials sector dominates, contributing nearly 30% of all observations. This is expected, given Australia’s resource-intensive market and the high concentration of mining, metals, and commodity-producing firms. Industrials and Consumer Discretionary follow with notable shares, while sectors such as Health Care, Energy, Consumer Staples, and Communication Services contribute moderate proportions. Information Technology and especially Real Estate represent only small fractions. Overall, the sample is unbalanced across industries, reflecting Australia’s resource-heavy market composition.

4.2. Variable Measurement

Table 1. Variable Description.

Variable	Definition/Measurement	Source
Greenwashing Score (GW)	Residual from the regression of ESG score on CO2 emissions	LSEG ESG + Emissions
ESG Score	Composite ESG performance score (0–100)	LSEG
CO2 Emissions	Scope 1 and 2 emissions (metric tons)	LSEG
Firm Risk (RV)	Realized volatility of daily stock returns	LSEG
Size (SIZE)	Natural log of total asset	LSEG
Profitability (ROA)	Net income/total assets	LSEG
Leverage (LEV)	Total debt/total assets	LSEG
Book-to-Market (BM)	Book value/market value of equity	LSEG
Cash Holdings (CASH HOLD.)		LSEG

summarizes the variables employed in the analysis, providing their definitions and the corresponding data sources from which each variable was obtained.

4.3. Greenwashing Proxy Construction

To quantify greenwashing, we follow Treepongkaruna et al. (2024) and regress firms’ overall ESG and E pillar score (E) on their actual carbon emissions:

$${ESG}_{i,t}=\alpha +\beta \cdot \left({CO2}_{i,t}\right)+ \gamma \sum {Control}_{i,t}+{ϵ}_{i,t}$$

where

• ESG_i,t = reported ESG score;
• CO2i,t = Scope 1 and 2 emissions;
• ϵ_i,t = residual = GW;
• Positive residuals (ϵ > 0) → ESG score exceeds what emissions justify → potential greenwashing;
• Negative residuals (ϵ < 0) → ESG scores align with or understate emissions performance → less likelihood of greenwashing.

Greenwashing can be measured in numerous ways: textual, audit-based, and event-based (Lublóy et al., 2024). However, each approach presents meaningful limitations that weaken its ability to detect environmental misreporting, particularly in settings like Australia’s emissions-intensive economy. Textual measures, for example, attempt to infer greenwashing by analyzing sentiment, tone, or linguistic complexity in sustainability reports (Gorovaia & Makrominas, 2024). While useful for capturing narrative manipulation, these methods depend heavily on natural-language algorithms and can easily misinterpret verbosity driven by regulatory compliance rather than strategic deception (Wang et al., 2025). Audit- or assurance-based indicators, although seemingly objective, are constrained by limited global coverage and inconsistent assurance quality across providers, making them unreliable for broad empirical analysis. Event-based proxies, such as environmental violations, lawsuits, or regulatory penalties, identify the most severe cases but fundamentally suffer from under-detection: they capture only instances that escalate into publicly observable failures, missing the far more common patterns of subtle, ongoing disclosure inflation.

These limitations create a compelling need for a more economically grounded and consistently measurable approach. The residual-based greenwashing measure used in this study directly addresses this gap. By estimating the portion of environmental disclosure that cannot be explained by CO₂ emissions and key firm characteristics, the method isolates the discrepancy between reported and expected environmental performance. In doing so, it identifies environmental claims that are disproportionately favourable relative to a firm’s emissions footprint, precisely the type of inflation that textual or event-based methods frequently overlook. Unlike binary event indicators, the residual measure offers a continuous scale capable of capturing both minor exaggerations and substantial misrepresentations. Our approach to constructing a greenwashing proxy aligns with the broader framework of ESG disclosure based on materiality, as outlined by Choi et al. (2023). Specifically, greenwashing is proxied by identifying discrepancies between the ESG information a firm chooses to disclose and the ESG topics deemed material to its operations. This approach is especially well-suited to the Australian context, where a significant share of listed firms operate in carbon-intensive sectors such as Materials, Energy, Utilities, and Industrials. In these industries, genuine environmental performance is fundamentally tied to production processes and emissions intensity. As a result, narrative-based methods may misclassify firms that “talk green” but remain operationally carbon heavy, while event-based methods may fail to detect misconduct that does not culminate in a regulatory breach. CO₂-adjusted ESG residuals overcome these weaknesses by grounding greenwashing measurement in operational reality, making them more sensitive to the structural characteristics of Australia’s resource-driven economy. Therefore, the residual-based approach not only offers methodological advantages but also provides a more accurate and context-appropriate tool for identifying greenwashing in Australia’s emissions-intensive corporate landscape.

