A. Stakeholder Pressure and Reputational Costs

A large body of research highlights that firms face powerful incentives to respond to stakeholder pressure because of the substantial reputational costs associated with failing to meet societal expectations (Fombrun (Reference Fombrun1996), Delmas and Toffel (Reference Delmas and Toffel2008)). Evidence shows that sanctions through market and nonmarket channels may be initiated by diverse stakeholders, such as investors, journalists, NGOs, employees, and communities (Klassen and McLaughlin (Reference Klassen and McLaughlin1996), Dyck et al. (Reference Dyck, Volchkova and Zingales2008), and Krüger (Reference Krüger2015)). When corporate practices are exposed as socially or environmentally irresponsible, firms tend to suffer a loss of legitimacy and trust in the eyes of key stakeholders, evidenced by declines in market value (Klassen and McLaughlin (Reference Klassen and McLaughlin1996), Bolton and Kacperczyk (Reference Bolton and Kacperczyk2021)) and heightened vulnerability to regulatory scrutiny or activist campaigns (King and Lenox (Reference King and Lenox2000), Christensen, Hail, and Leuz (Reference Christensen, Hail and Leuz2021)). Firms are therefore motivated to undertake measures, such as enhancing transparency, adopting governance reforms, or investing in sustainability initiatives to address these reputational concerns (Delmas and Toffel (Reference Delmas and Toffel2008), Flammer (Reference Flammer2013)).

Reputational pressures are particularly salient when the issues involve environmental consequences. Studies show that activist groups and NGOs play an important role in uncovering harmful corporate environmental practices and mobilizing pressure that compels firms to adjust their policies and strategies (Reid and Toffel (Reference Reid and Toffel2009)). Journalists and media coverage further magnify these effects by broadcasting environmental controversies to wider audiences (Dyck et al. (Reference Dyck, Volchkova and Zingales2008)). Employees may also act as whistleblowers when internal practices conflict with sustainability commitments (Böke et al. (Reference Böke, Brau, Glaeser and Jeong2024)). In addition, firms tend to be penalized in capital markets when controversies arise (Krüger (Reference Krüger2015), Krueger, Sautner, and Starks (Reference Krueger, Sautner and Starks2020), and Bolton and Kacperczyk (Reference Bolton and Kacperczyk2021)), while being rewarded when they proactively adopt environmental initiatives (Flammer (Reference Flammer2013)). Recent studies demonstrate that stakeholder monitoring helps enhance firm social performance (Li et al. (Reference Li, Mai, Wong, Yang and Zhang2026)) and that individual characteristics of CEOs and board directors significantly affect corporate environmental outcomes (Hsu et al. (Reference Hsu, Li and Pan2026), Li et al. (Reference Li, Xu and Zhu2025)). Taken together, this literature suggests that reputational concerns amplified by diverse stakeholders have significant implications for corporate environmental practices and outcomes.

B. Institutional Background

The adoption of anti-SLAPP statutes empowers stakeholders to voice their concerns more freely (Pring and Canan (Reference Pring and Canan1996), Norman (Reference Norman2010)). What distinguishes these laws is their procedural design, which directly reduces the legal and financial burdens of public participation. First, anti-SLAPP statutes typically authorize the early dismissal of meritless lawsuits via special motions to strike, preventing drawn-out legal battles that might otherwise drain defendants’ time and resources. This safeguard directly addresses the kind of protracted litigation faced by organizations such as Greenpeace International, which has repeatedly been targeted by energy firms seeking to silence environmental campaigns—cases later dismissed as baseless but costly to defend (see https://www.greenpeace.org.au/news/greenpeace-international-begins-groundbreaking-anti-slapp-case-to-protect-freedom-of-speech/). Second, anti-SLAPP laws deter frivolous litigation by requiring losing plaintiffs to cover defendants’ legal fees upon a successful motion (e.g., California’s CCP §425.16(c)). Third, they generally pause discovery proceedings while a motion is pending, preventing plaintiffs from strategically inflating litigation expenses to pressure a settlement. This measure strikes at the core coercive tactics of SLAPP suits, where plaintiffs exploit procedural costs, financially and emotionally, to punish critics.Footnote ¹ Finally, anti-SLAPP protections are intentionally broad, covering a wide range of public expression, including traditional speech, online commentary, activism, and other civic discourse, thereby reinforcing their continued relevance in today’s communication landscape.