4.4. Empirical Model

To examine H1, which posits that greenwashing reduces perceived firm risk in the short term by enhancing legitimacy and conveying false-positive signals of stability, we estimate the following baseline regression model:

$${RV}_{i,t}={\gamma }_0+{\gamma }_1{GW}_{i,t}+{\gamma }_2\sum {Control}_{i,t}+{\epsilon }_{i,t}$$

Here, RV_i,t denotes the realized volatility of firm i at time t, serving as the dependent variable (firm risk). ∑Control_i,t represents the set of control variables. GW_i,t is the main explanatory variable. A negative coefficient γ₁ would support H1 by indicating that higher levels of greenwashing are associated with lower short-term perceived risk.

To examine H2, which posits that the negative impact of greenwashing on firm risk fades over time as credibility diminishes under increased scrutiny and exposure, we estimate the following regression model:

$${RV}_{i,t}={\gamma }_0+{\gamma }_1{GW}_{i,t-1}+{\gamma }_2\sum {Control}_{i,t}+{\epsilon }_{i,t}$$

Here, RV_i,t denotes the realized volatility of firm i at time t; GW_i,t−1 is the one-period lagged greenwashing score at time t − 1, our primary explanatory variable. A statistically insignificant or less negative coefficient γ1 compared to H1 would support H2, indicating that the short-term risk-reducing impact of greenwashing weakens over time.

In the first step, where the greenwashing proxy is constructed, a difference GMM model is used to regress ESG scores on emissions, as this specification removes time-invariant heterogeneity from the residuals. In the second step, a system GMM is used to provide more efficient estimations amid variables that exhibit a random walk structure. The default lag structure is employed (i.e., instruments begin at lag 2), consistent with standard GMM practice. Collapsed instruments are used to limit instrument proliferation, thereby reducing overfitting and producing a more reliable and stable Hansen J test. A Windmeijer correction is applied to obtain more accurate standard errors for the two-step estimator. Standard errors are clustered at the firm level, reflecting the panel structure of the data.

Given regulatory changes and evolving economic conditions over time, year fixed effects are incorporated through year-dummy indicators to control for context shocks and time-specific trends.

5. Empirical Analysis

5.1. Descriptive Statistics, Graphical Analysis, and Correlation

The descriptive statistics in

Table 2. Descriptive statistics of the variables.

Variables	Obs.	Mean	Std. Dev.	Min	Max
Dependent variables
RV	1780	0.519	0.365	0.0000	5.820
Independent variables
GW	1565	22.25	22.73	−31.05	133.25
Control variables
ROA	1780	−0.035	0.337	−8.421	0.846
LEV	1780	0.438	0.314	0.003	5.301
SIZE	1780	6.661	2.051	−2.210	12.175
BM	1780	11.714	165.054	−9.005	4159.596
CASH HOLD.	1780	0.161	0.195	0.0005	1

provide important insights into the characteristics of the firms in our sample and offer preliminary evidence relevant to our hypotheses.

First, the dependent variable, RV, exhibits a moderate average value (0.519) but a substantial spread (0.365), suggesting considerable heterogeneity in firm-level risk exposure. This supports the need to examine potential drivers of firm risk, including greenwashing behaviour.

Second, GW displays a positive mean (22.25), implying that on average, firms overstate their ESG performance relative to their actual carbon emissions. However, the wide standard deviation (22.73) and extreme range (from −31.05 to 133.25) reveal significant variability in firms’ ESG disclosure practices. This dispersion suggests that while some firms genuinely align ESG performance with emissions, others exhibit potential greenwashing behaviours, an important factor when testing H1 and H2.

Third, the control variables demonstrate notable diversity. The slightly negative mean ROA (−0.035) indicates that, on average, many firms operate at low profitability levels, while the high maximum values suggest a subset of highly profitable firms. Leverage averages 43.8%, indicating a moderate financial risk; however, firms with leverage ratios above 5 represent potential outliers that may influence volatility. Similarly, the BM shows an extremely high variance, suggesting heterogeneity in firms’ valuation profiles. Meanwhile, cash holdings average 16% of total assets, implying liquidity differences that could affect risk management strategies.

These descriptive patterns highlight the importance of controlling for firm-specific financial characteristics when evaluating the relationship between greenwashing and firm risk. Furthermore, the substantial dispersion in both GW and RV strengthens the rationale for our hypotheses H1 and H2. Thus, the variability observed in Table 2 provides an empirical foundation for testing how greenwashing behaviour shapes firm risk dynamics within the Australian corporate landscape.

Figure 3 shows the average GW scores from 2014 to 2023. Scores remained relatively stable (21–23) for most of the period, indicating a persistent gap between reported ESG performance and actual environmental practices. However, a sharp decline in 2023 (average GW ≈ 19) signals potential structural changes:

• 2014–2017: Slight decline in GW, reflecting increasing awareness of sustainability disclosures.
• 2018–2022: Plateau in GW scores around 22–23, suggesting firms maintained consistent ESG signalling without substantial improvements in emissions alignment.
• 2023: Marked reduction, potentially attributable to ASIC’s anti-greenwashing enforcement measures, tighter climate-risk reporting regulations, and investor scrutiny.

The relatively high and stable GW scores prior to 2023 support the hypothesis that firms leveraging greenwashing maintained lower perceived risk in the short term, as investors may have treated inflated ESG credentials as credible signals of stability. The 2023 decline suggests that this relationship may weaken over time. As regulatory scrutiny and stakeholder awareness increase, the ability of greenwashing to reduce perceived firm risk diminishes, aligning with the hypothesis that the negative effect of greenwashing on firm risk fades when credibility is challenged.

Table 3. Pearson Correlation.

	RV	GW	SIZE	ROA	LEV	CASH HOLD.	BM
RV	1
GW	−0.33 ***	1
SIZE	−0.51 **	0.66 ***	1
ROA	−0.43 ***	0.22 ***	0.42 ***	1
LEV	−0.07 ***	0.049 **	0.17 ***	−0.04	1
CASH HOLD.	0.28 ***	−0.22 ***	−0.50 ***	−0.36 ***	−0.24 ***	1
BM	0.12 ***	−0.02	−0.11 ***	−0.02	−0.04	0.03	1

Notes: **, *** represent statistical significance at 5% and 1%, respectively.

presents the pairwise correlation coefficients among the study variables, offering preliminary insights into the relationships between GW, firm risk RV, and control variables.

The negative and significant correlation between RV and GW (−0.33), coupled with the strong positive association between GW and firm size (0.66) and ROA (0.22), suggests that larger, more profitable firms tend to exhibit higher levels of greenwashing while simultaneously experiencing lower realized volatility. This supports Hypothesis 1 (H1), which posits that greenwashing initially acts as a risk-mitigating mechanism by enhancing legitimacy and signalling stability.

Similarly, RV is negatively correlated with firm size (−0.51) and ROA (−0.43), implying that larger and more profitable firms are inherently less risky. Interestingly, cash holdings are negatively associated with GW (−0.22) but positively linked to RV (0.28), suggesting that firms retaining higher liquidity engage less in greenwashing yet still face higher volatility, likely due to operational or strategic uncertainty.

LEV exhibits relatively weak correlations across variables, indicating its limited direct influence on GW or RV within the sample. Finally, the BM shows minimal associations, consistent with its lower relevance to greenwashing practices or short-term risk exposure.

5.2. Regression Analysis

Table 4. Greenwashing Index Estimation with Composite ESG Score.

Dep. Variable	Indep. Variable	Coefficient
ESG	CO2	−0.000002 *** (0.00)
	ROA	−3.753448 (1.83)
	LEV	2.839214 (2.44)
	BM	−0.007241 *** (0.00)
	SIZE	1.727807 ** (0.75)
	CASH HOLD.	−2.280936 (1.91)
	Obs.	1385
	AR (2) Stat.	0.59 (0.56)
	Hansen J Stat.	34.93 (0.33)
	Model efficacy	√

Note: ** and *** represent statistical significance at 5% and 1% level, respectively. √ Indicates models are valid.

presents the results of the greenwashing index estimation using the composite ESG score as the dependent variable. The model is estimated using the GMM approach, which effectively addresses potential endogeneity and unobserved heterogeneity. GMM is particularly appropriate in this context, as ESG performance and firm characteristics are likely to exhibit simultaneous relationships. For example, firms’ emissions can affect ESG scores, while ESG adoption may also influence emissions over time. In all models, standard errors are reported in parentheses for coefficient estimates, while p-values are reported in parentheses for the AR(2) and Hansen J statistics.

The results show that carbon emissions have a significant negative effect on ESG scores (p < 0.01), indicating that firms with higher emissions are perceived as less sustainable. Firm size demonstrates a positive and statistically significant relationship (p < 0.05), suggesting that larger firms tend to report higher ESG scores, likely due to greater visibility and more comprehensive disclosure practices. The BM shows a negative and significant effect (p < 0.01), implying that firms with higher growth prospects or lower book values tend to score higher on ESG. By contrast, ROA, leverage, and cash holdings do not exhibit statistically significant effects, indicating their limited role in influencing reported ESG scores.

Importantly, the residuals from this regression are extracted and used to construct the GW score. This residual-based measure captures the misalignment between reported ESG performance and actual environmental outcomes (e.g., carbon emissions). A higher residual implies potential greenwashing, where firms achieve disproportionately high ESG scores despite poor environmental performance.

The model diagnostics confirm its robustness: the Hansen J-statistic (p = 0.33) suggests that the instruments used are valid, while the AR(2) test (p = 0.56) indicates the absence of second-order serial correlation. Together, these results validate the specification and support the reliability of the estimated greenwashing index.

By employing GMM estimation,

Table 5. Greenwashing and Firm Risk.