The contemporary legal framework addressing SLAPP lawsuits originates from California’s pioneering anti-SLAPP statute, codified as California Code of Civil Procedure §425.16 in 1992. Since then, similar anti-SLAPP provisions have been enacted across various jurisdictions, including Texas (Texas Civil Practice & Remedies Code §27.001 et seq.), New York (N.Y. Civ. Rights Law §70-a), and Washington (RCW 4.24.510). As of 2019, over 30 states in the U.S. have adopted such statutes. However, the scope of these laws varies widely: some protect only speech directed at government entities, while others extend coverage to any expression concerning matters of public interest. This variation has prompted considerable legal debate, with courts frequently testing the boundaries of what qualifies as protected speech (Chen et al. (Reference Chen, Li and Li2025)). The controversy has become especially pronounced in the environmental context. Major oil and agribusiness firms, for instance, file defamation and racketeering claims against environmental NGOs and media outlets reporting on climate concerns (see https://www.nytimes.com/2025/06/22/climate/oil-industry-anti-slapp-climate-lawsuits.html/). Parallel developments in Europe, such as litigation between Energy Transfer and Greenpeace, highlight the global resonance of these disputes and the growing international discussions on anti-SLAPP protections (see https://verfassungsblog.de/greenpeace-slapp-energy-transfer/).

From an institutional perspective, the common thread of these anti-SLAPP efforts is clear. That is, anti-SLAPP statutes recalibrate the legal balance of power, reducing the ability of well-resourced plaintiffs to suppress criticism through litigation tactics. The widespread adoption of anti-SLAPP statutes reflects their institutional significance in safeguarding constitutionally protected speech from coercive litigation practices. By embedding free-speech protections into civil procedure, anti-SLAPP laws fundamentally reduce the real and perceived costs of public criticism. These protections are particularly relevant in environmental contexts, where stakeholders, such as NGOs and journalists, often rely on legal protections for free speech to investigate and publicize issues such as corporate pollution and greenwashing (Pring and Canan (Reference Pring and Canan1996)). A direct implication is that, as discussed in Norman (Reference Norman2010), these statutes encourage more open dialogue and create a durable infrastructure for stakeholder oversight. Recent empirical evidence confirms these effects by showing that anti-SLAPP laws significantly amplify stakeholder voice in the corporate domain and elicit meaningful corporate responses in disclosure (Chen et al. (Reference Chen, Li and Li2025), Lee et al. (Reference Lee, Ng, Yoo and Zhang2026)), allocating resources at the workforce and capital markets (Jung et al. (Reference Jung, Wang, Wei and Zhang2021), Griffin et al. (Reference Griffin, Hong, Jung and Moon2024)), and improving transparency (Griffin et al. (Reference Griffin, Hong, Jung and Moon2024), Guernsey et al. (Reference Guernsey, Serfling and Yan2025)).

C. Theoretical Predictions

We contend that the institutional design of anti-SLAPP statutes carries significant implications for corporate environmental performance. By reducing the legal and financial risks associated with public participation and strengthening the voices of stakeholders, these laws raise the reputational costs of environmental negligence. Environmental issues differ from other areas of corporate responsibility in their visibility, long-term social impact, and often irreversible consequences for affected communities and ecosystems (Flammer (Reference Flammer2013), Bolton and Kacperczyk (Reference Bolton and Kacperczyk2021), and Li et al. (Reference Li, Xu and Zhu2025), (Reference Li, Mai, Wong, Yang and Zhang2026)). Events such as pollution incidents, toxic releases, and climate-related harms frequently trigger intensive attention from stakeholders (Dyck et al. (Reference Dyck, Volchkova and Zingales2008), Krüger (Reference Krüger2015), Krueger et al. (Reference Krueger, Sautner and Starks2020)). Under anti-SLAPP protections, these concerns can be raised with less fear of retaliation (Pring and Canan (Reference Pring and Canan1996), Norman (Reference Norman2010)), thereby increasing the likelihood that harmful practices will be publicly exposed. Anticipating this scrutiny and the reputational damage it may entail, firms are more likely to proactively improve their environmental practices. We therefore expect that the enactment of anti-SLAPP statutes is associated with improvements in corporate environmental performance.