Contemporaneous GW (H1)			One-Lag GW			Two-Lag GW (H2)
Dep. Variable	Indep. Variable	Coefficient	Dep. Variable	Indep. Variable	Coefficient	Dep. Variable	Indep. Variable	Coefficient
RV	GW.ESG	−0.005945 *** (0.002)	RV	L1. GW.ESG	−0.009049 *** (0.003)	RV	L2. GW.ESG	−0.005755 *** (0.002)
	ROA	−0.1009694 (0.09)		ROA	−0.1116502 (0.07)		ROA	−0.1007269 (0.07)
	LEV	−0.115005 (0.20)		LEV	−0.048623 (0.14)		LEV	−0.025399 (0.14)
	BM	0.0002181 *** (0.00002)		BM	0.0001626 *** (0.00002)		BM	0.0001674 *** (0.00002)
	CASH HOLD.	0.2501393 *** (0.10)		CASH HOLD.	0.1769943 (0.12)		CASH HOLD.	0.310039 *** (0.11)
	Obs.	1565		Obs.	1541		Obs.	1381
	AR (2) Stat.	−0.85 (0.38)		AR (2) Stat.	−1.15 (0.25)		AR (2) Stat.	−1.40 (0.16)
	Hansen J Stat.	34.15 (0.16)		Hansen J Stat.	31.45 (0.21)		Hansen J Stat.	29.13 (0.26)
	Model efficacy	√		Model efficacy	√		Model efficacy	√

Note: *** represents statistical significance at 1% level. √ indicates models are valid.

presents the relationship between greenwashing and firm risk under two dynamic specifications. Under H1, contemporaneous greenwashing scores exhibit a negative and statistically significant association with realized volatility (p < 0.01), suggesting that firms engaging in higher levels of greenwashing appear less risky in the short term.

Interestingly, when incorporating one-lagged greenwashing scores, the negative impact on risk becomes even stronger relative to the contemporaneous model. This finding likely reflects the data reporting lag inherent in ESG scores and carbon emissions, as most ESG and emissions data are disclosed with a delay of one reporting cycle. Investors and analysts, therefore, react with a lag to a firm’s overstated ESG performance, resulting in a more substantial perceived risk-reducing effect when one lag is considered.

However, when extending the model to include two-lagged greenwashing scores, the statistically significant magnitude of the effect declines sharply. This result supports H2, indicating that the risk-mitigating benefits of greenwashing are temporary. Over time, as regulatory scrutiny increases and stakeholders gain access to more accurate emissions data, the credibility of inflated ESG claims erodes, and the market corrects its earlier mispricing of risk.

Overall, the findings highlight a dynamic temporal relationship:

• In the short run, greenwashing lowers perceived firm risk.
• With a one-period lag, the effect peaks, reflecting delayed investor reactions.
• Beyond this horizon, however, the benefits dissipate, consistent with increased transparency and stronger anti-greenwashing enforcement in Australia.

Both models pass diagnostic tests: Hansen J-statistics indicate valid instruments in contemporaneous and one-lagged models, while AR (2) results confirm the absence of second-order autocorrelation. We excluded firm size from the control variables, as its inclusion alongside ROA contributed to instrument overidentification; removing it improved the J-statistics and overall model validity. Across all three models, the control variables demonstrate consistent patterns. Profitability (ROA) and LEV exhibit negative but statistically insignificant effects, suggesting that they play a limited role in explaining firm risk compared to the dynamics of greenwashing. In contrast, the BM is positive and highly significant across all specifications, indicating that value firms tend to experience higher volatility than growth firms. Cash holdings exhibit a positive and significant relationship in the contemporaneous and two-lag models, but remain insignificant in the one-lag model. This implies that firms with higher liquidity generally face greater volatility, potentially due to strategic uncertainty. Overall, BM and cash holdings emerge as important structural drivers of firm risk, while profitability and leverage have minimal influence.

Given the high correlation between size and greenwash score, we employed a GMM model reported in

Table 6. Greenwashing and Firm Risk (size is exogenous).

Contemporaneous GW (H1)			One-Lag GW			Two-Lag GW (H2)
Dep. Variable	Indep. Variable	Coefficient	Dep. Variable	Indep. Variable	Coefficient	Dep. Variable	Indep. Variable	Coefficient
RV	GW.ESG	−0.00444 ** (0.002)	RV	L1. GW.ESG	−0.01128 * (0.006)	RV	L2. GW.ESG	−0.004117 * (0.002)
	ROA	−0.10692 (0.09)		ROA	−0.12610 * (0.06)		ROA	−0.09512 (0.07)
	LEV	−0.13563 (0.19)		LEV	−0.06473 (0.14)		LEV	−0.06450 (0.13)
	BM	0.0001648 *** (0.00003)		BM	0.00016 *** (0.00003)		BM	0.000106 *** (0.00002)
	CASH HOLD.	0.03411 (0.10)		CASH HOLD.	0.13018 (0.15)		CASH HOLD.	−0.02790 (0.11)
	Size	−0.04472 ** (0.02)		Size	0.00474 (0.05)		Size	−0.057402 ** (0.02)
	Obs.	1565		Obs	1541		Obs	1381
	AR (2) Stat.	−0.89 (0.37)		AR (2) Stat.	−1.17 (0.24)		AR (2) Stat.	−1.40 (0.16)
	Hansen J Stat.	37.05 (0.09)		Hansen J Stat.	26.63 (0.43)		Hansen J Stat.	29.13 (0.26)
	Model efficacy	Partially Valid		Model efficacy	√		Model efficacy	√

Note: *, **, and *** represent statistical significance at 10%, 5% and 1% level, respectively. √ Indicates models are valid.

below that uses size as an exogenous variable. This approach allows us to study the effect of including the size of the firms and allows us to test our results for any omitted variable bias.

Incorporating firm size as an exogenous variable into our models does not alter the core conclusions. The contemporaneous coefficient for GW.ESG becomes comparatively more negative and remains highly significant (β = −0.0044, p < 0.01), indicating that firms exhibiting higher levels of greenwashing face lower perceived risk in the short term. However, the Hansen test produces a borderline result (p = 0.09), suggesting potential concerns regarding instrument strength. The findings for the one-lag and two-lag specifications are similarly consistent: the risk-reducing effect intensifies with the first lag and subsequently diminishes with the second lag, with both estimates significant at the 10% level. The Hansen J statistics remain within acceptable bounds, supporting the validity of the model. Overall, this variation in the model acts as a robustness test and reinforces the primary results.

5.3. Robustness Check: Using the Environmental (E) Pillar

To test the robustness of our findings, we reconstruct the greenwashing index using the Environmental (E) pillar score instead of the composite ESG score. This approach allows us to examine whether our results are sensitive to the choice of ESG measurement, focusing more narrowly on a firm’s stated environmental performance.

Table 7. Greenwashing Index Estimation with E Pillar.

Dep. Variable	Indep. Variable	Coefficient
E Pillar	CO2	−0.00000343 *** (0.00)
	ROA	−0.1154556 (0.50)
	LEV	−0.9468757 (1.45)
	BM	−0.00054 (0.004)
	SIZE	0.6881573 (0.501)
	CASH HOLD.	−0.7034515 (1.80)
	Obs.	1385
	AR (2) Stat.	−0.80 (0.43)
	Hansen J Stat.	10.54 (0.23)
	Model efficacy	√

Note: *** represents statistical significance at 1% level, respectively. √ indicates models are valid.

reports the GMM estimation using the E pillar score as the dependent variable. Consistent with the baseline results in Table 4, carbon emissions remain negatively and highly significant (p < 0.01), indicating that firms with higher emissions are perceived as less environmentally responsible. Other firm-level controls, such as ROA, leverage, size, cash holdings, and BM, are statistically insignificant in this specification, suggesting that emissions are the primary driver of reported environmental performance.

The residuals from this regression are used to construct the alternative greenwashing index (GW.EPillar), which captures discrepancies between stated environmental credentials and actual emissions. Diagnostic tests confirm the reliability of the model: the Hansen J-statistic (p = 0.23) validates the instrument set, and the AR(2) statistic (p = 0.43) confirms the absence of second-order serial correlation.

Table 8. Greenwashing From E Pillar and Firm Risk.

Contemporaneous GW (H1)			One-Lag GW			Two-Lag GW (H2)
Dep. Variable	Indep. Variable	Coefficient	Dep. Variable	Indep. Variable	Coefficient	Dep. Variable	Indep. Variable	Coefficient
RV	GW.EPillar	−0.003334 *** (0.001)	RV	L1. GW.EPillar	−0.0034642 ** (0.002)	RV	L2. GW.EPillar	−0.002649 (0.002)
	ROA	−0.110749 (0.08)		ROA	−0.0928285 (0.07)		ROA	−0.0902801 (0.08)
	LEV	−0.0869504 (0.21)		LEV	−0.0055118 (0.14)		LEV	−0.007308 (0.14)
	BM	0.0001998 *** (0.00002)		BM	0.0001634 (0.00002)		BM	0.0001657 (0.00002)
	CASH HOLD.	0.280439 (0.010)		CASH HOLD.	0.3079485 ** (0.12)		CASH HOLD.	0.3664208 (0.14)
	Obs.	1565		Obs.	1541		Obs.	1381
	AR (2) Stat.	−0.85 (0.40)		AR (2) Stat.	−1.16 (0.25)		AR (2) Stat.	−1.37 (0.17)
	Hansen J Stat.	40.0 (0.05)		Hansen J Stat.	45.15 ** (0.05)		Hansen J Stat.	47.30 *** (0.01)
	Model efficacy	√		Model efficacy	√		Model efficacy	Partially Valid

Note: *** and ** represent statistical significance at 1% and 5% level, respectively. √ indicates models are valid.

examines the impact of greenwashing, as measured by the E-pillar score, on RV under three scenarios: contemporaneous effects (H1), one-lag effects, and two-lag effects (H2).