We further argue that two complementary channels help explain how firms adapt to amplified stakeholder voice under anti-SLAPP laws. First, they may increase investment in environmental initiatives within operations as a forward-looking strategy. By allocating more financial and technological resources to green innovation, pollution abatement, and waste management, firms can credibly signal their long-term commitment to sustainability in alignment with stakeholders’ expectations (Aghion et al. (Reference Aghion, Dechezleprêtre, Hemous, Martin and Van Reenen2016)). Although such investments typically involve significant cost and uncertain returns (Xu and Kim (Reference Xu and Kim2022)), they often deliver substantive improvements in corporate environmental outcomes (King and Lenox (Reference King and Lenox2000), Berrone and Gomez-Mejia (Reference Berrone and Gomez-Mejia2009)). In this sense, green investments represent both a signal of commitment and an operational shift toward more robust environmental risk mitigation strategies in response to greater transparency and stakeholder monitoring.

Second, firms may reinforce environmental governance to institutionalize and sustain their sustainability efforts. As the foundation for lasting change, governance embeds environmental accountability into decision-making processes and aligns managerial incentives and organizational routines with long-term sustainability objectives (Delmas and Toffel (Reference Delmas and Toffel2008), Berrone and Gomez-Mejia (Reference Berrone and Gomez-Mejia2009)). In response to intensified stakeholder scrutiny under anti-SLAPP statutes, firms may adopt governance mechanisms that support more disciplined and forward-looking environmental strategies. Such mechanisms include improving board oversight of sustainability issues, embedding ESG criteria in executive compensation and supply-chain partner selection, and enhancing employee environmental training (Berrone and Gomez-Mejia (Reference Berrone and Gomez-Mejia2009), Li et al. (Reference Li, Xu and Zhu2025), (Reference Li, Mai, Wong, Yang and Zhang2026)). We contend that these governance measures help integrate environmental considerations into corporate decision-making and establish accountability systems that sustain the firm’s environmental effort. Together, environmental investment and governance represent strategic and structural adaptations for firms to respond effectively to heightened stakeholder scrutiny arising from anti-SLAPP protections (Porter and Van der Linde (Reference Porter and Van der Linde1995)).

A. Toxic Release Data

We measure corporate environmental performance using toxic chemical release, a direct and objective indicator for firms’ environmental footprint (Delmas and Toffel (Reference Delmas and Toffel2008), Akey and Appel (Reference Akey and Appel2021), Xu and Kim (Reference Xu and Kim2022), Duchin, Gao, and Xu (Reference Duchin, Gao and Xu2025), and Hsu et al. (Reference Hsu, Li and Pan2026)). Facility-level toxic release data are acquired from the EPA’s TRI, established under Section 313 of the Emergency Planning and Community Right-to-Know Act (EPCRA) (https://www.epa.gov/toxics-release-inventory-tri-program). The TRI covers chemicals that meet at least one of the following criteria: i) linked to cancer or other chronic human health effects, ii) associated with significant adverse acute human health effects, or iii) expected to cause significant adverse environmental effects. The covered list currently includes around 799 individually listed chemicals and 33 chemical categories (see https://www.epa.gov/toxics-release-inventory-tri-program/tri-listed-chemicals). Facilities that emit such chemicals are required to report their annual release quantities to the TRI. Since its inception in 1987, this reporting obligation has applied to facilities that i) employ at least 10 workers, ii) operate in specific 6-digit NAICS sectors, and iii) handle listed chemicals above defined threshold levels.Footnote ² From the TRI data set, we acquire comprehensive information on reporting facilities, including facility and parent company identifiers (e.g., names and DUNS number), reporting year, and the quantity of each listed chemical released into air, water, or land.