• H1 (Short-Term Effect)
  The contemporaneous coefficient for GW.EPillar is negative and highly significant (β = −0.003334, p < 0.01), suggesting that firms engaging in greater environmental greenwashing experience lower perceived risk in the short term. This finding aligns with H1, indicating that overstated environmental credentials enhance legitimacy and temporarily signal stability to investors.
• One-Lag Effect
  When introducing a one-period lag (β = −0.003464, p < 0.05), the risk-reducing effect strengthens slightly compared to the contemporaneous model. This result likely reflects reporting delays in ESG and emissions data, i.e., investors react to greenwashing signals after disclosures are made public, amplifying the short-term perception of stability. The AR (2) statistic (p = 0.25) supports model validity, and the Hansen J-statistic (p = 0.05) suggests borderline but acceptable instrument strength.
• H2 (Two-Lag Effect)
  In contrast, the two-lag model reveals that the impact of greenwashing fades over time (β = −0.002649, p > 0.10), becoming statistically insignificant. This supports H2, demonstrating that as transparency improves and regulatory scrutiny intensifies, markets adjust their risk assessment and greenwashing no longer provides protective reputational benefits. However, the Hansen J-statistic (p = 0.01) indicates potential over-identification concerns, suggesting results from this specification should be interpreted cautiously.
• Additional Insights
  Interestingly, cash holdings become significant in the one-lag model (p < 0.05) and remain positive in the two-lag model, implying that liquidity may interact with overstated environmental reporting to influence volatility dynamics in subsequent periods. Other control variables, including ROA, leverage, and BM, largely mirror the patterns seen in earlier models.

5.4. Robustness Check: Impact of COVID-19

Because our sample includes the COVID-19 period, we examine whether pandemic-related disruptions influence our findings. Following López Prol and Kim (2022), we re-estimate all models after excluding the COVID years (2020–2021). This allows a direct comparison between results based on the full sample and those based on a dataset that removes pandemic effects.

Table 9. Greenwashing and Firm Risk (Excluding COVID Years).

Contemporaneous GW (H1)			One-Lag GW			Two-Lag GW (H2)
Dep. Variable	Indep. Variable	Coefficient	Dep. Variable	Indep. Variable	Coefficient	Dep. Variable	Indep. Variable	Coefficient
RV	GW.ESG	−0.01047 ** (0.05)	RV	L1. GW.ESG	−0.01093 ** (0.005)	RV	L2. GW.ESG	−0.01447 ** (0.006)
	ROA	−0.17264 *** (0.64)		ROA	−0.20941 *** (0.06)		ROA	−0.23134 *** (0.06)
	LEV	−0.14800 (0.20)		LEV	−0.0673 (0.15)		LEV	−0.08666 (0.15)
	BM	0.000192 *** (0.00002)		BM	0.000123 *** (0.00003)		BM	0.0001026 ** (0.00003)
	CASH HOLD.	0.4085 (0.20)		CASH HOLD.	0.02681 (0.19)		CASH HOLD.	0.35023 *** (0.22)
	Obs.	1218		Obs.	1190		Obs.	1029
	AR (2) Stat.	−0.62 (0.54)		AR (2) Stat.	0.42 (0.67)		AR (2) Stat.	−1.70 (0.09)
	Hansen J Stat.	39.86 (0.03)		Hansen J Stat.	40.80 (0.02)		Hansen J Stat.	35.56 (0.06)
	Model efficacy	Partially Valid		Model efficacy	Partially Valid		Model efficacy	Partially Valid

Note: *** and ** represent statistical significance at 1% and 5% level, respectively.

presents the regression results examining the impact of greenwashing, measured using the ESG score, on RV under three scenarios: contemporaneous effects (H1), one-lag effects, and two-lag effects (H2). These results are based on the regression setup that excludes the COVID-19 years (2020–2021), allowing us to assess whether pandemic-related distortions influence the greenwashing–risk relationship.

• H1 (Short-Term Effect)
  After removing 2020–2021, the contemporaneous greenwashing coefficient becomes larger in magnitude and remains statistically significant (β = −0.01047, p < 0.05). This effect is almost double the pre-exclusion coefficient (−0.0059) reported in Table 5, indicating that outside of the extreme uncertainty of COVID-19, greenwashing reduces perceived volatility even more strongly. This supports H1 and suggests that during normal periods, overstated ESG claims are more effective at signalling stability and legitimacy to markets. However, the Hansen J-statistic (p = 0.03) indicates partial model validity, meaning instruments may be slightly weak when COVID years are removed, which is common when the sample size drops.
• One-Lag Effect
  The lagged effect remains negative and statistically significant (β = −0.01093, p < 0.05). This shows that investors continue reacting to greenwashing disclosures with a one-period delay, consistent with the information-processing lag in ESG reporting. Hansen test (p = 0.02) signals partial model validity, meaning the instruments are not strong enough once COVID volatility is removed.
• H2 (Two-Lag Effect)
  In contrast to our earlier results, the two-lag effect is now stronger and statistically significant (β = −0.01447, p < 0.05). This is the opposite of the pre-exclusion pattern. Instead of fading out, the risk-reducing effect of greenwashing persists for two periods when the COVID-19 years are removed. This suggests that during “normal” economic environments:

  ➢

  Greenwashing has longer-lasting reputational effects;

  ➢

  Markets are slower to adjust or uncover exaggerations;

  ➢

  Transparency and regulatory detection mechanisms operate more gradually.

However, given that the AR(2) statistic is borderline (p = 0.09), and the Hansen test (p = 0.06) suggests only partial instrument validity, we interpret this result with caution and do not consider the two-lag specification to be fully reliable.

6. Discussion

6.1. Substantive Interpretation

The results present a clear narrative on the interaction between greenwashing and firm risk in Australia from 2014 to 2023. Figure 4 below illustrates a visual conceptual framework that depicts how greenwashing influences firm risk across contemporaneous, one-lag, and two-lag dynamics, integrating the roles of ESG–emissions misalignment, investor perceptions, regulatory scrutiny, and evolving market responses. Using a residual-based greenwashing proxy derived from ESG–emissions misalignment, we find that firms with higher greenwashing scores exhibit significantly lower contemporaneous realized volatility. This supports legitimacy and signalling theory, which posits that overstated sustainability narratives temporarily reassure capital markets, foster investor confidence, and reduce perceived risk premiums despite weaker environmental performance.

However, this short-term effect is not durable. The results from lagged models reveal that the risk-reducing benefits of greenwashing are strongest with a one-period lag, likely due to delayed ESG and emissions reporting cycles and investor reaction times. By the two-period lag, the effect reduces in magnitude (becomes less negative), indicating that market participants eventually recognize discrepancies between ESG claims and environmental reality. The sharp decline in greenwashing scores in 2023 further aligns with this dynamic, likely reflecting the impact of ASIC’s anti-greenwashing regulations and heightened investor due diligence, signalling that narrative-driven advantages are diminishing in an evolving regulatory landscape.

Control variables provide additional context. The BM ratio is positively and robustly related to RV, indicating that value-type or distressed profiles carry a higher risk, independent of greenwashing. By contrast, ROA, leverage, and cash holdings are largely insignificant in the baseline GMM specifications, suggesting that narrative effects related to sustainability can matter for short-term risk beyond standard fundamentals. Correlations (Table 3) support this interpretation: greenwashing is more prevalent among larger and more profitable firms, which also exhibit lower volatility, precisely the profile of issuers best positioned to “stage” credibility through disclosure resources and investor relations.

6.2. Robustness and Measurement Validity

Our robustness exercises, the first that rebuilds the misalignment metric using the E-pillar rather than the composite ESG and the second that reconstructs the model with size considered, arrives at the same qualitative conclusions (Table 6, Table 7 and Table 8). Emissions remain a strong, negative determinant of stated environmental performance, and the residual E-based greenwashing score preserves the pattern of a negative contemporaneous association with risk and a smaller lagged effect. The E-pillar design sharpens the construct validity of the proxy because it focuses the “illusion–reality” gap squarely on environmental claims.

From an econometric standpoint, the reliance on GMM is justified by the panel structure, potential reverse causality (ESG narratives may both respond to and shape market risk), and firm-fixed unobservables. AR(2) tests are comfortably passed across specifications, and Hansen J values are acceptable in the main models (Table 4 and Table 5, all three columns). We note, however, that the lagged E-pillar model’s Hansen p-value (0.01) is borderline, signalling potential instrument proliferation or weak instruments in that variant. This does not overturn the pattern of results but does counsel caution and motivates instrument-set trimming in future estimations.

We further assess temporal robustness by removing COVID-19 years (2020–2021) from the sample, following López Prol and Kim (2022). This allows us to isolate the potential distortionary effects of pandemic-related volatility. The revised estimations (Table 9) confirm the resilience of the core relationship: greenwashing continues to exhibit a statistically significant, negative effect on realized volatility across contemporaneous and lagged models. This reinforces H1 and implies that market participants may interpret overstated ESG disclosures as stabilizing signals in “normal” periods. However, the partial validity of the Hansen tests (p-values ranging from 0.02 to 0.06) and a borderline AR(2) result in the two-lag specification (p = 0.09) raises caution about instrument reliability in the restricted sample. The potential for instrument proliferation or reduced power due to a smaller sample must be considered when interpreting the extended lag effects. Nevertheless, these findings enhance measurement validity by showing that the greenwashing–risk relationship holds even when external shocks are excluded, confirming that our proxy is not merely capturing pandemic-specific dynamics.