We next merge the TRI data with financial information on U.S. public firms from Compustat. Because the two databases lack a common identifier or linking tables, we rely on name-based matching combined with manual verification, following Akey and Appel (Reference Akey and Appel2021), Chen, Jing, Keasey, and Xu (Reference Chen, Jing, Keasey and Xu2026), Jing, Keasey, Lim, and Xu (Reference Jing, Keasey, Lim and Xu2024), and Li et al. (Reference Li, Xu and Zhu2025). Our matching procedure proceeds in three steps. First, we standardize firm names by removing common suffixes (e.g., “Corp,” “Limited,” “Ltd.”). We then apply a fuzzy string-matching algorithm based on the Levenshtein distance to link TRI parent names to Compustat firm names.Footnote ³ This approach measures the minimum number of single-character edits required to transform one string into another, allowing us to identify matches even when firm names differ slightly due to abbreviations, typographical errors, or formatting inconsistencies. Second, to address time-varying names in both data sources, we incorporate historical names from CRSP and extract historical names and addresses from 10-K, 10-Q, and 8-K filings using the SEC Analytics Suite available through WRDS and Loughran–McDonal database (https://sraf.nd.edu/sec-edgar-data/) (e.g., Loughran and McDonald (Reference Loughran and McDonald2014)). Third, we manually check every algorithmic match.Footnote ⁴ As an additional check, we compare headquarters locations, official company websites, and DUNS numbers to confirm each link.Footnote ⁵

TRI reports emissions at the facility-chemical–year level. We construct a firm-year measure of total toxic release by aggregating facility-level emissions across chemicals and facilities owned by the same firm in a given year. Our primary variable is total toxic pollution, defined as the sum of on-site releases (to air, water, and land) and off-site transfers for further treatment, disposal, or release (Jing et al. (Reference Jing, Keasey, Lim and Xu2024), Li et al. (Reference Li, Xu and Zhu2025), (Reference Li, Mai, Wong, Yang and Zhang2026)). Following Jing et al. (Reference Jing, Keasey, Lim and Xu2024) and Li et al. (Reference Li, Xu and Zhu2025), we exclude facilities that report zero toxic emissions throughout the 1990–2019 period.Footnote ⁶ This procedure leaves us with 20,150 facilities affiliated with 2,135 distinct publicly listed firms.

B. Sample Selection

The primary source for identifying the timing of each state’s anti-SLAPP adoption is the Reporters Committee for Freedom of the Press (RCFP) (https://www.rcfp.org/), which tracks the passage and statutory coverage of anti-SLAPP laws nationwide. To ensure accuracy, we cross-validate the enactment years with prior academic literature, including Chen et al. (Reference Chen, Jing, Keasey and Xu2026) and Li et al. (Reference Li, Xu and Zhu2025), who provide comprehensive timelines of anti-SLAPP implementation suitable for empirical research.

Our initial sample comprises all U.S. firms included in the merged Compustat and CRSP database from 1990 to 2019. We merge these firm-level records with the aggregated toxic release data and exclude firms that operate no reporting facilities. Each firm is then linked to the anti-SLAPP statute of its headquarters state (Chen et al. (Reference Chen, Li and Li2025)). We match the statute with firms’ headquarters for two main reasons. First, anti-SLAPP laws are enacted and enforced at the state level, and courts typically apply the statute of the state where the plaintiff’s principal place of business or domicile is located in defamation or related tort actions. According to the Restatement (Second) of Conflict of Laws (Sections 145 and 150.1), the state of the plaintiff’s principal place of business has the “most significant relationship” in determining the applicable law.Footnote ⁷ Consequently, when a firm faces public, media, and civil criticism, the applicable anti-SLAPP protections are generally determined by the headquarters state, and the headquarters state thus best captures the firm’s jurisdictional exposure to speech-related legal protections. Second, from an economic standpoint, a firm’s headquarters is the core of strategic decision-making and external communication. It hosts senior management, investor relations, and legal teams, and serves as the primary contact point for analysts, media, and regulators. Local media and residents in the headquarters state often have greater access to firm-specific information and stronger monitoring capacity, making them more responsive to changes in local speech protections. Thus, anti-SLAPP statutes in the headquarters state influence the likelihood that negative environmental information (e.g., pollution events or breaches of environmental regulation) is detected, reported, and disseminated.

To link corporate headquarters to states, we rely on firms’ historical headquarters locations. We first identify each firm’s headquarters state from the 10-K header files using the Loughran–McDonald database (e.g., Loughran and McDonald (Reference Loughran and McDonald2014)), which provides information between 1994 and 2019. We then supplement this information with data from Heider and Ljungqvist (Reference Heider and Ljungqvist2015) and the company header history file in the legacy Compustat/CRSP merged database, as compiled by Bena, Ortiz-Molina, and Simintzi (Reference Bena, Ortiz-Molina and Simintzi2022). Finally, we exclude firms in the financial and utility sectors and remove observations with missing control variables.

Recent finance and econometrics literature indicates that traditional two-way fixed effects Difference-in-Difference (TWFE-DiD) models can produce biased estimates when treatment effects vary across groups or overtime (Goodman-Bacon (Reference Goodman-Bacon2021)). In our setting, this bias may arise because firms in early-treated states can serve as controls for those in later-treated states, causing negative weighting and biasing the estimation of dynamic treatment effects (Goodman-Bacon (Reference Goodman-Bacon2021)). To mitigate this concern, we employ a stacked event-study design (Sun and Abraham (Reference Sun and Abraham2021), Barrios (Reference Barrios2022), and Wing, Freedman, and Hollingsworth (Reference Wing, Freedman and Hollingsworth2024)), which estimates treatment effects relative to adoption cohorts while avoiding comparisons between previously and newly treated units. Specifically, for each treatment year c in which at least one state enacted an anti-SLAPP statute, we form a separate cohort. The treatment sample for cohort c includes firm-year observations within the window [c − 5, c + 5] for firms headquartered in states that enacted anti-SLAPP statutes in year c. The control sample includes firm-year observations within the same window for firms headquartered in states that have not adopted an anti-SLAPP statute by year c + 5. We then stack all cohorts into a single panel for estimation, ensuring that identification relies solely on not-yet-treated states as controls.

In this way, after stacking observations of all cohorts, our final regression sample for the main analysis contains 118,287 cohort-firm-year observations from 1,600 unique firms between 1990 and 2019. We winsorize all continuous variables in each cohort at the 1st and 99th percentiles to mitigate the influence of outliers.

C. Empirical Design

We estimate the following cohort DiD regression using the stacked cohort sample constructed above to test the effect of the anti-SLAPP statutes on corporate toxic release:

(1) $$ {\displaystyle \begin{array}{c}{RELEASE}_{c,i,t}={\beta}_0+{\beta}_1 ANTI\_{SLAPP}_{c,s,t}+\sum {\beta}_k{controls}_{c,i,t}\\ {}+{\gamma}_{c,i}+{\rho}_{c,s}+{\delta}_{c,j,t}+{\varepsilon}_{c,i,t}\end{array}} $$

where c denotes the cohort, i the firm, t the year, s the firm’s headquarters state at year t, and j the 48-Fama French code for firm i at year t. In addition, $ {\gamma}_{c,i} $ , $ {\rho}_{c,s} $ , and $ {\delta}_{c,j,t} $ represent cohort–firm, cohort–state, and cohort–industry–year fixed effects, respectively. These fixed effects control for unobserved heterogeneity at the firm, state, and industry–year levels within each anti-SLAPP cohort. Specifically, cohort–firm fixed effects ( $ {\gamma}_{c,i} $ ) absorb time-invariant firm-level characteristics such as baseline environmental practices, managerial culture, and strategic orientation; cohort–state fixed effects ( $ {\rho}_{c,s} $ ) account for persistent differences across states, including civil litigation norms, political climates, and economic conditions that may influence both the adoption and impact of anti-SLAPP laws; cohort–industry–year fixed effects ( $ {\delta}_{c,j,t} $ ) capture common shocks or trends affecting firms within the same industry and year, such as sector-wide technological changes or macroeconomic fluctuations. Since the treatment in our setting varies by state, we cluster standard errors at the headquarters-state level, allowing for within-state correlation in the error terms and ensuring valid statistical inference (Imbens and Wooldridge (Reference Imbens and Wooldridge2009)).

Following Akey and Appel (Reference Akey and Appel2021), Chen et al. (Reference Chen, Jing, Keasey and Xu2026), Duchin et al. (Reference Duchin, Gao and Xu2025), Jing et al. (Reference Jing, Keasey, Lim and Xu2024), and Li et al. (Reference Li, Xu and Zhu2025), the dependent variable in equation (1), $ RELEASE $ , captures corporate environmental policies, proxied by total toxic pollution, including LOG(RELEASE) and sale-adjusted toxic emissions LOG(RELEASE/SALE), where LOG(RELEASE) is the natural logarithm of the amounts of total toxic pollution, and LOG(RELEASE/SALE) is the natural logarithm of sales-adjusted toxic pollution. $ ANTI\_ SLAPP $ is an indicator that equals to 1 if firm i is headquartered in state s that enacts an anti-SLAPP law in year t, and in all subsequent years. That said, zero values of $ ANTI\_ SLAPP $ indicate nonadopters or adopters’ preadoption years. A significantly negative coefficient on $ {\beta}_1 $ would indicate that the adoption of anti-SLAPP laws is associated with a reduction in corporate toxic emissions, consistent with our prediction that amplified stakeholder voice under anti-SLAPP motivates firms to improve their environmental performance.