6.3. Mechanisms and the Australian Setting

Why does greenwashing lower risk in the short run yet fade thereafter? Three complementary channels are plausible:

• Reputational Buffer: Firms with strong ESG narratives project resilience and sustainability leadership, attracting investor trust and lowering perceived tail risk in the short term.
• Information Frictions: ESG and emissions data are often reported with lags, delaying the market’s ability to detect inconsistencies between claims and reality.
• Institutional Learning: Over time, regulatory interventions, improved disclosure frameworks, and enhanced investor analytics reduce the scope for misrepresentation, diminishing greenwashing’s impact on perceived risk.

Australia’s sector mix (resource-heavy), exposure to commodity cycles, and the timing of climate-policy debates likely amplified the 2018–2022 plateau in misalignment and the 2023 reset. The descriptive drop in 2023 coheres with a market regime where narrative alone no longer suffices to quell perceived risk without verifiable environmental improvement.

6.4. Implications

For managers: The findings serve as a warning against treating ESG communications as a substitute for risk management. Greenwashing can buy time, but it is not a sustainable strategy. Resources are better deployed into measurable emissions reduction, data quality, and internal controls that will withstand escalating scrutiny.

For investors: Short-term calm around high-ESG storytellers may mask unpriced environmental liabilities. Incorporating simple misalignment diagnostics, such as residuals from ESG-on-emissions regressions, into risk models can help distinguish credible from cosmetic sustainability.

For regulators and standard setters: The sharp decline in average misalignment in 2023 suggests that enforcement and disclosure reforms are effective. Continued emphasis on scope coverage, assurance, and comparability should further reduce the scope for greenwashing and improve the informativeness of ESG signals for risk assessment.

6.5. Limitations and Avenues for Further Research

Our misalignment proxy is residual-based and thus inherits any measurement error in ESG scores and emissions (e.g., scope boundaries, estimation methods). While GMM mitigates endogeneity, the results may be sensitive to the number and strength of the instruments. Extensions could (i) examine heterogeneity by sector (e.g., resources vs. services); (ii) incorporate event-time analyses around enforcement or disclosure shocks; (iii) test alternative greenwashing constructs (textual sentiment vs. emissions); and (iv) trace long-horizon outcomes (spread dynamics, downgrades, litigation risk) once misalignment is revealed.

A key limitation of our emissions data is that it includes only Category 1 and 2 emissions. This omission may be particularly relevant for service-sector firms, for whom Category 3 emissions, associated with upstream supply chains and downstream product use, represent most of the total carbon exposure. Excluding these emissions may lead to systematic underestimation of firms’ true environmental footprint and may attenuate the ESG–emissions relationship used to derive our greenwashing proxy. Importantly, our greenwashing measure is based on residuals from the ESG–emissions model, making it less sensitive to absolute emission magnitudes; however, the potential downward bias in emissions measurement, particularly for service firms, could be a subject of further study.

7. Conclusions

This study contributes to the ESG literature by introducing a novel, firm-level measure of greenwashing, constructed as the residual from regressing ESG scores on actual CO₂ emissions, and linking it to realized volatility, a high-frequency, market-based measure of firm risk. This methodological approach offers a sharper distinction between authentic ESG performance and inflated sustainability claims, allowing us to isolate the effect of misrepresentation on perceived risk.

Our findings reveal three key insights. First, greenwashing is associated with a temporary reduction in firm risk, suggesting that overstated ESG performance may initially generate reputational benefits and signal stability to investors. Second, the effect peaks with a one-period lag, likely reflecting delays in reporting and investor processing. Third, the risk-reducing impact diminishes over time, implying that markets eventually detect misalignment between ESG claims and actual environmental outcomes. Robustness checks, using the environmental pillar and excluding the COVID-19 period, confirm the consistency and validity of these results.

Overall, our analysis suggests that greenwashing is not a sustainable risk management strategy. While it may lower perceived risk in the short run, it exposes firms to correction risks in the long term as transparency improves and regulatory scrutiny intensifies. For managers, this highlights the importance of aligning ESG communications with substantive performance. For investors and regulators, our framework provides a useful tool to detect ESG inflation and assess its implications for market risk.

The overarching message is simple: authentic environmental performance, i.e., not narrative polish, underpins durable risk reduction. As disclosure standards harden and data improve, the short-term cushion provided by greenwashing is likely to shrink further, and the costs of misrepresentation will rise. For managers, investors, and policymakers, the results suggest the need for credible decarbonization, improved measurement and assurance, and continued enforcement to ensure that ESG signals are informative for risk, thereby transitioning the market from an ESG illusion toward emissions reality.