The inclusion of control variables follows prior literature (Akey and Appel (Reference Akey and Appel2021), Jing et al. (Reference Jing, Keasey, Lim and Xu2024), Chen et al. (Reference Chen, Jing, Keasey and Xu2026), and Li et al. (Reference Li, Xu and Zhu2025)). Specifically, we include the logarithm of total assets (SIZE), return on assets (ROA), leverage ratio (LEV), firm age (AGE), cash holdings (CASH), cost of goods sold (COG), selling, general and administrative expense (SG&A), asset tangibility (PPE), research and development expenditure intensity (R&D/AT), an indicator for missing R&D data (MISSINGR&D), dividend payments (DIVIDEND), analysts coverage (ANALYST), and institutional ownership (INS). The definitions of all variables and data sources are summarized in Section A of the Supplementary Material.

D. Descriptive Statistics

Panel A of Table 1 summarizes the descriptive statistics of the variables used in equation (1), derived from the cohort-matched sample.Footnote ⁸ Firms report an average pollution level of 889.6 thousand pounds with a standard deviation of 3,127 thousand pounds. The log-transformed, sales-scaled measure of pollution (LOG(RELEASE/SALE)) has a mean of 2.997 and a standard deviation of 2.853. The mean value of ANTI_SLAPP is 0.035, comparable to Lee et al. (Reference Lee, Ng, Yoo and Zhang2026). Panel B presents the distribution of pollution by year. Average firm-level emissions, measured in thousands of pounds, decline steadily from the early 1990s to the late 2010s. Specifically, mean releases fall from about 1,130 thousand pounds in 1990 to 750 thousand pounds in 2019, indicating marked improvements in environmental efficacy. The log-transformed measures, both unscaled and sales-scaled, exhibit a similar downward trend.

TABLE 1 Descriptive Statistics and Sample Distribution

Panel C of Table 1 presents the breakdown of pollution by industry (Fama–French 48 industry groups). The data reveal substantial cross-industry heterogeneity. Toxic emissions are highly concentrated in a few resource- and energy-intensive sectors, such as Precious Metals, Steel Works, Petroleum and Natural Gas, Chemicals, and Business Supplies, where average emissions exceed 2,400 thousand pounds. In contrast, sectors such as Agriculture, Construction, and Candy & Soda record relatively lower emissions. The log-scaled measures (LOG(RELEASE/SALE)) further confirm that heavy manufacturing and extraction industries, as well as material-intensive sectors, dominate total reported releases.

A. Does the Adoption of Anti-SLAPP Laws Relate to Decreases in Toxic Releases?

We predict firms to improve environmental performance by engaging in lower toxic releases following the enactment of anti-SLAPP laws. Table 2 reports the estimation results of equation (1), which examines the association between anti-SLAPP enactments and firms’ toxic emissions.

TABLE 2 The Effect of Anti-SLAPP Laws on Corporate Toxic Emissions

Across all specifications, the coefficients on ANTI_SLAPP are significant and negative, suggesting that firms significantly reduce toxic pollution after the enactment of anti-SLAPP laws. This finding is consistent with our conjecture that stronger free-speech protection amplifies stakeholder voice related to environmental issues, and firms are motivated to improve corporate environmental performance. We find that the effect is also economically significant. In column 2, the coefficient on ANTI_SLAPP (i.e., −0.5308) implies that for an average firm in our sample, total toxic releases decrease by approximately 41% following anti-SLAPP adoption; in column 4, the corresponding coefficient (i.e., −0.2017) suggests that the firm’s sales-adjusted pollution intensity declines by about 18%.Footnote ⁹ Therefore, the implementation of anti-SLAPP statutes leads to a significant and economically meaningful improvement in corporate environmental performance, reducing both the absolute and relative levels of toxic emissions.Footnote ¹⁰

B. Dynamic Treatment Effects of Anti-SLAPP on Toxic Releases

To satisfy the DiD identification requirement, we assume that, in the absence of anti-SLAPP laws, firms in the treatment and control groups would have followed parallel trends in toxic emissions over time. To assess this assumption, we estimate dynamic treatment effects by replacing the anti-SLAPP indicator in equation (1) with a set of event-year dummies for treatment firms, spanning 5 years before to 5 years after adoption (t = −5 to t = 5), excluding year t = −1 as the reference period. The event-study specification, as shown in equation (2), provides both a visual and statistical test of the parallel-trend assumption and allows us to examine the timing of anti-SLAPP law effects on toxic emissions.

(2) $$ {\displaystyle \begin{array}{c}{RELEASE}_{c,i,t}={\beta}_0+\sum \limits_{l=-5,l\ne -1}^{l=5}{\beta}_l\times {Treat}_{i,c}\times {Year}_l+\sum {\beta}_k{controls}_{c,i,t}\\ {}\hskip-9em +\hskip0.35em {\gamma}_{c,i}+\hskip0.35em {\rho}_{c,s}+\hskip0.35em {\delta}_{c,j,t}+\hskip0.35em {\varepsilon}_{c,i,t}\end{array}} $$

Graphs A and B of Figure 1 present the dynamic treatment effects of anti-SLAPP enactments on LOG(RELEASE) and LOG(RELEASE/SALE) for a [c − 5, c + 5] window. The results suggest that trends for the treatment and control firms remain fairly parallel during the pre-event period, consistent with the parallel-trend assumption, while in the postevent period following anti-SLAPP enactment, firm toxic release decreases visibly.

FIGURE 1 Dynamics of the Changes in Toxic Release Around the Anti-SLAPP Enactment Years

Figure 1 shows the dynamic treatment effects (5-year lag and 5-year lead) of anti-SLAPP and time trends on annual toxic emissions. In Graph A, the dependent variable is LOG(RELEASE), while the dependent variable is LOG(RELEASE/SALE) in Graph B. Variables are defined in Appendix A.

C. Robustness Checks

A. Environmental Investment

Under amplified stakeholder voice following the anti-SLAPP laws, firms may increase investment in cleaner technologies, innovation, and pollution control initiatives to mitigate reputational and regulatory risks. We explore this mechanism along four dimensions: green innovation, environmental investment spending, pollution source reduction, and waste management activities.

1. Green Innovation

We begin by examining whether anti-SLAPP statutes promote corporate green innovation. To measure innovation outcomes, we use firm-level patent data compiled by Kogan, Papanikolaou, Seru, and Stoffman (Reference Kogan, Papanikolaou, Seru and Stoffman2017) (KPSS).Footnote ¹³ Green patents are identified based on classifications provided by Haščič and Migotto (Reference Haščič and Migotto2015), using the International Patent Classification (IPC) system maintained by the World Intellectual Property Organization (WIPO).Footnote ¹⁴ The dependent variable, GREEN_PATENT, measures the number of green patents that firm i applies in year t and are ultimately granted. We estimate equation (1) using a Poisson model.Footnote ¹⁵ Column 1 of Table 3 shows that firms affected by anti-SLAPP laws produce significantly more green patents. This result suggests that enhanced stakeholder voice and public accountability foster the development of cleaner technologies and environmental innovation.

TABLE 3 Mechanism Analysis Results: Environmental Investment

B. Environmental Governance

Our second mechanism focuses on environmental governance, which reflects the organizational structures and managerial systems through which firms integrate environmental considerations into corporate decision-making, aimed at improved environmental performance. We examine this mechanism across four dimensions, including the appointment of sustainability-related directors on the board, ESG-linked executive compensation, employee environmental training, and environmentally responsible supply chain management.

1. Sustainability Directors

We first examine whether treated firms are more likely to appoint a sustainability director. Appointing a sustainability director embeds accountability and expertise in environmental oversight, helping ensure sustainability goals are integrated into both strategic and operational decisions.

Director information is obtained from BoardEx. Following Chen et al. (Reference Chen, Jing, Keasey and Xu2026) and Fu, Tang, and Chen (Reference Fu, Tang and Chen2020), we classify a director as sustainability-related if their job title or description includes the terms of sustainability, sustainable, responsibility, ethics, or environment. We then create an indicator variable, SUS_DIR, which equals 1 for firm-years with at least one such director, and 0 otherwise. Column 1 of Table 4 shows a positive and statistically significant association between anti-SLAPP enactment and the likelihood of appointing a sustainability director. This result suggests that firms operating under stronger free-speech protection allocate more internal governance resources to environmental and ethical oversight. It also indicates that heightened stakeholder scrutiny encourages firms to formalize sustainability leadership roles, thereby signaling a stronger commitment to responsible management.

TABLE 4 Mechanism Analysis Results: Environmental Governance

A. Cross-Sectional Analyses

In this section, we examine whether the effect of anti-SLAPP statutes on firms’ toxic emissions varies across firms with different levels of stakeholder scrutiny and managerial incentive alignment. We argue that while enhanced protection of free speech amplifies public oversight, the extent to which firms respond depends on their external exposure to stakeholder pressure and internal incentive to pursue sustainable practices.

1. External Stakeholder Pressure

We first examine whether the effect is more pronounced among firms subject to greater external stakeholder pressure. Firms that rely heavily on government clients or operate in regions with active environmental NGOs arguably face stronger reputational and regulatory monitoring, making them more responsive to public criticism under anti-SLAPP protections.

To test this prediction, we employ two proxies for external stakeholder pressure: GOV_CUS is an indicator equal to 1 if the firm has government customers, and 0 otherwise, which captures exposure to government procurement oversight and the reputational constraints associated with serving public clients; ENV_EGO measures the average intensity of environmental NGOs in a firm’s headquarters state, reflecting the strength of local activist and advocacy monitoring. Both variables are measured before anti-SLAPP law adoption to mitigate endogeneity concerns related to concurrent institutional changes.

We interact these variables with the anti-SLAPP indicator and re-estimate equation (1) with these interaction terms included. Panel A of Table 5 presents the results. Columns 1 and 2 report regressions interacting ANTI_SLAPP with GOV_CUS, while columns 3 and 4 use ENV_EGO as the moderating variable. Across all specifications, the coefficients on the interaction terms are significantly negative, suggesting that, in line with our prediction, the reduction in firms’ toxic emissions following the enactment of anti-SLAPP laws is more pronounced among firms facing greater stakeholder exposure.

TABLE 5 Cross-Sectional Analysis Results

B. Does Anti-SLAPP Reduce Toxic Releases by Shrinking Firms’ Economic Activities?

We have by far documented that anti-SLAPP statutes are associated with lower toxic releases and improved environmental practices. However, a potential concern is that the observed decline in emissions could stem from reduced economic activity rather than genuine environmental improvement. As noted by Akey and Appel (Reference Akey and Appel2021), firms may lower reported emissions simply by scaling down production or operations. To address this concern, we test whether the enactment of anti-SLAPP laws leads to a significant decline in firms’ economic activities.

We employ five proxies to capture firm-level economic performance. First, we use the current-year production ratio from the TRI database. We aggregate facility-level ratios to the firm-year level to obtain PRODUCT_RATIO. Footnote ¹⁸ Second, NIGROWTH measures a firm’s annual growth rate of net income. Third, TFP stands for total factor productivity, which captures a firm’s efficiency in generating output given its labor and capital inputs. Specifically, we estimate TFP as the residual from regressing the natural logarithm of firm sales on the logarithms of the number of employees and net property, plant, and equipment, with regressions conducted separately by 2-digit industry and year. Fourth, we measure firm size by the number of employees (i.e., LOG(EMP)) to capture firms’ production scale and operational activity. Finally, FC, a financial-constraint index (Linn and Weagley (Reference Linn and Weagley2024)), reflects firms’ capability to sustain investment and operations—a decline in financial flexibility would indicate economic slowdown.

The results, reported in columns 1–5 of Table 6, consistently show that the coefficients on the anti-SLAPP indicator are statistically insignificant. These findings suggest that the implementation of anti-SLAPP laws does not significantly affect firms’ production activities or overall economic operations. Therefore, the observed reduction in toxic emissions likely reflects genuine environmental improvement rather than a contraction in firm scale.

TABLE 6 Do Anti-SLAPP Laws Shrink Economic Activities?